METAL ALKOXIDE-INDUCED CLEAVAGE OF SILICON-NITROGEN BONDS IN UNSYMMETRICALLY ALKOXYLATED DISILAZANES Thesis far the Degree of M. S. MICHIGAN STATE UNIVERSITY Ronald Eugene Goldsberry 1:966 LIBRARY 1955).; . . . Michlgan State University ABSTRACT METAL ALKOXIDE-INDUCED CLEAVAGE OF SILICON- NITROGEN BONDS IN UNSYMMETRICALLY ALKOXYLATED DISILAZANES by Ronald Eugene Goldsberry The purpose of this study was to synthesize unsym- metrically alkoxylated disilazanes and to investigate the ease with which sodium methoxide and aluminum isopropoxide cleaved silicon-nitrogen bonds as a function of the number of alkoxy groups bonded to silicon. Also, investigated was the possibility of obtaining linear silicon-nitrogen polymers from the cleavage of the silicon-nitrogen bond by the metal alkoxides. Two disilazanes, 1,2-dimethoxy-1,1,2,2—tetramethyldi- silazane (I) and 1-methoxy-1,1-dimethyl-2,2,2-triethoxy- disilazane (II) were prepared according to the following generalized equation, where R - CH3 or C2H5: (RO)3SiCl + (CH3O)(CH3)ZSiCl + 3NH3 -—*¢-2NH4C1 + (CH30)(CH3)ZSiNHSi(OR)3. Small amounts of the correSponding symmetrical compounds are also formed. Metal alkoxides were found to cleave the silicon- nitrogen bond in Compounds (I) and (II) to give alkoxy- silanes and compounds containing extended silicon-nitrogen frameworks. The latter compounds range from linear oligo- meric oils to highly crosslinked solids. Vapor phase chromatographic studies of the products of the various re- actions indicated that the cleavage of the silicon-nitrogen Ronald Eugene Goldsberry bond in the presence of metal alkoxides was a complex pro- cess. Sodium methoxide cleaved the silicon-nitrogen bond to the more alkoxylated silicon in compound (I), but cleaved the silicon-nitrogen bond with the less completely alk- oxylated silicon in compound (II). Aluminum isoprOpoxide did not appear to distinguish between the two silicon- nitrogen bonds in compound (I). However, for compound (II) the silicon-nitrogen bond involving the silicon with three ethoxy groups bonded to it was cleaved more readily by aluminum isopropoxide than was the correSponding bond to the singly methoxylated silicon. METAL ALKOXIDE-INDUCED CLEAVAGE OF SILICON- NITROGEN BONDS IN UNSYMMETRICALLY ALKOXYLATED DISILAZANES BY Ronald Eugene Goldsberry A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Chemistry 1966 DEDICATION To My Mother, Mrs. Constance Goldsberry ii ACKNOWLEDGMENTS The author is indebted to Professor W. E. Weibrecht for the helpful guidance and assistance offered during this investigation and during the preparation of this thesis. He is also especially grateful to his wife, Betty, whose patience, encouragement and assistance in preparing the.manuscript were invaluable to the completion of this degree. iii I. II. INTRODUCTION . TABLE OF CONTENTS NOMENCLATURE . . . . . ... ... . . . . . . HISTORICAL . . . . . . . . . . . . . . . . EXPERIMENTAL . . . . . . . . . . . . . . . . REAGENTS . . . . . . . . . . . . . . . . . PREPARATION OF HEXAMETHOXYDISILAZANE . . . 1. Preparation and Purification . . . . 2. Nuclear Magnetic Resonance Spectrum of Hexamethoxydisilazane . . . . . . PREPARATION OF 1-METHOXY-1p1-DIMETHYL- 2 I 2 l Z-TRIMETHOXYDISILAZANE o o o o o o o o . Preparation and Purification . . . . . Discussion of Analytical Data . . . . Infrared Spectra . . . . . . . . . . . Nuclear Magnetic Resonance Studies . . Summary . . . . . . . . . . . . . . UHhOONDI-l PREPARATION OF 1-METHOXY-1,1-DIMETHYL- 2,2,2-TRIETHOXYDISILAZANE . . . . . . . . . Preparation and Purification . . . . . Infrared Spectra . . . . . . . . . . . Nuclear Magnetic Resonance Studies . . Vapor Phase Chromatography . . . . . . Summary . . . . . . . . . . . . . . 01¢me SILICON ANAYLSIS . . . . . . . . . . . . . NITROGEN ANALYSIS . . . . . . . . . . . . PREPARATION OF ALKOXY SILANES . . . . . . PYROLYSIS OF 1-METHOXY-1,1-DIMETHYL-2,2,2- TRIMETHOXYDISILAZANE . . . . . . . . . . . 1. Pyrolysis of the Disilazane in a Sealed Tube .. . . . . . . . . . . . iv Page 13 13 13 13 14 14 14 16 17 17 21 22 22 24 24 25 26 27 27 28 29 30 TABLE OF CONTENTS (Cont.) III. 2. Pyrolysis of the Disilazane at Atomospheric Pressure . . . . . . 3. Pyrolysis of the Disilazane in the Presence of Sodium Methoxide . . . a. Experimental Procedure . . . . b. Nitrogen Analysis . . . . . . 4. Pyrolysis of the Disilazane in the Presence of Aluminum Isopropoxide. a. Experimental Procedure . . . . b. 8-Hydroxyquinoline Method for Determination of Aluminum . . PYROLYSIS OF 1-METHOXY-1,1-DIMETHYL-2,2,2- TRIETHOXYDISILAZANE . . . . . . . . . . . 1. Pyrolysis of the Disilazane in a Sealed Tube . . . . . . . . . . . 2. Pyrolysis of the Disilazane in the Presence of Sodium Methoxide . . . 3. Pyrolysis of the Disilazane in the Presence of Aluminum Isopropoxide SUMMARY . . . . . . . . . . . . . . . . . BIBLIOGRAPHY . . . . . . . . . . . . . . . APPENDICES O O O O I O I O O O O O O O O O I. INFRARED SPECTRUM OF COMPOUND (I) II. PROTON NMR SPECTRA OF DISILAZANES III. GAS CHROMATOGRAMS OF ALKOXY SILANES. Page 31 31 32 35 35 35 38 39 4O 4O 43 47 53 56 57 59 64 TABLE II. III. IV. VI. VII. VIII. IX. XI. XII. XIII. XIV. XV. LIST OF TABLES Distillation of 1-methoxy-1,1-dimethyl- 2,2,2-trimethoxydisilazane . . . . . . . . .Elemental analysis of 1-methoxy—1,l-di- methyl-2,2,2-trimethoxydisilazane . . . . Infrared absorption frequences of 1,1-di- methyl-2,2,2-trimethoxydisilazane . . . . Comparison of NMR peaks from the distilla- tion of 1-methoxy-1,1-dimethyl-2,2,2- trimethoxydisilazane . . . . . . . . . . Distillation of 1-methoxy-1,1-dimethyl- 2,2,2-triethoxydisilazane . . . . . . . . Physical constants of some alkoxy silanes. Pyrolysis of Compound (I) at 200°C . . . . Stoichiometry of the pyrolysis of Compound (I) with sodium methoxide . . . . . . . . Alkoxy silanes produced from the pyrolysis of Compound (I) with sodium methoxide . . Elemental analysis of the polymer from the, pyrolysis of Compound (I) with sodium methoxide . . . . . . . . . . . . . . . . Alkoxy silanes produced from the pyrolysis of Compound (I) with aluminum isopropoxide Elemental analysis of the polymers from the pyrolysis of Compound (I) with' aluminum isopropoxide . . . . . . . . . . roduced from the pyrolysis Alkoxy silanes at 200°C . . . . . . . . of Compound (II Alkoxy silanes produced from the pyrolysis of Compound (II) with sodium methoxide . . Elemental analysis of the polymer from the pyrolysis of Compound (II) with sodium methoxide . . . . . . . . . . . . . . . . vi Page 16 16 18 19 23 29 31 33 33 34 37 38 40 42 42 TABLE XVII. XVIII. LIST OF TABLES (Cont.) Page Alkoxy silanes produced from the pyrolysis of Compound (II) with aluminum isopropoxide 44 Elemental analysis of the crosslinked polymer from the pyrolysis of Compound (II) with aluminum isoprOpoxide . . . . . . . . 45 Elmental analysis of the oligomeric oil from the pyrolysis of Compound (II) with aluminum isopropoxide . . . . . . . . . . 45 vii LIST OF FIGURES FIGURE Page 1. Infrared spectrum of 1-methoxy—1,1-dimethyl- 2,2,2-trimethoxydisilazane . . . . . . . . . 58 2. Proton NMR Spectrum of Compound (I) . . . . 6O 3. Proton NMR spectrum of Compound (II) . . . . 61 4. Proton NMR Spectrum of the oligomeric oil from the pyrolysis of Compound (I) . . . . . 62 5. Proton NMR spectrum of the oligomeric oil from the pyrolysis of Compound (II) . . . . 63 6. Gas chromatogram of some known alkoxy silanes 65 7. Gas chromatogram of the decomposition mixture from heating Compound (I) at 200°C for 24 hours67 8. Gas chromatogram of the alkoxy silanes from the sodium methoxide catalyzed decomposition of Compound (I) . . . . . . . . . . . . . . 68 9. Gas chromatogram of the alkoxy silanes from the aluminum isopropoxide catalyzed decom- position of Compound (I) . . . . . . . . . . 69 10. Gas chromatogram of the decomposition mixture from heating Compound (II) at 200°C for 24 hours . . . . . . . . . . . . . . . . . . 70 11. Gas chromatogram of the alkoxy silanes from the sodium methoxide catalyzed decomposition of Compound (II) . . . . . . . . . . . . . . 71 12. Gas chromatogram of the alkoxy silanes from the aluminum isopropoxide catalyzed decom- position of Compound (II) . . . . . . . . . 72 viii I . IN TRODUCTI ON NOMENCLATURE The Committee on Nomenclature of the American Chemical Society and the Commission on the Nomenclature of Organic Chemistry of the International Union of Pure and Applied Chemistry have adopted a system for naming organosilicon compounds. Compounds derived from the structure H3SiNH2 are called Silylamines, with the use of appropriate pre- fixes to designate substitution. The system of prefix designation for nitrogen substitution becomes quite cumber- some when more than one nitrogen is attached to silicon. In these cases the amine grouping is designated as a sub- stitutent of the silane. The generic name silazane is given to the series H3Si(NHSiH2)nNHSiH3. Compounds of this series are called disilazanes, trisilazanes, etc. depending upon the number of silicon atoms in the molecule. Two spellings of this generic name are found in the literature, silazane and silazine, with the former being preferred. Compounds of the type (HZSiNH)n are given the generic name cyclosila- zanes, the prefix depending upon the number of silicon atoms in the ring. 2 HISTORICAL The study of silicon—nitrogen compounds embraces both inorganic and organic chemistry. Although Silicon-nitrogen compounds, such as Silylamines, differ markedly from their carbon analogs with respect to chemical behavior, the or- ganic nature of this class of compounds should not be under- emphasized. An excellent review article1 exists, which bridges both fields of chemistry and reviews publications through December, 1959, with a few important references from papers published in 1960. Trisilylamine and tri(methylsilyl)amine have been shown by infrared and Raman spectra2r3'4 and electron-diffrac- tion5 data to be co-planar molecules.6 Also, the Silicon- nitrogen bond in trisilylamine is shorter than the calcu- lated Silicon—nitrogen single bond distance but.longer than the calculated double bond distance. These factors have been interpreted in terms of the ability of the "lone pair" of electrons of nitrogen to be donated to the empty ‘g or- bital of the silicon (dw - pw overlap). . _ @- SiH3 SiH3 5 SiH3 SiH3 SiH3 . SiH3 \\.Ot7’ \\ éfi/i ‘\,§P/ N <—————> N <—-——> . The polarity of the Si-N bond in substituted trisilylamines is decreased and the basicity is lowered as compared with their methyl substituted isostructural counterparts.7 This may be observed for example in the formation of complexes 3 of silylmethylamines with trimethylboron;8 neither trisily- amine nor methyldisilykmine form a complex, dimethylsilyamine forms a weak complex and trimethylamine a more stable one. The compounds (CHasiH2)3 and (CHasiH2)2NCH3 do not form complexes with trimethylboron, but (CH3SiH2)N(CH3)2 does.9 The considerable difference between the electronega— tivities of silicon and nitrOgen contributes to a high thermalstability of the Si-N bond but, at the same time, is responsible for its susceptibility to solvolysisl° by polar solvents. In the infrared spectra of trialkylsilyl— amines, nitrogen-hydrogen bond stretching and deformation frequencies are similar to those found for normal amines.11 The Silicon-nitrogen asymmetric stretching frequency in the simple amines has been observed in the 900-1000 cm-1 region¥2r3r4 Other tabulations and published spectra are available.13I14I 15.16.17 The mass-spectral analysis of hexamethyldisilazane has been reported,18 as have the nuclear magnetic resonance 19 Spectra of hexamethyldisflazama ‘hexamethyl-cyclotrisilazane,13 and other silylamines.7 Synthesis of Silicon-Nitrogen Compounds The action of ammonia, primary and secondary amines on a halosilane results in the formation of a Silicon-nitrogen bond. Comparative yield data as a function of the nature of the halogen are not available in a sufficiently sterically hindered system for the differences in reactivity to be de- tected. From the data available, the bromo- and iodosilanes 4 appear more reactive toward a given amine than do the chlorosilanes.2° Owing to their availability, however, the chlorosilanes are most frequently employed. The halide released during the Course of the reaction is pre- cipitated as the amine hydrohalide. The reaction is reversible, with the halosilane being obtained from the silylamine and the amine salt.21 RasiCl + 2R'NH2 :——> R3SiNHR' + R'NH3C1 By analogy with the chemistry of the corresponding carbon compounds, the reaction of a monohalosilane with ammonia should yield a variety of silicon-nitrogen products. However, owing to the electronic and steric natures of the compounds involved, only one or, in a few cases, two re- action products are formed. When the non-halogen sub- stituents attached to silicon are hydrogen, there is a definite tendency for complete silylation of the amine2.2t23'24 When Silyl chloride is treated with ammonia, trisilyamine can be obtained in 80 per cent yield.22 H3SiCl + NH3 > (H381)3N + 3HCl The product of the reaction is dependent on the size of the alkyl substituent on Silicon. The tendency for complete silylation of the amine is decreased as the size of the groups attached to silicon is increased. With trimethyl- chlorosilane, only hexamethyldisilazane can be isolatedE5I2°I 27:23:29 An attempt to obtain trimethylsilylamine by the use of excess liquid ammonia failed. This failure was 5 probably caused by rapid condensation of the silylamine to the disilazane during attempted isolation. H I (CH3)3SiCl + NH3 -—> (CH3)3SiN-Si(CH3)3 Tris(trimethylsilyl)amine can, however, be formed by the use of the lithium3° or sodium6 salt of hexamethyldi- silazane. [(CH3)3Si]2NLi + (CH3)3SiCl ——> [(CH3)3Si]3N The preference for formation of silyamine, is again observed in the reaction of triethylchlorosilane with ammonia. In this reaction, triethylsilylamine is the major product, the disilazane being the minor product.31'32 (C2H4)3SiCl + NH3-——> (C2H5)3SiNH2 The disilazane may be obtained by treating the silyl- amine with triethylchlorosilane at a higher temperature.33 With higher trialkylchlorosilanes, only silylamines have been reported as products.3"~35 Trialkoxychlorosilanes react with ammonia and amines in the same general fashion as do trialkyl and triarylhalo- silanes. Both trimethoxy and triethoxychlorosilanes yield the corresponding disilazanes when treated with ammonia,36 while triiSOprOpoxy37 and tributoxychlorosilanes38I39 yield only the silylamines. A further increase in the size of the alkoxy group has no effect upon the products obtained. (CH3O)3SiCl + NH3'——> [(CH30)3Si]2NH (C4H90)3SiCl + NH3 -—> (C4H90)3SiNH2. 6 In recent years a considerable amount of research has been devoted to the synthesis of inorganic polymers with properties which would enable them to function as elastomers, resins, and lubricants at extreme tempertures. Such a polymer, methyl silicone, a polyorganosiloxane usually rep- resented as [(CH3)ZSiO]n, is a straight-Chain polymer con- taining alternating silicon and oxygen atoms. Since the N-H group is isoelectronic with oxygen, one might expect to be able to obtain substances analogous to the silicones in which the N-H groups replace oxygen atoms. A possible route to substances of this kind is illustrated by the re- action of dimethyldichlorosilane with ammonia according to the equation: n(CH3)ZSiC12 + 3rINH3 ——> [(CH3)ZSiNH]n + ZnNH4Cl Thus the synthesis of polymers, based on a silicon-nitrogen framework, that have optimum stability toward hydrolysis and high temperature and have a highly linear structure has been an active research area. One should note, however, that the hydrogen atom at- tached to nitrogen in these polyorganosilazanes should be capable of further substitution by other dimethyldichloro- silane molecules to give a crosslinked polymeric material. This is one feature which has stimulated the recent interest in Silicon-nitrogen polymers. Also, due to the considerable d7 - pr interaction between the Silicon and nitrogen atoms, arising from contributions of the'lone pair" of electrons on nitrogen to the empty g_orbitals of silicon, the silicon 7 framework is bonded together more rigidly than the flexible silicon-oxygen framework of the siloxanes. It is also well known that silicon-nitrogen compounds as a rule tend to split off ammonia in the presence of water and alcohols and give the analogous siloxanes. This tendency to hydrolyze is related directly to the number of silicon atoms bonded to the nitrogen atom and hence to the degree of dv-pw bonding in the molecule. For example, as one proceeds from a primary to a tertiary silylamine, the tendency towards hydrolysis becomes progressively smaller. This indicates that delocalization of the electron pair into Q orbitals of the silicon atom minimizes the attack on the nitrogen atom and leads to a more stable compound. Therefore, if one wants to obtain a hydrolytically stable Silicon-nitrogen polymer, one must "tie up" the electron pair on the nitrogen atom either by complexation with a metal ion or by tri-substituting the nitrogen atoms with silicon atoms. Steric factors are also important. The factors mentioned above are not the main diffi- culties encountered in the preparation of silicon-nitrogen analogs to the silicones. When one, in fact, carries out the reaction between dimethyldichlorosilane and ammonia under the normal conditions for obtaining polysiloxanes, one does not obtain a high polymer at all, but instead ob- tains cyclic compounds. Hexamethylcyclotrisilazane, [(CH3)ZSiNH]3, and octamethylcyclotetrasilazane [(CH3)ZSiNH]4, can be prepared in approximately 80 per cent yield by this 8. method.4° This strong tendency to form rings precludes the use of simple methods for obtaining organosilazane polymers. This ring-formation tendency is also found in siloxanes, but to a much smaller extent;41 the ring-chain equilibrium is understood and can be manipulated. Recently, however, by using special methods, several types of linear silicon- nitrogen polymers have been prepared. The polymers range from oils and greases to waxes and rubbers; some are ex- tremely stable while others decompose in moist air to liber— ate ammonia and form the corresponding siloxanes. Since dimethyldichlorosilane reacts with ammonia to form mixtures of hexamethylcyclotrisilazane and octamethyl- cyclotetrasilazane, plus a small amount of polymer, the silizane rings must be either Opened and the fragments joined to form a linear polymer or the rings themselves must be connected in some manner. Andrianov attempted to prepare polysilazanes by treatment of hexamethylcyclotri- silazane with potassium hydroxide.42 At 165°C, some cleav- age of the silicon-carbon bonds occurred with the formation of methane. Only highly crosslinked polysilazanes could be isolated from the reaction mixture. A basic catalysis involving a base in the ammonia system, i.e., sodium amide, is also unsatisfactory. It has been Shown that this reagent metalates the NH groups in the cyclic starting material rather than splitting the silicon-nitrogen bonds:43 Krfiger studied the use of ammonium halides, Which are acids in the ammonia system, as possible polymerization 9 catalysts.44 It had previously been noted by Anderson that silicon-nitrogen bonds could be broken in the presence of inorganic halides.45 Polymers were obtained which ranged from waxes to very tough elastomers which showed remarkable thermal stability and considerable hydrolytic stability. The silazane polymers were found to have approximately 2/3 of their nitrogen atoms completely substituted. This would indicate either of two structures: I II SiR2 SiR2 ’ N \\\II// or -N-SiR2-NH-SiR2-N-SiR2 I I I I SiR2 SiR2 SiR2 - ‘\\INH’/’/ .4n n where R = CH3. Structure I is the more probable structure, for it is well known that silazanes tend to form small rings.4°'47 However, there is no need to assume perfect ordering of the ring structures; and, therefore, the poly- mers are probably a mixture of structure I and structure II. Kruger also isolated two highly cross-linked polymers by heating freshly-distilled hexamethylcyclotrisilazane with 1% of ammonium bromide to 280°C, for several hours. A dark brown material resulted having properties between those of a rubber and those of a wax. The other polymer was pre- pared by carrying out the reaction of dimethyldichlorosilane with ammonia to obtain a mixture of cyclic trimer and tetra- mer, and then heating this crude reaction mixture at 280°C for several hours in the presence of 5% ammonium bromide as a catalyst. A White, waxy material resulted which looked 10 very much like polyethylene; both compounds were flexible, but not elastic.44 By mixing 55 grams of cyclotrisilazane and 3 grams of methyltrichlorosilane with 5% of ammonium bromide and heat— ing the mixture while bubbling ammonia through it,.Krfiger obtained an elastic, vwaxy; polymer which melted at 178- 190°C. This polymer was crosslinked through silicon and tertiary nitrogen. A polymer cross—linked through vinyl groups and ter- tiary nitrogen was prepared by Kruger by heating a mixture of 40 grams of cyclotrisilazane, 10 grams of trimethyl-tri- vinylcyclotrisilazane and 5% of ammonium bromide. A rub- ber Which swelled in carbon tetrachloride was Obtained. It was found by R. N. Minné48 and independently by K. Lienhard49 that a polymeric material could be obtained from the reaction of ethylenediammine with dimethyldichloro- silane. The two polymers obtained were reported as having the following structures: B CH3 CH3 -Si-NHCH2CH2NH- and -Si-N-CH2CH2-Si- én, n CH3 CH3 n Subsequent research in this area by D. Kummer5°t47 let to two more polymers: polymer I in xylene and polymer II in benzene. From analytical data on polymer I and poly- mer II it was found that neither of the two pure compositions, A or B, had been obtained. Instead,a mixture of both 11 structures is present in both polymer I and polymer II, and the proposed average structure is given by Kummer5° as: CH3 CHz-CHZ CH3 CH3 CHz-CHEN Si-N\<\ \N- Si-NCH2CH2NH—C —Si -N Si/// w‘\\\8 ///% CH3 CH3 CH3 CH3 3 n The presence of the five-membered rings in the polymer is supported by IR51 measurements and by double-resonance high resolution NMR.52 In no case could the pure linear compound of Minne be obtained, even with an excess of ethyl- enediamine. Attempts have been made to polymerize N,N',N"-nona- methylcyclotrisilazane.53 So far little success has been achieved, for it seems that the cyclic compounds are much too stable. The advantage of obtaining an N—methylsilazane is that cross—linking through the nitrogen atoms would be prohibited, and a linear polymer should, therefore, result. Finally, low—molecular-weight linear silicon-nitrogen polymers, were prepared by Weibrecht.54 In this method, sym-dimethoxy-tetramethyldisilazane and the corresponding tri and tetrasilazanes were prepared by the treatment with anhydrous ammonia. Prolonged heating of the disilazane in the presence of sodium methoxide, aluminum methoxide and isoprOpoxide, and potassium tertiarybutoxide led to liquid polysilazanes containing 5 to 6 silicon atoms, as well as to some other, high molecular weight silicon-nitrogen sub- stances of a more complicated nature. 12 Synthesis of Alkoxy Silanes The most important method of obtaining organosilicon alkoxides and phenoxides involves treatment of an organo— silicon chloride with an alcohol or a phenol: ESi-Cl + ROH -> ESiOR + HC1. Partial replacement of halogen is clearly possible in a polyhalogen compound: Et251C12 + EtOI-I —-—> EtzsiCl(OEt) + HC1. The alcoholysis reaction is reversible although the equilibrium normally lies well in favor of the products in the presence of an excess of the alcohol; therefore, it is desirable to remove the hydrogen Chloride either by boil- ing it out of the reaction mixture or by reaction with a base. More serious than the reversibility is the fact that if the hydrogen chloride is not removed, it can react with the alcohol to give the alkyl chloride and water. The latter can then hydrolyze the organosilicon alkoxide or chloride. Sodium alkoxides or phenoxides may be used in place of alcohols. They are not only the effective reagents but can serve as acceptors as well. A solution of sodium a1; koxflkain alcohol is a much more active reagent than the alcohol alone, and is particularly useful in introducing alkoxy groups in highly hindered systems. II. EXPERIMENTAL REAGENTS Silicon tetrachloride and dimethyldichlorosilane were supplied by Dow Corning and used without further purifica- tion. Methanol was refluxed over magnesium and absolute ethanol was used without further purification. All solvents were dried over sodium wire. PREPARATION OF HEXAMETHOXYDISILAZANE 1. Preparatign and Purification Hexamethoxydisilazane was prepared according to the method of Weibrecht and Rochow.54 In this procedure, silicon tetrachloride and a stoichiometric amount of methanol in solution were reacted to give trimethoxychlorosilane. Ammonia gas was added to this solution to give the desired product. In a typical experiment five—tenths of a mole of silicon tetrachloride was added to one liter of dry carbon tetra- chloride in a 3-liter, 3-necked, round-bottom flask equipped with a mechanical stirrer, reflux condenser, and dropping funnel. One and five-tenths moles of methanol was added dropwise and HCl evolution began immediately. After all of the methanol had been added, the reaction mixture was refluxed for 1.5 hours and the apparatus flushed with dry nitrogen to remove traces of hydrogen chloride. The 13 14 trimethoxychlorosilane. which was presumably the product of this reaction, was not isolated. Rather, ammonia gas was added directly to the hexane solution described above and ammonium chloride precipitation was observed immedi- ately. Concurrently, there was a rapid rise in temperature which caused the hexane to reflux without external heating. Ammonia addition was continued for five hours until the reaction mixture began to cool. The hexane solution was separated from the very large amount of NH4C1 by suction filtration. The ammonium Chlor- ide was then washed with 2 liters of hexane, and the two hexane fractions were combined. Undoubtedly, this washing of the ammonium chloride was one of the most crucial steps in maximixing the yield. Much of the desired product would otherwise almost certainly have been left clinging to the NH4C1. The hexane was distilled out of the mixture, and when the temperature had risen to 95°C, it appeared as though all of the hexane had been removed. The remaining liquid was fractionally distilled at atmOSpheric pressure using a 15 inch column packed with glass helices. The normal boil- ing point of hexamethoxydisilazane is 220°C. Forty and three tenths grams of material boiling between 220°-223°C were collected, which corresponded to a 63% yield of hexa- methoxydisilazane. 14 2. Nuclear Magnetic Resonance Spectrum of Hexamethoxydi- silazane The proton nuclear magnetic resonance spectrum of the hexamethoxydisilazane was obtained in 50% carbon tetrachlor- ide solution as solvent with tetramethylsilane as the ref- erence, using a Varian A-60 Spectrometer. The spectrum showed one peak which was assigned to the methoxy protons at O = 3.60 downfield from TMS. This indicates that all of the methoxy groups in the molecule are in the same chemical environment. This is entirely consistent with what would be expected for (CH3O)3-Si—NH-Si-(OCH3)3. PREPARATION OF 1-METHOXY-1,1-DIMETHYL-2,2,Z-TRIMETHOXYDISILAZANE 1. Preparation and Purification The l-methoxy-l,1-dimethyl-2,2,2—trimethoxydisilazane was prepared in a manner similar to that described for the preparation of hexamethoxydisilazane in the preceding sec- tion. A typical run is described below: Two moles of dimethyldichlorosilane was added to one liter of dry nrhexane in a 3-liter, 3-necked, round-bottom flask equipped with a mechanical stirrer, reflux condenser and dropping funnel. Similarly, two moles of silicon tetra- chloride was added to one liter of dry nfhexane. Two moles of methanol was added dropwise to the flask containing the 15 dimethyldichlorosilane, and six moles of methanol was added to the flask containing the silicon tetrachloride. Hydrogen chloride evolution began immediately. After all of the methanol had been added, the reaction mixtures were re- fluxed for 1.5 hours and in each case the reaction vessels were flushed with dry nitrogen to remove traces of hydrogen chloride. The methoxydimethylchlorosilane and trimethoxy— chlorosilane which were presumably the products of these reactions were not isolated. Rather, ammonia gas was added directly to the combined hexane solutions described above, and ammonium chloride precipitation was observed immediately. Concurrently, there was a rapid rise in temperature which caused the hexane to reflux without external heating. The hexane solution was separated from the very large amount of NH4Cl by suction filtration. Ammonia gas was again added to the hexane solution described above and the resulting mixture filtered by suction to remove the NH4Cl. This process was repeated until an excess of ammonia was present indicating the reaction was complete. In each case, the ammonium chloride was washed with hexane, and the hexane fractions were combined. The hexane was distilled out of the mixture, and when the temperature had risen to 95°C, it appeared as though all of the hexane had been removed. The liquid boiling above 95°C at atmospheric pressure corresponded to a 73% yield of crude disilazanes. This liquid was frac- tionally distilled at 10 mm. pressure, using a 15 inch, 16 insulated column packed with glass helices and yielded the fractions shown in Table I. Table I. Distillation of 1-methoxy-1,1-dimethyl-2,2,2- trimethoxydisilazane. Fraction BP°C (10 mm) BP°C (760 mm) I 47-49 163-165 II 53-55 170-175 III 65-71 185-190 IV 75-78 197—200 V 82-84 205 VI 86-90 210 VII 92—95 220 VIII higher boiling oily material 2. Discussion of Analytical Data Analytical data obtained for Fraction V were compared with the values calculated for 1-methoxy-1,1-dimethyl-2,2,2- trimethoxydisilazane. These data are shown in Table II. Table II. Elemental analysis of 1-methoxy-1,1-dimethyl- 2,2,2-trimethoxydisilazane Calculated % Found % Carbon 32.10 32.23 Hydrogen 8.49 8.54 Silicon 25. 25.10 Nitrogen 6.24 6.31 Molecular Weight 225.40 230. 17 The agreement between the calculated and found per— centages of the various elements leaves little doubt that this substance is indeed the desired 1-methoxy-1,1-dimethyl— 2,2,2-trimethoxydisilazane. 3. Infrared Spectra The infrared spectra of Fractions I through VIII were obtained on pure liquid films between NaCl discs and Nujol mulls using a Unicam SP-200 Spectrophotomer. (See Figure 1, Appendix I.) All of the spectra were essentially identical and most of the peaks in the spectra could be assigned by comparison to values reported by Smith55 for some alkoxy- silicon compounds and also the infrared absorption frequen- cies of hexamethylcyclotrisilane discussed by Kriegsmann.17 The infrared absorption frequencies, their relative intensities and tentative assignments are given in Table III. This infrared spectrum is entirely consistent with what might be expected for a molecule such as 1-methoxy-1,1- dimethyl-2,2,2-trimethoxydisilazane. All of the important frequencies have been observed. 4. Nuclear Magnetic Resonance Studies of Fractions I Through VIII The proton nuclear magnetic resonance spectra of all fractions were obtained as described in the preceding section. The spectrum of Fraction I shows only 2 peaks, one of which is assigned to methyl protons at O = 0.09 downfield from TMS 18 Table III. Infrared absorption frequencies of 1,1-dimethyl- 2,2,2-trimethoxydisilazane Frequency cm-l Intensity* .Assignment 3278 m symmetric N-H stretch 2900 S symmetric C-H stretch, methyl 2800 m symmetric C-H stretch, methoxy 1460 w 1405 w antisymmetric C-H deformation 1250 vs symmetric C—H deformation characteristic of Si-CH3 1190 s SiOCH3 rock 1100 vs C-O stretch 960 vs Si-N-Si stretch 850 vs Si-O stretch 810 vs 725 w * m = medium, 5 = strong, vs = very strong, and w = weak. 19 and the other due to methoxy protons at O = 3.50. Inte— gration gave a ratio of methyl protons to methoxy protons of 2:1. Since there was only one methyl and one methoxy peak, all of the methoxy groups must find themselves in the same environment. The NMR Spectrum and boiling point of Fraction I agree with those reported for 1,2-dimethoxy- 1,1,2,2-tetramethyldisilazane. The Spectra of Fractions II, III, IV, V, and VI all contain three peaks, one of which was assigned to methyl protons at O = 0.09 downfield from TMS and the other two, due to methoxy protons, one at O = 3.50 and the second at O = 3.60 both downfield. (See Figure 1, Appendix II.) The ratio of total methoxy to methyl protons and methoxy protons at O = 3.60 to those at O = 3.50 of Factions II through VI are given in Table IV. Table IV. Comparison of NMR peaks from the distillation of 1-methoxy—1,1-dimethyl-2,2,2-trimethoxydisilazane Fraction M29. MeO at O = 3.60 Me MeO at O = 3.50 II 1.3 1.4 III 1.5 1.8 IV 1.7 2.9 V 1.99 2.96 VI 2.1 3.8 The methoxy peak found at O = 3.60 and total methoxy becomes progressively larger in going from Fraction II to VI. 20 This illustrates that while there is only one type of methyl group present, structurally, there are two distinct types of methoxy groups. The Spectrum of Fraction VII showed one peak which was assigned to the methoxy protons at O = 3.60 downfield from TMS. This indicates that all of the methoxy groups are in the same chemical environment. The spectrum of Fraction VIII showed eight peaks. The methyl peak at O = 0.09 downfield from TMS is split into four peaks, while the methoxy peak at O - 3.50 is split into two peaks andthe methoxy peak at O‘= 3.60 is split into two peaks. This indicates that the methyl and methoxy groups both find themselves in four different chemical en- vironments. From the NMR data as well as the results of the ele- mental analysis and the infrared spectra, the following structures can be written for Fractions I through VIII: CH3 H CH3 I I I I CH3O - Si - N - Si - OCH3 I I CH3 CH3 II, III, and IV azeotropic mixtures of CH3 H CH3 I I I CH30 "' Si - N " Si " OCH3 I I CH3 CH3 and CH3 H OCH3 I I I CH30 - Si - N - Si - OCH3 I I CH3 OCH3 21 CH3 H OCH3 I I I v CH3O — Si - N - Si - OCH3 I I CH3 OCH3 VI Azeotropic mixture CH3 H OCH3 I I I I I CH3 OCH3 and OCH3 H OCH3 I I CH3O - Si - N - Si - OCH3 I I OCH3 OCH3 OCH3 H OCH3 I I VII CH3O — Si - N - Si - OCH3 I I OCH3 OCH3 VIII Low molecular weight mixed silazane poly- mers. 5. Summary A small amount of a substance collected at 47°C was shown by NMR to be the symmetrical compound, 1,2-dimethoxy- 1,1,2,2-tetramethyldisilazane. A small amount of another substance, boiling at 92—95°C, was identified as hexamethoxy- disilazane. Although azeotropic mixtures of the various disilazanes were obtained, it was possible to isolate a fraction which boiled at 82-84°C (b.p. = 205°C at 740 mm) in 35-40% yield. This compound will henceforth be designated as compound (I). 22 PREPARATION OF 1-METHOXY-1,1-DIMETHYL-2,2,2-TRIETHOXYDISILAZANE 1. Preparation and Purification The method of preparation of the 1-methoxy-1,1-dimethyl- 2,2,2-triethoxydisilazane was analagous to that used in the preparation of 1-methoxy-1,1-dimethyl-2,2,2-trimethoxydi- silazane. Two moles of dimethyldichlorosilane was added to one liter of dry n-hexane in a 3-liter, 3-necked, round-bottom flask equipped with a mechanical stirrer, reflux condenser and dropping funnel. Similarly, two moles of silicon tetrachloride was added to one liter of dry n-hexane. Two moles of methanol was added dropwise to the flask contain— ing the dimethyldichlorosilane and six moles of absolute ethanol was added to the flask containing the silicon tetra- chloride. Hydrogen chloride evolution began immediately. After all of the methanol and ethanol had been added, the reaction mixtures were refluxed for 1.5 hours and in each case the reaction vessels were flushed with dry nitrogen to remove traces of hydrogen Chloride. The methoxydimethyl- chlorosilane and triethoxychlorosilane which were pre- sumably the products of the reactions were not isolated. Rather, ammonia gas was added directly to the combined hexane solutions described above and ammonium chloride precipita- tion was observed immediately. This reaction, like the others, was also found to be exothermic. The hexane solution 23 was separated from the NH4C1 by suction filtration. Am- monia gas was again added to the hexane solution described above and the resulting mixture filtered by suction to remove the NH4Cl. This process was repeated until an excess of ammonia was present indicating that reaction was complete. In each case the ammonium chloride was washed with hexane, and the hexane fractions were combined. The hexane was distilled out of the mixture, and when the tem- perature had risen to 95°C it appeared as though all of the hexane had been removed. The liquid boiling above 95°C at atmospheric pressure corresponded to a 75% yield of the crude mixed disilazanes. This liquid was fraction- ally distilled at 10 mm pressure, using a 15 inch, insulated column packed with glass helices and yielded the fractions shown in Table V. Table V. Distillation of 1-methoxy-1,1-dimethyl-2,2,2- triethoxydisilazane Fraction BP°C (10 mm) BP°C(760 mm) I 47-49 163-165 II 67-71 190 III 73-75 195-197 IV 80-86 205-210 V 90-92 215 VI 107-110 245-250 VII 112-150 255-290 24 2. Infrared Spectra The infrared spectra of Fractions I through VI were obtained on pure liquid films between NaCl discs and Nujol mulls using a Unicam SP-200 SpectrOphotometer. All of the spectra were identical to the spectra of the 1,1-dimethyl- 2,2,2-trimethoxydisi1azane discussed in the previous sec- tion. This is consistent since the 1-methoxy-1,1-dimethyl- 2,2,2—trimethoxydisilazane and 1-methoxy-1,1-dimethyl- 2,2,2-triethoxydisilazane molecules are so similar. 3. Nuclear Magnetic Resonance Studies of Fractions I, V, and VI The proton nuclear magnetic resonance spectra were obtained as described in the previous section. The spectrum of Fraction I was identical to the spec- trum of Fraction I from the preparation of 1-methoxy-1,1- dimethyl-Z,2,2-trimethoxydisilazane. There were two peaks present, one of which is assigned to methyl protons at O = 0.09 downfield from TMS and the other due to methoxy protons at O = 3.50. Integration gave a ratio of methyl protons to methoxy protons of 2:1. The Spectra of Fractions V and VI showed nine peaks. (See Figure 2, Appendix II.) There was a Singlet at O = 0.9 downfield from TMS, assigned to methyl on silicon; a trip- let at O = 1.24 downfield, assigned to methyl protons in the ethoxy groups; a singlet at O = 3.50 downfield, due to methoxy on Silicon; and a quartet at O = 3.90 downfield, 25 attributed to methylene protons in the ethoxy groups. The ratios of methyl to ethoxy to methoxy protons in Fractions V and VI were 2:5:1 and 10:42:1, respectively. For Frac- tion V, this corresponds to a ratio of methyl to ethoxy to methoxy groups in the molecules of 2:3:1 and for Fraction VI a ratio of 10:25:1. The spectra further indicated that that both of the methyl groups as well as all three of the ethoxy groups were in the same chemical environment. 4. Vapor Phase Chromatography» The gas chromatographic analysis was done using an Aerograph A-90-P chromatograph equipped with a 20% SE 30, 60/80 chromosorb wax column. Fractions I and V showed one major peak. This indi- cates that these fractions were pure compounds. The re- tention times of these fractions were identical to the retention times of sym-dimethoxytetramethyldisilazane and 1-methoxy-1,1-dimethyl-2,2,2-triethoxydisilazane, respec- tively. Fractions II, III, IV and VI showed two major peaks. The retention time of the first major peakfor Fractions II, III, and IV was the same as the retention time of Fraction I. The retention time of the second major peak for Fractions II, III, and IV, and the first major peak for Fraction VI was identical to the retention time of Fraction V. The retention time of the second major peak for Fraction VI was the same as the retention time of hexaethoxydisilazane. 26 From the gas chromatographic analysis as well as the results from the NMR data, the nitrogen analysis and the infrared spectra, the following structures can be written for Fractions I through VI: CH3 H CH3 I I I I CH30 "' Si -' N " Si - OCH3 I I CH3 CH3 II, III, and IV azeotropic mixtures of CH3 H CH3 I I I CH3O - Si — N - Si - OCH3 I I CH3 CH3 CH3 H OC2H5 CH30 - Si - N — Si - OC2H5 CH3 OC2H5 CH3 H OC2H5 V CH3O - Si " N - Si - OC2H5 CH3 OC2H5 VI Azeotropic mixture of CH3 H I I CH3O - Si - N - Si - (OC2H5)3 I CH3 H (C2H50)3 ‘ Si ‘ N ' Si ‘ (0C2H5)3 5. Summary Fractional distillation of a mixture of crude disila- zanes from the preparation of 1-methoxy-1,1-dimethyl-2,2,2- triethoxydisilazane yielded a small amount of 1,2-di- methoxy-1,1,2,2-tetramethyldisilazane. A fraction was isolated which boiled at 90-92°C (b.p. 215 at 760 mm) in 27 42% yield, compound V, 1-methoxy-1,1-dimethyl-2,2,2-tri- ethoxydisilazane. This compound will be referred to as Compound (II). Azeotropic mixtures of the various disila- zanes were also obtained, including a mixture which con- tained hexaethoxydisilazane. SILICON ANALYSIS Porcelain crucibles were fired in a muffle furnace at 700°C to constant weight and stored in a desiccator over Mg(ClO4)2. Weighed samples (0.2 - 0.3 g) of the disilazane were added to the crucibles which were cooled in dry ice. Concentrated sulfuric acid (3-5 cc) was then slowly added to the samples. The crucibles were removed and allowed to warm to room temperature. They were placed in a muffle furnace and the temperature slowly raised to 900°C over a period of 12-16 hours. The crucibles and contents were re- moved from the furnace and stored over Mg(ClO4)2 until cool and reweighed to obtain the weight of Si02 formed. The per cent silicon was calculated using the expression 0.467 x wt SiOz sample weight X 100 ' %Si= NITROGEN ANALYSIS A weighed sample of disilazane (0.2—0.3g) was dis— solved in aqueous ethanol and several drOps of bromocresol purple added. Excess standard hydrochloric acid was added and back titrated with standard sodium hydroxide to the 28 indicator color of a blank that was run previously. The per cent nitrogen was calculated using the expression %N = Volume of standard acid x N of acid x 14 Sample weight X 100' PREPARATION OF ALKOXY SILANES The alkoxy silanes were prepared according to the method of Peppard, Brown, and Johnson.56 To avoid repeti— tion of experimental details, the data pertaining to the alkoxy silanes are presented in tabular form and, except as otherwise noted, the preparation of dimethoxydiethoxy- silane described below may be considered as typical. The reaction was carried out in a 500-ml, three-necked flask fitted with a mechanical stirrer, dropping funnel, and gas exit tube, all openings being protected by calcium chloride tubes. To 0.5 mole of silicon tetrachloride con- tained in the flask was added, with cooling and over a period of two and one—half hours, one mole of ethanol and one mole of methanol. The mixture was then transferred to the boiling flask of an all-glass fractionating column and heated rapidly to free it of dissolved hydrogen Chloride. Fractional dis- tillation of the remaining liquid yielded the following: Si(OMe)4 121o EtOSi(0Me)3 133-1350 ( EtO)ZSi(OMe)2 143-146O ( EtO)3SiOMe 155-157O Si(OEt)4 166.1o . 29 Physical constants of alkoxy silanes in order of in- creasing molecular weight are presented in the following table: Table VI. Physical constants of some alkoxy silanes Formula Name B.P.(lit) Ref. (CH3)ZSi(OCH3)2 dimethyldimethoxy- silane 82.20 - 57 (CH3)ZSiOCH3(OC2H5) dimethylmethoxy- ethoxysilane ---- (CH3)ZSi(OC2H5)2 dimethyldiethoxy- silane 113.8O — 57 Si(OCH3)4 tetramethoxysilane 121o - 58 (CH3)20CH35i(O-37C3H7) dimethylmethoxyiso- propoxysilane ---- (CH3O)3Si(OC2H5) trimethoxyethoxy- silane 133-135o - 59 (CH3O)ZSi(OC2H5)2 dimethoxydiethoxy- silane ' 143-146o - 59 (CH30)3Si(O-ifC3H7) trimethoxyiso- \ propoxysilane ---- (CH30)Si(OC2H5)3 methoxytriethoxy- silane 155-1570 - 60 Si(C2H50)4 tetraethoxysilane -—-- (C2H5O)3Si(O—ifC3H7) triethoxyisoprop- oxysilane ---- (CH30)(C2H5O)Si(O-i7C3H7)2 methoxyethoxydi- . isoprOpoxysilane ---- PYROLYSIS OF l-METHOXY-l,1-DIMETHYL-2,2,2-TRIMETHOXYDISILAZANE In an investigation of the properties of some alkoxy- silazanes, Buerger and Wannagat60 found that pyrolysis of 30 hexamethoxydisilazane at 200°C in a sealed tube led to the formation of cross-linked polymers, accompanied by the evo- lution of tetramethoxysilane. 1. Pyrolysis of the Disilazane in a Sealed Tube In order to study the pyrolysis of 1-methoxy-1,1- dimethyl-Z,2,2-trimethoxydisilazane, 20 ml of the disila- zane was heated at ZOO-225°C for 24 hours in a sealed Pyrex tube. The tube was opened after it was cooled in a dry ice bath. No attempt was made to isolate all of the pure components of the pyrolysis mixture. Rather, the re- tention times on a vapor phase chromatographic column were found for previously prepared pure samples of all of the possible alkoxysilanes and known mixtures of them. (See Figure 1, Appendix III.) The constituents of the pyrolysis mixtures were then identified by comparison of their re- tention times with the retention times of the pure compounds. From the areas under the peaks of the vapor phase chromato- grams of the mixtures (See Figure 2, Appendix II.), the approximate mole ratios of the constituents were estimated. All of the peaks could be assigned without question. In Table VII the alkoxysilanes found, are listed in order of increasing retention times on the column. The major por— tion of the mixture was unchanged starting material. From Table VII, it is seen that a small amount of the sample de- composed with the formation of dimethyldimethoxysilane and tetramethoxySilane. 31 Table VII. Pyrolysis of compound (I) at 200°C. Volatile Products Approx. Mole Ratios (CH3)2Si(OCH3)2 1 (CH3O)4Si 3 H -Si(CH3)20CH3- ‘Si (OCH3 )3 (CH3)20CH3-Si-N-Si(OCH3)3 QMH 2. Pyrolysis of the Disilazane at Atmospheric Pressure A sample of 1-methoxy-1,1-dimethyl—2,2,2-trimethoxy- disilazane was refluxed under nitrogen for 24 hours in a 100 ml. round-bottom flask equipped with a water-cooled reflux condenser and a calcium chloride drying tube. The mixture was cooled to room temperature and then analyzed by VPC as described in the previous section. A small amount of dimethyldimethoxysilane and tetramethoxysilane formed; however, the major portion of the starting material was re- covered unchanged. Again, the ratio of dimethyldimethoxy- silane to tetramethoxysilane was 1:3. 3. Pyrolysis of the Disilazane in the Presence of Sodium Methoxide The following series of experiments were performed to test the catalytic effect of various nonaqueous bases on the polymerization of 1-methoxy-1,1-dimethyl2,2,2-tri- methoxydisilazane. These experiments were all carried out in the absence of a solvent. The disilazane and the 32 catalyst, under examination, were mixed and the system heated to the reflux temperature of the disilazane. a. Experimental Procedure: In a typical experiment 0.110 mole of the disilazane was mixed with 1.85 x 10"-2 mole of sodium methoxide in the absence of a solvent and refluxed for 24 hours in a 100 ml., round-bottom flask equipped with a water-cooled reflux condenser. The con— denser was connected through two Leiden gas washing bottles to a reservoir of 50 ml. of 0.1160 N hydrochloric acid which took up any ammonia generated during the pyrolysis. Ammonia was given off during the heating process, and after 24 hours, the reaction mixture consisted of a clear, slightly yellow liquid. Upon cooling and standing for several hours, a crystalline solid separated from the liquid. The liquid fraction was decanted from these crystals and the two un— knowns were studied separately. Back titration of the hydrochloric acid with 0.201 N sodium hydroxide was used to determine the amount of ammonia produced. Table VIII shows the results obtained in two typical experiments. Ammonia in a 1:1 mole ratio to the quantity of sodium meth- oxide was formed regardless of the amount of disilazane. The NMR spectrum of the solid was identical to that of the disilazane except that the methyl protons were shifted 0.24 ppm up field. The solid was known to be the sodium compound of 1-methoxy-1,1—dimethyl-2,2,2-trimethoxy— disilazane from the results of a similar experiment by 33 Weibrecht and Rochow.54 Therefore, the solid was not further analyzed. Table VIII. Stoichiometry of the pyrolysis of Compound (I) with sodium methoxide I II Moles of disilazane 1.11 x 10"1 1.16 x 10"1 Moles of sodium methoxide 8.43 x 10.3 5.08 x 10-3 Moles of $01dum methox1de 7.59 x 10-2 4.379 x 10-2 Mole of diSilazane Moles of NH3 liberated -3 -3 during pyrolysis 8.02 x 10 4.97 x 10 MOleS Of ammonia 9.55 x 10‘1 9.78 x 10‘1 Moles of sodium methoxide The liquid fraction was analyzed by VPC. The volatile con- stituents of the liquid fraction were identified by com- parison of their retention times on a vapor phase chromato— graphic column with retention times of the pure compounds. (Figure 3, Appendix III.) Table IX lists the alkoxysilanes found in order of increasing retention times on the column. Table IX. Alkoxy silanes produced from the pyrolysis of Compound (I) with sodium methoxide W Metal Alkoxide Volatile Products Approximate Mole Ratios NaOCH3 (CH3)28i(OCH3)2 1 (CH30)4Si H 3 (CH3O)(CH3)ZSi-N-Si(CH3)2(OCH3) 2 -Si(CH3)2(OCH3) -Si(0CH3)3 1/3 34 The liquid fraction was finally distilled at 10 mm pressure until all of the volatile material had been re- moved and a polymeric material was left behind as a residue. This solid was found to be insoluble in water, dilute hydrochloric acid, dilute sodium hydroxide, methanol, ben- zene, hexane and carbon tetrachloride. It was decomposed slowly by hot, concentrated sulfuric acid. Nitrogen and silicon analyses have been obtained. However, the apparent formation of silicon carbide under the conditions usually used to determine per cent carbon and hydrogen led to non- reproducible carbon and hydrogen analyses. The presence of an infrared band corresponding to N—H stretching is taken as evidence that the silicon-nitrogen chains are not completely cross-linked through tertiary nitrogen. The analytical data are given in Table X. Table X. Elemental analysis of the polymer from the pyrolysis of compound (I) with sodium methoxide m % Nitrogen Calculated 100% linear 100% crosslinked 17.75 12.76 Found Percentages % Carbon 16.60, 16.42 % Hydrogen 4.66, 4.74 % Nitrogen 13.93, 13.98 % Silicon 36.48 These data show that the polymer is quite highly crosslinked. 35 b. Nitrogen Analysis: A weighed sample was decomposed under concentrated sulfuric acid in a Kjeldahl flask and gently heated for 2 hours. The polymer decomposed with SiOz as one of the products. An excess of concentrated sodium hydroxide was then added to the Kjeldahl flask and the distillate was collected in 50 ml. of standard hydro— chloric acid to which several drops of methyl red had been added. A blank was run and the sample was then titrated with standard sodium hydroxide to the indicator color of the blank. 4. Pyronsis of the Disilazane in the Presence of Aluminum Isopropoxide a. Experimental Procedure: 0.114 mole of the di- silazane and 3.9 x 10.3 moles of aluminum isopropoxide were mixed in the absence of a solvent and refluxed for 24 hours in exactly the same manner employed in the previous polymer- ization studies. The reaction mixture was a rosy orange color and a solid remained on the side of the flask. After cooling, the liquid fraction: was decanted from the solid and the two studied separately. Back titration of the standard hydrochloric acid with standard sodium hydroxide showed 2.97 x 10"3 moles of ammonia were liberated. This is approximately a 1:1 mole ratio of the ammonia liberated to the quantity of aluminum isopropoxide. The solid was washed with hexane to remove impurities. It was soluble in hydrochloric acid but insoluble in hexane, 36 carbon tetrachloride, benzene and tetrahydrofuran. A quali- tative test with the aluminon reagent indicated the presence of aluminum. The NMR of the solid was not taken since the solid was insoluble in all organic solvents attempted. How- ever, the infrared spectrum of the solid was identical to the starting material including a peak corresponding to the N-H stretching frequency. The amount of aluminum was deter- mined by the 8-hydroxyquinoline precipitation method. The per cent aluminum found was 3.09. If the aluminum salt ' Si(OCH3)3 A has the formula Al(N<: . )3, the calculated per . Sl(CH3)2-OCH3 cent aluminum would be 3.682. If the aluminum salt has the B . H3C\I;I /Sl(OCH3)3 formula ( C-O)2 Al(N\ H3C/’ Si(CH3)20CH3 , the per cent aluminum would be 7.32. The per cent aluminum found is low when compared with the calculated per cent aluminum for both possible structures. The found percentage is closer to the first proposed structure, but this possibility may be ruled out on steric factors alone. The agreement be- tween the found and calculated percentage for the second structure is clearer when it is obServed that the calculated percentage is approximately twice as great as the found per cent aluminum and the infrared spectrum shows a band corresponding to N-H stretching. Thus, a mixture of poly- silazanes similar to structure two with only one-half of the hydrogens on nitrogen replaced by aluminum may be proposed. 37 The liquid fraction was analyzed by VPC (See Figure 4, Appendix III) in order to identify the volatile constitu- ents. Table XI lists the alkoxysilanes found in order of increasing retention time on the column. Table XI. Alkoxy silanes produced from the pyrolysis of Compound (I) with aluminum isoprOpoxide v“— Volatile Products Approx. Mole Ratios (CH3)28i(OCH3)2 10 (CH3O)4Si 10 (CH3)(CH30)-Si—(O-ifC3H7) 2 (CH30)3‘Si‘(O'lfC3H7) 1 HH‘ -Si(CH )2(OCH3) = "Si‘OCH3 53 Finally, in an attempt to isolate a polymeric material, the liquid fraction was distilled at 10 mm pressure until all of the volatile material was removed. If the temper- ature is not controlled, a polymeric solid similar in prOp- erties to the one from the reaction of sodium methoxide and Compound (I) is obtained. Also it appeared as though sili- con carbide was formed when this polymer was analyzed for carbon and hydrogen. However, if the liquid is carefully distilled at a low temperature and at 10 mm pressure, a mixture ofcfligomeric oils is obtained. The oligomeric oils were soluble in a variety of common organic solvents. The NMR spectrum (See Figure 3, Appendix II) of the oils, taken in carbon tetrachloride as solvent, showed a methyl to methoxy ratio of 1 to 1.9, essentially the same as that 38 for the starting material. However, three structurally dif- ferent kinds of methyl (predominately one) and five struc— turally different methoxy groups (predominately two) were identified by the spectrum. The methyl to methoxy ratio in the oligomeric oils is consistent with the ratio of methoxydimethylsilyl to trimethoxysilyl groups shown to have been cleaved by the composition of the mixture of volatile pyrolysis products. The analytical data for the two dif- ferent polymers are given in {Table XII. Table XII. Elemental analysis of the polymers from the pyrolysis of Compound (I) with aluminum iso- propoxide W % Nitrogen Calculated 100% linear 100% crosslinked 10.52 11.63 Found Percentages Solid Polymer Oligomeric Oils % Carbon 12.4, 12.21 27.06, 26.96 % Hydrogen 4.11, 3.84 6.98, 7.19 % Nitrogen 11.48 9.15, 9.09 % Silicon 35.29 28.8 , 28.72 The infrared spectrum of the oligomeric oil indicated that essentially most of the N-H bond remained intact dur— ing the pyrolysis. Therefore, it appears on this evidence as though this oil is a linear organopolysilazane. The solid polymer is highly crosslinked. b. 8-Hydroxyquinoline Method for the Determination of Aluminum: Weighed samples (0.2-0.32 g.) of the sample 39 were added to 100 ml of distilled water. This solution was made acidic with concentrated hydrochloric acid and warmed to 60-70°C. The insoluble material was separated from the supernatant liquid by suction filtration. A slight excess of a 5 per cent 8-hydroxyquinoline solution in 2M acetic acid was added to the liquid allowing 1 ml of reagent for each 3 mg. of aluminum present. Slowly, 2M ammonium acetate solution was added until a precipitate formed and then an additional 25 ml was added to increase the buffer- ing action. The supernatant liquid was yellow from the ex- cess reagent. The solution was cooled for 30 minutes and then the precipitate was collected in a sintered glass crucible, previously dried to constant weight in the oven. The precipitate was washed with cold water until the wash- ings were colorless. The crucible and precipitate were dried for an hour at 120-140°C, cooled in a desiccator and weighed. The dried precipitate has the formula Al(C9H60N)3. The per cent aluminum was calculated using the expression %A1 = 27 x wt A1(C9H60N)l , 100 459 x sample weight PYROLYS IS OF l-METHOXY-l , 1-DIMETHYL-2 , 2 , Z-TRIETHOXYDISILAZANE The polymerization studies of this disilazane were run parallel to the studies on 1-methoxy-1,1-dimethyl-2,2,2- trimethoxydisilazane; therefore, experimental details will be eliminated. 40 1. Pyrolysis of the Disilazane in a Sealed Tube A sealed tube containing 20 ml of the disilazane was heated in an oven at ZOO-225°C for 24 hours. A clear yel— low liquid remained in the tube. The liquid was shown by VPC (Figure 5, Appendix III) to contain the following con- stituents. Table XIII. Alkoxy silanes produced from the pyrolysis of Compound (II) at 200°C Volatile Products Approx. Mole Ratio (CH3)2Si(OCH3)2 1° (CH3)2Si(0CH3)(0C2H5) 40 (CH3)2Si(OC2H5)2 3° 51(0CH3)2(0C2H5)2 5 Si(OCH3)(OC2H5)3 48 Si(OC2H5)4 98 Si(0Et)3 = 1.9 Si(CH;)éOCH3 1 From Table XIII, it is seen that the sample decomposed with the formation of alkoxysilanes. However, the major portion of the starting material was recovered unchanged. Distillation of the pyrolysis mixture to dryness without the formation of nonvolatile materials indicated that a solid polymer was not formed. 2. Pyrolysis of the Disilazane in the Presence of Sodium Methoxide Eleven one-hundredths of a mole of disilazane and 41 1.85 x 10-2 mole of sodium methoxide were refluxed for 24 hours in the absence of a solvent. The ammonia liber- ated during the reaction was trapped in a reservoir of standard hydrochloric acid. After 24 hours, the pyrolysis mixture consisted of a brownish-yellow liquid with a crys- talline solid in the bottom of the flask. After cooling, the liquid and solid were separated. Back titration of the hydrochloric acid with standard sodium hydroxide indicated 1.91 x 10_2 mole of ammonia was liberated. This is a 1:1 mole ratio of ammonia to the quantity of sodium methoxide. The solid was washed with hexane and identified on the basis of its sensitivity to moisture, proton NMR spectrum, which was identical to that of the pure starting material except for an up-field shift of the methyl proton resonances, and its IR spectrum. This solid is the sodium salt of 1—methoxy- 1,1-dimethyl-2,2,2-triethoxydisilazane. The volatile constituents of the liquid fraction were characterized by VPC (Figure 6, Appendix III). Table XIV lists the alkoxysilanes found in order of increasing re- tention times on the column. A polymeric solid was also isolated from this mixture which had all the properties of the polymer from the re- action of sodium methoxide with Compound I. The analytical data are given in Table XV. These data and the fact that the infrared spectrum of this polymer showed a band corresponding to N-H stretching indicate that while the polymer is crosslinked, it is not completely crosslinked through tertiary nitrogen. 42 Table XIV. Alkoxy silanes produced from the pyrolysis of Compound (II) with sodium methoxide —t W Volatile Products Approx. Mole Ratios (CH3)ZSi(OCH3)2 6 (CH3 )2 (CH3O)Si(OC2H5) 18 (CH3 )33i(0C3H3)2 24 (CH3O)4Si 6 (CH3o )3Si(oc3H3) 1 (CH30)251(0C2H5)2 5 (CH o)Si(oC3H5)3 12 ((3szO )4Si I 1° -SiLEH3)2(OCHi) -Si(OC2H5 )3 Mn» Table XV. Elemental analysis of the polymer from the pyroly- sis of Compound (II) with sodium methoxide %Nitrogen Calculated 100% linear 100% crosslinked 14.8 10.6 Found Percentages* % Carbon 14.74, 14.40 % Hydrogen 4.14, 3.93 % Nitrogen 14.42, 13.9 % Silicon 35.85 *Carbon and hydrogen results nonreproducible due to the possible formation of SiC. 43 3. Pyrolysis of the Disilazane in the Presence of Aluminum Isopropoxide A mixture of 0.092 mole of the disilazane and 1.651 x 10-3 mole of aluminum isoprOpoxide was refluxed for 24 hours. The ammonia liberated, 1.49 x 10”3 moles, was in a 1:1 mole ratio to the quantity of aluminum isopropoxide. The py- rolysis mixture consisted of a solid and a liquid. The solid was similar in all respects to the solid obtained from the reaction of aluminum isopropoxide with 1-methoxy— 1,1-dimethyl-2,2,2-trimethoxydisilazane. The solid was insoluble in all attempted solvents and the aluminon test was positive. The infrared spectrum showed a band cor- responding to the N-H stretching frequency and the per cent aluminum found was 3.27. The calculated percentage for the following structure is 6.569: CH3 - /,Si(OC2H5)3 /CO A]. N\ CH3 2 31(0CH3)(CH3)2 The calculated percentage is approximately twice as great as the found percentage. Since the infrared spectrum con- tains a band corresponding to N-H stretching, it is pro- posed that the solid is composed of a mixture of polysila- zanes similar to structure B with one-half of the hydro- gens on nitrogen replaced by aluminum. The volatile con- stituents of the liquid fraction were analyzed by VPC (See Figure 7, Appendix III) and in Table XVI the alkoxysilanes, found, are listed in order of increasing retention times on the column. 44 Table XVI. Alkoxy silanes produced from the pyrolysis of Compound (II) with aluminum isopropoxide Volatile Products Approx. Mole Ratios ( CH )3Si(0CH3)3 4 (C CH3 )2(0CH3)Si(0C2H5) 20 (CH3)2Si(OC2H5)2 20 (CH3O)4Si 4 (CH3 o)3Si(0C3H5)3 1 (CH3o)Si(0C3H3)3 (C3H5 O)4Si 12 (C3H5 o)3Si(o-i _-C3H7) 12 (CH3O )(C 2H50)Si(0‘ifC3H7)2 20 Starting material 40 (CH3o)Si(CH3)3-[NHSi(CH3)3]3NHSi(0CH3)(CH3)3 40 -Si(CH3)3 (0CH3) "SiICC2H5)3 AHA In addition to the volatile pyrolysis products from the liquid fraction, high molecular weight silicon-nitrogen com- pounds were formed. When the liquid was distilled at 10 mm pressure and at a high temperature, an unreactive solid was formed which was insoluble in acid, base, and organic sol- vents. Table XVII gives the analytical data found for this polymer. The presence of an infrared band corresponding to N-H stretching is taken as evidence that the Silicon nitrogen chains are not completely crosslinked through tertiary nitrogen. When the liquid was distilled at 10 mm pressure 45 Table XVII. Elemental analysis of the crosslinked polymer from the pyrolysis of Compound (II) with aluminum isoprOpoxide Found Percentages* % Carbon 15.85 % Hydrogen 4.85 % Silicon 38.15 *Carbon and hydrogen analyses are not fully reliable. Table XVIII. Elemental analysis of the oligomeric oil from the pyrolysis of Compound (II) with aluminum isopropoxide Found Percentages % Carbon 29.41, 29.28 % Hydrogen 7.21, 7.26 % Nitrogen 10.77, 10.69 % Silicon 29.20, 29.34 46 and with a small amount of heat a mixture of oligomeric oils was obtained. The NMR data (See Figure 4, Appendix II) obtained for the oligomeric oils showed there were more than two structurally different methyl groups as indicated by three peaks, two Sharp and one broad. The ethoxy and methoxy proton resonances were also broad. A methyl to ethoxy to methoxy ratio of 2:3:1 was found. These ratios are also consistent with the vapor phase chromatographic analysis of the mixture of alkoxysilanes formed during the pyrolysis. The infrared spectra of these oligomeric oils indicate that essentially most of the N-H bonds remained intact during the pyrolysis. The elemental analysis data for this polymer are given in Table XVIII. It appears on this evidence as though this oil is a linear organopolysilazane. III. SUMMARY Prolonged heating of Compounds (I) and (II) at 200°C results in slight decomposition to the corresponding alkoxy silanes with most of the starting material remaining un- changed. Hexamethyldisilazane and 1,2-dimethoxy-1,1,2,2- tetramethyldisilazane have been shown to be thermally stable,54 while hexamethoxydisilazane when heated to 200°C, decomposes almost completely.6° Sodium methoxide and aluminum isoprOpOXide cleaved the silicon-nitrogen bond in Compound (I) with the formation of predominantly dimethyldimethoxy and tetramethoxy silane, the corresponding sodium or aluminum salt of the disilazane and high molecular weight polymers. For sodium methoxide, the number of trimethoxysilyl groups lost was three times that for methoxydimethylsilyl. For aluminum isopropoxide, there was an equal number of both groups lost. Since the cleavage is thought to involve a nucleophilic attack by the alkoxide ion on the electrophilic silicon atom and also, due to the electron withdrawing power of oxygen atoms bonded to silicon, it is reasonable on electrostatic grounds to expect the trimethoxysilyl group to be the predominant Cleavage product. Therefore, one would predict that the susceptibility of a particular silicon atom to nucleophilic attack by alkoxide should increase as the number of alkoxy groups bonded to that silicon increases. However, this 47 48 reasoning is based on electrostatic reasoning, alone, and does not take into account steric hindrance due to the groups bonded to silicon or the relative nucleophilic strengths of the attacking alkoxide ions. Steric hindrance appears to be least important and electrostatic factors predominate for attack by sodium methoxide on Compound (I), whereas, steric effects to a significant extent determined the site of attack by aluminum isopropoxide. The weaker base, aluminum isopropoxide, in the presence of the di— silazane and under mild conditions gave a somewhat linear polymer, but only solid polymers, crosslinked through nitro- gen, could be isolated with sodium methoxide. Compound (II) contains the slightly bulkier ethoxy groups bonded to silicon. The polymers and salts formed from the cleavage reaction with sodium methoxide and aluminum isopropoxide were similar to those formed from the cleav- age of Compound (I). However, a larger number of alkoxy- silanes were produced from the cleavage of Compound (II). This is due to the fact that this disilazane contains two different kinds of alkoxy groups and the unequivocally established fact that alkoxy groups bonded to silicon ex- change readily.61 Therefore, the final product is a com- plicated mixture of all of the possible rearranged alkoxy— silanes. Steric factors seemed to have been important in both cases involving the attack of the alkoxy group on the more hindered silicon atom. 49 The Cleavage of the dimethylmethoxysilicon-nitrogen bond by the more basic methoxide ion was 3/2 as great as the Cleavage of the triethoxysilicon-nitrogen bond. The methoxide ion preferred to attack the less hindered silicon atom. The more discriminating aluminum isopropoxide showed a slightly greater preference for attack at the more electrOphilic but hindered silicon atom and the ratio of dimethylmethoxysilyl groups cleaved to triethoxysilyl groups is 3:4. The difference in nucleophilicity between sodium methoxide and aluminum isopropoxide probably accounts for the difference in the degree of condensation observed for these two bases. The solid polymers were insoluble in all attempted solvents and gave unreliable carbon and hydrogen analyses due to the formation of silicon carbide. This evidence indicates that these solids are highly crosslinked poly- mers of undetermined structure. Probable structures were determined for the liquid polymers on the basis of their elemental analysis, physical prOperties, NMR and IR data. Structures involving a linear silicon-nitrogen framework along with the retention of at least a fraction of the N-H bonds contain too high a per- centage of nitrogen to be in agreement with analysis. The per cent nitrogen calculated for the completely cross- linked case is also higher than the found value. Similarly, linear polymers with completely silylated nitrogen must also be rejected. These lead to theoretical nitrogen 50 percentages which are too low. Therefore, these extremes must be rejected as probable structures unless the oligomeric oils are considered to be a mixture. The simplest formula calculated from the elemental analyses for Compound (I) is (Si3C7H205N2). IR data in- dicated the presence of a N—H bond and NMR data show that the methyl to methoxy ratio is 1:2. The following structure in Which one-half of the total nitrogen is completely substituted is proposed as a probable structure: 0CH3 0CH3 H '7 I I I Si N —— Si N I I I , 0CH3 SiOCH3 0CH3 J_ (CH3 )2 J n It is assumed that the methyl and methoxy groups can be interchanged along the chain. The empirical formula for the polymer from Compound (II) is (Si4C10H3705N3). The IR data show that the nitrogen is not completely substituted and the ethoxy to methyl to methoxy ratio was shown to be 3:2:1 from NMR data. These data and the large body of evidence that silazanes tend to form 6 and 8 membered rings make plausible a mix— ture of the average structures shown on page 51. The results obtained in the alkoxide cleavage experi- ments suggest that the much touted d7 - p7 interaction be— tween silicon and nitrogen, while obviously important in some systems, Should not be considered as the silicon- nitrogen Chemist's panacea. For unsymmetrically alkoxylated I «~- Si __,.__:3. (OEt)3 51 (He)3 N Si Si(Me)3 0Me H . N _—-—— (He)3 (QEt)2 H -_’ Si N -- 52 disilazanes a large d7 — pv interaction would have the net effect of weakening the silicon-nitrogen bond to the less (rather than more) highly alkoxylated silicon. The experi- mental results reported in this thesis are more in agree- ment with what would be predicted for an, at best minimal d7 - pw contribution to the bonding and can be explained in terms of the largely uncompensated electron withdrawing power of alkoxy groups bonded to silicon. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. BIBLIOGRAPHY Fessenden, R. and J. S. Fessenden, Chem. Rev., 61, 361 (1961). Ebsworth, E. A., et al. , Spectrochim. Acta, 13, 202 (1958) Kriegsmann, H. and W. Forster, Z. anorg. u. allgem. Chem., 298, 212 (1959). Robinson, D. W., J. Am. Chem. Soc., 39, 5924 (1958). Hedburg, K. 0., J. Am. Chem. Soc., 11, 6491 (1955) Goubeau, J. and J. Jiméhez—Barbera, Z. anorg. u. allgem. Chem., 303, 217 (1960). Ebsworth, E. A. and N. Sheppard, J. Inorg. Nucl. Chem., 9, 95 (1959) Burg, A. B. and E. S. Kuljian, J. Am. Chem. Soc., 12, 3103 (1950) Ebsworth, E. A. and H. J. Emeléus, J. Chem. Soc., 1958, 2150. Anon., Chem. Eng. News, No. 46, 47 (1960). Emeléus, H. J. and N. Miller, J. Chem. Soc., 1239, 819. Ebsworth, E. A., §t_g&,, J. Chem. Soc., 1258, 1453. Cerato, C. C., et al., Chem. Phys., 22, 1 (1954). Fessenden, R. and D. F. Crowe, J. Org. Chem., 2;, 598 (1960). George, P. D., et al., J. Am. Chem. Soc., 58, 786 (1936). Goldschmidth, A. and J. R. Wright, U.S. Patent 2,758,126 (1956). Kriegsmann, H., Z. Elektrochem., 61, 1088 (1957). Sharkey, A. G., et al., Anal. Chem., 22, 770 (1957). Holzman, G. R., et al., J. Chem. Phys., 25, 172 (1956). 53 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 54 Tansjo, O. L., Acta Chem. Scand., 55, 35 (1959). Birkofer, L. and A. Ritter, Ann., 612, 22(1958). Burg, A. B. and E. S. Kuljian, J. Am. Chem. Soc., .12. 3103 (1950). Stock, A. and K. Somieski, Ber., 5g, 740 (1921). Sujishi, S. and S. Witz, J. Am. Chem. Soc., 15, 4631 (1954). LTD Midland Silicones, British patent 737,229 (1955). Mjérne, 0., Svensk Kem. Tid., 52, 120 (1950). Osthoff, R. C. and S. W. Kantor, Inorg. Synthesis, 5, 55 (1957). Sauer, R. 0., J. Am. Chem. Soc.,,éfi: 1707 (1944). Sauer, R. O. and R. H. Hasek, J. Am. Chem. Soc., 55, 241 (1946). Wanno at U. and H. Niederprfim, Angew. Chem., 1;, 574 (1959? Bailey, D. L., et al., J. Am. Chem. Soc., 15, 435 (1948). Gierut, J. A., et al., J. Am. Chem. Soc., 55, 786 (1936). Andrianov, K. A., et al., Izvest. Akad. Nauk. S.S.S.R., Otdel. Khin. Nauk, 1958, 47. .Larsson, E., Kgl. Fysiograph. Sallskop. Lund, Foch., 28,1 (1958). Larson, E. and R. E. Marin, Acta Chem. Scand., 5, 1173 (1951). Rosnati, L., Gaza. Chem. ital., 78, 516 (1948). Schwarz, R. and F. Weigel, Anor. u. allegem. Chem., 268, 291 (1951). Larson, E.,Acta Chem. Scand., 5, 1084 (1954). Larson, E., Kgl. Fysiograph. Sallskop. Lund, F6ch., Inorganic Synthesis, 5, 61 (1957). Rochow, E. G., Chem. of the Silicones, Second Ed., Wiley, (1951), p. 81. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 55 Andrianov, K. A. and G. Ya. Rumba, vysokomolekul. Soedin, g, 1060 (1962). Wannagat, U., Private communication. Krfiger, K. and E. G. Rochow, Angewandte Chemie, 5, 458 (1962). Anderson, H. H., J. Am. Chem. Soc., 13, 5802,(1951). Kummer, D, ONR Technical Report, Contract No. Nonr- 1866 (13), (1963). Eaborn, C., Organosilicon Compounds, Academic Press, Inc., New York, p. 343. Rochow, E. G. and R. N. Minné, J. Am. Chem. Soc., _8_2, 5625 (1960) F. A. Hen lein and K. Lienhard, Makromal Chem., 52, 218 (1959 . Kummer, D., ONR Technical Report, Contract No. Nonr- 1866 (13), 1962). Kummer, D. and E. G. Rochow, Z. anorg. u. allgem. Chem., in press. Kummer, D. and J. D. Baldeschwieler, J. Phys. Chem., '51, 98 (1963). Liehard, K., ONR Technical Report, Contract No. Nonr—1866 (13), (1963). Weibrecht, W. E. and E. G. Rochow, J..Organometallic Chem., _5_, 520-525 (1966). Smith, A. L., Spectrochim. Acta, 55, 87 (1960). Peppard, D. B., W. G. Brown and W. C. Johnson, J. Am. Chem. Soc., 55, 73 (1946). M. G. Vorarkov, Zhur. Obshchei. Khim., 55, 907 (1959). Forbes, G. S. and H. H. Anderson, J. Am. Chem. Soc., .33, 1703 (1944). Friedel, C. and J. M. Crafts, Annales de chimie et de physique, [4] g, 5(1866). (Chem. Rev., 5;.) Wannagat, U. and Burger, Zeit. ffir anorg. u. allgem. Chemie., 319 (1963). Eaborn, C. and R. A. Shaw, J. Chem. Soc., 2027 (1954), 1420 (1955. - APPENDICES 56 APPENDIX I INFRARED SPECTRUM OF COMPOUND (I) 57 58 .QONMHHmHU Imxonumafluulm.N.NIH%£uOEHUIH.Hlmx0£uOEIH mo Esuuoomm CommumcH .H mudmflm 325:3»; 0mm 00m 000, OON— CO": 000— 000— OOON 000m 000? OOOW O I .Ifillq III 4-. 1-|I.‘J-II|I.-|4 Ili “- - . .,_ 1 1 1 1 1 O, 7 _. _.Y - 1-- -1111- l - MIC—H If ' " ‘4’ I -1, aoueugwsueu -1..II.--IlI-..~.IIII I-I-1 Om” 1 11 . om II.-- 1 . .- t . -.- It III: I.-. I. . On I- 1 1 w :10". T.IIII1 on o .. .. 1 ow . .,. L . “wig.“ _._',‘.:““‘,, 1*:1-0- 4 1 1 . ._ 1 . 1. 1 . ov 3L}- . . .7 III.- . .I . ov Om ILIfi. 1 .1 00.4 M‘ "'3'. Om I I r. I I I 1.1.. .__4_-._...__-.__ _-.- _ -_ __ I M ’Q haw—4)- ”o«».-Q — On I 4.. II-II} /-\ IIII- Om . 11.1-1.1}... - - 1 . 1 Om I1 .. - IIHII-l . 1 1 1 1 CG fIL-II-III .--l 9 . -.-I,-... .. .- M I.I-I. ..i I-.III III~II| II.” III . "-.. I..- “III. I I . -. II-I.. om -. |lIl1--I|.. III-I CG “and 4.3531..th “nagvgmunofixoofiv no BEBE REE 1 1 1 1 1 1 1 1 1 _ 1 _ 1 _ _ 1 _ I1 1 1 OO— 1:; -I_- 1 1 . _I iIP-II. 1-- I -I.l .I -e I“ II- LII- ._ Cr: II III. - l I--.-I I .I - ”INJré-ICQWF 211...}. ,. file-12m; -~._I : o. m m ..- O r- v m . _ APPENDIX II PROTON NMR SPECTRA OF DISILAZANES 59 60 .AHV UCSOQEOU mo Esnpommm mzz cououm .N muswflm 2.: 0— 0.2. 1- km 616585523188nt 1. n2...- 61 .AHHV UCSOQEOU wo fishgoomm mzz cououm .m musmflm 2.: 9 3m 9mm. £8 1 ldl d . 114%.. : 1.. 1 1 _. 1. wt... ___ .118:E§%GB:%B1 62 .AHV UGDOQEOU mo mfimhaoumm 0:0 Eoum HHO Oenwfiomflao 0:0 mo Eduuommm mzz CODOHm .w musmflm §_ "'0' nth. ohemno-«-ovd< + ”homuovammnammxouovAonmo1 EEmdofifififié 63 It 0 nth 3.. 1 q mfimhaonmm 050 Eoum HHO OHHwEomHHO may mo Irene-3-82 4 giganlgxooac gab; .AHHV UGDO moo mo ESHuommm mzz couonm .m muomflm pnm use a J 63 It 0 or... mammaonhm wnu Eonm HHO «d... b q .AHHV Canadnoo mo OHHOEovflao map No Eduuommm mzz couomm .m ousmflm nnm awn 4 a Jeane-«-82 + 2510851153833 .6..— SS uni-85° APPENDIX III GAS CHROMATOGRAMS OF ALKOXY SILANES 64 65 A.CHEV OEHB GOHDGmumm ma ¢H ma NH HH OH m w b o e 6213mm one 165$ 002 1005 «Amzvfimn 002 m0 OusuxHE 4 .mmcmHHm axoxam QBOCM 080m mo Emnmoumaouno mmw .0. mndmflm .66 A.GHZV mEflB COHucmpmm ma mH vH ma NH HH OH m m b o .031 ‘ c «AwmovemmAOum «Aumovflmflomz «AumovemNAOmz AumovemeAsz mo wusuxfle < .mmcmaflm mxoxam QBOCx 080m mo EmnmoumEoHno mmo .m wudmflm 67 A.CHSV OEHB COHucmpmm 0H mH VH NH NH HH OH m w h m (I), .mnson wN How UoOON um AHV UCDOQEOU mCHummfl Eoum musuxHE COHuHmomEOOOU 0:» mo Emumoumfiounv mmw .b muanm IIJ.)I{$\1 68 9H 0H wH NH NH .AHV CGSOQEOU mo coHuHmomEOO IOU UmmmHmumo mononumE on» Scum meMHHm mxome 0:“ mo EMHmONMEOHSO mmo A.GHEV mEHB coHpnmuom HH 0H .w musmHm m w b m m 69 A.GHZV OEHB GOHDGOumm OH m 11 m h mH mH VH NH NH HH .AHV UGSOQEOU mo COHpHmomEOOwC CwNmHmvmu mUonmoumomH ESCHEsHm 03p Eoum mOGMHHm mxome may no Emumouweounu mmw .m musmHm I I | _ ‘1‘:- 70 NH NH vH NH NH A.GHSV GEHB GOHOCODOM HH OH m w b N m v N N H (331171) 321) .muson vN Mom UoOON um AHHV UGDOQEOU mCHummn Scum wuduxHE coHuHmomEoqu mng mo Emumoumfionflo mam .OH OusmHm 71 NH NH vH NH NH A.CHSV mEHB COHucmumm HH OH m N J h . HH UCSOQEOU mo COHpHmomfiwomw UmnhHmumo mponguOE EDHUOm map Scum mmcmHHm mxome 030 m0 Emumoumfiougo mmw .HH mnsmHm )x1x1 I 72 A.CHSV OEHB GOHpcmumm NH NH VH NH NH HH OH m N b q - . ((11711 1-. . HH UGUOQEOO mo aoHuHmo EOW Imp UmumHmumo OUonmonm IomH ESCHESHM 03p Eoum r OCMHHm mxome man mo Edam IoumEOHSO mmw .NH musmHm MICHIGAN STATE UNIVERSITY LIBR I I I IIII IIIIIIII'I‘I‘“ 3 129313061 5169