! A STUDY OF THE EQUIUBRIA AND RATES OF EXCHANGE OF SECONDARY AMINES WITH SILYLAMINES Thai: for the 509m of AL 5. MICHIGAN STATE UNIVERSH’Y Charies A. Roth 1966 —‘ LIBRARY T H k: 515 ' ' M1ch1 gm S ta ta University __._.. ABSTRACT E¢UILIBRIA AND RATES OF EXCHANGE OF SECONDARY AMINES WITH SILYLAMINES by Charles A. Roth The labile exchange of electronegative groups on organosilicon compounds has long been recognized as an important difference between silicon and carbon chem- istry. One example of this is the exchange of the nitrogen moiety of an aminosilane with another amine. RBSiNRé + HNRg, :5: RasiNRg + HNR2' There are numerous examples in the literature pertaining to the use of this reaction to prepare new silylamines. Generally, a readily available silylamine containing a low boiling amino group such as -NH2, -NHCHB, etc. is refluxed with another amine to give the desired silyl- amine. The volatile amine which is displaced is driven off. In this investigation the equilibrium constants for the reactions of a number of aliphatic and aromatic amines with several dimethylaminosilicon compounds in benzene were measured. it was found that the ability of an aliphatic amine to displace the (CH3)2N- group bonded to silicon was governed by several factors. Firstly, for aliphatic amines the steric effect of the groups on nitrogen appears to be very important. secondly, primary amines are more readily exchanged on silicon than secondary amines. Thirdly, aromatic amines exchange much more completely with a dimethylamino group on silicon than do aliphatic amines. This is probably due to dw-pq bonding of the aromatic nitrogen with the unfilled orbitals of silicon, giving a lower energy state in the resulting silylamine. Electron donating groups in- crease the extent to which exchange occurs. Finally, it is interesting to note that lesubstituted anilines do not exchange. The rate of exchange was also measured for some amines with a few silylaminee. The reaction was found to be second order overall and is probably first order with respect to both amine and silylamine. Although the mechanism has not been rigorously established, a four centered intermediate of the type shown below has been used to rationalize the data. ,CH Eat-IL 3 RE SCH , . 3 N - H R Both the equilibrium measurements and the rate studies were conducted by means of proton magnetic resonance spectrosCOpy. A STUDY OF THE EQUILIBRIA AND RATES OF EXCHANGE OF SECONDARY AMINES WITH SILILAMINES By Charles A. Roth A-THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Chemistry 1966 ACANCWIEDGEMENTS The author wishes to express his appreciation to Professor W.E. Weibrecht for his constant encourage- ment and assistance throughout this work. The author wishes to thank the Dow Corning Corp- oration for their financial assistance. The author is also deeply indebted to Drs. John L. Speier and John W. Ryan of Dow Corning who for five years encouraged and instructed him in the area of organo- silicon chemistry. Finally the author wishes to express his gratitude to his wife Barbara and sons Scott and Mark for their devotion and love and for the many hours lost as a family. ii TABLE OF CLNTENTS Acknowledgements Table of Contents List of Tables List of Figures I II III IV VI VII VIII IX XI Objectives Background Discussion of Results and Conclusions A. Equilibrium Studies B. Reverse Equilibrium Reactions C. Kinetic Studies D. Summary Experimental A. Materials B. Preparation of Phenyldimethylsilyldi- methylamine ' 0. Preparation of Trimethylsilyldimethyl- amine D. Preparation of Trimethylsilylpiperidine E. Preparation of Trimethylsilyl-tggt- butylamine Analyses of Materials A. Organic Amines B. Silylamines Method of Conducting Equilibrium Experiments Treatment of Data for Kinetic Reactions Discussion of Error Kinetics of the Reaction of Trimethylsilyl- dimethylamine with tert-Butylamine in the Absence of Solvent at 58° Kinetics of the Reaction of Trimethylsilyl- dimethylamine with pert-Butylamine at 58° Kinetics cf the Reaction of Vinyldimethyl- silyldimethylamine with tert-Butylamine at 38° in Benzene iii Page ii iii vi (DR) 20 22 25 27 27 28 29 ' 51 32 33 33 33 35 45 47 51 56 TABLL OF CONTENTS - Continued Page XII Kinetics of the Reaction of Trimethylsilyl- dimethylamine with Piperidine at O0 and 38° with no Solvent 59 XIII Kinetics of the Reaction of Trimethylsilyl- dimethylamine with Aniline at 58 0° in Ben- zene 63 XIV Bibliography 66 iv Table II III IV IIST CF TABLES Page Equilibrium Data for the Exchange at 38° of Amines with Trimethylsilyldi- methylamine 9 Equilibrium Data for the Exchange at 38° of Amines with Phenyldimethylsilyl- dimethylamine lO Equilibrium Data for the Exchange at 38° of Amines with Vinyldimethylsilyl- dimethylamine ll Dissociation Constants of Organic Amines in Water at 25° 13 Equilibrium Data for the Exchange at 38° of Dimethylamine with Substituted Trimethylsilylamines 21 Figure IIST OF FIGURES Page Proton Magnetic Resonance Spectra of Dimethylamine, Trimethylsilyldimethyl- amine and Tetramethylsilane in Benzene 37 Proton Magnetic Resonance Spectrum of the Reaction of Trimethylsilyldimethyl- amine with Benzylmethylamine 38 Proton Magnetic Resonance Spectra of the Kinetic Study of the Reaction of Aniline with Trimethylsilyldimethyl amine 41 Concentration of Trimethylsilyldimethyl- amine versus Time 44 Kinetics of the Reaction of Trimethyl- silyldimethylamine with teat-Butyl- amine at 38°in the Absence of Solvent 53 Kinetics of the Reaction of Trimethyl- silyldimethylamine with Eggt-Butyl- amine at 38° in Benzene 55 Kinetics of the Reaction of Vinyldi- methylsilyldimethylamine with 3332- Butylamine at 38° in Benzene 58 Kinetics of the Reaction of Trimethyl- silyldimethylamine with Piperidine at 0° and 38° in the Absence of Solvent 62 Kinetics of the Reaction of Trimethyl- silyldimethylamine with Aniline at 38° in Benzene 65 vi -1- I. Objectives The exchange of an amine with a silylamine is a useful technique for the preparation of new silyl- amines. ESiNR2 + HNRé =-"—‘ ESiNRé + HN‘R2 For synthetic use this reaction is forced to the right by removing the amine which is liberated. Cne objective of this study was to measure equilibrium constants of this reaction and to show both the effects of the substituents on the amine nitrogen and the effect of various groups attached to silicon. A second objective was to measure the rate constants for representative examples of this reaction. Such studies should lead to an understanding of the nature of this reaction and the effect that the struc- tures of the reagents have upon the course of the reaction. II. Background In 1949 Larsson et al. (1) demonstrated that N- (triethylsilyl)ethylamine underwent an exchange reaction when heated with primary amines according to the equation: 5 2NR -—- EtasiNHR + H This discovery led to a widely used method of pre- Et SiNHEt + H NEt 2 paring many new silylamines from readily available silylamines. Iarsson reported (2) that Me231(NHMe)2 "decomposed" on heating to give methylamine as one of the products. This was undoubtedly a condensation of the type 2 MezSi(NHMe)2 77- WeNHMe2SiNMeSiM82NHMe + MeNH2 + some {seasiNMe}x From this it is clear why the compound MeBSiNH2 eluded isolation. Rather, the reaction of MeESiCl with ammonia led to the condensed product hexamethyl- disilazane (5). 2 Me 0101 + excess NH3'—+ 2[M8501NH2]“ MeabiNHoiMe 3 3 + 2 NHACl + NH5 Since Larsson's work, there are many literature references to the preparation of silylamines by this route. An excellent review of the chemistry of silicon- nitrogen compounds including a section on silylamine- amine exchange appeared in 1961 (4). Reference is made to more than fifty silylamines prepared by this exchange reaction. Abel and Bush (5) found that symmetrically sub- stituted ethvlenediamines reacted with a big-amino- silane to give cyclic compounds. HRNCH 2CH2NRH + Me281(NHEt)2 -“ H C'—-—-CH + 2 H NEt 2| , 2 RN NR \Si/ (1) Me2 2 At room temperature compound (I) formed a polymer which reverted to the starting material upon heating. (1) _25:_. ‘_______ polymer 200° The exchange of (I) (R is Ethyl) with tetrakis-di- methylaminosilane formed a spirane. Et Et CHZ-N\ ,N-CH 2 (I) + (Me N) Si -* I Si | + 2 Me Si(NMe ) 24 r. \ 2 22 VH2-N N-CH2 Et Et 2 I Andrianov (6) has examined the reaction of aniline with octamsthylcyclotetrasilazane and hexamethylcyclo- trisilazane and found in each case that the products included N,N',N"-triphenylhexamethylcyclotrisilazane. In this reaction silicon-nitrogen bond cleavage took place with the formation of a six-membered ring from an eight-membered ring. Me 812 l \ 4 HN NH M92 ' ' ,,Si\. Me28i\ [SiMe2 PhN RPh NH 4 Me Si SiMe 2 \'N ,. 2 + 12 HZNPh -—~ Ph Me2 HN-Si-NH ! l + 12 RH5 + unidentified Measi SiMe2 I I products HN-SivNH Fink (7) has shown that heating R2Si(NHR')2 or allowing it to exchange with a primary amine in which R or R' are bulky groups led to a cyclodisilazane. 0 2 R281(NHR )2 -‘ _ u R2°i NR I + 2 R'NH R'N - SIR2 or _.s 2 2 R281(NHR')2 + 2 R"NH2 The exchange is reported (4,8) to be acid catalyzed but it has also been shown to proceed in the absence of a catalyst (9). Ammonium salts are commonly used as catalysts. , In 1961 Fessenden showed (9) that hexamethyldi— silazane lost ammonia when refluxed with some secon- dary aliphatic and aromatic amines. In addition, the reversible nature of the reaction was demonstrated. NeablNH01M85 + 2 HNBu2 -—* 2 hejbiNBu2 + NH5 press. Me SihBu + excess NH “ he SiNHSiMe + 5 2 5 23°, 12 hrs. 5 5 N-Trimethylsilylpyrrole and ammonia likewise formed pyrrole and hexamethyldisilazane. Data were also obtained concerning the effect of solvent on the extent of exchange of piperidine with hexamethyldisilazane. % Yield of Solvent N-trimethylsilyl piperidine Ether O p-Dioxane 35 Toluene . 54 Decalin 54 Quinoline 61 Several reports concerning silylamine-amine exchange reactions appear to be anomolous. Hexamethyl- disilazane (II) is reported not to exchange with di- ethylamine, because no products boiling above the starting materials were found (10). Later (9) II (b.p. 126°) was shown to be separated only with difficulty by distillation from N,N-diethylaminotri- methylsilane (b.p. 128°). Ammonia was a product of thisreaction. Another example of an unusual result in the aminolysis of silicon compounds involved the use of N-substituted anilines. N—Methylaniline did not exchange with hexamethyldisilazane (9) and N-ethyl- aniline did not exchange with hexamethylcyclotrisilazane (6). Contrary to these reports Tanslo (11) found that Eggs-(ethylamino)propylsilane does exchange with N-methylaniline. n-PrSi(NHEt)3 + 2 PhMeNH -—s n-PrSiNHEt(NMePh) + 2 HZNEt The divergences in these results may in some way be due to the extent to which the substituted aniline can approach the silicon-nitrogen bond. In 1962 Sergeeva (12) found that hydrazines will exchange with silylamines. Up to four amino groups on silicon could be exchanged, giving a series of compounds of the structure R4_xSi(NHNRé)x. This was the first example in which a silicon hydride i.e. R is H was present in the exchange of an amine with a silylamine. In 1965 Goldin and Ivanova (13) studied this aminolysis reaction with other hydrazines. Alkoxy groups attached to silicon are not attacked during these amine-silylamine reactions as shown by the work of Larsson (14) in which benzylamine reacted with tributoxy-N-ethylaminosilane. (BuO)BSiNHEt + H2NCH2Ph —-x (Buo)3SiNHCH2Ph + HZNEt Amides react with amino groups on silicon. Wannagat (15) has shown that hexamethyldisilazane and organic amides or imides, including urea, liberate ammonia and form an N-silylated amide. Q C' 3 5 + 2 H2NCR —~ 2 MeBSiNHCR + NH Klebe and Bush (16) quantitatively measured the Me SiNHSiMe 3 equilibria of substituted trimethylsi1ylacetanilides with acetamide. Q X g Q X CHaCNH2 + gSéMe3 =2 CH5 NHSiMe3 + CHBCNH@ CH3 An electron withdrawing group (X) shifted the equi- librium as written above to the right while an electron donating group had the opposite effect. This study used nuclear magnetic resonance spectros- copy to measure the concentration of the various species at equilibrium. It was decided to use this same method to evaluate the equilibrium constants in the reaction of amines with silylamines, described below. III. Discussion of Results and Conclusions The equilibrium between aliphatic primary and seccndary amines or aromatic amine with each of three dimethylaminosilanes (RMe2SiNMe2, RaMe, Ph and CH=CH2) was studied by means of nuclear magnetic resonance spectrosc0py and the equilibrium constants for the reactions were calculated. "' I H 0 It RM€2blNM92 + HNR R =2 RMeBSiNR R + HNM62 With a few exceptions the equilibria were studied in benzene solutions. In most cases good reproduci- bility of the equilibrium constants was achieved at various concentrations of the reagents and also independently of the direction from which the equilibrium was approached. Using 'H n.m.r. the rates of the above reaction for a representative number of amines were also measured. A. Equilibrium Studies The data for the equilibrium studies can be found in Tables I, II and III. The equilibrium constant Kc for the reaction is dependent upon several factors. In order to simplify the discussion consider first the aliphatic amines. Primary amines exchange more completely than second- ary amines (9). The present work substantiates this \0 Table 1 L Equilibrium Data for the Exchagge it 38 of Amines with Trimethylsilyldimethylamine. 11185;31NP.-182 + R2NH : MejSiNR2 + Hittea A B C D Stoichiometry R2NH Moles moles AzB Me3S1NTe2 HNMe KC at equil. at equIl. 1:1 EteNH 1.71 0.29 0.026 2:1 " 3.57 0.43 0.035 1:1 (iso-Pr)2NH 1.69 0.31 0.034 1:1 tert-BuNH2 1.42 0.58 0.17 2:1 " 3.14 0.86 0.21 1:1 " 1.48 0.32 0.13’ 1:1 piperidine 1.06 0.94 0.78 2:1 " 2.75 1.25 0.76 1:1 " 1.29 0.71 0.30' 1:1 PhCH)(Me)NH 1.09 0.91 0.70 2:1 3 2.78 1.22 0.69 1:1 aniline 0.65 1.35 4.32 2:1 " 2.28 1.72 4.62 1:1 p-chloroaniline 0.52 1.48 8.19 2:1 " 2.18 1.82 8.45 1:1 m-chloroaniline 0.61 1.39 5.20 2:1 " 2.27 1.75 4.75 1:1 m-toluidine 0.54 1.46 7.16 2:1 " 2.20 1.80 7.38 1:1 p-toluidine 0.43 1.57 13.3 1:1 Ph2NH 1:1 H-methylaniline * No solvent No evidence of exchange Table II -10- E u ibrium Data for the Exchange at 38° of Amines with Phenyldimethylsilyldimethylamine. L‘ ' ad ‘ (" A B C Stoichiometry RaNH . Moles A:B PhMe2SiNMe2 at quil. 1:1 EtaNH 1.62 1:1 (iso-Pr)2NH 1.85 2:1 " 3.81 1:1 PhCh2(CH5)NH 1.18 2:1 " 2.91 1:1 tert-BuNH2 1.49 2:1 " 3.32 1:1 aniline 0.70 2:1 " 2.32 1:1 p-toluidine 0.37 2:1 " 2.08 1:1 m-toluidine 0.54 2:1 " 2.20 HNMe2 D moles at equil. 0.38 0.15 0.19 0.82 1.09 0.51 0.68 1.30 1.68 1.55 1.92 1.46 1180 KC 0.05 0.007 0.003 o.+83 0.448 0.117 0.105 3.49 6.80 16.9 21.2 7.30 7.35 -11- Table III Equilibrium pgta for the Exchange at 38° 9; Amines with Vinyldimethylsilyldimethylamine. ViMeZSiNMe2 + R2NH ==2 ViMeZSiNR2 + HNMe2 A B C D Stoichiometry R2NH .Moles Moles Kc A:B ViMe SiNMe HNMe at e uil. at eqail. 1:1 Et2NH 1.64 0.36 0.047 1:1 piperidine 0.99 1.01 1.014 1:1 " 1.01 0.99 0.953 1:1 aniline 0.70 1.30 3.45 1:1 m—toluidine 0.51 1.49 8.53 2:1 " 0.20 1.80 7.35 1:1 p-toluidine 0.45 1.55 12.0 1:1 tert-BuNH 1.37 0.64 0.21 2 -12- in showing that Eggt—butylamine exchanged by a factor of about ten times more completely than diethyl- or di-igg-propylamine. This was true regardless of the nature of the group on silicon. For similar reasons N-methylbenzylamine gave a much higher value of K0 than might have been expected. Since by definition, the exchange of dimethylamine with RMe2SiNMe2 would yield an equilibrium constant of 1.0, substitution of only one methyl hydrogen with phenyl in N-methylbenzylamine only slightly reduced the value of Kc' The values of Ko for the two silylamines used with piperidine seem to be inordinately high, compared to those obtained for other secondary amines. In piperidine, however, the two alkyl moieties on nitrogen are not free to rotate due to the cyclic structure of the compound. Therefore, these groups apparently do not have a large deactivating effect. A second factor which undoubtedly also influences the reaction is the relative base strengths of the amines. Table IV gives the dissociation constants in water for the amines used in this study. In discuss- ing the effect of the base strengths of the amines, substituted anilines as well as the aliphatic amines were considered. Only meta and para substituted anilines (substituents = CH5 and Cl) were used in order that the steric effect of these groups on the -15- Table IV Dissociation Constants of Organic Amines in Water at 25 Amine pKa Ka L MezNH 3.28 5.2 x 10"4 Et2NH 3.02 9.6 x 10"4 (iso-Pr)2NH 1.95 1.1 x 10'} tert-BuNH2 3.55 2.8 x 10'4i piperidine 2.79 1.6 X 10-3. N-benzylmethyl- - 4.42 3.8 X 10-5 amine _10 aniline 9.42 3.8 X 10 m-chloroaniline 10.54 2.9 K 10'11 p-chloroaniline 10.07 8.5 x 10'11 m-toluidine 9.31 4.9 x 10"10 p-toluidine 8.92 1.2 x 10"9 Ph2NH 13.12 7.6 x 10'14 Némethylaniline 9.30 5.0 X 10'10 -14- energy of the transition state would be of the same order of magnitude. The values of the equilibrium constants for all of the amines studied were grouped into two sets,,those greater than 1.0 and those less than 1.0. The aliphatic amines with pKa values ranging from 2.8 to 3.5 fall into the set having Kcl.0. However, only qualitative conclusions can be drawn because attempts to establish a linear relationship between Kc or log Kc and Ka or pKa failed. Thus, it was concluded that both the steric nature of the in- coming amine as well as its base strength relative to the amine attached to silicon effected the position of equilibrium. The ability of the nitrogen lone pair electrons to enter into bonding with the vacant d orbitals of silicon suggests that amines which can facilitate such bonding by making the electron pair more available sterically as in piperidine or through conjugative effects such as in the anilines ought to increase the magnitude of the equilibrium constant. Electron donating substituents on aniline in- creased the equilibrium constant. A plot of log Kc/Ko’ ‘ K is the equilibrium constant for the reaction of aniIine with RMeésiNMez -15- versus 0" (See below) for the meta- and para-toluidines and aniline with the three silylamines yielded a straight line. a 4 L A J n -2 -l O I 2 3 log K/Ko for the reaction: X X where R is CH5- e , CH2=CH- a and phenyl o . —l6- The cf values used for the substituents are: Substituents (X) 0' p-chloro 0.227 m-chloro 0.373 p-methyl -0.170 m-methyl -0.069 hydrogen 0.0 The chloroanilines were not consistent with this linear free energy relationship. more data must be collected concerning the electronic effects of sub- stituents, but it is possible to venture a few pre- dictions based on the available data. Only hypercon- Jugative structures can be written for the toluidines. Thus the resonance contribution for this group is relatively small and a linear relationship of log k/Ko can be obtained. However, resonance structures in- volving the amino group can be written for the chloro- anilines. Therefore, the Hammett relationship fails. Taft (17) has shown that the Hammett equation usually does not hold when the substituents are capable of direct conjugation with the reaction site. This strongly supports the existence of a pseudo double bond (dfl-p") between the silicon and the nitrogen of the arylaminosilane. The negative value of‘P indicates that the reaction is favored by a high electron density on the nitrogen of the aniline (18). An attempt was -17- made to determine the effect of the strongly electron withdrawing nitro group on the exchange reaction. This was foiled by the fact that both meta- and para- nitroaniline are only sparingly soluble in benzene. Since all other exchanges were measured in benzene at a concentration of 1.5 I silylamine, it was not possible to include these two compounds. The exchanges were performed in nitrobenzene but the separation of the peaks necessary for a measurement of the equil- ibrium constant was considerably less in nitrobenzene than in benzene and no reliable results could be obtained. Others (19) have shown that this exchange reaction is solvent dependent. This work has also shown that the equilibrium constants for the exchange of aliphatic amines w1th silylamines in nitrobenzene were markedly different from those obtained for identical reactions in benzene as the solvent. Generally, higher values of Ko were obtained in nitro- benzene. This is probably due to increased solvation of the amine by nitrobenzene. N-substituted anilines showed no exchange with dimethylaminosilanes. The base strengths of these amines cannot be used as criteria in this case since N-methylaniline is a stronger base than aniline. The observation that N-substituted anilines do not exchange has also been reported by others (6.9). -18- In addition to the amines listed in Tables I, II and III, exchange reactions involving di-n-butylamine, di-n-prOpylamine and ethyleneimine were also studied. In these three cases, however, the 'H n.m.r. spectra were toocomplex to interpret. The equilibrium constants obtained for the reac- tions of vinyldimethylsilyldimethylamine were not significantly different from those obtained with trimethylsilyldimethylamine. From these data it was concluded that the vinyl group exerts on influence similar to that of methyl on the silicon-nitrogen bond. The equilibrium constants obtained in the reactions of phenyldimethylsilyldimethylamine were somewhat lower than those obtained for trimethylsilyldimethylamine when aliphatic amines were used. The usual effect of a phenyl group on silicon is that day-per bonding of the TT electrons of the ring and the vacant d orbitals of silicon causes the silicon to become less electro- negative due to contributions from structures I and II. C3 C3 ’ " {—9 " Qsee 8.1...me I II Thus, the normal electron releasing character (positive inductive effect) of the MeBSi group is compensated by the electron donating ability of the phenyl group of PhMeZSi. The result is a strengthening of the silicon- -19- nitrogen.bond and a reduction in the ease with which exchange takes place. On the other hand, a less basic amine would prefer to displace the dimethylamino group. It was found that indeed aniline and substituted anilines exchanged with phenyldimethylsilyldimethylamine to the same extent as trimethylsilyldimethylamine. -20- B. Reverse Equilibrium Reactions Although this reaction of amines with silylamines was long thought to be an equilinrium reaction, Fessen- den (9) was first to experimentally show this to be true. He found that hexamethyldisilazane and di-n- butylamine liberated ammonia by the reaction: 1. , . , _ u _ MeabiHNSiMe5 + 2(n Bu)2NH -* 2we SiN(n Bu)2 + NH5 3 Upon treating trimethylsilyldibutylamine with ammonia in a bomb at room temperature for 12 hrs., hexamethy- ldisilazane was formed. 2 Me SiN(n-Bu)2 + NH -*- MeBSiHNSiMe5 r HN(n-Bu)2 3 3 In order to establiSh the validity of the equili- brium constants determined in the present study a number of trimethylsilyl— N,N-dialky1amines were syn- thesized and allowed to react with dimethylamine in benzene. 1v185SiN'R2 + HNMe2 ;=2 111858.1NMe2 + HNR2 The compounds prepared were: MeasiNEt2 meBSiNBu2 [he- SiNHCMG , . 5 ‘5 N19 551N -—-‘ MejSiN: > + HNMe2 A 100 m1. flask fitted with a condenser and drying tube was charged with trimethylsilyldimethylamine (0.1 mole,1l.7 g.), piperidine (0.1 mole, 8.6 g.) and a small amount of NHucl. The solution was refluxed for 14 hours and distilled at 50 mm Hg. The following fractions were taken: Fraction wt,(g,) Temp. n35 1 1.4 55 g 2 1-6 65 1.4592 3 5.6 66 1.4402 4 4.7 66 1.4401 Small amounts of dimethylamine and piperidine were found in the Dry Ice trap. Fractions 5 and 4 (10.5 g., 0.064 m., 64%) were analyzed by v.p.c. and found to be greater than 98% pure. These fractions had n35 1.4402, 825 0.857, RD (calcd.) 0.517 RD (found) 0.315. For this compound Birkhofer reports (25) b.p. 1610 and ngo 1.4425. 5331. Calcd. for 08H198iN: '% Si, 17.85: h N, 8.91. Found: % Si, 17.57: % N 8.91. The 'H n.mr. spectrum showed a sharp CHasi peak at 2.0 cps., a broad (CH2)3 resonate at 87 cps. and also a broad (CH2)2N peak at 165 cps. These were in the ratio of 9.1/5.9/4.0 the theoretical values being 9.0/6.0/4.0. -32- E. Preparation 9; Trimethylsily-tert-butylamine A 500 cc., three necked f18$k fitted with a condenser, stirrer, thermometer, and addition funnel was charged with 200 cc. pentane, 60 g. triethylamine (0.6 moles), and 56.6 g. Eggt-butylamine (0.5 moles). At room temperature there was slowly added with stirring 54.5 g. trimethylchlorosilane (0.5 moles). The reaction mixture was stirred one hour and filtered. The amine hydrochloride was washed several times with pentane and the combined filtrates distilled through a 1.5 X 27 cm. column packed with glass helices. After removing the solvent there was obtained 26 g. (56 % yield) of pure trimethylsilyl-tggt-butylamine b. 118- 20°, ngs 1.4049-51, 025 0.7625, RD(calcd) 0.524 and RD(found) 0.522. 5321. Calcd. for C7H1981N: % Si, 19.52; % N, 9.65. Found: % Si, 19.4; % N,9.65. The 'H n.m.r. spectrum of this compound in benzene showed two singlets one at 9 cps. (CH531) and another at 64 cps. (CHEC) in the ratio of 9/9.1. The theoreti- 001 values are 9.0/9.0. -55.. v, Analyses pg Materials A. Organic amines B. All liquid amines were analyzed for purity by vapor phase chromatography (v.p.c.) and by comparing their refractive indices with published values. The amines were stored over potassium hydroxide pellets to maintain dryness. Para-toluidine was recrystallized from water and para-chloroaniline and meta- and para- nitroaniline were recrystallized from ethanol. These amines were dried in 13233 and their melting points agreed favorably with literature values. Silylamines After distillation the silylamines were also analyzed for purity by v.p.c. and by comparing their physical properties with published values. 1. Silicon analyses The silicon analyses were performed by convention- al methods by the Analytical Department of the Dow Corning Corporation, Midland, Michigan. 2. Nitrogen analyses The silylamines were analyzed for nitrogen by the non-aqueous titration method of Fritz (24). Weighed samples (0.2-0.5g.) were placed in 50 to 50 cc. of glacial acetic acid and several dr0ps of a methyl violet indicator solution added‘. This solution was ' A solution of approx1mately 0.1 g. methyl violet in 10 cc. of chlorooenzene then titrated with a solution of 11010,“ in glacial acetic acid previously standardized with potassium hydrogen phthalate. The amount of nitrogen was calculated by means of the expression Vol. HClO4 (1.) X Norm. of HClO4 X 14.01 X 100 % N = Sample Weight -35- VI. Method 2; Conducting Equilibrium Experiments With few exceptions the equilibrium studies were conducted as follows: 8 mm X 200 mm Pyrex glass tubes sealed on one end were charged with 2.00 millimoles of a silylamine by means of a syringe. Exactly 1.00 ml. (2.00 millimoles) of a 2.0 M solution of an amine in benzene was then added together with a few grains of (NH4)2804. When the ratio of reagents was raised from 1:1 to 2:1, 4.00 millimoles of the silylamine was used with 1.00 ml. of the amine solution. An addition- al amount of benzene was then added to maintain a constant molarity of the silicon species. The tubes were shaken thoroughly and sealed with rubber septa. After standing a minimum of 12 hours at room temp- erature, a sample was withdrawn with a syringe and. placed directly into an n.m.r. sample tube. The 'H n.m.r. spectrum was recorded at a sweep width of 250 cps. This allowed both the peaks of CH Si at 3 about 0 cps. and the CH N of the dimethylaminosilane 5 and dimethylamine at 150 to 150 cps. to be recorded with maximum separation. In all cases the spectra were redetermined after 24 hours and no significant change was noted. It was found that in benzene solutions the (CH3)2NSi showed a resonance at 145 cps. while (CH3)2NH was found at 155 cps. This separation -36.. of 10 cps. was sufficient to easily integrate the peaks and determine the amount of each moiety present. Generally, the amount of Me2NH present was found by simple proportion relative to the total amount of CHESi, although the position of the CHasi peak varied slightly depending on the remaining groups on silicon. This allowed the total CHESi peaks to be integrated and set equal to the initial concentration of silyl- amine. The number of protons and hence the number of moles of dimethylamine could be determined. Figure G.shows the 'H n.m.r. spectra of (a) Me2NH, (b) MeeNH containing tetramethylsilane (TMS) and (c) Me2NH containing TMS and MeasiNMe2 all in benzene. A typical spectrum obtained in the equilibrium studies is reproduced in Fig. 2. Reverse Equilibrium Reactions —_8 NH + MeasiNH2 .F— Me3 A solution of dimethylamine (Matheson, dried in Me SiNM92 + HNR 2 2 a KOH train) in benzene was prepared and found to be 2.28 M by titration with 0.1721 N H0104 in acetic acid. The correct amount of this solution and also the silylamine was loaded into a sample tube by means of a 1.00 cc. syringe. Additional benzene was then added to bring the molar concentration of the silicon .. . . fim-.--1-v4-q_.----.-4-io__---..1-.-:i-- .3i -4 -4 .e. 4- Ho Lu 3 ..i. ..... i i . , _ _ _ . _ . 21.17, _: .833 _ . _ d - _ . _ , cerium... I... 2.350 c. :24»: 1 4! 4 ii 4 N\1 1...): 0:32.00 5 mZF » 19...»: {I 3 . <1 < n 7 1g 5- glib-s . 'uinnuouvtsoil! nnnnn Si in e h . I. .................. -1350... ........o .. £29.: 9 mt». 12...... I..-:Il:llll if!!! I) . Jill L 11 Jul...- ’ (.4 Irl In! -I:I-----:-...-----l---li.lu _ w _ .m.L a 1. Jul 5 1 a .6 . Ila!- II 8. Jul-sol! ' an l- 3. r---.i.+m .- . -. . 1 .3: . :4 -. --LJ- -LL --> -v- .---- - _ — . AT U. LDDIE)D\P bibrnbhti rhthbbtbL—DD?D h 4‘ . a or $7 a; 3 ed T, i.-- . ”no 6 an I s m. . . i; en“ “ . is -58- . £19214© : |\ /I Afeznjsguflw5?u +515... IS 4.... nIo :zAflw :¢.:0© + $192.54.; .5 m .eE -39- Species to the same value used in studying the reverse of this reaction. A few crystals of (NH4)2SO4 were added and the system allowed to equilibrate. The position of equilibrium was then measured in a manner analogous to that described earlier. The values of KC and Kgl are given in Table 5. The good agreement of K0 for this reaction and Kc for the reverse reaction (Cf. Table l) verify the condition of equilibrium. An attempt was made to prepare N—anilinotrimethyl- silane and p-chloroanilinotrimethylsilane in order that they might be included in this series of reactions with dimethylamine. Both of these syntheses were fraught with difficulties and pure materials were not obtained. In both cases the isolation of the anilinosilane free from traces of the aniline was difficult. Kinetic Measurements The kinetics of the excnange ESiNMeZ ++ HNRé= LISiNRé + HNIvSe2 for a number of amines and silylamines was followed by 'H n.m.r. The 10 cps. separation of the methyl protons of (CH5)2NSiE and (CH5)2NH allowed the disappearance of the former and the appearance of the latter to be measured at various time intervals. This was done by placing the correct amount of silylamine into an n.m.r. sample tube after maximizing the magnetic field of the instrument according to standard procedures (25). The .40- tube was placed into the probe and the offset control adjusted so that the (CH5)2NSiE protons were as far to the left of the chart paper as possible. Adjustment of the instrument required sufficient time for the sample to reach the temperature of the probe. The correct amount of amine was then added to the n.m.r. tube and the time (To) recorded. After shaking the solution vigorously the tube was replaced in the probe. The spectrum amplitude was then adjusted for maximum peak heighth. A region of about 20 cps. (ca. 5 cps. on either side of the peaks) containing the methyl protons of (CH3)2NSi and (CH5)2NH was recorded. The offest adjustment was then increased 10 cps. (upfield) and again scanned at various time intervals from To. Repeating this procedure resulted in spectra of the. type shown in Fig. 3 . Fig.9 shows the kinetic results of such a measure- ment for the exchange: MeBSiNMe2 + H2N© :‘i‘MeBSiNHQ + HNM62 The concentration of MeBSiNMe2 at each point can be determinded in two ways. The area of each peak can be measured by means of a planimeter and related to the (area) concentration at equilibrium. The second method of determing concentrations involves relative peak height measurements. The peak height ratio N of ESiNMeg/HNMe2 can be found and the absolute concentra- i . . 3.252; 2,8 Mm om. ms o.v n... on mu N~ m. 8.. m. 0.. e m 4 ~ . . #44...\14.<44.-.4A.\41414. 44.... 44.4. 44 .4 d4 44 .. i n . x 2 \ s.\ .. x l . . e _ n \ W 1.1 \n .. .. o r . w m k . l o, , n. . w ,.... NC (a; ._ ,lnaoss. .1 1:; M i n l on. Q. on , \ g c 6.. o.o o... o Mimi ” undow mzz. «moo: - mien m2.» 530.. 4:92: + .mlmrfov Wm. .zf 4.1.1925me .42- tion of ESiNMe2 can be determined from a knowledge of the initial concentration of E SiNMe2. Thus if Ao moles of ESiNMe2 are initially present and At moles present at time t then (Ao - At) . Ct moles of HNMe2 must have formed. Since the ratio 5 SiNlVl82/ Hlee2 or At / CtsN can be measured and At + 015 = A0 then: Ct = A0 - At 01‘ A =N(AO-At) t Solving this eXpression for At then gives the concentra- tion of E SiNMe‘2 at time t. A normalization of the peak heights per proton need not be made since both peaks in question contain an equal number of protons, namely six. The two types cf measurements were compared in several cases. Fig. 4 shows the disappearance of Me SiNMe2 done by both methods for the reaction of 5 MeBSiNMe2 with aniline. The excellent correlation of these two methods prompted the abandonment of the tedious area measurements and ad0ption of the peak height ratio method exclusively in later experiments. Such peak height measurements are valid when peaks corresponding to the same nuclei in the same chemical environment are being compared since identical nuclei have identical spinspin relaxation times (26). Legrow (27) showed that excellent kinetic data can be obtained in this fashion when he examined the base catalyzed rearrange- -45- ments of silylcarhinols. B: Ph2MeSiC(0H)Ph2 -—-> Ph2MeSiOCH2Ph2 The CH3Si peak is shielded significantyly differently in each compound to allow kinetic measurements by 'H n.m.r. 22:: m! :. Oh cm on O¢ on ON 0. o mhzwimmawduz .5..ng vaun— e mhzm2mmam «1.1925me no 20200 v .0: (es-norm: '(‘Homss'i‘um .45- VII. Treatment_gf Data for Kinetic Reactions The above discussion described how the concen- tration of silylamine was determined at various time intervals. These data were then treated in a manner used for second order equilibrium kinetics for a system of two components (289. For a reaction of the type: 1) A+BI§4C+D 2) - %% - k(AB) - k'(CD) If A0 and B0 are the initial concentrations then dA 2 5) - d? . kA(Bo- A0 + A) - k'(Ao- A) This expression can then be integrated by the method of partial fractions giving (Ao - As) (A - A6 + Q) 4) 1n ’ . (k - k')Qt (A - A6) (A0- A6 + Q) where 2 2 Q: X (B0 - Ao ) + 4A B K (K-l) °° . k -0[K(B -Ao)+2Ao]+.)(K-1) K = -—- and A6 = k' 2(k - 1) In all experiments equal concentrations of A and B or ESiNMe2 and HNR2 were used. Therefore, equation 4) becomes K - K A K - 1 A K + 1 2A K m < fr>[( )+ our >] =(k-k') of t (K 4 VK)[A(K - 1) - 10(1E - 1)] K - 1 -46- or %% = k(a - x)2 - k'x2. Using reaction variable notation i.e. x = A0 - A, a = A0 and x8 . Ao - Ae the following form of the expression is obtained x(a - 2x6) + ax 1n e = k a(xe - x) x 2a(a - xe) t 8 Thus a plot of 1n x(a - 2x8) + axe ‘ 8(Ie ’ X) versus time t 2a(a - Xe) xe should give a straight line with slope equal to the second order rate constant k. ‘ Throughout the remainder of this dissertation this expression is referred to as In Z/A. -47- VIIIJ Discussion q£_Error Since nuclear magnetic resonance spectroscOpy was used to measure both the equilibrium constants and the rates of reaction, the discussion of error centers itself upon the inherent errors of the insurument. Equilibrium Studies The temperature at which equilibrium measure- ments are made will alter the equilibrium constant. In these experiments the temperature was necessarily that of the probe of the A-60. This was checked frequently and found not to vary more than two degrees from 58°C. With such small temperature changes the expression giving the dependence of he on temperature, 0 AH 1n no a -—- + Constant RT indicates that this is a small scource of error, provided that AH° is not large. The solutions used to measure the equilibrium constants were examined at 12 to 16 hours after mixing. To show that the equilibrium condition had indeed been achieved many were rechecked after 24 hours and some after four days. No significant change was noted in any of those tested. From this it was con- 448- cluded that 12 to 16 hours at room temperature was sufficient time for the solutions to reach equilibrium. The largest error present in this study was the area measurement of peaks by the automatic integrator of the A-60. The accurate measurement of a peak by n.m.r. is most difficult and subject to the largest error when that peak is very small. The area deviation commonly associated with n.m.r. integrals is i 5%. In the equilibrium studies reported here the constant CH3S1§ peak was related to the methyl protons of di- methylamine. Thus, if a reaction formed very little dimethylamine, this peak was very small and difficult to measure unequivocally. When the equilibrium was far to the right the amount of dimethylamine was large. The expression for calculating the equilibrium constant is: [MeaNI-I] ESiNRZ] ° 3 [Me2NSiE] [HNR2] However, since by the equation for the reaction, each mole of dimethylamine which is formed also forms a mole of SSiNR2 this reduces to: [MeeNH] 2. KC : [Mestia] [HNRZ] .49- Consequently a variation in a rather large value of the dimethylamine concentration appreciably alters the value of Kc' The Optimum conditions for having the smallest error in the equilibrium constant occur when the numerator and denominator are nearly unity. That is when the peak areas for both Me2NH and MeeNSiE are nearly equal. The reactions in which no solvent was used are probably least reliable since the dimethylamine was not very soluble in these solutions. Evidence to this effect was observed by a slight pressure of di- methylamine on the ampules. Benzene was a good solvent for these reactions since dimethylamine is soluble in it and the aromatic protons were far enough downfield in the n.m.r. not to interfere with the spectrum of the reactants. The measurement of exact amounts of reagents by means of a syringe is probably reproducible within one per-cent. Kinetic Studies In addition to the errors discussed above the kinetic determinations were effected by descrepancies in measuring the time of each succesive point. In an attempt to reduce this contribution to the error it was found that at a sweep time of 250 sec. (full chart) -50- 15-20 sec. were required for the pen to scan the two peaks in question. The pen was started approximately 10 sec. prior to the time of measurement thus reducing this error to about i 10 sec. In some cases the magnetic field collapsed after a portion of the kinetic run was made and the experi- ment had to be terminated. -51- IX. Kinetics g; the Reaction gf Trimethylsilyldimethyl- amine with tert-Butylamine in the Absence g; Solvent 335800 SiNMe2 + H2NCMe5 r—' MeaSiNHCMe3 The exchange reaction in the absence of a solvent Me + HNM62 5 led to a lower equilibrium constant. Values of 0.17, 0.18 and 0.21 were obtained when the reaction was per- formed in benzene. With no solvent a value of 0.126 was found for Kc“ 0f the initial 5.77 moles/l. of MeBSiNMe2, 2.78 moles/1. was present at equilibrium. The following data were obtained for this system. Time Me3SiNMe2 (min.) (moles/l.) 1n Z/A‘ 0 3.77 0 3 5.68 0.0065 5 5.62 0.0110 10 5.59 0.0154 18 5.49 0.0217 22 5.45 0.0252 25 5.42 0.0279 50 5.57 0.0528 36 5.50 0.0401 40 5.28 0.0425 45 5.22 0.0495 ‘ See footnote bottom of page 46 -52- 50 5.16 0.0575 55 5.15 0.0618 60 5.08 0.0700 65 5.04 0.0775 70 5.01 0.0857 When the terms in the last row were plotted versus time a straight line was obtained as shown in Fig. 5. The lepe of this line was found to be 1.94 X 10"5 l. moles-lsectl. -55- 62.3 ms. _ ._. ON 00 on O¢ on ON 0. . O 1 q q 4 a q 4 J 4 4 q 4 T u . 0 “AVG .330 00.0 «.4192: + a..zo.o:zmm...:9 P8fl):—OO .. . OI . M1Mh ~:zo...:9 4 «1.292%...5. . .6... m .071 .. -54- I. Kinetics 2f the Reaction 9f Trimethylsilyldimethyl- amine with tert-Butylamine at 58° C in Benzene 5 :2 MeaSZUQHCMe3 At equilibrium 1.076 moles/l. of Me5BiNMe2 remained Me§SiNMee + H2NCMe + HNM62 of the initial 1.515 moles/1.. This kinetic experiment gave a value for Kc of 0.18 which is in agreement with the two values of 0.17 and 0.21 obtained in the equilibrium studies. The data collected for the reaction are compiled below. Time MeBSiNMe2 (min.) (moles/1.) 1n Z/A‘ 0 1.515 0 5 1.492 0.0101 8 1.459 0-0550 10 1.417 0.0461 15 1.594 0.0581 25 1.554 0.0952 55 1.275 0.1560 75 1.159 0.2640 The plot of the terms in the last column versus time gave a straight line with slope k equal to 6.4 X 1 1 10'5 1. moles- sec.’ . See Fig. 6. * See footnote bottom of page 46. -55- 5.2.262; n ON 00 on O? O J G ‘ .5 d I ‘ 1 «3.6.2: 4...:9ozzlmemxo. 4|." 4:20.318 + 4.10.26.119 .on A . 00.0 0.6 0.0 0N.0 .nN.0 V/Z ”I -56- XI. Kinetics 2f the Reaction 2; Vigzldimethzlsilyldi- methylamine with tert-Butzlamine 32 58° 0 i2 Benzene 2N0Me5 r— CH 5 0f the initial 1.510 moles/1. of ViMe2SiNle2 for CH2:CHMe2SiNM92 + H aCHMe2SiNHCMe + HNM92 2 this reaction there was found 1.055 moles/1. remaining at equilibrium. This results in an equilibrium con- stant of 0.19 which is in agreement with the value of 0.21 obtained in the equilibrium study. The data for this reaction are shown below. Time ViMe2$iNl62 (min.) 9(moles/l.) ln Z/A‘ 0 1.510 0 4 1.455 0.0251 6 1.411 0.0469 10 1.588 0.0588 12 1.575 0.0675 15 1.559 0.0752 19 1.545 0.0849 25 1.298 0.1152 50 1.261 0.1402 55 1.258 0.1589 40 1.208 0.1862 45 1.185 0.210 50 1.165 0.256 ‘ See footnote bottom of page 46. -57- 55 1.148 0.257 60 1.159 0.270 When the last column was plotted versus time a straight line with 310pe equal to the second order 1 1 rate constant k of 7.75 X 10'5 1. moles- see? was obtained. See Fig. 7. -58- .. 2.... .2...» on no on at 0' on on nu ON d 1‘ d d 1 d I «1:92: +£55.12 5.6. were «are Mum £1802“: . 5192.153 :0 "~10 L N 0.0 .06 26 m.. . :6 fl V o .6 «N6 and XII. Kinetics of the Reaction of Trimethylsilyldimethyl- amine with Piperidine at Q: and 580 with fig solvent I\41e.fis~iNI\(;e2 + HN/ ;_—_-¥ “'eBSiND + Hvale2 In an attempt to determine the energy of act- ivation for this reaction the exchange of piperidine with trimethylsi1y1dimethy1amihe was examined in the absence of any solvent. Unfortunately, because of mechanical difficulties with the n.m.r. instrument, measurements could only be taKen at two temperatures. 00 and 580 0. Thus, there is considerable ambiguity in the activation energy. A value of 4.9 Koal. mole -1 has been estimated on the basis of the available data. After these data were obtained, a second error was discovered. It will be noted that the intercept is not zero for either the Kinetic measurements. Apparently some error was made in either the initial amount of one of the reagents or in the measurement of xe. the amount of MeasiNIMe2 which has reacted. Since both plots intercept the ordinate at 0.012 the error was‘probably consistent in both experiments. Therefore the important aspect of these measurements, the slope, will be correct. A: 0°C Initially the concentration of Me -60- SiNMe2 was 5.864 moles/1.; at equilibrium it was 2.420 moles/l. Time Me3SiNMe2 _Lmig.) (moles/1.) ln Z[A‘ 5.86 0 4 5.67 0.0155 5.62 0.0175 11 5.57 0.0215 15 5.48 0.0289 19 5.46 0.0506 25 5.45 0.0514 29 5.59 0.0568 55 5.28 0.0476 45 5.24 0.0519 48 5.21 0.0550 55 5.18 0.0585 65 5.06 0.0755 A plot of 1n Z/A yielded a straight line with slope k equal to 1.5 x 10"5 1. moles- 1 sec? ‘ See footnote bottom of page 46. 1 See Fig. 8. _A_§ 58°C —61- Initially the concentration of MeasiNMe2 was 5.86 moles/1.; at equilibrium it was 2.50 moles/1.. plots are shown in Fig. 8. ‘ See footnot bottom of page 46. Time MeBSiNMe2 (min.) gmoles/l.) 1n Z/A‘ 0 5.86 0 5 5.55 0.0227 5 5.48 0.0286 7 5.42 0.0558 9 5.56 0.0595 11 5.51 0-0445 15 5.24 0.0517 15 5.22 0.0559 20 5.15 0.0644 25 5.05 0.0749 50 5.00 0.0824 35 2.95 0.0959 A plot of 1n Z/A yielded a straight line with slope k equal to 4.6 X 10'51. moles-lsectl. The above two ~62- A.z_3 92:. 2. cm on cc on cm 2 o . “0.6 ob . comm 36 wl .O-.OZ / . V . .5253 .3192: + A ”25.129914. Oz: +§o.z,m..£o. .2... m .07.. -65- XIII. Kinetics g; the Reaction 2; Trimethylsilyldi- methylamine with Aniline at 58° C in Benzene MeasiNMe2 + H2N. ;=! M9381M© + HNM62 It was found that at equilibrium there was 0.557 moles/l. of MeBSiNMe2 remaining. Since initially there was 1.515 moles/1., xe becomes 0.978 moles/1.. The following data were collected from these values. Time MeESiNMe2 _§min.) ‘Lmoles/l.) 1n é/A' O 1.515 0 2 1.458 0.055 4 1.518 0.098 5 1.204 0.171 6 1.180 0.188 8 1.115 0.241 10 1.022 0.526 15 0.955 0.400 16 0.918 0.448 19 0.804 0.651 50 0.678 0.989 40 0.657 1.185 Plotting the data from the last column versus time (Fig. 9) gave a straight line with sloPe k equal to the second order rate constant 5.0 X 10"4 1. moles-leec:l ‘ See footnote bottom of page 46. -64- sec.'l. For this Kinetic experiment the area of the peaks as well as the peak heights were determined. The data for this comparision are found in Fig. 5. The excellent agreement of the two methods will be noticed. The actual :H n.m.r. spectrum from which this data was compiled is reproduced in Fig. 5. -65- Do on «1:92: + ©§6wnze «2.3 m2; ow On ON J i 4 1 «.1 ©zux + "36.2% {£9 m 0.... O. «.0 0.. .0; “I V/Z 10. 11. 12. 15. 14. -66- XIV. gibliography E. Iarsson and 0. Mjorne, Svensk Kem. Tidskr. Q1, B. Iarsson and B. Smith, Acta Chem. Scand., 5 487 (1949). C. Eaborn, "Organosilicon Compounds", Butterworths Scientific Publications, Iondon, 1960, chap. 11. R. Fessenden and J. Fessenden, Chem. Rev., él, (4) 561-88 (1961). E. Abel and R. Bush, J. Organometal. Chem., 5, 245- 52 (1965). K. Andrianov and G. Rumba, Zh. Obshch. Khim., 52, (6). 1995-97 (1962). W. Fink, Helv. Chim. Acta, 42, (2), 498-508 (1964). E. Larsson, Swedish Patent 130,574 (1950); CA 52, 6654 (1951). R. Fessenden and D. Crowe, J. Org. Chem., gg, (11), 4658-41 (1961). S. Ianger, S. Connell and 1. Wender, J. Org. Chem., £2. (1). 50 (1958). L. Tanslo, Acta Chem. Scand., 55 29—54 (1959). Z. Sergeeva and S. Tszyan-lan', Zh. Obshch. Khim., 29. (6),1987-93 (1962). G. Goldin and N. Ivanova, ibid., 55, (5), 911-15 (1965). E. Iarsson, Kgl. Fysiograf. Sallskap. Innd, Forh., 2E. 159-44 (1954). 150 16. 17. 18. 19. 20. 21. 22. 25. 24. 25. 26. 27. 28. -57- U. ‘Nannagat and J. Pump. waterr. Chemiker-Ztg., eg, (10), 319-20 (1962). J. Klebe and J. Bush, First Internationa Symposium on Organosilicon Chemistry, Prague, 1965. R. Taft Jr. in M. Newman, "Steric Effects in Organic Chemistry", John Wiley and Sons, Inc., New York, N.Y. 1965, chap. 15. L. Hammett, "Physical Organic Chemistry", chraw Hill Book Co., New York, N.Y., 1940, pp. 197. R. Pike, J. Org, Chem., gé, (1), 255-56 (1961). J. Ryan and A. Smith, Dow Corning Corp., Midland, Mich., Unpublished Results. L. Sommer, "Stereochemistry, Mechanism and Silicon", McGraw Hill Book 00., New York._N.Y., 1965. A. Vogel, w. Cresswell and J. Leicester, J. Phys. .Chem. 2g, 174-77 (1954). Birkhofer et al., Chem. Ber. 25, 5804 (1960). J. Fritz, "Acid Base Titrations in Non-Aqueous Solvents", G. Frederi0k Smith Chem. 00., Columbus, Ohio, 1952. R. Bible Jr., "Interpretation of N.M.R. Spectra", Plenum Press, New York, N.Y., 1965, pp. 119. J. Pople, w. Schneider and H. Bernstein, "High Resolution Naclear Magnetic Resonance", McGraw Hill Book 00., New York, N.Y., pp. 77. G. Legrow, Ph.D. Dissertation, University of Toronto, 1964. A. Frost and R. Pearson, "Kinetics and Mechanism", John Niley and Sons, Inc., New York, N.Y., 1961, chap. 8. NH llll! H H HI ‘lllll‘ 68 9437