ABSTRACT INVESTIGATIONS OF THE EQUILIBRIUM REACTIONS OF SILYLPHOSPHINES WITH AMINES AND THE REACTION OF TRIMETHYLSILYLDIPHENYLPHOSPHINE WITH NICKEL HALIDES BY Ronald Eugene Goldsberry The purpose of this study was to synthesize phenyl and methyl substituted trimethylsilylphosphines and to investi~ gate their equilibrium reations with amines. The possibility of obtaining stable silylphosphine adducts with nickel halides was also investigated. Two silylphosphines, trimethylsilyldiphenylphosphine (I) and trimethylsilyldimethylphosphine (II) were prepared ac- cording to the following equations where R = CH3 or Ph: 2Ph2PCl + 2N3 > thPPth + 2NaCl SS S (CH3)2PP(CH3)2 + 2(Q-Bu)3P -—¢ (CH3)2PP(CH3)2 + 2(ngu)3P RZPPRZ + 2M(Na or Li) > ZRZPM RzPM + (CH3)3SiCl > R2PSi(CH3)2 + MCl. The equilibrium constants for the reactions of (I) and (II) with amines as indicated by the following equation were obtained as a function of the type of amine used in the reaction, )I.I|.I ll. 2 Ronald Eugene Goldsberry Measz2 + HNR; < > Me3SiNR§ + HPRz. For both (I) and (II), the primary aromatic amines were observed to exchange to a slightly greater extent than the primary aliphatic amines which exchanged to a greater ex- tent than the secondary aliphatic amines. The values of the equilibrium constants also indicated that for a particu- lar amine, (I) exchanged to a greater extent than (II). The following explanations were offered to help ration- alized the results of the equilibrium studies. First, as is expected, the silylamines formed from secondary aliphatic amines are less stable than those from primary aliphatic amines. Similarly, silylphOSphines with bulky phenyl groups are less stable than silylphosphines with methyl groups. Second, the primary aromatic aminosilanes are more stable than their aliphatic analogs possibly because of the extra stabilization present in the aromatic aminosilanes from conjugation of the silicon—nitrogen w-bond (dw-pw) with the aromatic ring. In contrast to this type of stabiliza— tion in aromatic aminosilanes, such effects in the aromatic phosphinosilane (I) appear to be not as effective. 'Third, the silylamines as a class of compounds are more stable than the silylphosphines. This may be rationalized by suggesting that the silicon-nitrogen bond is more stable than the silicon-phosphorus bond. The adducts NiX2(PthSiMe3)2 were formed from the re- actions of (I) with nickel halies (Br, I). The structures 3 Ronald Eugene Goldsberry of these adducts have been designated as diamagnetic square- planar. Also obtained as products of these reactions were the adducts Ni(PPh2)2[PPhZSi(CH3)3]2 (III) and NiX2(HPPh2)2 or 3 (IV). The adduct (III) resulted from the cleavage of the Si-P bond in (I) and (IV) resulted from the hydrolysis of (I). INVESTIGATIONS OF THE EQUILIBRIUM REACTIONS OF SILYLPHOSPHINES WITH AMINES AND THE REACTION OF TRIMETHYLSILYLDIPHENYLPHOSPHINE WITH NICKEL HALIDES BY Ronald Eugene Goldsberry A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1969 DEDICATION To My Wife, Betty and Our Parents ii I, u q, u- . , .2' ACKNOWLEDGMENTS The author is indebted to Professor K. Cohn for the helpful guidance and assistance offered during this in- vestigation and during the preparation of this thesis. He is also grateful to Professors Hammer and Weibrecht for their assistance during the investigation. Finally, the author is grateful to his fellow graduate students whose suggestions and friendship were invaluable to the completion of this degree. iii ““ “"‘" J 0. .0 TABLE OF CONTENTS Page INTRODUCTION . . . . . . . . . . . . . . . . . . . 1 Nomenclature . . . . . . . . . . . . . . . . 1 HISTORICAL . . . . . . . . . . . . . . . . . . . . 2 EXPERIMENTAL . . . . . . . . . . . . . . . . . . . 25 Reagents . . 25 Preparation of Trimethy181lyldimethylamine . 25 Characterization of Product . 26 Preparation of Phenyldimethylsilyldimethylamine 27 Characterization of Product . . . . . 28 Preparation of Diphenylphosphine . . . . . . 28 Characterization of Product . . . . . . 30 Preparation of Monophenylphosphine . . . . . 30 Characterization of Product . 32 Preparation of Tetraethyldiphosphinedisulfide 33 Preparation of Trimethylsilyldiethylphosphine 34 Characterization of Product . . . . 35 Preparation of Tetramethyldiphosphinedisulfide 36 Preparation of Trimethylsilyldimethylphosphine 38 Characterization of Product . . 41 The Cleavage of Tetrahydrofuran by Lithium Dimethylphosphide . . . . . . . . 41 Characterization of Product . 43 Preparation of Trimethylsilyldiphenylphosphine 45 Characterization of Product . . . . 47 Reaction of Trimethylsi1yldiphenylphosphine with Nickel Iodide in Benzene . . . . 47 Characterization of Product . . . . . . 48 Reaction of Trimethylsilyldiphenylphosphine with Nickel Iodide in the Absence of Solvent 50 Characterization of Product . . . . 52 Reaction of Trimethylsilyldiphenylphosphine with Nickel Bromide in Benzene . . . 55 Reaction of Trimethylsilyldiphenylphosphine with Nickel Bromide in the Absence of Solvent 55 Characterization of Product . . . . . . 56 iv TABLE OF CONTENTS (Cont.) Silicon Analysis . . . . . . . . . . . . . Nitrogen Analysis . . . . . . . . . . . . . Nickel Analysis . . . . . . . . . . . . . . Infrared Spectroscopy . . . . . . . . . . . Nuclear Magnetic Resonance Spectroscopy . . Vapor Phase Chromatography (Glpc) . . . . . Ultraviolet and Visible Spectrophotometry Magnetic Moment Measurements . Equilibrium Reactions of Trimethylsilylphosphines with Amines . . . . . . Method of Conducting Equilibrium Experiments Reaction of Trimethylsilyldiphenylphosphine with N—methylbenzylamine . . Reaction of Trimethylsilyldiphenylphosphine with Diethylamine . . . . . . . . . Reverse Equilibrium Reactions of Diphenyl- phosphines with silylamines . . . . . . DISCUSSION OF RESULTS AND CONCLUSIONS . . . . . SUMMARY . . . . . . . . . . . . . . . . . . . . BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . APPENDIX I. INFRARED SPECTRA OF COMPOUNDS . . . APPENDIX II. PROTON NMR SPECTRA OF COMPOUNDS . 68 69 82 85 90 94 TABLE II. III. IV. VI. VII. VIII. IX. X. XI. XII. XIII. XIV. LIST OF TABLES Page Synthesis of silicon—nitrogen compounds from dihalosilanes and higher halosilanes 8 Synthesis of silicon-phosphorus compounds from alkali metal phosphides and halosilanes 10 Equilibrium data for the exchange at 38° of amines with dimethylaminosilanes . . . . . 14 Kinetic data for the exchange of amines with dimethylaminosilanes . . . . . . . . . . . 15 Reactions of monosilylphosphine with boron trichloride, diborane, monobromodiborane, and boron trifluoride . . . . . . . . . . . 23 Distillation of phenyldimethylsilyldimethyl— amine . . . . . . . . . . . . . . . . 28 Distillation of diphenylphosphine . . . . . 30 Distillation of monophenylphosphine . . . . 32 Distillation of trimethylsilyldiethylphosphine 36 Distillation of trimethylsilyldimethyl— phosphine . . . . . . . . . . . . . . . . . 40 Distillation of products from the cleavage of tetrahydrofuran by lithium dimethylphosphide 43 Distillation of trimethylsilyldiphenyl— phosphine . . . . . . . . . . . . . . 46 Experimental data for the preparation of Ni12 complexes of Me3SiPPh2 . . . . . . . . 49 Absorption spectra for NiI2L2 complexes . . 53 vi LIST OF TABLES (Cont.) TABLE XV. XVII. XVIII. XIX. Page 1H nmr data for the exchange of trimethyl- silyldiphenyl and trimethylsilyldimethyl— phosphines with amines . . . . . . . . . . 64 Equilibrium data for the exchange of tri— methylsilyldiphenylphosphine with amines . 66 Equilibrium data for the exchange of tri— methylsilyldimethylphosphine with amines . 67 Equilibrium data for the exchange of diphenyl— phosphine with substituted silylamines . . 68 Basicity constants of organic amines in water at 25°. . . . . . . . . . . . . 72 Dissociation constants of some organic phosphines in ethanol—water mixtures . . . 77 vii ‘unnflr—”“a7’ n-I Figure LIST OF FIGURES APPENDIX I Page Infrared spectrum of trimethylsilyldimethyl— amine . . . . . . . . . . . . . . . . . . . 90 Infrared spectrum of diphenylphosphine . . . 91 Infrared spectrum of trimethylsilyldimethyl— phosphine . . . . . . . . . . . . . . . . . 92 Infrared spectrum of the product from the cleavage of tetrahydrofuran with lithium dimethylphosphide . . . . . . . . . . . . . 93 APPENDIX II 1H nmr spectrum of trimethylsilyldimethylamine 94 1H nmr spectrum of tetramethyldiphosphine . . 95 1H nmr spectrum of trimethylsilyldimethyl— phosphine . . . . . . . . . . . . . . 96 1H nmr spectrum of the product from the clea— vage of tetrahydrofuran with lithium dimethyl- phosphide . . . . . . . . . . . . . 97 1H nmr spectrum of trimethylsilyldiphenyl- phosphine . . . . . . . . . . . . . . . 98 1H nmr Spectrum of N—methylbenzylamine . . . 99 1H nmr spectrum of the equilibrium mixture from the reaction of trimethylsilyldiphenyl— phosphine with N—methylbenzylamine .. . . . 100 viii “ac—'j LIST OF FIGURES (Cont.) Figure Page 8. 1H nmr spectrum of the equilibrium mixture from the reaction of dimethylamine with trimethyl— silyldiphenylphosphine . . . . . . . . . . . 101 9. 1H nmr spectrum of the equilibrium mixture from the reaction of tert—butylamine with tri— methylsi1yldiphenylphosphine . . . . . . . . 102 10. 1H nmr spectrum of the equilibrium mixture from the reaction of diethylamine with trimethyl— silyldiphenylphosphine . . . . . . . . . . . 103 11. 1H nmr spectrum of the equilibrium mixture from the reaction of diethylamine with trimethyl- silyldiphenylphosphine (100 Mc) . . . . . . 104 ix INTRODUCTION 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}:2 Compounds derived from the structure HasiNHz are called silylamines, with the use of the appropriate prefixes to designate substitution. Similarly, compounds derived from the structure HSSiPHz are called silylphosphines. This system of prefix designation for nitrogen or phosphorus substitution becomes quite cumbersome when more than one nitrogen or phosphorus is attached to silicon. In these cases, the amine or phosphine grouping is designated as a substituent of the silane. The generic name silazane is given to the series Hasi(NHSiH2)nNHSiH3 and likewise the name silphosphane is given to the series H3Si(PHSiH2)nPHSiH3. Compounds of this series are called disilphosphanes, trisilphosphanes, etc. depending upon the number of silicon atoms in the molecule. Compounds of the type (stiNH)n and (HZSiPH)n are given the generic name cyclosilazanes and cyclosilphosphanes respec- tively, the prefix depends upon the number of silicon atoms in the ring. HISTORICAL The chemistry of silicon-nitrogen compounds has been investigated quite thoroughly. It is evident that silicon— nitrogen compounds differ markedly from their carbon analogs with respect to their chemical behavior. There have been numerous publications which dealt with this topic, and there are two excellent articles which review current activity in the field of silicon—nitrogen compounds.3:4 However, the chemistry of silicon—phosphorus compounds is not as well investigated, presumably because of the dif— ficulty in preparing these compounds. Recently, however, the interest in the chemistry of these compounds has in— creased, with a concurrent development of better methods for their synthesis. The interests in these compounds are varied, but the majority of the research compared their chemistry to the chemistry of the analogous nitrogen compounds. The chem- istry of nitrogen and phosphorus may be expected to be simi— lar because of similarity in the outer electronic structures of the gaseous atoms. Nitrogen does have some properties in common with the heavier elements of the group; however, it shows highly individual behavior in many important respects. For example, elemental nitrogen 2 COnsis ts of gaseous N2 molecules, whereas the heavier elements form tetraatomic molecules (3.3., P, P4) and are solids at room temperature.5 There are several cases where sufficient differences exist between nitrogen and the succeeding members of the group to detract from the usefulness of regarding nitrogen as the prototype for the group. For example, stoichiome- trically analogous halides, oxides and oxo acids of nitrogen and phosphorus are almost completely unrelated to each other whereas,those of phosphorus,arsenic and antimony are similar to each other. Therefore, it is of interest to determine the similarities and differences of the silicon-nitrogen compounds compared to the silicon compounds of phosphorus, arsenic and antimony. Formatux10faadv—pw bond has been postulated as a characteristic of the Si-N bond.6 Trisilylamine~and tri— (methylsilyl) amine have been shown by infrared and Raman spectra7:3,9 and electron—diffraction data10 to be co-planar molecules.11 Compared with the calculated (p-p) single bond in Si—N calculated from the sum of the covalent radii (variously assessed as 1.87 or 1.80 g) and a (pw-pw) double bond in =Si=N- (1.62 R), an interatomic distance of 1.73 i 0.01 R has been found for a series of Si-N compounds.12:13 These factors have been interpreted in terms of the ability of the lone pair of electrons of nitrogen to be donated to the empty d_orbital of the silicon (dv=pw overlap). 4 The polarity of the Si-N bond in substituted trisilyl— amines is decreased and the basicity is lowered as compared with their methyl substituted isostructural counterparts.14 This may be observed for example in the formation of com- plexes of silylmethylamines with trimethylboron;15 neither trisilylamine nor methylsidilylamine forms a complex. Di- methylsilylamine forms a weak complex and trimethylamine forms a more stable one. The character of the silicon-phosphorus bond is pres- ently being debated. Davidson, et al.16 have suggested that the heavy atom skeletons of the molecules (SiH3)3P and (SiH3)3As may be planar or nearly so due to the simi- larities of their vibrational spectra to those of tri- silylamine (H381)3N. An extended Hfickel molecular orbital treatment by Cowley17 has led to a claim that the most stable configuration for trisilylphosphine should be planar. These results may be rationalized by postulating that the silicon-phosphorus bond does have some dw—pw character. However, some authors have taken an agnostic view of the supposedly large (p-d)—w contributions to silicon not only for the silicon-phosphorus bond, but for the Si-Ge and Si-N bonds as well. Randall and Zuckerman18 calculated spin-spin coupling constants for several compounds and related these values to the amount of s—character in a bond. For silicon—nitro- gen compounds, the amount of s-character in a nitrogen hy— brid increases with dw-pw bonding. This is based on the 5 aSSEunption that the dw—pw bond is formed from an orbital on nitrogen having perhaps pure :p character.19 These nmr results obtained by Randall and Zuckermann led them to take a skeptical view on the proposed ability of phosphorus and nitrogen to donate their lone pair of electrons to the empty d-orbitals of silicon. However, Ebsworth20 has since pointed out that there may be considerable overlap between a w-type d—orbital of silicon and a nitrogen lone pair which is SP3 in character, so that the amount of s-character in a nitrogen hybrid does not necessarily have to increase with dw—pw bonding. Zuckermann's skepticism of the dv-pw bond is also supported by force constant data which depicts single— bond character for the Si-N and Si-P bond521:22 rather than double bond character. Finally Beagley, g£_al.23 has shown by electron diffraction data that the heavy atoms of tri— silylphosphine, (SiH3)3P and trisilylarsine (SiH3)3As are not coplanar as might be expected if appreciable dw-pw char- acter existed. The bond length was determined to be 2.25 R while the predicted single bond length calculated by adding the required Pauling covalent radii is 2.27 A. These ex- periments may suggest that the ability of phosphorus to donate its p electrons to atoms which contain empty d-orbitals (that is, silicon) is not experimentally verifiable. -f'rv.‘ 4-H- .. -."—' V' m - "6 Synthesis of Silicon-Nitrogen and Silicon—Phosphorus Compounds The most common method of establishing a bond between silicon and nitrogen is by/the action of ammonia or primary 'and secondary amines on halosilanes. Data which relate the comparative yields of silylamines produced when sterically hindered halosilanes and amines are allowed to interact with each other as a function of the halogen on the halosilane is not available; therefore, the differences in the reactivity EFFI_._‘«'*‘-‘ w— Of the halogens on the halosilane have not been established with any degree of certainty. From the data available, the bromo- and iodo—silanes appear to be more reactive toward a given amine than do the chlorosilanes.24 Because of their availability, however, the chlorosilanes are most frequently employed in synthesis. The halide ion produced during the reaction is precipitated as the amine hydrohalide. The re- action is reversible and the halosilane can be obtained from the silylamine and the amine salt.25 > R3SiNI'IR' + R'NHsCl RasiCl + 2R'NH2 <___ The product of the reaction of a monohalosilane with ammonia or an amine depends upon the electronic and steric nature of the reactants and products. A silylamine, disilazane or a trisilylamine may be formed. When the non—halogen substitu— ents attached to silicon are hydrogen, there is a definite tendency for complete silylation of the amine.25:27l28 For example, when silylchloride is treated with ammonia, tIiSiLlylamine can be obtained in 80 percent yield.26 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. When trimethyl— chlorosilane is allowed to react with ammonia, only hexa- methyldisilazane can be isolated.29:3°:31:32133 In con- trast to the results obtained with trimethylchlorosilane, the preference for formation of silylamine is observed in the reaction of triethylchlorosilane with ammonia. In this reaction, triethylsilylamine is the major product, the disilazane is the minor product.34135 with higher tri— alkylchlorosilanes, only silylamines have been reported as products.36137 The reaction of a dihalosilane with ammonia or an amine results in the formation of a silylamine, a silazane, or a polysilazane (see Table I) depending upon the nature of the reactants and the reaction conditions. With dihalosilanes in which there is little hindrance, the cyclosilazanes and the polysilazanes pre P(SlH3)3 + (H3S1)2PI + H3sipI2 + (S1H3)4PI The most convenient and general method of obtaining silicon-phosphorus compounds involves the reaction of alkali metal (lithium, sodium and potassium) phOSphides with halosilanes (chloro, bromo and iodo). There is not sufficient data available te determine which of the alkali metal phosphides is the most reactive to a given halosilane; however, the reactions of lithium phosphides with chloro- silanes are the ones most frequently used. This is probably because of the relative ease of obtaining the lithium phos— phides from butyl lithium. As already mentioned, the chloro— silanes are employed more frequently because of their avail- ability. A mixture of lithium phosphides are obtained when phos- phine is treated with butyl lithium according to the follow- ing equation:52 Ph3 + n—BuLi 35995» LiPH2 + LizPH + L13P 4". -§ fir“ w-u~_u.;-— 10 Wherl these lithium phosphides are allowed to react with tri— methylchlorosilane, the mono—, di- and tri—substituted silyl- phosphines are obtained.51 Cyclosilphosphanes are obtained when diethyldichlorophosphine, Etzsiclz, is added to the mixture of lithium phosphides.51 This reaction is described by the equation: (Phosphide mixture) + Etzsiclz EE%§£> ; i I; /P\ | Etzsi’// :::SiEt2 + EtZSi SiEtZSiEtz \ \. P P,//’ H The reactions of other alkaliphosphides with halosilanes are reported in Table II along with the products obtained from these reactions. Table II. Synthesis of silicon-phosphorus compounds from alkali metal phosphides and halosilanes. Alkali Halosilane Metal Products Ref. Phosphide MeasiCl LiPEt2 Me3SiPEt2 53a Mezsic12 LiPEt2 Me3Si(Cl)PEt2 + (EtzP)2SiMe2 53a Sicl4 LiPEt2 Si(PEt2)4 53b MeasiCl NaP¢2 MessiP¢2 54 HasiBr KPH2 P(SiH3)3 55 Me3SiF KPH2 (Measi)2PH, (Messi)3P 56 “ l . . ‘ gliv .0 11 Finally, silicon—phosphorus compounds may be obtained from the reaction of silyl-lithium compounds with halo phosphines.57 This method is limited, however, to triphenyl and other phenylated silyl—lithium compounds. (Whereas tri- alkylsilyl lithium compounds have at most only a transitory existence, the stability of phenylated silyl—lithium com- pounds has been attributed to delocalization of the negative charge over the aromatic portion of the molecule).58:59 The reactions involving silyl—lithium compounds are further complicated from the fact that only small yields are ob- tained. The silyl-lithium compounds are most easily pre— pared in solvents like tetrahydrofuran (THF) and other cyclic ethers.6° However, the silyl—lithium compounds react with the THF under the conditions at which the reaction is run.61 Therefore, the yield of the silylphosphine is reduced. Exchange Reactions of Amines with silylamines In 1949 Larson. et al?7demonstrated that N(triethyl- silyl)ethylamine underwent an exchange reaction when heated with primary amines according to the equation: EtasiNHEt + H2NR > Et3SiNHR + HZNEt Since Larsson's work, new silylamines have been prepared from readily available silylamines by this route. Abel and Bush62 found that symmetrically substituted ethylenediamines reacted with a big—aminosilane to give cyclic compounds ac— cording to the equation: .5; “'3... .WT fs' - ~ >_u- “gs-.- ' ' -..v‘._;.r .- . .~ 12 HRNCHZCHZNRH + MeZSi(NHEt)2 > HZC-——CH2 + 2H2NEt RN NR (R = C2H5) \Si/ Mez (I) The exchange of (I) with tetrakis—dimethylaminosilane led to the formation of a spirane according to the equation: Et Et I I ' CHz-N -CH2 2(1) + (MezN)4Si-———> I §:>Sf//N I + 2Me28i(NMe2)2 CH2 "' N—CHZ I I Et Et Fink63 demonstrated that heating RZSi(NHR)2 or allowingit to exchange with a primary amine in which R or R' are bulky groups led to a cyclodisilazane. 2RZSi(NHR')2 R28i — NR' or > | | + 2R'NH2 R'N - SiR2 2R2Si(NHR')2 + 2R"NH2 Fessenden64 showed that hexamethyldisilazane lost am- monia when refluxed with some secondary aliphatic and aromatic amines. In addition, the reversible nature of the reaction was demonstrated. This reaction is described by the equations: MeasiNHSiMe3 + ZNHBUZ _"_>’ 2MeasiNBuZ + NH3 Me3SiNBu2 + excess NH3 ggfisslz hrs > MessiNHSiMe3 + HNBu2 There appears to be an anomaly involved in the exchange reactions of silylamines with N-substituted anilines. It was 13 rePOrted that N—methylaniline did not exchange with hexa- methyldisilazane and that N—ethylaniline did not exchange with hexamethylcyclotrisilazane.65 In contrast to these reports, Tanslo66 reported that trig-(ethylamino)propyl- silane does exchange with N—methylaniline according to the following equation: g_-I>1:Si(NHEt)3 + 2PHMeNH > 9_—prSiNHEt(N1vlel>h)2 + 2H2NEt. The differences in these results have been rationalized by considering the extent to which the substituted aniline can approach the silicon-nitrogen bond.67 Klebe and Bush68 quantitatively measured the equilibria of substituted trimethylsilylacetanilides with acetamide. CH3 x C=O O O x 0 II I II II CH3CNH2 + .NSiMe3 < > CH3CNHSiMe3 + CH3CNH. When (X) was an electron withdrawing group, the equilibrium was shifted to the right; an electron donating group shifted the equilibrium to the left. The equilibrium constants were measured by means of 1H nmr spectroscopy. The equilibrium constants and rates of exchange for the reactions of aliphatic and aromatic amines with each of three dimethylaminosilanes (RMeZSiNMez, R = Me, Ph and CH=CH2) were also determined by means of 1H nmr spectroscopy.67 Table III gives some of the values of the equilibrium con- stants calculated for these reactions and Table IV gives some of the rate constants. 14 Tablxa III. Equilibrium data for the exchange at 38° of amines with dimethylaminosilanes Aminosilane Amine KC Me3SiNMe2 EtzNH 0.026 " (iso—Pr)2NH 0.034 " .Eggg-BuNHZ 0.17 " PhCH2(Me)NH 0.70 " mftoluidine 7.16 " pfitoluidine 13.3 PhMeZSiNMez EtzNH 0.05 " (igg-Pr)2NH 0.007 " PhCH2(CH3)NH 0.483 " EggnguNHz 0.117 " pftoluidine 16.9 " mftoluidine 7.3 ViMeZSiNMez EtZNH 0.047 " Eggt—BuNHZ 0.21 " mftoluidine 8.53 " pftoluidine 12.0 15 Tdblxa IV. Kinetic data for the exchange of amines with dimethylaminosilanes Aminosilane Amine ngp. k (1 mole_1 slc-l) MeasiNMez Eggt—BuNHz 38 6.4 x 10'5 " " (a) 38 1.9 x lo"5 " piperidine (a) o 1.5 x 10‘5 " " (a) 38 4.6 x 10'5 " aniline 38 5.0 x 10“4 viMezsiNMe2 Eggt—Buunz 38 7.8 x 10'5 aNo solvent. The study of these exchange reactions led to the fol— lowing experimental results. First the primary aliphatic amines exchanged more readily on silicon than secondary amines. Second, primary aromatic amines exchanged more com— pletely with a dimethylamino group on silicon than did ali— phatic amines. Finally, electron donating groups on the aro- matic ring were observed to increase the extent of the ex— change. These results have been rationalized by suggesting that for the aliphatic amines the steric effect of groups on the amine nitrogen as well as the base strength of the amines are important factors which effect the magnitude of the equi- librium constants. With regard to the aromatic amines, the results were rationalized by postulating dv=pv bonding of the aromatic nitrogen with the unfilled d orbitals of silicon. An interaction of this type would give a lower energy state 16 in the resulting silylamine. The effect that the groups on silicon had on the equilibrium constant or the effect of the solvent on the reaction was not completely determined. The kinetic study indicated that the aromatic amines ex- changed faster than the aliphatic amines with Silyldimethyl- amines. The reaction was observed to be second order over- all and was believed to be first order with respect to each reactant. A postulated four centered intermediate of the type shown below has been used to rationalize the kinetic data. =i_/CH3 _§ ?\\CH3 >13-.. With respect to the Silylphosphines there has been no reported attempt to study the exchange reactions of these compounds with amines or other phosphines prior to this work. Reactions of the Silicon-Phosphorus Bond The polarity of the silicon-phosphorus bond contributes to a high thermal stability of the bond but, at the same time, is responsible for its susceptibility to cleavage by polar compounds. Therefore, most reactions of silicon— phosphorus compounds with non-metal and transition-metal halides have resulted in the cleavage of the bond. However, there have been some examples reported in the literature in which stable adducts have been formed. 17 The silicon-phosphorus bond is cleaved by water to produce the corresponding phosphine. The alkaline hydrolysis of Silylphosphines has been used for the quantitative deter- mination of the phosphine group.69 (EtZP)4Si + 4H20 > Si(0H)4 + 4Et2PH OH Halogens and interhalogens caused cleavage of the sili- con-phosphorus bond to produce the corresponding organo- phosphorus and silicon halides.7O In the case of the inter- halogens the more electronegative halogen attached to silicon as eXpected- A typical reaction is illustrated by the following equation: thPSiMes + C12 > thPCl + MeasiCl Diphenylboron chloride and trimethylsilyldiphenylphos- phine produced diphenylborondiphenylphosphide according to the equation:70 thPSiMe3 + thBCl > PhZPBth + Me3Si c1 Iododimethylarsine caused cleavage of the silicon- phosphorus bond in the silylphosphine to release iodotri- methylsilane.7O Instead of the other expected product, diphenylphosphidodimethylarsine, tetraphenyldiphosphine and tetramethyldiarsine were obtained. Me3SiPPh2 + MegASI ——> Me3SiI + PhZPAsMe2-——> (MezAS)2 + (Ph29)2 ET??? “T“r- LZ‘P 18 Abel71 has found that certain compounds cleaved the silicon-phosphorus bond by insertion. The silicon-phos- phorus bond was cleaved by carbon dioxide according to the equation: 0 O II II Me3SiPPh2 + C02 —-—-—> Me3SiOCPPh2 or MeasiCOPth I II Two structures are possible for the product (I) and (II). The authors did not differentiate between the two alterna- tives by physical evidence, but favored structure (I) on the basis, of a C-O stretching frequency at 1682 cm—1 which agreed closely72 with the corresponding nitrogen com— pound analogous to (I). Carbon disulfide reacted similarly to carbon dioxide to produce the compound Me3SiSCSPPh2, the structure of which was not determined.71 Both ethyl and phenyl isothiocyanates reacted with tri— methylsilyldiphenylphosphine according to the equation:71 Me3SiPPh2 + RNCS -—> Me3SiN(R)CSPPh2(III) or Me381C(S)N(R)PPh2(IV) Two structures are possible for the products (III) and (IV) and both have been claimed for various insertions of iso- thiocyanates in similar systems.73 Structure (III) was the favored on the basis of results of methanolysis which pro- duced methoxytrimethylsilane and NH(R)C(S)PPh2. The NH linkage in HN(R)C(S)PPh2 was 'suggested by. the infrared spectrum. . 13...}.PHV’V‘fi’2-n 3!.— .. ‘ 19 The product of the interaction of trimethylsilyldi— phenylphosphine and ketene (see the equation below) showed a strong carbonyl stretching mode in the infrared spectrum at 1671 cm_1 indicating insertion of the carbon—carbon double bond rather than the carbonyl group. Measipph2 + CH2=C=O -—e Ph2PCH2(CO)SiMe3 or (V) Ph2P(CO)CH2SiMe3 (VI) Spectroscopic data was used to differentiate between the two possible structures (V and VI). The carbonyl stretch- ing mode which was observed at 1671 cm“1 is considerably higher than that reported79 for the monsilyl ketones (the carbonyl group in V would be essentially in a monosilyl ketone environment). Also, the chemical shift of the Me3Si protons at 0.00 ppm and the methylene doublet (JP H = 6.5 cps) virtually rule out structure (V). Although Silylphosphines do not appear to undergo in— sertion into the carbonyl group of ketene, they do insert under mild conditions into the carbonyl group of hexafluoro- acetone. The reaction is described by the equation: Me3SiPPh2 + (CF3)2CO > thPC(CF3)ZOSiMe3 The three possible structures for the product of this re- action are listed below. -ir__ in: ‘.. - bu. _ —.--. o-ui >-' [.w" 20 CF3 CF3 Ph2P - é - OSiMe3 pth-o-- é - SiMe3 CF3 CF3 VII VIII 0 CF3 thP - C - SiMeS CFa Ix L Nuclear magnetic resonance data (1H and 19F) indicated the presence of a mixture of two components. The minor product E in the mixture was believed to have structure (VII). The chemical shift of the Si(CH3)3 protons was virtually iden- tical to the chemical shift of -0.23 ppm reported75 for the very closely analogous compound HC(CF3)20SiMe3. These pro— tons showed the expected absence of coupling with fluorine and phosphorus nuclei. The nmr data indicated that the major product of the reaction was structure (IX) rather than (VIII). rThe 1H nmr spectrum showed P(V)-H coupling and a coupling constant of 21 cps was observed for the coupling between fluorine and phosphorus nuclei in the 19F nmr spec- trum. Structure IX contains phosphorus in the +5 oxidation state and also has interatomic distances short enought to agree with the nmr data; whereas, structure (VIII) is not compatible with the nmr data. The formation of structure (IX) was rationalized by suggesting that the ester (VIII) underwent an Arbuzov rearrangement to (IX). 21 The fission of the silicon—phosphorus bond by transition- metal halides produced phosphido-complexes of the metals in some instances and in others the metal halide was reduced to the metal. Similar results have been reported for the reaction of disilthianes with metal halides. Abel76 found that thiosilanes and disilthianes did not form stable co— ordination complexes with transitional and post-transitional metal halides, but invariably underwent fission of the silicon-sulfur bonds to give alkyl/arylmercaptides, sulfides or sulphonium derivatives of the metals. Copper (1) chloride, bromide and iodide was obServed to dissolve in trimethylsilyldiphenylphosphine, possibly to form some variety of phosphine—copper-halide coordination complex.7o Subsequent heating of the copper chloride solu- tion produced chlorotrimethylsilane and copper diphenyl- phosphide. The reaction is described by the following equa- tion: Ph PSiMe ) Cuc1(?) > [Ph PCu] + Me SiCl 2 2 n 2 n 3 Silver chloride, bromide and iodide were all instantly reduced to silver metal with the concomitant formation of tetraphenyldiphosphine and the corresponding halogenotri- methylsilane7° (see equation below). thPSiMea + AgX > Ag + Ph4P2 + Me3SiX (x = Cl,Br,I). Mercury(II) chloride, bromide and iodide were reduced to mercury with the formation of tetraphenyldiphosphine and i 4.14:..- .Ve 22 the trimethylsilyl halide.7o Nickel chloride cleaved the silicon—phosphorus bond in trimethylsilyldiphenylphosphine to produce nickel diphenyl- phosphide70 (see equation below). The molecular weight determination of the complex in benzene indicated a degree of polymerization of about 5—6, but the low solubility and the extreme air sensitivity of the complex made precise measurements difficult. Niel2 + thPSiMe3 > [(PhZP)2Ni]n + MeZSiCl The reaction between MeasiPPhZ and the pentacarbonyl halides of manganese and rhenium under mild conditions led to the known77 dimeric species [M(CO)4P(C6H5)2]2 73. The overall reaction is represented by the following equation: 2M(CO)5Br + 2Me3SiPPh2 ——> [M(CO)4PPh2]2 + 2MeZSiBr + 2c0 (M = Mn and Re) A repetition of this reaction under more vigorous con- ditions for prolonged periods resulted in the predominant formation of the trimeric, tricarbonyl species [M(CO)3PPh2]3. The reaction may be represented by the equation: 3M(CO)5Br + 3Me3$iPPh2 ——> [M(CO)3PPh2]3 + 3Me35iBr + 6co. Silylphosphines react with boron acceptors to form adducts of variable thermal stability. At high temperatures, the adducts decomposed to give halosilanes and phosphino— boron compounds. 23 Monosilylphosphine reacted with BF3, BC13,BZH6 and BZH5Br to form adducts which decomposed to give uncharac- terized glass-like polymers and monohalosilanes.79:3°:81 Table V gives the adducts which were formed, their formation and decomposition temperatures, and also the products from the decomposition reaction. Table V. Reactions of monosilylphosphine with boron tri- chloride, diborane, monobromodiborane and boron- trifluoride. H3SiPH2 + BX3 > H3SiPH2°BX3 > H3SiX + (HZPBX2)X A B C D E Adduct Formation Decomposi- Decomposition 3x3 (c) T (0c) tion T (00) products of Adduct of Adduct BF3 H3SiPH2'BF3 —134 -96 (HZPBF2)x (Hasi)3P BF3 BCl3 HasiPHz'BCla - 78 ~23 (H2P3c12)X BzHe H3SiPH2°BH3 - 78(pres.) —25(pres.) (HZPBHZ)x BszBr H3SIPH2.B2H5BI - 78 -45 SiH3PH2BH2Br BHZBr H3SiPH2'BH2Br - 45 —23 (HZPBH2)X Noth and Schragle82 investigated the adducts of boron accep- tors with trimethylsilyldiethylphosphine. The reaction of diborane and BF3 with this silylphosphine in ether gave 1:1 adducts which were thermally stable up to 80°. Above 80°, 24 the MeasiPEtz'Bfla adduct decomposed to give the trimer (EtzPBH2)3; the MeasiPEtz‘BFa adduct decomposed above 1000 to give a mixture of unidentified phosphinoborane polymers. similarly, BC13, BBr3, (i—Pr)ch1, (BuO)zBCl and (Me2N(2BCl formed adducts with MeasiPEtz which were stable at room temperature. The adducts decomposed above 1200 to the cor- responding trimethylhalosilane and dimeric diethylphosphino- borane. The reactions are described by the general equation: 2Me3SiPEt§BX2XE ——> 2Me3SiX' + [EtzPBX2]2 (x = Cl, Br, ipr, 0Bu, MezN; x' = Cl, Br) The tendency for formation of B-P addition compounds of the type RasiPR'sza has been arranged in the following order in the series of boron compound investigated on the basis of the inductive, steric and mesomeric effects of the substituents on boron.82 BBr3 > BCl3 “'BH3 > BF3 > (C3H7)2BC1 > [Me2N12Bcl With respect to the silyl substituent on the phosphine, the Silylphosphines formed adducts with boron Lewis acids under similar conditions to those needed for the formation of the unsilylated phosphine adducts. It has been suggested that the differences in their thermal stabilities results from alternative decomposition routes rather than from changes in donor strength.79 "‘k" .2- "-‘vw EXPERIMENTAL SECTION m Trimethylchlorosilane, phenyldimethylchlorosilane (Dow Corning Corporation), trimethylsilyldiethylamine (Aldrich Chemicals), diphenylchlorophosphine and mono— phenyldichlorophosphine (Alfa Inorganics, Inc.) were puri- fied by distillation. Phosphorusthiochloride (K and K Laboratories, Inc.), and tri-n—butylphosphine (Aldrich Chemical Company, Inc.) were used without further purifica— tion. Also anhydrous nickel bromide and iodide (City Chemical Corporation) and commercial methyl bromide (J. T. Baker Chemical Company) were used without further puri— fication. All solvents and liquid amines were dried over calcium hydride and distilled. The solid amines were re— crystallized from aqueous ethanol and dried in a desiccator over magnesium sulfate. Preparation of Trimethylsilyldimethylamine This compound was prepared according to the method of Beattie.83 According to this procedure, excess anhydrous dimethylamine was added to trimethylchlorosilane to give the desired product. 25 26 In a typical experiment, a 500-ml three—necked round-bottomed flask fitted with a condenser, stirrer and gas inlet tube was charged with 300 ml of xylene (distilled from sodium) and a 108.6 9 sample (1 mole) of trimethyl— chlorosilane. Anhydrous dimethylamine was bubbled into the solution at room temperature while stirring vigor— ously. When the mixture became too thick to stir, it was filtered to remove the solid dimethylamine hydrochloride. The solid was washed several times with xylene and the washings were combined and added to the original filtrate. Additional dimethylamine was added to the solution and the resulting mixture was again filtered. This procedure was repeated until dimethylamine hydrochlonuh no longer pre— cipitated. The material was then distilled at atmospheric pressure (746 mm) through a 4 ft Vigreaux column. The two best fractions (bp 80-860) were redistilled through a 3 ft spinning band teflon column. Characterization of Product The glpc suggested that the product was greater than 99% pure. It had the following properties: bp 83-840, (746 mm); 2352.1'3954' (The infrared spectrum exhibited a _1 _ very strong absorption at 1250 cm Wthh may be related . . - 94 to a symmetric C-H deformation characteristic of Sl-CH3. (See Figure 1, Appendix I) The 1H nmr spectrum of the liquid dissolved in benzene showed two absorptions, one of which was assigned to methyl protons on silicon at 27 6 = -0.33 ppm and the other assigned to methyl protons on nitrogen at 5 = —2.5 ppm. The ratio of methyl protons on silicon to methyl protons on nitrogen was 3.0:2.0. (See Figure 1, Appendix II.) Preparation of Phenyldimethylsilyldimethylamine The phenyldimethylsilyldimethylamine was prepared in a manner similar to that described for the preparation of trimethylsilyldimethylamine in the preceeding section and also similar to the method described by Roth.67 In a typical reaction, a one liter three—necked round—bottomed flask equipped with a condenser, stirrer and gas inlet tube was charged with 600 cc of pentane and a sample of 341.2 g (2.0 mole) of phenyldimethylchlorosilane. Dimethyl— amine was bubbled into the solution at room temperature while stirring vigorously. The solid dimethylamine hydro— chloride was filtered when the mixture became too thick to stir, and washed several times with hexane. The washings were added to the original filtrate and the solvent was then removed from the product by means of a rotating evapor— ator. The remaining material was distilled at a pressure of 30 mm through a 27 x 1.5 cm column packed with glass helicies. The fractions that were collected are listed below. 9—35.95... 28 Table VI. Distillation of phenyldimethylsilyldimethylamine Fraction Wt (g) (°C atB35 mm‘Hg) E35 2 1 24 103 1.4950 2 85 105 1.4958 3 101 105 1.4958 4 28 105 1.4949 Characterization of Product Glpc suggested that fractions 2 and 3 were pure. These fractions have the following properties: bp 105° (30 mm); 2?5 g 1.4958; £254 0.904. The 1H nmr spectrum showed two absorptions, one of which was assigned to methyl protons on silicon at 6 = —0.26 ppm and the other to methyl protons on nitrogen at 5 = —2.4 ppm. The ratio of the two different methyl groups was 1.0:1.0. Preparation of Diphenylphosphine Diphenylphosphine was prepared according to the method of Kuchen and Buchwald.54 In this procedure,sodium diphenyl- phosphide, prepared from diphenylchlorophosphine, was hydro- lyzed with ethanol to give the diphenylphosphine. Although the initially formed product was the tetraphenyldiphosphine, the phosphorus—phosphorus bond was cleaved by the action of excess sodium to give the sodium salt. A typical run is described below: "“.“‘"““"> ‘1’ 29 A 250-ml pressure-equalized dropping funnel and nitrogen inlet tube were inserted through one of the side arms of a carefully dried 500—ml three-necked round—bottomed flask. A stirring rod was inserted through the main mouth of the flask and the side arm was equipped with a water condenser (fitted with a drying tube) and a thermometer. The flask was flushed with a slow stream of nitrogen before and during 1. _.._....- the reaction and 5 g (0.217 mole) of sodium chunks and 200 ml of dibutyl ether were placed in the flask. While stir- 1111. £3343 ring gently, the contents were heated slowly to 105°. The heating mantle was then removed. When the temperature had fallen to 99°, the stirrer was stopped and the suspension was allowed to cool to 23°. The mixture was then heated again with stirring so that the solvent was refluxing (140°) and diphenylchlorophosphine (13 g, 0.059 mole) was added drop by drOp. After 4 hours, the suSpension was again cooled to 23° and transferred to another 500—ml three-necked flask equipped in the previously described manner. The suspension was cooled to 0° and a 35 ml sample of ethanol was added over a 15 minute period. The mixture was then refluxed at about 140° for 30 minutes. After this time, water was added until two distinct phases were present. The ethereal phase was separated in a 1-1 separatory funnel and dried with anhydrous calcium chloride. The ether was removed by means of a rotating evaporator and the remaining liquid was frac- tionally distilled at 5 mm pressure through a 15 x 0.5 inch column packed with glass helicies. The following fractions were collected: 30 Table VII. Distillation of diphenylphosphine. . B.P. 25 Fraction (°C at 5 mm Hg) p_ g 1 125—130 1.6247 2 132—135 '1.6273 3 135—136 1.6263 Characterization of Product All three fractions had similar properties; however, glpc suggested the purity of fraction three to be greater than 99%. This fraction had the following properties: bp 135°(5 mm); p?5 g 1.6263 [lit.54 bp 165° (25 mm), p?5 g 1.6263]. The infrared spectrum showed an absorption at 2350 cm_1 which may be related to a P-H stretching mode. (See Figure 2, Appendix I.) The 1H nmr spectrum of the sample dissolved in benzene showed a doublet which may be assigned to the proton on phosphorus (JPH = 168 cps) at 0 = —4.28 ppm. The spectrum also exhibited absorptions which may be assigned to the phenyl protons these absorbed in the same region as the solvent at about -7.2 ppm. Preparation of Monophenylphosphine Monophenylphosphine was prepared as described by Pass and Sschindlbower.84 This method involved the reduction of phenyldichlorophoSphine with lithium tetrahydroaluminate in 31 ether. The product was obtained after hydrolysis with 1:1 hydrochloric acid. An additional product, tetraphenylcyclo- tetraphosphine is obtained during the reduction if the addi- tion of the dichlorophosphine to the lithium tetrahydro- aluminate is too fast. This is because of the reduction of the unreacted dichlorophenylphosphine with the already formed monophenylphosphine. One mole (38 g) of lithium tetrahydroaluminate was added to 500 ml of ether in a dry 1—1 three—necked round- bottomed flask equipped with a stirrer, 500—ml pressure— equalized dropping funnel, nitrogen inlet tube and condenser. The flask was flushed with a slow stream of nitrogen and the contents were cooled to 5°. While stirring vigorously, a 304 g (1.7 moles) sample of monophenyldichlorophosphine diluted with 200 ml of ether was added in such a manner that a temperature of +50 was not exceeded. At the end of the addition, the reaction mixture was refluxed for 30 min— utes at 35°. The mixture was then cooled in an ice bath and as much 1:1 hydrochloric acid solution was added with stirring in order to give two distinct liquid phases. The two layers were separated in a 1-1 separatory funnel and the upper organic layer was dried over anhydrous calcium sulfate. The ether was removed by distillation at atmospheric pressure and the remaining product was fractionally dis— tilled at atmospheric pressure through a 15 x 0.5 inch in- sulated column packed with glass helicies. The following fractions were collected: I—mm ass-r. 32 Table VIII. Distillation of monophenylphosphine L. Fraction (0c .3547 mm 3.) 22° 2 1 50-70 1.5011 2 70-140 1.5149 3 140-152 1.5717 4 152-155 1.5760 5 155 1.5647 Characterization of Product ‘On the basis of the p?° 2g ir, 1H nmr spectra and glpc, fraction 4 was shown to be the most pure fraction. Frac— tions 3 and 5 were also shown to be mostly monophenyl— phosphine contaminated with impurities. Because all of the three fractions contained water, the product was allowed to remain in contact with calcium hydride for 48 hours. The final product had the following properties: bp 158-1590 (742 mm); 320 _1_)_ 1.5794 [lit.84 bp 160° (760 mm), _n_2° p 1.5794]. The infrared spectrum exhibited an absorption at 2350 cm_1 characteristic of a P-H stretching mode. The 1H nmr spectrum of the product dissolved in carbon tetrachloride exhibited a doublet which may be assigned to the protons on phosphorus (HHI‘ = 200 cps) at 0 = -3.95 ppm. 33 Preparation of Tetraethyldiphosphinedisulfide Tetraethyldiphosphinedisulfide was prepared by the re- action of ethylmagnesium bromide with thiophosphoryl chlor— ide.85 This compound is a very versatile intermediate for preparing compounds with two ethyl groups on phosphorus. Most other methods of preparation of such compounds give large amounts of mono- and tri-ethylated products. The grignard was prepared by reacting magnesium metal with ethyl bromide in ether. A 36 g (1.5 moles) sample of magnesium turnings was added to 500 ml of anhydrous ether in a 1—1 three—necked round—bottomed flask equipped with a mechanical stirrer, reflux condenser and dropping funnel. An 163.5 (1.5 moles) sample of ethyl bromide dissolved in 100 ml of ether was added drop by drop to the suspension of magnesium in ether with stirring. The reaction started spontaneously and the remainder of the ethyl bromide was added at a rate such that the solution boiled gently under reflux. The solution was then heated to 35° for one hour and cooled to 23°. A thermometer was inserted in the flask and an 86 g (0.5 mole) sample of thiophosphoryl chloride dissolved in 100 ml of ether was added drop by drop to the grignard solution over a period of 4—5 hours. The mixture was stirred and the temperature maintained at 22—230 by means of a water bath. A thick white precipitate formed during the course of the addition. The resulting slurry was heated to 35° for one hour to complete the reaction. The reaction products were . 4391..., 34 then carefully decomposed with as much 10% by volume sul- furic acid solution in order to give two clear phases. The top ethereal layer was separated from the aqueous layer in a 1-1 separatory funnel and dried with CaSO4. The ether was removed by means of a rotating evaporator. The residual oil which eventually solidified into colorless platelets upon cooling to 23° was suspended in 100 ml of methyl alcohol. The platelets were filtered and then recrystallized from a 1:1 acetone water mixture. The product was dried over phosphorus pentoxide in a vacuum desiccator. The crystal had a mp of 76-780 (literature 76-770)?5 Attempted Preparation of Trimgthylsilyldiethylphosphipg Trimethylsilyldiethylphosphine was prepared by a modi- fication of the method by Fritz and Poppenburg.52 This method involved reaction sodium diethylphosphide with trimethylchlorosilane in dibutyl ether. The sodium diethylphosphide was prepared as described by Issleib and Tzschasch.86 A 500—ml pressure—equalized dropping funnel and nitro- gen inlet tube were inserted through one of the side arms of a carefully dried 2—1 three-necked round—bottomed flask. A stirring rod was inserted through the main mouth of the flask and the other side arm was equipped with a water cooled reflux condenser. The flask was flushed with a slow stream of nitrogen and a suspension of 67 g (2.8 moles) of sodium sand in one liter of dried dibutyl ether was heated to the .. case—4' ‘an- 115-: 35 reflux temperature of the solvent (140°) and stirred by means of a mechanical stirrer. Tetraethyldiphosphinedi— sulfide (60.5 g, 0.25 mole) diluted with 500 ml of dibutyl ether was added drop by drop over a 2 hour period. After heating the mixture for a total of 16 hours, it was cooled to 23° and the suspension of sodium diethylphosphide and sodium sulfide were transferred in an inert atmosphere to another 2—1 flask. The excess sodium metal remained behind in the form of shiny lumps. In a manner analogous to that employed in the preparation of the sodium diethylphosphide, trimethylchlorosilane (108.6 g, 1 mole) dissolved in 100 ml of dibutyl ether was added drop by drop to the reflux- ing, stirred suspension of sodium diethylphosphide. After refluxing for an additional hour, the mixture was filtered and washed under nitrogen. The dibutyl ether was removed by distillation at atmospheric pressure. The remaining liquid was fractionally distilled at 10 mm pressure. The following fractions were obtained: (See Table IX.) Characterization of Product The boiling point, glpc, 1H nmr spectrum and index of refraction indicated that a pure fraction was not ob- tained. However, a fraction was obtained that had the fol— lowing properties: bp 50-530 (10 mm); 2?° p 1.4396 [lit.52 bp 70—72° (20 mm)]. This fraction is believed to have the following com— ponents; Me3SiPEt2, EtzPSH, and EtzPPEtz. The 1H nmr 36 spectrum was too complex to interpret because of coupling of the protons from the ethyl group with phosphorus. Table IX. Distillation of trimethylsilyldiethylphosphine. Fraction (°C 13de Hg) 32° 2 r 1 43-45 1.4238 ) 2 45-47 1.4263 E 3 47-50 1.4312 ‘ 4 50-57 1.4396 L 5 67-72 1.4430 6 83-85 1.4658 Preparation of Tetramethyldiphosphinedisulfide Tetramethyldiphosphinedisulfide, like the corresponding ethyl analog, is a valuable intermediate for the preparation of compounds containing two methyl groups on phosphorus. This compound was prepared by reacting methylmagnesium bromide with thiophosphoryl chloride.87 The methyl grignard was prepared by the reaction of methyl bromide in ether with magnesium turnings. A 2—1, two—necked round—bottomed flask containing 1 l of dried anhydrous ether was equipped with an inlet tube dipping below the surface of the ether and an outlet tube protected by a phosphorus pentoxide drying tube. The 37 flask and contents were then cooled in an ice-salt bath to 0°. Methyl bromide cooled to 20° was introduced through the inlet tube until approximately 400 g of methyl bromide was added. The flask was weighed again and the amount of methyl bromide was found to be 392.9 g. In a 3—1 three- necked flask equipped with a mechanical stirrer, a reflux condenser, and a 500—ml pressure-equalized separatory fun- nel, a 93.9 g (3.9 moles) sample of magnesium turnings was added together with 700 ml of anhydrous ether. While stir- ring, the methyl bromide was added slowly to the suspension of magnesium in ether. The reaction started spontaneously and the remaining methyl bromide was added at a rate such that the ether was gently refluxing. At the end of the addition,the solution was heated to 35° for one hour. A thermometer was inserted through one of the side arms of the flask and the solution was cooled to 0-5°. Thiophosphoryl chloride (184 g, 1.08 moles) dissolved in 100 ml of ether was added drop by drop over a period of five hours to the methyl grignard. A thick white precipi- tate formed during the course of the addition. After the addition, the resulting slurry was pmued into a 5-1 beaker containing approximately 600 g of ice. Sulfuric acid (10%) was then added with stirring over a period of 20 minutes. The product was filtered and the solid was washed with 4 l of cold water in 4 batches. It was then recrystallized from 3 l of ethyl alcohol and dried over phosphorus pentoxide in a vacuum desiccator. The final product (62.2 g, 62%) 38 obtained was a white crystalline solid: mp 225—2270 [lit§7 223-2270]. Preparation of Trimethylsilyldimgthylphosphing Trimethylsilyldiethylphosphine has been prepared by the reaction of lithium diethylphosphide with trimethyl- chlorosilane in ether solution.53a The lithium diethylphos— phide may be prepared by the reaction of an ether solution of phenyl lithium with diethylphosphine.88 However, the dialkylphosphines are most conveniently prepared by the re— at“ “t"?— duction of the corresponding tetraalkyldiphosphinedisulfides 89 An alternative with lithium hydroaluminate in ether. method for the preparation of trimethylsilyldimethylphosphine which eliminated the handling of the volatile dimethylphos— phine involved the preparation of lithium dimethylphosphide from tetramethyldiphosphine. The latter was prepared by the reduction of tetramethyldiphosphinedisulfide with tri- butylphosphine.9° The reaction of trimethylchlorosilane with lithium dimethylphosphide was most conveniently carried out in a vacuum system without solvent at —78°. A carefully dried 500-ml three-necked round—bottomed flask equipped with a thermometer, nitrogen inlet tube and magnetic stirrer was fitted to an insulated Vigreaux column 6 inches long. A standard distilling head with a thermometer and a 200 mm Liebig condenser was connected to the column. The receiving flask was a 250—ml two—necked round-bottomed flask which was equipped with a nitrogen inlet tube and a 39 vacuum adapter. The end of the vacuum adapter had a drying tube. The receiving flask was surrounded by a Dewar con- taining a slush bath of dry ice and isopropyl alcohol. Com- mercial tributylphosphine (81.52 g, 0.40 mole) and tetra— methyldiphosphinedisulfide (37.4 g, 0.20 mole) were placed in the distilling flask and the system was thoroughly flushed with nitrogen. It was very important to maintain 5 a positive pressure of nitrogen in the system throughout the ( distillation. The mixture was stirred for five minutes and ' slowly heated until the temperature of the distilling flask was about 250°. As the temperature rose, the mixture be- came homogeneous and the product boiling at 120° was col— lected; yield 19.7 g (81%, 0.16 mole). (See Figure 2, Appen— dix II for flralH nmr spectrum of this compound.) The receiving flask was flushed with nitrogen through the inlet tube on the receiving flask and stoppered. In a glove bag filled with nitrogen 2.2 g (0.32 mole) of lithium chips and 100 m1 of diethyl ether were added to the tetramethyldi- phosphine. The flask was removed from the glove bag and fitted with a condenser containing a drying tube while main- taining a slow flow of nitrogen throughout the system. The mixture was Stirred and heated for eight hours during which time a white solid was formed. The receiving flask was then fitted with a vacuum—stopcock adapter and attached to a vacu— um system. The ether was removed in vacuo and the flask was .surrounded by a Dewar of liquid nitrogen (—196°). Commercial trimethylchlorosilane, bp 57° (40 g, 0.37 mole, 47 ml) which 40 had been previously distilled, was then distilled in vacuo at —196° into the flask, and the system was then allowed to warm slowly to —1600 (petroleum—ether slush) and then again allowed to warm slowly to -78° (dry-ice—isopropyl alcohol). During this time, a slow reaction took place between the lithium dimethylphosphide and the trimethyl— chlorosilane. Periodically the stopcock on the reaction flask was closed and the flask was removed from the vacuum _“‘?’ system and shaken to insure complete mixing of the reac— 'r tants. After three hours, the system was slowly warmed to —450 (chlorobenzene slush) and all volatile products were distilled in vacuo to a trap held at -196°. The impure trimethylsilyldimethylphosphine which remained in the re- action flask with the solid LiCl was permitted to warm to room temperature and distilled in vacuo into a 100 ml flask held at -196°. Finally, the trimethylsilyldimethylphosphine trapped in the 100 m1 flask was distilled at 20 mm pressure in a nitrogen atmosphere to yield the following fractions: Table X. Distillation of trimethylsilyldimethylphosphine. Fraction B.P. (°C 20 mm Hg) _?5 2 1 26-28 1.4622 2 28—33 1.4622 3 33-33 1.4622 4 33-34 1.4622 41 Characterization of Product The 1H nmr and infrared spectra (see Figure 3, Appendix II and Figure 3, Appendix I) for fractions 1-4 were very similar, and the absorptions observed were identical to those that would be expected for the compound trimethylsilyl- dimethylphosphine. Elemental analysis and characterization I of the product by glpc were difficult to obtain because of ' the reactivity of the product to moisture and oxygen. The product collected (25.0 g, 58% yield) had the following F‘J' (.0 . properties: bp 33—34° (20 mm); p?5 Q_1.4622. The 1H nmr spectrum showed a doublet (J PH 4.5 cps) at 5 = -0.13 ppm which was assigned to the methyl protons on silicon and a doublet (J PH = 2.5 cps) which was assigned to methyl protons on phosphorus at 0 = -1.0 ppm. The infrared spec— trum showed a strong absorption at 1250 cm-1 characteristic -1 _ of Si-CH3 and a weak absorption at 420 cm which was as— signed to the Si-P stretch.91 The Cleavage of Tetrahydrofuran by Lithium Dimethylphosphide Issleib and Tzschasch have reported that lithium diethyl- PhOSphide, Li+[(CzH5)2P]_ decomposes when allowed to remain in contact for long periods of time with tetrahydrofuran to Give a yellow, uncharacterized product.92 In contrast to these results, Hewertson and Watson93 suggested that lithium diethylphosphide does not undergo a chemical reac— tion with tetrahydrofuran, but that the solvent causes 42 ionization of the Li-P bond. We have found that lithium dimethylphosphide, in apparent contrast to the results re— ported for lithium diethylphosphide, undergoes a fairly rapid, nearly quantitative reaction with tetrahydrofuran to give (CH3)zP-CHz-CHz—CHz-CH2-0_ Li+. This compound was not characterized, but further allowed to interact with tri- methylsilylchloride, (CH3)3SiCl, which gave (CH3)2P(CH2)4OSiMe3 in greater than 90% yield. In a manner almost identical to the procedure described for the preparation of lithium diethylphosphide, a mixture of 10.9 g (0.089 mole) of tetramethyldiphosphine and 1.84 g (0.267 mole) of lithium metal were refluxed in 100 ml of tetrahydrofuran. The color of the mixture became bright yellow. After heating for six hours, excess lithium was removed by filtration. A 28.9 g (0.267 mole) sample of tri- methylchlorosilane dissolved in 50 ml of tetrahydrofuran was then added drop by drop to the filtered mixture. This resulting mixture was then heated until refluxing for an additional 30 minutes to insure completion of the reaction. It was filtered and the solvent was removed by distillation at atmospheric pressure. The remaining liquid was frac- tionally distilled at 20 mm pressure. The following frac- tions were obtained: 43 Table XI. Distillation of product from the cleavage of tetrahydrofuran by lithium dimethylphosphide. Fraction B.P. °C (20 mm) 1 47-50 50—53 59-60 86—89 89—105 OSCDIBOJN 106-108 12.3““ “-72- Characterization of Product Fraction six was shown by glpc to be greater than 99% pure. The total weight of this fraction was 33 g (91% yield). The infrared spectrum displayed the expected ab- sorptions for the molecule (CH3)3SiO(CH2)4P(CH3)2. (See Figure 4, Appendix I.) A strong peak at 1250 cm_1 was assigned to a symmetric c-H deformation characteristic of Si-CHa. Another very strong peak at 1100 cm_1 was assigned to C—0 stretch characteristic of SiOC.94 other absorptions were recorded at (cm—1, intensity): 2950, 2850(vs); 2800, 2725, 1475, 1450(w); 1425(m); 1375(w); 1285(m); 1050(3); 1000(w); 965(m); 935, 900(5); 875, 850(va); 750(5); 710(m). The 1H nmr spectrum showed a peak at 0 = —0.08 ppm which was assigned to the methyl protons on silicon, a 44 doublet (J PH 2.5 cps) which was assigned to methyl pro- tons on phosphorus at 0 = —0.87 ppm, broad peaks at 5 = -1.42 ppm which were assigned to the methylene protons attached to phOSphorus and broad peaks at -3.58 ppm which were assigned to the methylene protons attached to oxygen. (See Figure 4, Appendix II.) Integration gave a ratio of the methyl protons on silicon to the methyl protons of phos— phorus of 3.0:2.0. The 31P nmr spectrum showed a single absorption at 5 = +53.6 ppm relative to H3PO4. Apal. Calcd for C9H23PSiO: C, 52.40; H, 11.16; P, 15.04. Found: C, 52.30; H, 11.08; P, 14.79. Molecular weight: theoretical, 206; found, 226. Although both SiOC and POC have absorptions at 1100 cm"1 the 31P nmr and 1H nmr showed definitely that the compound? was of the type RzPC instead of RZPOC. Compounds of the latter type, have 31P chemical shifts in the region of —100 ppm relative to H3PO4 and 1H chemical shifts at about —1 ppm with coupling constants of about 7 cps. Whereas compounds of the former type have 31P chemical shifts in the region of +50 ppm and 1H chemical shifts around —0.9 ppm with coupling constants around 2.8 cps.95,96 Therefore the band at 1100 cm”1 can be assigned to SiOC and indicates the formula of the compound to be (CH3)3SiO(CH2)4P(CH3)2. The reactions which were observed may be described ac- cording to the equations. (1) (CH3)2P—P(CH3)2 + 2Li + 2 CHz-CHZ -> 2 (CH3)2P(CH2)4O— Li+ CHE/CH2 45 (2) (CH3)2P(CH2)4O_ Li + (CH3)3SiCl > (CH3)2P(CH2)4OSi(CH3)3 + LiCl. The cleavage of dioxane by lithium diphenylphosphide has been suggested previously.97 In a similar manner, it has been shown that alkyl and aryl phosphides can cleave ethylene oxide.98 Preparation of Trippthylsilyldiphenylphosphine " "“7711. n.54‘ Trimethylsilyldiphenylphosphine has been prepared in yields above 60% by the reaction of trimethylchlorosilane with sodium diphenylphosphide in butyl ether.54 The sodium diphenylphosphide was prepared from commercially available diphenylchlorophosphine. Although the initially formed product was the tetraphenyldiphosphine, the phosphorus— phosphorus bond was cleaved by the action of excess sodium to give the sodium salt. a 250—ml pressure-equalized dropping funnel and nitrogen inlet tube was inserted through one of the side arms of a carefully dried 500-ml three-necked round—bottomed flask. A stirring rod was inserted through the main mouth of the flask and the other side arm was equipped with an Allihn Condenser fitted with a drying tube. The flask was flushed with a slow stream of nitrogen and a suspension of 15.22 g (0.65 mole) of sodium sand in 300 ml99 of p—butyl ether (distilled from sodium) was heated so that the solvent was refluxing and was stirred by means of a mechanical stirrer. Diphenylchlorophosphine (33 g, 0.15 mole, 26.8 ml) diluted 46 with 75 ml of butyl ether was added drop by drop over a 1— hour period. After heating the mixture for a total of four hours, it was cooled to room temperature and the suspension of sodium diphenylphosphide and NaCl was transferred under a stream of nitrogen to a 1—liter three-necked round-bottomed flask. The excess sodium metal remained behind in the form of shiny lumps. Using equipment similar to that employed in the preparation of sodium diphenylphosphide, trimethyl- chlorosilane (49.5 g, 9.45 moles, 59 ml) dissolved in 100 ml of butyl ether was added drop by drop to the refluxing, stirred suspension of sodium diphenylphosphide over a two— hour period. After refluxing for an additional hour, the mixture was filtered and washed under nitrogen by means of a filtering stick similar to the one described by Halah.1°° The solvent was removed by distillation at atmosphereic pressure, and the residual oil was fractionally distilled at 3 mm pressure in a nitrogen atmosphere. The following frac- tions were collected during a typical distillation: Table XII. Distillation of trimethylsilyldiphenylphosphine. Fraction B.P. °C, 3mm p?5 Q 1 137-141 1.6001 2 142-144 1.6008 3 144-145 1.6013 4 145-146 1.6025 47 Characterization of Product Fractions 1 and 2 (24.0 g, 62%) were shown to be pure by glpc and had the following properties: bp 126—1270 (1 mm); 3125 p 1.6001 [lit5.4 bp 126-1270 (1 mm) 325 2 1.6000]. The 1H nmr spectrum exhibited a doublet (J PH = 4.9 cps) which was assigned to methyl protons on silicon at 0 = -0.14 ppm. (See Figure 5, Appendix II.) The infrared spectrum showed a peak at 440 cm—1, characteristic of the silicon-phosphorus bond.91 Reaction of Trimethylsilyldiphepylphosphine with Nickel Iodide in Benzene All preparations were carried out under nitrogen. The anhydrous nickel iodide was dissolved slowly in a refluxing benzene solution of the trimethylsilyldiphenylphosphine (four moles of the ligand to one mole of the nickel iodide). The resulting solution was intensely colored. The excess nickel iodide was removed by filtration and the resulting complex was crystallized by the addition of p—hexane. A typical reaction procedure is described below. A 100-ml three-necked round—bottomed flask was equipped with a reflux condenser (fitted with a drying tube contain— ing phosphorus pentoxide), a nitrogen inlet tube and a stopper. The empty flask was weighed and a sample of 2.0 g (7.7 m moles) of trimethylsilyldiphenylphosphine was added to 50 ml of dry benzene after the flask had been flushed 48 Witll dry nitrogen. A sample of 0.59 g (1.9 m moles) of nidkel iodide was added and while stirring, the mixture was heated so that the solvent was refluxing (about 86°). The mixture turned green after 15 minutes and after heating the mixture for a total of three hours, it was filtered to remove excess nickel iodide. Green and brown precipitates formed when the solution was cooled and dry hexane added. The precipitate, now red, was washed several times with hexane and dried under vacuum. The filtrate, which was green, was cooled and a green precipitate formed. This precipitate was recrystallized from benzene and dried under vacuum. Finally, a third crop of crystals were obtained when the final filtrate was allowed to cool for 12 hours. Characterization of Product Three solids of different colors were isolated in the reaction of nickel iodide with trimethylsilyldiphenylphos— phine. Table XIII gives the reaction times, solvent sys- tems, mp, and analytical data for the different solids isolated. The red solid isolated was diamagnetic in the solid state and its infrared spectrum showed a band at 2350 cm—1 which may be assigned to a P—H stretching mode. Hayter1°1 obtained Ni12(¢2PH)2 by the reaction of NiI2 with 02m in benzene. This compound was also diamagnetic in the solid state and its elemental analysis was very close to that ob— tained for the red solid isolated in the reaction of N112 .Q’JPK .' I 49 wb.® wH.® wo.h Hb.® mn.v wo.v v.mv m®.mv whalvhfi NH cmmnw wo.h Hw.h be.» bo.oH mm.v Hm.m w.mv mv.om omHIme m 05am Mumn wo.h mH.oH bv.h wv.oH mm.w nm.m v.mv wo.®v oomlwmm m pom mafia ooamo bosom anmo bosom UonU bosom poamo chom Uono ocsom ma cofiu Uflaom EOUHHHm R HTMUHZ R monogamonm R smmouomm R Gonnmv R [060% MO HOHOU Nammflmmwz wo mmxmamEoo «HHZ mo QOHumHmmmnm mnu now want Hmucmeflnwmxm .HHHN manna 50 Witkl trimethylsilyldiphenylphosphine. The absorption spec- tra for the two compounds were also very similar. Therefore, it was concluded that the red solid isolated was the compound obtained by Hayter. The diphenylphOSphine resulted from the hydrolysis of trimethylsilyldiphenylphosphine. Similarly, the dark blue solid isolated was diamagnetic and its infrared spectrum exhibited a band at 2350 cm-1 characteristic of a P-H stretching mode. Its elemental analysis was in agreement with the analysis Hayter obtained for the five-coordinate compound Ni(HP¢2)3I2, C, 49.8; H, 3.8; Ni, 7.2; P, 10.8. These results suggest, therefore, that the dark blue solid was not the four-coordinate com— plex Ni(Me3SiP¢2)Iz or Ni(HP¢2)2I2 but the five-coordinate complex previously obtained by Hayter, NiIz(¢2PH)3. As with the red solid, it may be suggested that the diphenyl— phOSphine resulted from the hydrolysis of trimethylsilyl- diphenylphosphine. The characterization of the green solid will be dis— cussed in the next section. Reaction of Trimethylsilyldiphenylphosphine with Nickel Iodide in the Absence of Solvent Because the reaction of trimethylsilyldiphenylphosphine with nickel iodide in benzene presented difficulties with reSpect to hydrolysis, the reaction was carried out in a vacuum system without solvent. Trimethylsilyldiphenylphos— phine was added to the nickel iodide in a nitrogen atmosphere -l'E-VE-_‘lh‘.. " 1. 1n.— . E3. .0 51 andl the mixture was allowed to react in vacuo for eight hours. The volatile materials were removed by distillation in vacuo and collected at —196°. The green residue which resulted from the reaction was recrystallized from benzene. In a typical reaction, a 100—ml three—necked round— bottomed flask equipped with standard tapered joints was fitted with a vacuum—adapter and stoppers. All the glass- ware was previously baked in an oven at 150° for 24 hours before use. Nickel iodide (0.59 g, 1.9 mmoles) was added in an inert atmosphere to the flask and the flask was then connected to a vacuum system and evacuated for 24 hours. The flask was then transferred to a glove bag containing a nitrogen atmosphere and 2.0 g (7.7 mmoles) of trimethyl— silyldiphenylphosphine was added. The flask was again con— nected to the vacuum system. The mixture was stirred by means of a magnetic stirrer while the volatile products were distilled in vacuo to a trap held at -196°. The mix— ture changed to a green slurry after it had been stirred at room temperature for eight hours. The flask was trans— ferred from the vacuum system to a glove bag and one of the stoppers was replaced with a nitrogen inlet tube and the adapter with a drying tube filled with phosphorus pentoxide. The flask was then removed from the glove bag and a nitro— gen atmosphere was maintained in the system throughout all succeeding operations. The green slurry was dissolved in 15 ml of dry benzene and filtered to remove excess nickel iodide. Dry hexane (10 ml) was added to the filtrate and 52 tkms solution was cooled in an ice bath. Dark green crystals precipitated from the benzene-hexane solution. These crys- tals were collected, washed with cold hexane and dried under vacuum for 12 hours. Characterization of Product The green solid isolated in this experiment was the F‘ same as the one isolated in the preparation of the complex ( in benzene. The elemental analysis obtained for the complex L was as follows: L Calcd for NiIz[Me3SiP®2]2: C, 43.4; H, 4.59; Ni, 7.08; Si, 6.78. Found: C, 45.69; H, 4.08; Ni, 6.71; Si, 6.18. The complex dissolved readily in benzene and dichloromethane to give green solutions. These solutions decomposed on standing in solvent for long period of time. The complex was not readily oxidized, but the ligand (Me3SiP¢2) hydro— lyzed in the presence of moisture. Magnetic measurements using Hg[Co(NCS)4] as a standard indicated that the complex was diamagnetic in the solid state. The infrared spectrum of the complex taken as a nujol mull showed a strong ab- sorption at 1250 cm-1 characteristic of Si-CH3. Also, the absence of a peak at 2350 cm-1 suggested that the diphenyl- phosphine complex was not present. The electronic spectra were taken in benzene solution in the region of 300—1000 mu. The solution spectrum for the green solid was nearly .0 .0 53 ideultical to the solution spectra observed for the cor— responding complexes of dibutylphenylphosphine and diphenyl— phosphine. (Table XIV). The spectrum consisted of: (If a very intense absorp- tion band at 380 mu which was assigned to 3d-4p transitions.”2 (2) A very intense band at 460 mu and (3) a band at 560 mu which was probably a ligand field band. Table XIV. Absorption spectra (1 in mu) for NiIsz complexes Complex A Amax (mu) 8 Ref (Et3P)2Ni12 373 4690 103 459 2900 610 485 (BuzPhP)2Ni12 370 4040 102 600 490 (ButhP)2NiIZ 320 4350 102 400 3460 500—600 sh 925 374 (HthP)2NiI2 380 4840 101 460 2340 550 490 (MessithP)2NiIZ 380 4873 460 2352 560 489 54 TY“: small amount of volatile material isolated from the re— action Was characterized by 1H nmr and infrared spectroscopy. The 1H nmr spectrum showed an absorption at 0 = —0.80 ppm which may be assigned to methyl protons on silicon. The ir spectrum showed an absorption at 1250 cm_1 characteris- tic of SiCH3 and a peak at 320 cm—1 characteristic of 81—1. This compound was not very stable. It appeared to be de- composed by light. These results indicated that the volatile material isolated from the reaction of trimethylsilyldi— ‘Tf—t—‘E—‘fi'; flr—— phenylphosphine with nickel iodide was trimethylsilyliodide. The analytical data for the solid complex combined with the characterization of the small amount of volatile material as MeasiI are consistent with a nickel(II) complex of the formula NiI2[MeasiP¢2]2. Analytical data for carbon and hydrogen may be rationalized by suggesting the presence of some Ni(P¢2)2(Me38iP¢2)2. This may be due to an elimina- tion reaction of the type observed by Issleib1°4 according to the following equation: 4M€381P¢2 + N112 __9 ZMessiI + Ni(M63SiP¢2)2(P¢2)2 The absorption spectrum of this complex is consistent with a square—planar structure. Particularly, no absorption was observed in the spectrum in the region of 900 mu (emax = 300-4000) characteristic of tetrahedral complexes of the . 102 type N112(BuPh2P)2. .0 .0 55 Rgaction of Trimgthylsilyldiphenylphosphine with Nickel Bromide in Benzene The reaction procedure was the same as the one described previously for the reaction of nickel iodide with MeasiP¢2 in benzene. The following is a typical experiment. A 100—ml three—necked round-bottomed flask fitted with a reflux condenser, nitrogen inlet tube and a stopper was charged with a 0.51 g (2.3 mmoles) sample of nickel bromide and a 2.33 g (9.23 mmoles) sample of trimethylsilyldiphenyl— phosphine and then heated to the reflux temperature of the solvent (about 86°). The nickel bromide dissolved slowly in the benzene and a green color appeared immediately. After heating the mixture for two hours, it was filtered in an inert atmosphere to remove excess nickel bromide. The green filtrate was cooled in an ice bath to 0° and hexane was added. The green crystals which precipitated were collected on a filter and washed with hexane. The crystals were dried over phosphorus pentoxide in a vacuum desiccator. The fil- trate was concentrated, cooled and hexane was added. The brown solid which precipitated was collected on a filter and dried under vacuum. Reaction of Trimethylsilyldiphenylphosphine with Nickel Bromide in the Absence of Solvent This reaction was analogous to the corresponding re— action with nickel iodide without solvent. In a typical experiment a 0.505 g (2.31 mmoles) sample of nickel bromide 56 was added to a 100—ml three—necked round-bottomed flask (previously dried in an oven) fitted with stoppers and a vacuum stopcock-adapter. The flask was connected to the vacuum system and evacuated for 24 hours. Trimethylsilyl— diphenylphosphine (2.382 g, 9.24 mmoles) was then added to the flask in a glove bag filled with nitrogen. The system i.-. was evacuated again and stirred by means of a magnetic stirrer for 24 hours. The volatile material from the re— action was distilled in vacuo and trapped in a flask held ; at —196°. A green slurry remained in the reaction flask. L1 One of the stoppers and the vacuum stopcock-adapter were replaced by a nitrogen inlet tube and a drying tube filled with phosphorus pentoxide. A nitrogen atmosphere was main— tained in the system throughout the remainder of the reac— tion. The green solid remaining in the reaction flask was dissolved in 50 ml of benzene and filtered to remove excess nickel bromide. Hexane (15 ml) was added to the benzene solu— tion and it was then cooled in an ice bath to 0°. The green solid which crystallized was collected on a filter and dried under vacuum. Characterization of Product The two solids isolated in the reaction of trimethyl— silyldiphenylphosphine were soluble in benzene and methylene chloride. They appeared to be analogous to the corresponding nickel iodide complexes (MeasiP02)2NiI2 and (HP¢2)2NiI2. 57 The infrared spectrum of the brown solid taken as a nujol mull showed an absorption at 2350 cm"1 characteristic 1 was not observed of P—H. A strong absorption at 1250 cm— suggesting that Me3SiP¢2 was not present in the complex. The ir data may be rationalized by suggesting that the ligand, MeasiP¢2, was hydrolyzed to HP02. Therefore, the brown solid is suggested to be the five—coordinate diphenyl— phosphine complex (Ni(HP¢2)3Br2) since this was the only product Hayter1°1 obtained from the reaction of diphenyl- phosphine with nickel bromide. The properties of the green solid were similar to those of the corresponding nickel iodide complex. The infrared spectrum showed a Si-C stretching mode at 1250 cm—1 characteristic of Si—CH3. The spectrum did not show an absorption in the region of 2350 cm—1 suggesting that diphenylphosphine was not present. The solution spectrum taken in benzene in the region of 300—1000 mu showed two bands: (1) a very intense band at 400 mu and (2) a band at 560 mu. This was very similar to the spectrum of the square—planar complex, Ni(Bu2¢P)2Br2, prepared by Venanzi.102 The magnetic measurement indicated that the complex was diamagnetic in the solid state. The small amount of volatile material from the reaction was characterized by 1H nmr and infrared spectroscopy. The 1H nmr spectrum showed an absorption at 0 = —0.62 ppm which was assigned to methyl protons on silicon. The ir spectrum 1 showed an absorption at 1250 cm“ characteristic of SiCH3 and a peak at 410 cm_1 characteristic of Si—Br. These Iesnlts suggest that an elimination reaction had taken place to give the compound, Ni(Me35iP¢2)2(P¢2)2. Therefore, as with the nickel iodide complex, it is suggested that the complex is the square—planar Ni(MeasiP¢2)2Br2 with a small amount of the elimination product [Ni(P¢2)2(MeasiP¢2)2] present. Silicon Analysis pg Porcelain crucibles were fired in a muffle furnace at 700° to constant weight and stored in a desiccator over Mg(ClO4)2. Weighed samples (0.2-0.3 g) of the silylamine 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° over a period of 12—16 hours. The crucibles and contents were removed from the furnace and stored over Mg(ClO4)2 until cool,and reweighed to obtain the weight of Sioz formed. The per cent silicon was calculated using the expression 0.467 x wt SiO ' = __—_._—__l % 81 sample weight X 100' Nitrogen Analysis The silylamines were analyzed for nitrogen by the non— aqueous titration method of Fritz.1°5 Weighed samples (0.2-0.3 g) were placed in 30 to 50 cc of glacial acetic ! 59 :5 acid and several drops of a methyl violet indicator solution added.* This solution was then titrated with a solution of HClO4 in glacial acetic acid previously standardized with potassium hydrogen phthalate. The amount of nitrogen was calculated by means of the expression Vol. HClO4(l) x Norm. of Hc104 x 14.01 % N = Sample Weight X 100. Nickel Analysis The procedure described by Coskran1°6 was used to ob— tain the nickel analysis. A sample of 30-50 mg of the nickel compound was placed in a glass stoppered bottle and 5 1 ml of distilled water was added. The solution was slowly decomposed by 5 ml of concentrated HNO3, evaporated to near — dryness and then cooled slightly. A 70% solution of HClO4 was added slowly and the solution was again evaporated nearly to dryness. Finally, the solution was diluted with 70 ml of water, filtered and then washed with 30 ml of water. The pH of the solution was adjusted to 6-7 with NH4OH and an ethanol solution consisting of 1% dimethyl— glyoxime (DMG) was added. The volume of DMG added was determined by the relation that 1 ml approximately equals 0.0025 g of Ni. A 2 ml excess of DMG solution was added to insure complete precipitation. The solution was then *A solution of approximately 0.1 g of methyl violet in 10 cc of chlorobenzene. heated so that it was almost boiling and the pH adjusted to 9—10 with NH4OH. The solution was then allowed to stand at 23° for one hour after which time it was filtered and washed with cold water. It was also washed with 50% ethanol in case too large of an excess of DMG was added. Finally, it was dried at 110° for 1 hour. The per cent nickel was calculated using the expression. 0.2032 x wt. of Ni(DMG)2 % N1 = sample weight Infrared Spectroscopy Infrared spectra were obtained as pure liquid films, CC14 solutions and nujol, hexachlorobutadiene and flurolube mulls between Nacl or KBr discs. They were also obtained on gaseous samples and KBr pellets. The instruments used included the Unicam S—P-2003 Perkin-E1mer—237B and 301 spectrophotometers. Nuclear Magnetic Resonance Spectroscopy The 1H nmr spectra were taken in approximately 10% solutions, with benzene as solvent and tetramethylsilane as an internal reference. The spectra were obtained on the Varian HA-100 and A—60, and Jedkn C—60 spectrophotometers. The 31P nmr spectra were taken in approximately 50% solu— tions, with benzene as solvent and 85% H3PO4 as an external reference on a Varian HA—100 spectrophotometer at 40.4 Mc. Va or Phase Chromato ra h (Glpc) The VPC analyses were obtained with an Aerograph A—90-P and F & M 810 research chromatographs with helium as the carrier gas. Each instrument was equipped with a 5 ft 20% SE—30, 60/80 chromosorb wax column. Ultraviolet and Visible Spectrophotometry The absorption spectra were obtained with a Cary Model 14 spectrophotometer in benzene solutions at 25°. Magnetic Moment Measurements The magnetic susceptibilities of the finely divided solids were measured at room temperature on a Gouy—type balance using Hg[Co(CNS)4] as a standard. Eguilibrium Reactions of Trimethylsilylphosphine With Amines Method of Conducting Eguilibrium Experiments.- All operations were conducted in a dry glove bag filled with nitrogen. All equipment and glassware were baked in an oven at 120° for 48 hours to eliminate water. A sample of trimethylsilylphosphine was added to a previously weighed nmr tube. Benzene was added to make the mole frac— tion percentage of the silylphosphine equal to 10. A sample of the amine was added and the nmr tube, which was then sealed with a nmr tube cap, was shaken thoroughly. After standing for a minimum of 24 hours at room temperature, t a; the proton nmr spectrum was recorded at a sweep width of 250 cps with a Varian A-60 nmr spectrometer. An HA—100 Varian nmr spectrometer was used to record the spectra with peaks having chemical shifts too small to be resolved by the Varian A—60. The spectrum was recorded six times to eliminate errors in the instrument and the peaks represent- ing the various species were integrated with a planimeter in order to determine the relative amounts of each moiety 9 present. The area of the peak in question was taken as an average of the six trials. This procedure was repeated twice to insure reproducible results. A typical example of a determination of an equilibrium constant follows: Reaction of Trimethylsilyldiphenylphosphing with NeMeth lbenz lamine CH3 I MeasiP¢2 + H-N — CH2 —. l 1 CH3 l Measi—N‘CHZ + HP¢2 - This reaction was run using relative concentrations of amine to silylphosphine of 1:1 and 2:1. In each case 1 mmole of silylphosphine was used and its concentration in the solution was adjusted to 10 mole per cent by adding the appropriate amount of benzene. The 1H nmr spectrum of the compounds in the reaction mixture was recorded on a Varian A-60 nmr spectrophotometer at a sweep width of 250 cpS. The separation of the HNCH3 from the SiNCH3 peak was 5 cps and the separation of the HNCH2 peak from the SiNCH2 peak was 16 cps. Therefore, the amounts of the amine moieties present in the equilibrium mixture could be deter- mined directly by relating the areas under the HNCH2 and SiNCHz peak to the number of moles of each one present. The areas under the HNCH3 and SiNCH3 peaks were also meas- ured and related to the amounts of the amine moieties present as a method of checking the results obtained from integration of the previously mentioned peaks. The amount of diphenylphosphine present at equilibrium was equal to the amount of the silylamine present at equilibrium. The initial number of moles of the free amine was related to the total area of the two amine peaks (HNCH2 and SiNCHz). The initial number of moles of the silylphosphine was re— lated to the initial amount of free amine by the stoichiometry of the reaction. Therefore, the equilibrium number of moles of silylphosphine present was equal to the initial amount present minus the number of moles of silylamine formed. Figure 6, Appendix II shows the 1H nmr spectra of CH3 TMS + H-N — CH2-. and Figure 7, Appendix II the 1H nmr spectrum of CH3 I MessiP¢2 + H—N — CH2. + TMS. The 1H nmr data for other equilibrium reactions are given in Table XV. s W, .A J _. Table XV. . . / RZNH + Me3S1PRé = Me3siNR2 + HPRé (R 1H nmr data for the exchange of trimethylsilyl- diphenyl and trimethylsilyldimethyl phosphines with amines. (Recorded on the Varian A—60 spectrometer.) Ph or Me) HNR2 ySiNR2 0(SiNR2)-0(HNR2) (ppm CH3 CH3 HN-CHzPh SiN—CHzPh 0.08 (Figure 7, Appendix II) CH3 CH3 HN-CHzPh SiN-CHzPh 0.27 HN(CH3)2 SiN(QH3)2 0.17 (Figure 8, Appendix II) H H2Nc(CH3)3 l SiNC(CH3)3 HH II SiNC MeasiNR; + HPRZ (R=Ph or Me) Second, the reactions of phosphines with silylamines were investigated to determine if the reactions described by (1) could be run reversibly. Third, quantitative measure— ments of the equilibrium constants for the reactions of ali— phatic and aromatic amines with Silylphosphines were taken. Fourth, interactions of nickel iodide and bromide with (I) were examined to determine if stable adducts could be ob— tained. The phosphines and silylamines used in these investi- gations were prepared by well established procedures which require no additional elaboration. However, the syntheses of Silylphosphines have not been investigated as thoroughly. As a result of these studies, the following conclusions have been reached concerning the preparation of silylphos- phines with phenyl or alkyl substituents on phosphorus. 69 70 “ENE alkyl or phenyl substituted phosphinosilanes are best prepared by the action of a chlorosilane on an alkali metal phOSphide rather than by the reaction of a chlorophOSphine with a silyllithium compound. Non-cyclic ethers are preferred as solvents for the preparation of the alkali metal phosphides. Cyclic ethers, for example tetrahydrofuran and dioxane, enhance the forma— L tion of the alkali metal phosphides compared to non-cyclic 1 3,—1- -‘ l - av: - ethers; however, the cyclic ethers are also cleaved by the metal phosphides under the conditions at which the silyl— 'wrer phosphines are prepared. The Silylphosphines (I) and (II) reacted with several selected aliphatic and aromatic amines to give the corre—- sponding trimethylsilylamines and diphenylphosphine or di- methylphosphine respectively. These products were not isolated but were identified by comparing the values of their 1H nmr chemical shifts and coupling constants with those values of the known compounds. The reversibility of the exchange reaction was con- firmed by the fact that dimethylamine and a silylphosphine could be identified as products when trimethylsilyldimethyl- amine was allowed to react with a phenylphosphine according to equation (2) (2) MeasiNMez + HPPh2 > Me3SiPPh2 + HNMez. Although these exchange processes favored the formation of the silylamine and phosphine, the reaction was shifted to 71 the lright to favor the silylphosphine and dimethylamine by removing the volatile amine by distillation in vacuo. The data for the equilibrium studies can be found in Tables XVI and XVII. In all the exchange reactions in- vestigated in these studies, the substituents on silicon remained constant and the solvent was benzene. Thus, ‘flflw neither the effect of the groups on silicon nor the effect of the solvent on the reaction was determined. 1 .‘u‘ .1. .1 . . . " .‘g ..1 l The substituents on phosphorus and nitrogen were varied; therefore, the equilibrium constants could be de— termined as a function of the nature of the groups on these elements. The effects that the groups on nitrogen had on the position of equilibria in the reactions described by equation (1) for Silylphosphines (I) and (II) were similar and will be discussed first. of the factors which may be considered, we will focus our attention on two, the steric and electronic effects of the substituents on the amine. Primary aliphatic, secondary aliphatic and primary aromatic amines were used in the exchange reactions. For the ali— phatic amines, the primary amines exchanged more completely than the secondary amines. ~Therefore, the steric effect is an important factor that appears to influence the position of equilibria. Factors that make the electron pair on nitrogen more available for dw-pw bonding, such as the in- ductive effects of the alkyl amine groups or conjugative effects in substituted anilines may also influence the position of equilibria. Table XIX gives the basicity 72 I Tablxa XIX. Basicity constants of organic amines in water at 25° 67 - Aming_ pr Kb (iso—Pr)3NH ' 1.95 1.1 x 10'3 EthH 3.02 9.6 x 10" MezNH 3.28 5.2 x 10" I tert-BuNH2 3.55 2.8 x 10“ 3‘ N-benzylmethylamine 4.42 3.8 x 10‘5 3 l ‘p-toluidine 8.92 1.2 x 10'9 l} m—toluidine 9.31 4.9 x 10‘10 g—toluidine 9.60 2.5 x 10'10 constants for some amines in water. These constants repre— sent the ability of the amine to donate their lone pair of electrons to hydrogen ions. Therefore, we assumed that the ability of amines to donate these electrons to d or- bitals of silicon was proportional to the basicity of the amines. So, the stability of the silylamine should increase as the basicity of the amine increases. This effect, in turn should lead to the equilibria described in equation (1) being shifted to the right, and a higher value of the equi— librium constant should result. For the aliphatic amines, the secondary amines are more basic than the primary amines. According to basicity arguments, this would suggest that the secondary amines should exchange more completely than primary 73 amines“ However, the reverse is true. For example, di— isopropylamine is the most basic of the aliphatic amines but exchanged to a smaller extent than all others. These results are rationalized by suggesting that the steric nature of the amines influences the equilibrium constants to a greater extent than do the factors which make the elec- tron pair on nitrogen more available for dv-pw bonding. FL O?fi »‘ -. The results from the equilibrium studies of the aromatic amines are more difficult to rationalize. The results from the equilibrium studies of the aliphatic amines suggested lfnh‘o :L‘.’ that the basicity of the amines influenced the position of equilibria only to a minor extent. Consequently, if only steric factors are considered, the values of the equilibrium constants for the primary aliphatic and aromatic amines should be similar because the steric nature of the two amines are similar (CH3-. —NH2 and (CH3)3CNH2). However, the differences in the basicity between the aromatic and primary aliphatic amines Kb tert-BuNH2 2.8 x 10-4 —-———-Kb 71 a = ~ ~ -9 z 105 E 0‘11 me 1.2x10 is greater than the difference in the basicity between the secondary and primary aliphatic amines Kb diisopropyl amine = 1.1 x 10-3 Kb tert-BuNH2 2.8 x 10" < 10 . Therefore, if basicity is a factor, the difference between the equilibrium constants obtained with primary aromatic and primary aliphatic amines should be greater than the 74 difference between the equilibrium constants obtained with two different aliphatic amines. Thus, if both basicity and steric factors are considered, the primary aliphatic amines would be expected to have larger values of the equilibrium constants than those of the aromatic amines. However, the values of the equilibrium constants for the aromatic amines are found to be approximately equal to those for the ali— I) phatic amines (Kc for pggp—BuNHZ = 6.50, Kc for pftoluidine g I = 7.01). The values obtained may be attributed to the extra E stability of the aromatic aminosilane compared to the §' 5. primary aliphatic aminosilane from dv-pw contributions and the further stabilization of the Si-N bond in the aro- matic aminosilane from the conjugation of this bond with the aromatic ring H I Measi 2:. N- G -CH3 . This rationalization is based on the assumption that the Si-N bond in the aromatic aminosilanes does have some dw-pw character even though the lone pair of electrons on the aromatic amine has a smaller tendency to be donated to a silicon d orbital than does the lone pair of electrons on a primary aliphatic amine. The net effect of this con- jugation is to apparently make the aromatic aminosilane as stable as the primary aliphatic aminosilane and in turn - make the values of the equilibrium constants for the two amines equal. 75 The values of the equilibrium constants obtained for pftoluidine are low compared to those for the aromatic amines. This effect ("ortho effect") has been observed in the determination of the effect of substituents on the aromatic ring in aromatic acids and amines. This effect is not well understood but it has been suggested that it has to do with the nearness of the groups involved, but is more than just steric hindrance.107 The equilibrium constants are thermodynamic functions which reflect only the energies of the initial and final states; thus, the differences in the values of the equi— librium constants are due chiefly to differences in stabili- ties of the silylphosphines and the silylamines. In other words, the more stable the silylamines relative to the silylphosphines, the larger the values of the equilibrium constants for reaction described by equation (1). The change in the values of the equilibrium constants as a function of the change in the stability of the silylamines (keeping the stability of the silylphosphine constant) were discussed previously. The change in the equilibrium constants as a function of the change in the silylphosphine will now be discussed (the stability of the silylamine is held constant). A comparison of the values of the equi- librium constants in Table XVI to those in Table XVII shows that for any given amine the values of the equilibrium constants for the phenyl phosphinosilanes are larger than those for the methylphosphinosilanes. These results are .r ,_‘__.,( *4 76 rationalized in the follOWing manner: Obviously, the efr fects of the substituents on phosphorus are responsible for the different stabilities of the silylphosphines. The ef— fect may be a steric factor or an electronic effect or both. Rationalizations based on steric effects appear to be in agreement with the experimental determinations of the values of the equilibrium constants. The phenyl groups are larger than the methyl groups; therefore, on the basis of steric effects it would be expected that the phenyl phosphinosilanes would be less stable than the methyl phosphinosilanes. Be- cause the values of the equilibrium constants are larger in the case of the phenyl phosphinosilanes, arguments based on steric factors appear to be valid. In order to determine the electronic effects of the substituents on phosphorus, the base strengths of the pho- phines must be determined. The ability of phosphorus to donate its electrons into d orbitals of silicon may be related to the basicity of the phosphine using arguments analogous to those employed in the discussion of 3 bond— ing of amines. The dissociation constants of some phosphines are listed in Table XX.108 The table does not give the value of the pK for di— phenylphosphine; however, it can be seen from the table that the methyl phosphines have larger pK values than the phenyl phosphines. Therefore, it may be reasonably assumed that dimethylphosphine has a larger pK value than diphenyl- phosphine. This conclusion about the basicity of '1‘. II 7. 1 77 Table XX. Dissociation constants of some organic phosphines in ethanol—water mixtures Phosphine pKa Me3P 7.2 Et3P 6.7 MezPhP 4.2 EtzPhP 4.0 Ph3P 2.6 Me2PH 3.9 + + aR3PH = R3P + H . diphenylphosphine suggests that the dimethylphosphine group would favor donating its lone pair of electrons into d orbitals of silicon to a larger extent than the diphenyl— phosphine group. This factor by itself would cause the methylphosphinosilane to be more stable. But, if the Si—P bond in the methylphosphinosilane has w-bond character, it would be expected that the Si-P bond in the phenylphosphino— silane would also have v—bond character but to a lesser degree because the phenyl groups on phosphorus probably lower the ability of the phosphorus to donate its lone pair of electrons to the silicon d orbitals. Consequently, the Si-P w-bond in the phenyl phosphinosilanes would be conjugated with the aromatic rings attached to phosphorus Me3Si ::.P-( <:::> )2 . This conjugation with the two . ’9 78 aromatic rings should stabilize the phenylphosphinosilanes compared to the methylphosphinosilanes. For example, the phenylaminosilane, MeasiN- @ , was found to be much more stable from conjugative effects than the methylamino— silane, MeasiNMe2.67Combining this electronic factor and the steric factor, the values for the equilibrium constants of the phenylphosphinosilanes should be smaller than those of methylphosphinosilanes. Or, in other words, the phenyl— phosphinosilanes should be more stable than the methylphos— phinosilanes. The fact that the values of the equilibrium constants for the aromatic phosphinosilanes are larger rather than smaller than those for the methylphosphino— silanes suggested that dw-pw bonding did not influence the relative stabilities of the two silylphosphines to a large extent. These conclusions about the lack of ability of phosphorus to donate its p electrons to empty d orbitals of silicon are also supported by other authors.1sl 21.22.23 The values of the equilibrium constants for the reac— tions of trimethylsilyldimethylphosphine were also compared to those obtained for the reactions of trimethylsilyldi— methylamine with other amines (Table III). Generally, the values of the equilibrium constants for the silylphosphines are larger than those for the silylamines. Also, the equi— libria for the reaction of methylphosphinosilane with di- methylamine favors the formation of the methylaminosilane. These results suggested that the silylamines are more stable 79 than the silylphosphines and also that the relative strength of the silicon-nitrogen bond is greater than that of the silicon-phosphorus bond. Although these results do not un— equivocally negate dw-pw bonding in silicon—phosphorus compounds, they may suggest that the silicon—phosphorus bond is not as stabilized by dv—pv bonding as well as the silicon-nitrogen compounds. The calculations in this investigation are based on the fact that the reactions described by equation (1) are in equilibrium. In order to establish the validity of the equilibrium constants determined for the forward reactions (Me3SiPR2 + HNRé), the values of the equilibrium constants were determined for the reverse reactions (see Table XVIII). The values obtained for these reverse reactions are in good agreement with those for the forward reactions, suggesting that the data is valid. The results from the equilibrium studies which suggested that the lone pair of electrons on phosphorus in the silyl— phosphines were available for bonding motivated the study of the interaction of trimethylsilyldiphenylphosphine with nickel halides. This silylphosphine reacted with nickel iodide and nickel bromide under anhydrous conditions and at room temperature to give green solids. The formulas of these solids have been designated as NiX2[PPhZSiMe3]2 where X is Br or I. These solids were not obtained as pure products but as mixtures of NiX2[PPhZSiMe3]2 and Ni(PPh2)2(PPhZSiMe3)2. The elemental analyses of these “HT-““TTfll-r 97'?“ ~r "~ 9 iily‘cim -‘T'Qrw-ma s... . .__. -.’~. agar-“’55 Wit .. 80 solids suggested that the mixtures were predominately the Nix2(pph28iMe3)2 adducts. The Ni(PPh2)2(PthsiMe3)2 was obtained from the cleavage of the silicon—phosphours bond by the nickel halide according to the following equation: 4Me38iPPh2 + Nix2 -> 2Me35iX + Ni(PPh2)2(PPh2SiMe3)2. This nickel phosphide complex is very similar to the dia— magnetic complex obtained by Issleib1°4by a similar reac— tion of Nixz with diphenylphosphine: 4HPPh2 + Nix2 > Ni(PPh2)2(HPPh2)2 + 2HX. The reaction of the nickel halides with trimethyl- silyldiphenylphosphine in benzene at about 86° gave pre- dominately the previously described101complexes NiX2(HPPh2)2 or 3. The diphenylphosphine resulted from the hydrolysis of the ligand, Me3SiPPh2, by trace amounts of water. The diphenylalkyl and triphenyl nickel halide complexes, NiX2(Ph2PR)2, are paramagnetic and have been assigned in some cases tetrahedral structuresLO2However, the closely related complexes obtained in this investigation, NiX2(Ph2PSiMe3), where R is Measi, are diamagnetic and their absorption spectra in benzene solution are consistent with square—planar structures. In particular, the complexes show no absorption bands near 900 mu similar to those present in the paramagnetic complexes, NiX2(Ph2PR)2. A closer examination of the magnetic susceptibility measure— ments for the diphenylalkyl complexes in benzene solutions W 'F“. 81 suggests that the magnetic moments are lower than expected for the presence of two unpaired electrons. This behavior was rationalized by first assuming that the intermediate values of the magnetic moments were due to the following equilibrium: > (BuPh2)2NiX2 <—— (BuPh2)2NiX2 Diamagnetic Paramagnetic (”eff = 3.2 D.M.). If it is also assumed that the diamagnetic and paramagnetic species in benzene have square-planar structures, the magnetic susceptibility of the benzene solutions can be ac- counted for by postulating a diamagnetic ground state for the planar complex with an easily accesible paramagnetic excited state. Thus, if the separation between the two d orbitals of the metal (dxy and dX2_y2) is less than kT, an equilibrium as described above will arise. The fact that only diamagnetic species are observed for the complexes NiX2(PhPZSiMe3)2 suggests that the separation of the two d orbitals of the metal is larger than kT. This would dictate that the ligand, MeasiPth, has greater ligand field strength than the thPR or PhaP ligands in order to increase the separation of the d orbitals above kT. It is suggested that the large inductive effect of the Measi group compared to the alkyl groups is a possible ex- planation of the increase in field strength of the Me3SiPPh2 ligand over the RPth ligands. Thus, the square-planar diamagnetic structure of the NiX2(PPhZSiMe3)2 complexes can be rationalized. SUMMARY Trimethylsilylphosphines react with amines to give the corresponding trimethylsilylamine and phosphine. This is an equilibrium reaction which favors the formation of the silylamine over the silylphosphine. The values of the equilibrium constants for these reactions, which were measured by 1H nmr spectroscopy, varied with the substituents on the amine and phosphine. For a particular silylphosphine, the equilibrium constants were measured for reactions of the silylphosphine with secondary aliphatic, primary aliphatic and primary aromatic amines. (The equilibrium constants for the reactions in— creased in the following order with respect to the amine: secondary aliphatic amines < primary aliphatic amines g primary aromatic amines. These results were rationalized by suggesting that both the steric nature of the amines and factors which make the electron pair on nitrogen more available for dw—pv bonding are important and influence the values of the equilibrium constants. The steric fac- tors influenced the values of the equilibrium constants of the aliphatic amines to a larger extent and the elec- tronic factor was most important in influencing the values of the equilibrium constants of the aromatic amines. 82 O 83 The equilibrium constants for the reactions of a given amine with one of the two trimethylsilylphosphines MeasiPth or Me3SiPMe2, may be arranged in the following order, MeasiPth < MeasiPMez. The values of the equilibrium constants relative to the two silylphosphines were rational— ized in terms of the size of the substituents on phosphorus (trimethylsilyldiphenylphosphine is sterically less stable than trimethylsilyldimethylphosphine). .7 The electronic factor which was expected to effect the equilibrium constants suggests that the silylphosphines are not stabilized by dw—pw bonding to the same extent that the silylamines are stabilized by this type of interaction. Finally, the values of the equilibrium constants strongly support the fact that the relative strength of the silicon— nitrogen bond is greater than that of the silicon—phos— phorus bond. Complexes of the type NiL2X2 (L = MeasiPth, x = Br, I) were obtained when trimethylsilyldiphenylphosphine was allowed to react with nickel halides. The complexes NiX2(HPPh2)2 and Ni(PPh2)2(PPhZSiMe3)2 were also obtained as products of this reaction. The NiX2(HPPh2)2 resulted from the hydrolysis of the ligand, MeasiPth, and the Ni(PPh2)2(MessiPPh2)2 resulted from the cleavage of the Si—P bond in Me3SiPPh2. In contrast to the paramagnetic tetrahedral or gig— square—planar structure of most NiX2(RPPh2)2 complexes, the NiX2(Me3SiPPh2)2 has been assigned a diamagnetic square- 84 planar structure. The difference between the two complexes have been assigned to the greater ligand field strength of the MeasiPth ligand compared to the RPPh2 ligand which in turn has been related to the larger positive inductive ef- fect of the Me3Si— group. BIBLIOGRAPHY .0 '0 BIBLIOGRAPHY ~ 1. E. J. crane, Chem. Eng. News, 24. 12, 333 (1946). 2. E. J. Crane, Chem. Eng. News, 32“ 4517 (1952). 3. R. Fessenden and F. Fessenden, Chem. Rev., 61, 361 (1961). 4. U. Wannagat, "Advances in Inorganic Chemistry and Radiochemistry," Vol. 3, Academic Press Inc., 1961. 5. F. Cotton and G. Wilkerson, "Advanced Inorganic Chemistry,” John Wiley and Sons, Inc., New York, 1966, p. 229. 6. K. Hedburg, J. Amer. Chem. Soc., 71” 6491 (1955). 7. E. Ebsworth, et al., Spectrochim. Acta, 13“ 202 (1958). 8. H. Kriegsmann and W. Forster, Z. Anorg. U. Allggm. Chem., 298, 212 (1959) 9. D. Rflfinson, J. Am. chgg, Soc., 80, 5924 (1958). 'w 10. K. Hedburg, J. Am. Chem. Soc., 11, 6491 (1955). 11. J. Goubeau and J. Jimenez—Barbera, Z. Anorg. U. Allgem. Chem., 323, 217 (1960). 12. M. Yokai, et al., J. Am. Chem. Soc., 12“ 4139 (1953). 13. H. Kriegsmann, Z. Elektro Chem., 61, 1088 (1957). 14. E. Ebsworth and N. Sheppard, J. Inorg. Nucl. Chem., 2, 95 (1959). 15. Burg and E. Kuljian, J. Am. Chem. Soc., 12, 3103 A. (1950). 16. G. Davidson, et al., Spectrochim. Acta, 22¢ 67 (1966). o 17. A. Cowley and W. White, Abstr. of 153rd Meeting, Am. Chem. Soc., 1967, L 145. 85 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 86 E. Randall and T. Zuckerman, Chem. Comm., 20, 723 (1966). E. Randall, g£_gl., J. Am. Chem. Soc., 88“ 622 (1966). E. Ebsworth, Chem. Comm., 22” 530 (1966) E. Ebsworth, p£_gl., Sppctrochim. Acta, 13“ 202 (1968). G. Davidson, gp_gl., Spectrochim. Acta, 22” 67 (1966). B. Beagley, et al., Chem. Comm., 21“ 601 (1967). L. Tansjo, Acta Chem. Scand., 12” 35 (1959). ! i: L. Birkofer and A. Ritter, App,, 612“ 22 (1958). E L A. Burg and E. Kuljican, J. Am. Chem. Soc., 72, 3103 i (1950). "V i‘ A. Stock and K. Somieski, §g£., 22x 740 (1921). E; S. Sujishi and S. Witz, J. Am. Chem. Soc., 22“ 4631 (1954). LTD”Midland Silicones, British patent 737,229 (1959). O. Mjorne, Svenski Kem. Tid., 62” 120 (1950). R. Osthoff, and S. Kantor, Inorg. Synthesis, 5“ 55 (1957). R. Sauer, J. Am. Chem. Soc., 66” 1707 (1944). R. Sauer and R. Hasek, J. Am. Chem. Soc., 68“ 241 (1946). D. Bailey, gp_gl., J. Am. Chem. Soc., 22, 435 (1948). J. Gierut,_g£_al., J. Am. Chem. Soc., §§/ 786 (1936). E. Larson, Kgl. Fysiograph. sallskop. Lund. F6ch., 28x 1 (1958). E. Larson and R. Marin, Acta Chem. Scand., 2/ 1173 (1951). A. Stock and K. Somieski, Ber., 22: 740 (1921). S. D. Brewer, et al., J. Am. Chem. Soc., 22/ 3888 (1948). C. P. Haber, U.S. Patent 2,553,314 (1951); chem. Abstracts, 22x 6777 (1951). 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. K. Hyawa, et al., KO 1 Ra aku Zasshi, 52” 1445 (1956); Chem. Abstracts, 53“ 4176 £1959). L. H. Sommer, et al., J. Am. Chem. Soc., 76” 1030 (1954). E. Larson and L. Bjellerup, J. Am. Chem. Soc., 15, 995 (1953). O. Ruff and E. Geisel, B23., 38“ 2235 (1905). L. Tansjo, Acta Chem. Scand., 13” 35 (1959). H. H. Anderson, J. Am. Chem. Soc., 14, 1421 (1952). G. Fritz, Z. Naturforsch., 86, 776 (1953). G. Fritz, Z. Anorg. U. Allgem. Chem., 280, 332-345 (1955). S. Gekhale and W. Jolly, Inorg. Chem., 3, 1141 (1964). S. Gekhale and W. Jolly, Inorg. Chem., 4, 596 (1965). B. Aylett, et al., J. Inorg. Nucl. Chpm., 1, 187 (1955). G. Parshall and R. Lindsey, Jr., J. Am. Chem. Soc., 81, 6273 (1959). a) G. Fritz and G. Poppenburg, Angew. Chem., 22“ 208 (1960). b) G. Fritz and G. Poppenburg, Angew. Chem., 12“ 297 (1963). W. Kuchen and H. Buchwald, Chem. Ber., 22“ 227 (1959). E. Amberger and H. Beeters, Angew. Chem. Internat. Edit., ,1", 52 (1962). A. Bruker, et al., Doklody Akad. Nauk. S.S.S.R., 135, 843 (1960). L. Maier, Helv. Chem. Acta, 46, 2667 (1963). H. Gilman and D. Wittenberg, Quart. Rev. (Londoplj 13, 116 (1959). H. Gilman and H. Winkler, in H. Zeiss (Ed.), Organo- metallic Chemistry, Reinhold, New York, 1960, pp 270—345. H. Gilman, et al., J. Am. Chem. Soc., 82, 403 (1960). H. Gilman, et al., J. Am. Chem. Soc., §2/ 5933 (1958). E. Abel and R. Bush, J. Organometal. Chem., 3, 245 (1965). ~ 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88 W. Fink, Helv. Chim. Acta, 41, (2), 498 (1964). R. Fessenden and D. Crowe, J. Org. Chem., 16, (11), 4638 (1961 ). “"' K. Andrianov and G. Rumba, Zh. Obshch. Khim., 32, (b), 1993 (1962). w L. Tanslo, Acta Chem. Scand., £3, 29 (1959). C. Roth, Master's Thesis, Michigan State University, East Lansing, Michigan (1966). L. Klebe and F. Bush, First International Symposium on Organosilicon Chemistry, Prague, 1965. G. Fritz and G. Poppenburg, Angew. Chem., 15, 297 (1963). E. Abel, et al., J. Chem. Soc., 2371 (1968). E. Abel and I. SabherWal, J. Chem. Soc., 1105 (1968). R. Cragg and M. Lappert, J. Chem. Soc., (A), 82 (1966). T. George, et al., J. Chem. Soc., 2157 (1965). A. Brook, et al., J. Am. Chgg; Soc., 82, 5102 (1960). A. Janzen, Canad. J. Chem., 43“ 3063 (1965). E. Abel and C. Jenkins, J. Organometal. Chem., 14, 285 (1968). R. Hayter, J. Am. Chem. Soc., 86, 823 (1964). E. Abel and I. Sabherwal, J. Organometal. Chem., 12' 491 (1967). J. Drake and J. Simpson, Inorg. Chem., 6, 1984 (1967). J. Drake and J. Simpson, Chem. Comm., 21x 249 (1967). C. Russ and A. MacDiarmid, Angew. Chem. Internat., g, 418 (1966). H. thh and W. Schragle, Chem. Ber., 98“ 352 (1965 I.-R. Beattie, et al., J. Inorg. Nucl. Chgm., 23” 55 (1961). Pass and Sschindlbauer, Monatsch., 92, 151 (1959). K. Issleib and A. Tzschasch, Chem. Ber., 22, 705 (1959). K. Issleib and A. Tzschasch, Chem. Ber., 23, 1852 (1960). G. w. Parshall, Org. Syn., 4;, 102 (1965). 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 89 K. Issleib and A. Tzschasch, Chem. Ber., 92” 1125 (1959). K. Issleib and A. Tzschasch, Chem. Ber., 92, 704 (1959). L. Maier, J. Inorg. Nucl. Chem., 24, 275 (1962). H. Linton and E. Nixon, Spectrochim. Acta, 15” 146 (1959). K. Issleib and A. Tzschasch, Chem. Ber., 92, 118 (1959). W. Hewertson and H. R. Watson, J. Chem. Soc., 1490 (1962). A. L. Smith, Spectrochem. Acta, 16, 87 (1960). V. Mark, C. Dungan, M. Crutchfield and J. Van Wazer in M. Grayson and E. J. Griffith (Ed.), "Topics in Phos- phorus Chemistry,“ Vol. 5, John Wiley and Sons, New York, 1967, p. 227 ff. H. Vahrenkamp and H. N6th, J. Organqmgtallic Chem., 12, 281 (1968). D. Wittenburg, D. Aoki and H. Gilman, J. Am. Chpm. Soc., QQ, 5933 (1958). K. Issleib and H. J. Mobius, Chem. Ber., 94, 102 (1961). W. Jolly, "Inorganic Synthesis," Prentice-Hall, Inc., Englewood Cliffs, N. J., 1960, p. 161. D. Halah and J. P. Fackley, Jr., Inorganic Synthesis, 10, 29 (1967). R. Hayter, Inorg. Chem., 2, 933 (1963). Coussmaker, et al., J. Chgm. Soc., 2705 (1961). Jensen, Z. Anorg. Chem., 265, 229 (1936). K. Issleib and E. Wenschuh, Z. Anorg. Allgem. Chem., 305, 15 (1960). J. Fritz, "Acid Base Titrations in Non-Aqueous Solvents,‘ G. Frederick Smith, Chem. Co., Columbus, Ohio, 1952. K. Coskran, Ph.D. Thesis, Iowa State University, Ames, Iowa (1968). R. Morrison and R. Boyd, "Organic Chemistry,“ Allyn and Bacon, Inc., Boston, Mass., 1961, p. 455. K. Issleib and H. Bruchlas, Z. Anorg. U. Allgem. Chem. 316, 1 (1962). APPENDIX I INFRARED SPECTRA OF COMPOUNDS 90 .mafifimH>£u08HUHMHHmamnumfiflnu mo Enuuommm pmumuwcH .H madman 3:53.23: 0 Omo 00m 000. OON. 00v. 006— OOG— OOON 0000 000? 000m 0 oi 0.! J 3 u s 0.1. ONW. " 8 u i 0 col oma (K 0‘ On On Ow Ow Oh Oh Om 00 00 Om 00F . In, IlllluL 00. F' I'lllly) 9.05.: 252533: J M— V— Mr N— = O— m w \- O D v . m .wcflcmmosmamcmnmflp mo Enuuommm ownmnmcH .N mnsmflm 3.. E 2.2:: 000V 0000 O O O O m N aouemmsum 0? 0m 00 Oh Om Om 92 cow .wcflsmmosmamnuwfifloawaflmahnuTEflHu mo Eduuummm pwumnwcH :5 i _ 1...: 62382 95. RN. RH- Qfim mZO¢U_<< mZO¢U=2 .m musmflm (0X) BDNVlllWSNVUl 93 .cmuswouphg Imnuwu mo 0mm>mmao 0:0 Seam nosooum onu mo Eonuommm pmnmumcH :2” {EU >UZNDOw-m ca. 3. ed. 205.: as on 206.: o... .v onsmflm (£4) Danst‘Nval APPENDIX II PROTON NMR SPECTRA OF COMPOUNDS .msflfimamsuwfifivamaHmamnumfifluu mo Esnuowmw HE: ma In wusmflm A < t I}. _ (<1 < .1. m w. H y .i o . 4 4. i 9 “7 i T m 1 .i __ .mcflnmmonmfloahnuwfimuuwu mo Esuuommm HEQ ma .N musmflm 95 W-Ww—_ -m “"UF .moasomosoessudaac seasons spas chasm Ioupmnmuuwu mo mmm>mwao map Eonm nonpoum mnu mo Ednuommm HE: mu .3 ondmflm 5.... ii i. 97 98 Figure 5. 'H nmr SPECTRUM 0F Me SiPPh2 3 TMS amlur J 99 .mcHEmHNNGwhamnuofilz mo Eduuowmm MEG ma A. 121: _ o.» _ .w wusmfim X OK .TGHEmHRNswnamzumelz sufl3 mcflnmmonmamcmnmfloamaflmamnuwefluu sofluommu one EOHM muzuxfle ESHHQHHHDUT may mo Esuuowmm HE: ma .b musmflm 100 I. ii...) . .ceasonosonSscsoacamaanamsumaenu sues mcaamamsndaao mo concommu mnu Eonm mmsuxfls ESHHQHHH560 0:9 mo Eduuoomm HE: me .w mudmflm 101 -- “Pastas. v-~-.W w.- .- V. T: "It ~ -17. .7. .ocflsmmosmawcmnmflpamHHmamaumfifluu nufl3 mGHEMH>u5QIunu mo coHuommn 030 Eouw musuxfle Ednnnflafisqm 0:5 mo Eduuommm MEG ma .m musmxm 102 .measonosoamsmsoaoamaamassumaaeu spas msaamamsumao mo cofluommu may Eoum musuxflfi ESHHQHHHSUT msu mo Esuuommm MSG me .OH mudmflm ms: 103 |\ airfare / MI ' Izmzo. Io . zmzomzo EmmIo N N m N N m :n_n_I+ $25 22 H :2 E + snazm m2 '72 W776?" mus—arm“ "‘.__.. ‘_ ”v5: - a!~—m,. _. 4. 5' a ' ii _: .. 104 C_|_-_|3CH2NH 0:3CH2NSi \ ( WWW») 1. Figure 11. 1H nmr spectrum of the equilibrium mixture from the reaction of diethylamine with trimethylsilyl- diphenylphosphine (100 Mc). v--. .. Inch-«0‘.-- (film