Synthesis and Reactivity Studies of Zwitterionic Silenes and 2-Silenolates

Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 1034

BY TAMAZ GULIASHVILI

ACTA UNIVERSITATIS UPSALIENSIS UPPSALA 2004 ! "" #$% & ' ( & & )' '* +' , - ('*

. ' +* ""* / ' 0 / & 1, / 2/ * 3 * "#. 60 * * ISBN 92%%2!"6"2

+' ' ' & 2 2 2 2 ' ( & /7 8 ** /927 9: 8 ' /9:7 92 8 * & ' 8 ' 2 2 2 8, ' 2 ' , ' ' ( ' ( /* + 8, 2 2 2 & ' & ; ; '< < , ' #2 = * +' ' ' ' && & /7 * +' ' , 8 , ' #2 ( >: ? * +' , ' #2 , ' ;5-< 23 ,'' ' 2' & ' *+' 2 2 2 , ' ' * @ ' , ' 2 2 2 , ' ' ' ( A '(' * +' 2& ' & & ' , ' ' & ' ; '< ( * 3 ' & & /7 2 ' & '(' ( & ' * 0 & 2 , ' 2 ' & ( 2 * +' 2 ' 2 ' ( * +' & & 2 , & ' 8 B2 ( '* +' 2 ' / ;C/ 7 #D*6E< /2 ( ' ' /7 ;;/< 7 *4 ! F<* +' 2 ( / A , ' A ' >: ? , ' #2 *

/ 1, / 2/ 23 0 / ' B2 ( '

! " " # " $% &''" " ()*&+ ,- "

G + 8 . ' ""

5// "2 # B 5/ 42%%2!"6"2 $ $$$ 2!%" ;' $HH *A*H I 7 $ $$$ 2!%"< KNOWLEDGE IS CHIMERA - FOR BEYOND IT LIES OTHER KNOWLEDGE, AND THE INCOMPLETENESS OF WHAT IS KNOWN RENDERS THE KNOWING FALSE.

--THE THOMAS COVENANT CHRONICLES- List of Papers

This thesis is based on the following papers, referred to in the text by its Roman numerals.

I Evidence for Formation of Silenes Strongly Influenced by Re- versed Si=C Bond Polarity. Ibrahim El-Sayed, Tamaz Guliash- vili, Rita Hazell, Adolf Gogoll, Henrik Ottosson, Org. Lett. 2002, 4, 1915-1918

II Formation of Transient Zwitterionic Silenes and Their Diels- Alder Reactions with 1,3-Dienes: Mechanism and Reactivity. Tamaz Guliashvili, Arnaud Martel, Ibrahim El-Sayed, Andreas Fischer, Henrik Ottosson, Manuscript.

III The First Unsuccessful Attempt to Add Alcohols to a Silene. Tamaz Guliashvili, Henrik Ottosson. Manuscript.

IV The First Isolable 2-Silenolate. Tamaz Guliashvili, Ibrahim El- Sayed, Andreas Fischer, Henrik Ottosson, Angew. Chem. Int. Ed. 2003, 42, 1640-1642.

V Potassium 1-Dialkylamino-2,2-bis(trimethylsilyl)-2-Silenolates: Heavy Enolates with Remarkable Stability. Tamaz Guliashvili, Arnaud Martel, Andreas Fischer, Patrik Akselsson, Henrik Ottosson, Manuscript. Contents

1. Introduction...... 7 2. An Overview of the Chemistry of Silenes ...... 10 2.1. Structural and Spectroscopic Properties...... 10 2.2 Chemical Reactions of Silenes...... 13 2.3. An Overview of Methods for Preparation of Silenes ...... 17 2.3.1 [2+2] Cycloreversion of Silacyclobutanes...... 17 2.3.2 Elimination / Metalation Techniques...... 18 2.3.3. Photochemical and Thermolytic Formation of Silenes from Acylpolysilanes ...... 19 3 Zwitterionic Silenes: Thermolytic Formation and Trapping Reactions.....21 3.1 Attempted Generation of Zwitterionic Silenes using the Sila-Peterson Olefination Reaction ...... 21 3.2 Zwitterionic Silenes from Carbamylpolysilanes ...... 22 3.3 Reactions of Zwitterionic Silenes with 1,3-Dienes ...... 30 3.4 Reactions of Zwitterionic Silenes with Alcohols ...... 35 4. Novel Metal 2-Silenolates: Influence of the Counterion on Structure and Reactivity ...... 41 4.1. A Novel and Convenient Pathway to Stable Metal 2-Silenolates .....41 4.2 Trapping experiments and metal exchange reactions...... 43 4.3 Formation and X-ray structural characterization of the first crystalline metal 2-silenolate [18]crown-6 salt ...... 44 4.4 Potassium 1-Dialkylamino-2,2-bis(trimethylsilyl)-2-Silenolates: A Class of Remarkably Stable 2-Silenolates...... 46 Summary and Outlook ...... 50 Summary in Swedish ...... 51 Acknowledgement ...... 55 References and notes...... 56 Abbreviations

1-Ad 1-Adamantyl Alk Alkyl AO Atomic orbital Ar Aryl Bn Benzyl B3LYP Becke’s three-parameter hybrid HF/DFT method nBu n-Butyl tBu tert-butyl CCSD(T) Coupled cluster theory with singles, doubles and perturbative triples excitations DEPT Distortionless enhancement by polarization transfer DFT Density functional theory Et Ethyl h Hour HF Hartree-Fock g-HMBC Gradient enhanced heteronuclear multiple bond correlation HOMO Highest occupied molecular orbital HSQC Heteronuclear single bond quantum coherence IED Inverse electron demand INEPT Insensitive nuclei enhancement by polarization transfer lp Lone-pair LUMO Lowest unoccupied molecular orbital Me Methyl Mes 2,4,6-Trimethylphenyl min Minute MO Molecular orbital NMR Nuclear magnetic resonance Ph Phenyl iPr iso-Propyl THF Tetrahydrofurane TMS Trimethylsilyl TLC Thin layer chromatography o-Tol o-Tolyl TS Transition state 1. Introduction

The diversity and richness of organic chemistry stems from the ability of carbon to form single, double, and triple bonds with itself and with other elements of the periodic system. Low-coordinated carbon compounds such as alkenes, alkynes, ketones, and carbenes hold central positions in organic chemistry. Since silicon is the element most closely related to carbon, it was an endeavor throughout the 20th century to synthesize Si counterparts of these low-coordinated compounds, and to explore their structure and reactivity. For a long time it was considered that heavy Group 14 and 15 elements cannot form multiple bonded compounds because of the poor overlap of their p-orbitals and the poor size match between the atoms of different Group 14 and 15 elements.1 The so-called “double-bond rule” stated that compounds with multiple bonds involving heavy main-group elements are unstable.2 However, through the landmark work of Gusel’nikov and Flowers in the late 1960s the first transient Si=C bonded compound, a silene, was detected through pyrolytic [2+2] cycloreversion of silacyclobutanes.3 In 1981, the double-bond rule was finally proven incorrect when stable and isolable compounds containing Si=C (1, Scheme 1),4 Si=Si (2, Scheme 1),5 and P=P 6 bonds were reported by the groups of Brook, West, and Yoshifuji, respectively. Since then, synthesis of multiple bonded compounds for most of the elements in groups 13 to 16 has been reported. Until now, a number of stable and isolable Si unsaturated compounds including silenes,7 1- 8 9 10 silaallene, silaimines (R2Si=NR'), silanethiones (R2Si=S), silaaromatic 11 12 compounds, silylenes (R2Si:), and very recently even a disilyne (3, Scheme 1)13 have been synthesized. As a result a series of novel reactions have been developed that lead to these low-coordinated Si species. The de- velopments in silene chemistry have been well summarized in several review articles.14

CH(TMS)2 Mes Mes (TMS) HC TMS OTMS 2 Si CHMe2 Si C Si Si Si Si Me2HC Si CH(TMS) TMS 1-Ad Mes Mes 2 (TMS)2HC 1 23 Scheme 1

7 The early days of unsaturated Si compounds chemistry were marked by theoretical and gas-phase experimental studies, whereas the investigations in the past twenty years have progressed toward reactions in condensed phases. We now believe that the stage is set for developing and exploring transformations of low-coordinated Si compounds in applied research fields. Al- though silenes due to their high kinetic reactivity will never receive the same important status in organosilicon chemistry as the alkenes do in organic chemistry, there will be an increasing number of new compounds synthesized, as well as new reactions introduced, through Si=C bonded species. Recent achievements support this statement. The most direct application of silenes in organic synthesis has recently been demonstrated by Steel and co- workers.15 They showed that silacyclohexenes formed from [4+2] cycloadditions between silenes and various 1,3-dienes provide access to 1,2-diols and lactones via a phenyl-triggered Fleming-Tamao oxidation, and the outcome of this study can be considered as a novel 1,4-functionalization of 1,3- dienes. Transition metal complexes containing silenes and silylenes are found to be relevant to a number of catalytic processes, and some of these complexes have already proved useful in various catalytic systems. For instance (i) iridium silylene complexes are catalysts for the hydrosilylation of ketones under mild conditions,16 and (ii) dehydropolymerization process of polycarbosilanes, a polymer precursor used in the Yajima process17 for fab- rication of silicon carbide ceramic fibers, appears to involve metal Ș2-silene complexes.18 Applications in polymer chemistry may also be envisioned, even though the use of the stable heavier Group 14 and 15 element containing multiple bonded compounds for inorganic polymer synthesis is virtually unex- plored.19 One class of polymers that relates to a multiple bonded Si species is silicones. It was F.S. Kipping who in 1920 coined this name by analogy with a ketone as he believed that the product of the reaction of dichlorosilanes with base is a Si=O bonded compound, a “silicon ketone”.1a It was later found that the Si=O bond spontaneously polymerizes to give cyclic or linear siloxanes.1d The term “silicone” was however carried over to the actual polymeric products, that now are the basis of a billion-dollar industry. With regard to other polymers of heavy alkenes, Gates and co-workers very recently reported the successful free radical polymerization of a P=C bonded compound to give poly(methylenephosphine), a new polymer with phospho- rus(III) atoms in the main chain.20 We believe that it might now also be possible to find silenes that can be polymerized under certain conditions to give polymeric materials with alternating Si and C centers in the main chain. Most known Si=C bonded compounds are highly moisture and air sensitive, as well as rather unselective in various chemical transformations. This fact could be a reason for the lack of applications of silenes in organic synthesis. To make Si=C bonded compounds interesting as synthetic reagents it is thus important to temper their reactivity. In this regard, quantum chemical

8 computations can serve as a useful tool to design silenes with desired properties. Such calculations can be used to test for silenes that are less moisture and air sensitive, and less vulnerable to dimerization than most presently studied silenes. Computations indicate that silenes with reversed Si=C bond polarization, so-called reverse polarized or zwitterionic silenes, are interesting candidates in this regard.21 Our experimental work on the structure and reactivity of such zwitterionic silenes now support the previous computational work. In the present thesis, recent results obtained in the synthesis and investigation of chemical properties of silenes and metal 2-silenolates with reversed Si=C bond polarity are summarized. It is shown how the reactivity of silenes and 2-silenolates can be tuned by introducing two ʌ-electron donating substituents at the C end of Si=C bond and also demonstrated that these new compounds with reversed Si=C bond polarity participate in various chemical transformations in highly selective way, in contrast to the Si=C bonded compounds with natural bond polarization, that are rather unselective in the same type of chemical reactions. The hope is that these findings represent a step toward the introduction of silenes in organic synthesis. Zwitterionic silenes seem to be more useful in organic synthesis applications than naturally polarized silenes because they are stable toward dimerization, less vulnerable to nucleophilic addition, and more selective in cycloaddition reactions.

9 2. An Overview of the Chemistry of Silenes

2.1. Structural and Spectroscopic Properties When compared to carbon, ʌ-bonds to silicon and other heavy Group 14 elements are not very strong and compounds with such bonds are kinetically reactive. It is known that the increased bond length to sp3-hybrized Si makes this atom more vulnerable to nucleophilic attack when compared to C and this is even more true for sp2-hybrized Si.22 The ʌ-bond strength in silenes is weaker than in alkenes: it has been estimated to be 28 - 46 kcal/mol for par- 23 ent silene H2Si=CH2 by measurement of the bond rotational barrier, and the quantum chemically calculated values are in agreement.24 For comparison, the C=C and Si=Si ʌ-bond energies are 65 and 22 - 30 kcal/mol, respectively.25 Unlike the C atoms in alkenes, the Si atoms in Si=Si double bonds are slightly pyramidal as the sum of the valence angles at Si (ȈSi) is 345 – 360 o (Scheme 2).26

G G G G Si Si CC Si C Si C

Scheme 2

Compounds with Si=C bonded are usually planar (Scheme 2), but those that are highly influenced by reverse Si=C bond polarization (Schemes 2 and 3) were predicted by computations to distort from the planar structure and become pyramidal at Si.23b Our recent experimental results support this finding (vide infra).27 In the early days of silene chemistry most structural data came from IR, UV, and microwave spectroscopy. Until now, only a few X-ray crystal structure characterized silenes have been reported, and much structural information has instead been obtained from ab initio and DFT calculations. Important structural features of silenes are the SiC bond distance (r(Si=C)), and sum of the valence angles at Si (ȈSi). These parameters also serve as indicators of how strongly the Si=C bond is influenced by reverse polarization. The reverse polarization of the Si=C bond is effected through ʌ-electron donating groups at the C end of the Si=C moiety.21,27 The silenes can be called zwitterionic silenes if their electronic structure is dominated by resonance structures II and III (Scheme 3).

10 R1 X R1 X R1 X R1 R3 Si Si Si Si R Y R Y R2 Y 2 2 R2 R4 I II III

Reverse polarized silenes Naturally polarized silenes

R1 - R4 = SiR3, H, Alk, Ar. R1, R2 = SiR3, H, Alk, Ar.

X, Y = NR2, OR, SR. Scheme 3

The r(Si=C) values for various silenes are given in the Table 1. Both experimentally measured and theoretically determined bond lengths are shown for comparison. A good correlation is observed between theoretical and experimental values and hence the r(Si=C) can be predicted by calculations with high accuracy. As seen in Table 1, strongly reverse polarized silenes, i.e. zwitterionic silenes, have Si-C single bonds, as can be expected when II and III (Scheme 3) dominate the electronic structure, in contrast to the par- 28 ent silene, H2Si=CH2, which has a measured r(Si=C) of 1.7039 Å. The central Si in a zwitterionic silene is also pyramidal because the negative charge is localized to this atom. In a recent quantum chemical study by Ot- tosson,21b correlations between q(Si) and r(Si=C) as well as ȈSi were found. These results show that hybridization of the Si atom changes from sp2 in naturally polarized silenes to sp3 for fully zwitterionic silenes. The energy for inversion (Einv) of the Si atom is another interesting property of zwitterionic silenes. In silenes of the general formula Z2Si=C(NH2)2, Einv has been estimated to be sensitive to the Si substituents because it decreases in the following manner (B3LYP/6-31+G(d,p) energies in kcal/mol): Z = F (31.8) > Z = Cl (26.3) > Z = OH (25.9) > Z = CF3 (23.5) > Z = NH2 (14.8) > Z = Ph (6.3) > Z = H (4.3) > Z = H3Si (3.5) > Z = Me (0.4). These findings can be used to computationally design zwitterionic silenes with a chiral Si atom. The most suitable substituents on the unsaturated Si for generation of the chiral zwitterionic silenes or 2-silenolates was found to be two perfluori- nated alkyl groups of different sizes.21b Other structural information of Si=C bonded species can be obtained from 1H and 13C NMR spectroscopy, but 29Si NMR should be considered as the main experimental technique for structural characterization of Si=C bonded compounds. The 29Si isotope, with spin I = ½ is the only naturally occurring NMR active isotope but the natural abundance of the 29Si isotope is merely about 4.7%. Unfortunately, the low abundance and the low sensitivity of the 29Si nucleus cause some difficulties for a direct acquisition of 29Si NMR spectra. Usually more than 30 seconds are necessary between pulses in order to allow the 29Si nuclei to relax. To overcome these problems it is preferable to use a pulse sequences that utilizes magnetization transfer from J-coupled 29 protons (Si-H, Si-CH3, etc) to the Si nuclei, so as to enhance the Si NMR

11 signal. The most commonly used technique for magnetization transfer from the nearby protons to Si nuclei is INEPT31 and also DEPT.32 The 29Si and 13C NMR chemical shifts of silenes strongly depend on the nature of substituents at Si and C. Usually 29Si NMR chemical shifts of sp2- hybridized Si in silenes are found downfield from the positions normally observed for sp3-hybridized Si. In silenes bearing two TMS groups at the Si, t as in the Brook type silene TMS2Si=C(OTMS)R (R = Bu, 1-Ad, CEt3, Mes), the Si nuclei resonates in the region 34 - 54 ppm.4a,33,34 On the other hand, silenes with two methyl groups at Si are characterized with downfield 29Si 29 t 35 NMR signals, for example į Si = 144 ppm in Me2Si=C(SiMe3)SiMe( Bu)2. The Si atoms in Si=C bonded compounds that are affected by reverse polarization, such as 4-silatriafulvene,36 lithium,37 and potassium 2-silenolates,27b are very shielded and therefore have high field shifted 29Si resonances (į 29Si = -71 ppm in 4-silatriafulvene, and -82 and -78 ppm in 2-silenolates t + + TMS2Si=C(OM) Bu with M = Li , K respectively ).

Table 1. Experimental and calculated SiC bond distances of some formally Si=C bonded compounds.

SiC bond distance (Å) Silene Measured Calculated (method) (method)

1.704 1.713 H2Si=CH2 (millimeter wave spectroscopy)28 (B3LYP/6-31+G(d))21b

1.692 1.713 (microwave Me Si=CH 2 2 spectroscopy)29 (B3LYP/6-31+G(d))21b

1.764 TMS Si=C(OTMS)Ad 2 (X-ray crystallography) 4b -

- 1.870 TMS Si=C(OTMS)NMe 2 2 (B3LYP/6-31G(d))27a 2.162 2.024a (X-ray crystallography) 30 (B3LYP/6-311+G(d,p)) C6H4[N(R)]2C=Si[N(R)]2C6H4 30

1.926 1.938 (X-ray TMS Si=C(OK)tBu 2 crystallography) 27b (B3LYP/6-31+G(d))27b a Calculated on smaller model systems.

12 1 With regard to coupling constants, all JSi=C cluster around 85 Hz, i.e. lar- 1 3 ger than the JSi-C between sp hybridized silicon and methyl carbons (47 - 48 Hz). The observed range for 13C chemical shifts in silenes (from 77 to 274 ppm) is larger than for alkenes (from 95 to 155 ppm). In Brook-type silenes t 2 TMS2Si=C(OTMS)R (R = Bu, 1-Ad, CEt3, Mes), the sp hybridized C atom has į13C = 197 – 214 ppm,4a,33,34 whereas the sp2 hybridized C atoms in 2- silenolates has į13C at 246 and 274 ppm with lithium and potassium as coun- tercations, respectively.37, 27b With regard to the UV characteristics, simple silenes with small substi- tients have the ʌʌ* transition in the region 245 - 260 nm. For example, the parent silene H2Si=CH2 is characterized by a UV absorption at 258 nm. A bathochromic shift is usually observed by the two silyl groups at the sp2- hybridized silicon. For example, Brook type silenes TMS2Si=C(OTMS)R have strong UV absorption bands around 340 nm.4a,33,34 Similarly, another t highly substituted silene ( BuMe2Si)(SiMe3)Si=C(2-Ad) (2-Ad = 2- adamantylidene) has a UV transition at 322 nm.38 While the same bulky silyl groups if present at the unsaturated C-end have only a small influence on the ʌʌ* transition compared to the parent silene H2Si=CH2. Thus for 39 Me2Si=CTMS2 Ȝmax = 278 nm. Although spectroscopic characterization of Si=C bonded compounds is the main experimental means for studying their structure and reactivity, quantum chemical computations are still a main guiding tool for exploring structure and reactivity of unsaturated Si compounds because most silenes are transient species, and in most cases it is impossible to observe their formation at ambient temperatures. For example, only in a recent millimeter wave spectroscopy study by Borgey, Burger and co-workers28 was the ground state geometry of H2Si=CH2 finally established, and the search for the millimeter wave transition in the 180 - 473 GHz frequency region had to be guided by state-of-the-art ab initio calculations of the rotational constants at the CCSD(T) level of theory. However, a treatise on the computational methods used to investigate silenes will not be given here. The reader is instead referred to the recent books of Cramer and Jensen.40 The calculations discussed in the present thesis were predominantly performed using Becke’s three-parameter hybrid HF/DFT method, abbreviated as B3LYP.

2.2 Chemical Reactions of Silenes Most silenes are highly reactive compounds. They usually dimerize to disilacyclobutanes in head-to-tail or head-to-head manner, depending on the nature of the substituents.41 Isolation of disilacyclobutanes is usually taken as evidence for formation of silene intermediates. Silenes that are stable to- wards dimerization have bulky groups and may also have ʌ-electron donating groups at the C end of the Si=C bond. The reaction mechanism is differ-

13 ent for head-to-tail and head-to-head dimerizations. For instance, Brook-type silenes 4 undergo head-to-head dimerization via a biradical intermediate 5 (Scheme 4). Disproportionation takes place in 5 when R' = H so that 6 is formed,42 but when R = alkyl, the product is the 1,2-disilacyclobutane derivative 7.43 The preference for head-to-head dimerization of Brook type silenes can be explained from MO calculations on model systems related to 4. In these silenes the largest coefficients of both the HOMO and LUMO are found at the Si atom, in contrast to naturally polarized silenes, and the observed regioselectivity appears to be more favourable.21a Naturally polarized silenes 8 with small substituents at the Si and C atoms undergo head-to-tail dimerization to give 1,3-disilacyclobutane derivative 9, most probably preceded by formation of a dipole-dipole complex (Scheme 5).

OTMS

(TMS)2Si CRR'2 (TMS)2Si CRR'2 lk OTMS =A OTMS R 7 TMS TMS OTMS TMS Si CRR'2 2 Si TMS Si CRR' 2 R TMS CRR'2 = TMS H OTMS OTMS 45 TMS TMS Si CR'2 TMS Si CRR'2 TMS OTMS 6 Scheme 4

TMS G GR Si R

TMS R' R' Si(TMS)2 + (TMS)2Si R R G G TMS R' 8 Si 9 R' TMS Scheme 5

Other chemical transformations of silenes include [4+2] cycloadditions with 1,3-dienes and Į,ȕ-unsaturated esters, [2+2] cycloadditions with alkenes, alkynes, ketones, imines, and azo-compounds, and nucleophilic addition of alcohols, water and other nucleophilic reagents, and these reactions are well summarized in several review articles on Si=C bonded compounds.14 Most silenes undergo [4+2] cycloadditions with 1,3-dienes, but often side- products arising from [2+2] and ene reactions are formed in substantial amounts. The relative ratio of [4+2] (11, Scheme 6), [2+2] (13), and ene

14 (12), adducts strongly depends on the polarity of Si=C bond, the nature of 1,3-diene as well as the reaction temperatures. For example, in Brook-type t silenes TMS2Si=C(OTMS)R (R = Bu, 1-Ad), the ene pathway becomes more favored in the reaction with 2,3-dimethyl-1,3-butadiene and the ene product is obtained in 40% total yield.44 On the other hand, the same type of silenes in the reaction with 1,3-butadiene give about 20% [4+2] cycloadduct and 80% of [2+2] adducts as an approximate 1:1 mixture of the two stereoi- somers.44 The proportion of cycloadduct 11 increases slightly relative to 13 with raised temperature (Scheme 6). Metal 2-silenolates, which are closely related to reverse polarized silenes, give exclusively [4+2] cycloadducts with 1,3-dienes.45 The regioselectivity of [4+2] cycloadditions between silenes and unsymmetrical 1,3-dienes also depends to some extent on the polarization of the Si=C bond. For example, reaction of isoprene with the naturally polarized silene Me2Si=C(SiMe3)2 mainly leads to 4-methyl-1-silacyclohexenes but also to [2+2] adducts,46 the t same diene with Brook-type silenes TMS2Si=C(OTMS)R (R = Bu, 1-Ad) also gives preferably 4-methyl-1-silacyclohexenes but also ene adducts.46 On t the other hand, lithium 2-silenolates TMS2Si=C(OLi) Bu give exclusively 4- methyl-1-silacyclohexenes.45

R R R2 R4 1 3 R Si R Si 1 3 + + R1 Si C R3 + R Si R R R2 4 R1 R3 H 2 4 R2 R4 10 11 12 13 Scheme 6

Among the chemical reactions of silenes, the 1,2-addition of nucleophiles is the most thoroughly investigated both computationally and experimentally.47-49 Wiberg proposed that addition of alcohols to silenes involves re- versible formation of the silene-alcohol complex (Scheme 7), followed by proton transfer from the attacking alcohol to the C atom.46 Recently, Leigh and co-workers investigated the rates of addition of alcohols to several silenes using nanosecond laser-flash photolysis and found that silene-alcohol complex is reversibly formed in the first step and the proton transfer from the attacking alcohol or second alcohol molecule occurs in the second and the rate determining step.48 This mechanism is very different from the mechanism of alcohol addition to C=C bonds, where the rate determining step is the attack of an electrophile.50 The alcohol and water addition to naturally polarized silenes is a strictly regiospecific process. The alkoxy or hy- droxy group in the product is always attached to the Si end of Si=C bond. A computational investigation on the mechanism of alcohol and water addition to the Si=C bond revealed that the reaction path, where the nucleophilic re-

15 agent attacks the Si end of Si=C bond, has a small activation barrier and the reaction is strongly exothermic.47a

H R'' R'' H O O

R2CSiR'2 + R''OH R2CSiR'2 R2CSiR'2

H OR''

R2CSiR'2 Scheme 7

If, however, the nucleophile attacks the C end the barrier is very high and the reaction is moderately exothermic. Recently, Apeloig and co-workers reported a computational DFT study of the mechanism of water and alcohol addition to various silenes.47b They found that the transition state energy for water or alcohol addition to silenes depends on the polarity of Si=C bond. Reactions of naturally polarized silenes exhibit very low or even negative TS barriers (-3 to 8 kcal/mol), but substituents that reduce the Si=C bond polarity significantly increase the activation barrier for water addition. The au- thors conclude that the strategy for designing novel, kinetically stable silenes should therefore involve both kinetic stabilization of silenes by bulky substituents and reduction of Si=C bond polarity by means of ʌ-electron donating groups attached to the C atom. Kira and co-workers found that the regioselectivity of nucleophilic addition of methanol to 4-silatriafulvene (14, Scheme 8) is completely opposite to that of alcohol addition to naturally polarized silenes (Scheme 7).36 Thus, the alcohol proton is added to the Si atom of Si=C bond of 4-silatriafulvene to form 16. The reactions of silenes with alcohols are of special interest since most silenes are easily trapped by alcohols, and isolation of the addition products serves as an evidence for formation of silenes.

t t Bu t Bu t SiMe2 Bu SiMe2 Bu Si Si t t t SiMe Bu t SiMe Bu Bu 2 Bu 2

14 15 MeOH

OMe H t Bu t Si SiMe2 Bu t Bu SiMe tBu 2 16 Scheme 8

16 Scheme 9 summarizes the main chemical transformations of silenes, and it can be seen that a variety of Si containing compounds can be generated via Si=C bonded species.

* Si O Si O R Li * n Si OR H Si C Si C H Si O RLi R1 R O O = H C R2 R 2 + R 1

Si C Si Me Si - 3 S N iY = -N + HX Y TMS Si Si C Si NN X H

Si Y = Cl, OMe, NMe2

* Si N Si N * n Scheme 9

While cycloadditions and alcohol additions to naturally polarized silenes have been extensively investigated experimentally, there are only few papers that deal with the same transformations of zwitterionic silenes.27a,36

2.3. An Overview of Methods for Preparation of Silenes 2.3.1 [2+2] Cycloreversion of Silacyclobutanes The first transient silene 18 ever generated was formed through pyrolytic [2+2] cycloreversion of silacyclobutanes 17 (Scheme 10).3 Due to the high temperatures employed (700 - 1000 K) this approach is suitable only for generation of transient silenes bearing small substituents at Si. It was found that [2+2] cycloreversion of silacyclobutanes proceeds via an initial homo- lytic cleavage of the C-C bond, followed by ethene release.51 This biradical mechanism was also supported by ab initio calculations,51c and has implica- tions for the stereochemistry of the silacyclobutane decomposition because biradicals 21 and 23 lead to a loss of the stereochemistry of the reactants 20

17 and 22, respectively.52 This can be rationalized by the C-C bond rotational processes in biradicals 21 and 23 which competes with the ȕ-scission yielding the silene 19 together with E- and Z-2-butene (Scheme 10).

' + Me Si Me2Si CH2 H2CCH2 Me 17 18

Me Me ' Me

Me Si Me Me Si Me Me2Si CH2 + Z 20 Me 21 Me 19 Me

Me Me ' Me Me Si Me Me Si Me Me2Si CH2 + E Me Me Me 22 23 19 Scheme 10

A similar approach can be employed to convert disilacyclobutanes into silenes. 1,3-Disilacyclobutanes require drastic conditions (up to 780 °C),53 while even highly substituted 1,2-disilacyclobutanes are known to revert under very mild conditions into silenes.41e-l

2.3.2 Elimination / Metalation Techniques Another approach to silenes is based on a salt elimination technique and was developed by Wiberg and co-workers.54 They generated silene 26 from 24 through 25 as intermediate (Scheme 11). This method was applied to generate stable silenes that cannot be formed by photolysis and/or thermolysis of acylpolysilanes (vide infra). In a similar manner, some silenes could be gen- 55,56 erated by means of OH/SiR3 or F/SiR3 elimination reactions.

OR H OR Li R'Li TMS Si TMS Si TMS Si TMS -R'H TMS - LiOR TMS 24 25 26 Scheme 11

A relatively new synthetic approach to silenes was established independently by the groups of Ishikawa (Path A, Scheme 12)41h,57 and Apeloig (Path B, Scheme 12).38,58 The key step in both methods is the formation of intermedi-

18 ate Į-alkoxypolysilanes 28 and 31 from the reaction of methyl lithium with acylpolysilanes 27 or silyl lithium reagent with non-enolizable ketone 30 respectively, followed by 1,2-elimination of metal silanolates and formation of silenes 29 and 32. These reactions are Si analogues of the Peterson olefination reaction and are often referred to as the Sila-Peterson olefination reactions. Silenes formed through the Sila-Peterson olefination reaction have a common substituent pattern because they have two silyl groups at the Si end. One advantage of this method is that alkyl, aryl, and vinyl substituents can be introduced at the C end. However, it seems difficult to generate silenes with ʌ-electron donating groups at the C end (vide infra). Almost all silenes from the Sila-Peterson olefination reaction are transient species that usually dimerize in a head-to-head fashion, yielding 1,2-disilacyclobutanes or linear polysilanes, and only in the case when the groups at the C and Si ends become sufficiently bulky is the silene stable toward dimerization.

Path A Li TMS O TMS O TMS Me TMS Si + CH3Li Si Me Si -TMSOLi TMS Ph TMS TMS Ph TMS Ph 27 28 29

Path B

OLi TMS + (TMS)3SiLi TMS O Si -TMSOLi Si TMS TMS TMS 30 31 32 Scheme 12

2.3.3. Photochemical and Thermolytic Formation of Silenes from Acylpolysilanes The photochemical rearrangement of acylpolysilanes 33 into silenes 34 through a sigmatropic [1,3] Si ĺ O shift (Scheme 13) was discovered by Brook and co-workers.41e,59 Depending on the bulk of the R group the generated silenes are either reactive intermediate species (R = Me, Et, iPr, CH2C6H5) or they are in thermal equilibrium with their head-to-head dimers (R = tBu) and in the latter case the formation of silene 34 is observable spec- troscopically.41e,59 In case the R group is sufficiently bulky, such as with R =

19 1-Ad, silene 34 is stable and isolable and the first X-ray structurally characterized silene was prepared through this procedure.4

OTMS TMS O TMS OTMS hQ or ' ' (TMS) Si R TMS Si Si 2 TMS R (TMS) Si R TMS R 2 OTMS 33 34 35 Scheme 13

The thermolytic isomerization of acylpolysilanes usually gives mixture of compounds. However, if this reaction is carried out in the presence of a scavenger, such as an alcohol or a 1,3-diene, the reactions are much more clean and give the corresponding trapped products in quantitative yield. Brook and co-workers demonstrated that thermolysis of 33 with R = tBu in the presence of 1-phenylpropyne gives 2-silacyclobutene in 72 % isolated yield.59 Recently Ishikawa and co-workers also showed that silenes were formed from various substituted tris(trimethylsilyl)acylsilanes through heating them up to 140 ºC in benzene solutions in the presence of suitable scav- engers. Thus, 34 with R = tBu, 1-Ad or Mes were generated by thermolysis of acylpolysilanes 33 and trapped with ketones yielding the corresponding siloxetanes.60 The pathways for the formation of silenes shown in schemes 12 and 13 are most interesting in connection to the results presented in the next chap- ters. However, there are also other procedures for generating Si=C bonded compounds, such as photolysis of disilanes,61 and rearrangements of silyl- carbenes into silenes.39,41,62 Since these methods are not commonly applied in the synthesis of stable silenes, we will not discuss further the details of these procedures.

20 3 Zwitterionic Silenes: Thermolytic Formation and Trapping Reactions

3.1 Attempted Generation of Zwitterionic Silenes using the Sila-Peterson Olefination Reaction Silenes with reversed Si=C bond polarization should bear two ʌ-electron donating groups at the C end of the Si=C bond and therefore in the first instance, silene 37 was selected as a synthetic target (Scheme 14). Dimeriza- tion of 37 with R = R' = Me is endothermic by 14 kcal/mol according to B3LYP/6-31G(d) calculations, and if formed, 37 should therefore exist in monomeric form. The Sila–Peterson olefination reaction was attempted for generation of 37. It was believed that reaction of tris(trimethylsilyl)silyl lithium or potassium with N,N,N’,N’-tetraalkylureas would proceed through intermediate 36 (related to 28 and 31 shown in Scheme 12), followed by elimination of the metal silanolates (TMSOLi or TMSOK) and release of silene 37.

TMS M NRR' rt TMS TMS NRR' TMS Si M + O O Si X Si NRR' THF TMS NRR' TMS NRR' TMS TMS NRR' -TMSOM 36 37 R, R' = Alk, M = Li, K Scheme 14

However, no silene formation was observed in the course of these reactions. Presumably, the reaction leads to an intermediate carbene, which although internally stabilized through conjugation with the nitrogen lone pairs, undergoes fast unselective reactions yielding several unidentified products. The reasons for the failure of the Sila-Peterson olefination to produce silenes related to 37 are not yet fully understood and must await further experimental and computational studies.

21 3.2 Zwitterionic Silenes from Carbamylpolysilanes The attention was instead turned to 2-amino-2-siloxysilene 39a (Figure 1, Scheme 15). It also has two strongly ʌ-electron donating groups at the C atom and should fulfill the criteria of a zwitterionic silene. Carbamylpolysi- lanes 38a - g were selected as starting compounds for synthesis of silenes 39a - g.63,27a It was believed that 38a - g could undergo sigmatropic [1,3] Si ĺ O shifts of a TMS group under thermolytic or photolytic condition in similar manner as shown in the scheme 13. However, the first experiments were unsuccessful and no rearrangement was observed upon thermolysis of 38a up to 180 °C. It is noteworthy that, Lickiss and co-workers previously attempted to generate silenes 39a and 39c photochemically from the same precursors without success.63 To find a reason for this observation, we explored through quantum chemical calculations the [1,3] Si ĺ O shift of TMS group in 38a leading to 39a. These calculations were performed at B3LYP/6-31+G(d)//B3LYP/6-31G(d) level (Figure 1).

26.8 TS ) l mo / l 18.6 kca ( TMS OTMS

E Si TMS NMe2 TMS 39a O 0.0 TMS Si TMS NMe2 38a Figure 1. The relative energies of transformation of carbamylpolysilane 38a into silene 39a calculated at B3LYP/6-31+G(d)//B3LYP/6-31G(d) level.

The transition state energy for transformation of 38a to 39a is 26.8 kcal/mol, and 38a is 18.6 kcal/mol more stable than 39a. This implies that 39a should have a low population at the reaction temperatures, and when formed it will undergo a rapid back transformation into the more stable reac- tant 38a. An attempt to monitor the equilibrium between 38b and 39b through 1H-NMR spectroscopy at 90 °C was unsuccessful since no changes were observed in the spectrum of a solution of 38b. We only observed broadenings of the TMS proton signals, indicating that the equilibrium between 38b and 39b is too fast to be detectable by 1H NMR spectroscopy and that the population of 39b is very low. The same thermolysis reactions were then repeated in presence of 2,3-dimethyl-1,3-butadiene as trapping reagent, and it was found that the reactions proceed smoothly and lead to formation

22 of single products. The 1H, 13C, 29Si NMR spectroscopic, and X-ray crystallographic studies revealed that the products of these reactions are exclusively [4+2] cycloadducts (Figure 2 and Scheme 15). However, it was rather unexpected that cycloadduct 45c, instead of 46c, was identified by X-ray crystallography (Figure 2.)

Figure 2. ORTEP drawing of silacyclohexene 45c.

The mechanism that could account for formation of 45 involves rearrangement of 39 to 42 via carbene 40 or silylene 41 intermediates, but these pathways were not supported by B3LYP calculations performed on simpli- fied model reactions (Scheme 16). The transition state barrier separating silene 39 from carbene 40 and silylene 41 is 69.5 and 66.5 kcal/mol respectively, and if carbene 40 or silylene 41 is formed, they would form a mixture of E- and Z-isomeric silenes 42 and 43, respectively (Path A, Scheme 15). The [4+2] cycloaddition reaction of these isomeric silenes 42 and 43 with 2,3-dimethyl-1,3-butadiene would also form two diasteromeric cycloadducts 44 and 45, and since only 45 was formed, it is concluded that rearrangement of 39 to 40 and 41 did not take place in the course of the reaction. In addition, no formation of [1+2] adducts arising either from addition of 40 or silylene 41 to 2,3-dimethyl-1,3-butadiene was observed, also excluding formation of 40 and 41 as intermediates.

23 OTMS TMS Si NR Path A TMS 2 40

TMS TMSO TMS O CD3C6D5 or C6D6 TMS OTMS TMSO NR2 TMS Si Si Si + Si TMS NR2 TMS TMS TMS NR2 TMS NR2 65 - 180 ° C 39 42 43 38 OTMS Si NR2 TMS TMS 41 Si Si TMSO NR2 TMSO TMS 38a - 46a R = R' = Me TMS TMS TMS NR2 38b - 46b R = Me R' = Ph 45 44 38c - 46c R = R' = Ph 38d - 46d R = R' = iPr 38e - 46e NRR' = 1-pyrrolyl 38f - 46f NRR' = 1-piperidyl 38g - 46g R = R' = nBu

Path B

TMS O CD3C6D5 or C6D6 TMS OTMS TMS Si Si TMS NR2 Si TMS NR2 65 - 180 ° C TMS OTMS TMS NR2 38 39 46

45 Si TMSO NR2 TMS TMS Scheme 15

There is also a second argument that could speak against the isomerization of 39 into 42 and 43. The Si=C bond distance in 42a is 1.823 Å, according to B3LYP/6-31G(d) calculations, which is closer to a Si-C single bond rather then to a Si=C double bond. Moreover, the energy difference between Z- and E-isomers (42a and 43a) is only 0.6 kcal/mol at the same level of theory, favoring the Z-isomer. Since 42a has a considerable Si-C single bond character the Z/E-isomerization at the reaction temperatures (65 – 180 °C) could be feasible, and when trapped the two isomeric silenes would yield the [4+2] cycloadducts 44 and 45, respectively. Formation of the single cycloadduct 45, and not a mixture of 44 and 45, indicates that silene 39 is trapped by 2,3-dimethyl-1,3-butadiene to afford 46 which rearranges to 45 (Path B, Scheme 15). Unfortunately, no other experimental evidence is available so far to support the conclusion that 39 does not undergo isomerization into 42 and 43. It is instead reasonable to assume that cycloadduct 46

24 is formed which should be unstable due to an anomeric interaction between the lone pair of the N atom (lp(N)) and the ı* C-O bond orbital. This interaction weakens the C-O bond so that 46 would easily rearrange to the more stable adduct 45 (Path B, Scheme 15). Indeed, quantum chemical calculations at B3LYP/6-31G(d) level showed that the C-O bond length in 46a is 1.453 Å but the same C-O bond length is 1.434 Å if the NMe2 group in 46a is replaced by an H atom. This indicates that the lp(N) causes the lengthen- ing of the neighboring C-O bond (Table 2) . In addition, the N atom in 46a is more planar (ȈN = 347.5°) in comparison to the case when the the OSiMe3 group in 46a is replaced by an H atom (ȈN = 339.9 °).

OSiH3 H Si Si 3 NH 69.5 2 20.2 SiH3 11.1 SiH H SiO 3 O 27.8 H3Si OSiH3 3 NH2 H3Si Si Si Si H3Si NH2 H3Si SiH3 SiH3 NH2

0.0 18.2 OSiH -5.0 66.5 3 48.5 Si C NH2 H3Si SiH3 33.5 Energy in kcal/mol calculated at B3LYP/6-31+G(d)//B3LYP/6-31G(d) level.

Scheme 16

Table 2. Selected geometrical parameters of 46a and model [4+2] cycloadducts calculated at B3LYP/6-31G(d) level.

TMS Si R' TMS R

Cycloadduct R(C-O) [Å] R(C-N) [Å] ȈN (deg.) R = NMe R' = OTMS 2 1.453 1.465 347.5 (46a)

R = NMe2 R' = H - 1.478 339.9 R = OTMS R' = H 1.434 - -

A further support for the rearrangement of 46 to 45 comes from B3LYP/6- 31+G(d)//B3LYP/6-31G(d) calculations performed on the model system

25 (reaction pathway AĺBĺCĺD, Figure 3). As can be seen in Figure 3, the model cycloadduct D, related to the observed cycloadduct 45, is more stable than initially formed model cycloadduct C, related to the cycloadduct 46 by 31 kcal/mol. Moreover, the transition state energy required for this transformation is only 17.4 kcal/mol calculated at B3LYP/6-31+G(d)//B3LYP/6- 31G(d) level. )

l endo: 31.2 TS2 /mo l 27.8 TS1 kca ( E 18.2

H Si OSiH SiH3 O 3 3 0.0 Si + H3Si Si + H3Si NH2 TS3 -6.2 SiH3 NH2 A B -23.7

H Si Si OSiH 3 3 -54.7 H3Si NH2

C Si H3SiO NH2 H3Si SiH3 D Figure 3. The relative energies for transformation of the model silylamide A, related to 38, in to the model cycloadduct D, related to 45. Calculated at the B3LYP/6- 31+G(d)//B3LYP/6-31G(d) level.

The transition state geometry for the transformation of model cycloadduct C into D is shown in Figure 4, and the proposed mechanism of the rearrangement of 46 to 45 is shown in scheme 17.

Figure 4. The transition state geometry for rearrangement of the model cycloadduct C into D of Figure 3. Calculated at the B3LYP/6-31G(d) level. Bond lengths in Å.

26 NR2 TMS TMSO R N R2N 2

TMSO Si TMS Si TMS Si TMS TMS TMS OTMS

Si Si TMS NR2 TMSO NR2 TMS OTMS TMS TMS

46 45 Scheme 17

Recently, a related rearrangement has been found by Ishikawa and co- workers (Scheme 18).64 They observed that heating of cis-silacyclobut-3-ene B, afforded the trans-silacyclobut-3-ene E, via silyl substituted cycloprop-2- ene intermediate C (Scheme 18).

TMS OTMS TMS tBu TMS OTMS H tBu TMS t TMSO TMS Si C Si Bu Si TMS t Bu 140 ° C 250 °C A t But B H Bu H E

190 ° C

TMS TMS OTMS TMSO t Si TMS Si Bu 1,3 TMS shift TMS H tBu 250 ° C H tBu t C D Bu Scheme 18

However, the mechanism found for this rearrangement seems to be stepwise and this is in contrast with the concerted transformation of 46 into 45 (Scheme 17).

27 Table 3. Influence of N-substitution in 38 on reaction time required for a complete transformation of 38 into cycloadduct 45.a Carbamylpolysilane Temp. Time (h) Yield,b % Solvent (ºC ) (isolated)

TMS O TMS Si 180 48 93 (88) toluene TMS NMe2 (38a)

TMS O TMS Si 180 48 93 (87) toluene TMS N (38f)

TMS O TMS Si 180 48 89 ( - ) toluene i TMS N( Pr)2 (38d)

TMS O TMS Si 180 3 93 (85) toluene TMS N (38e)

TMS O TMS Si 100 6 95 (93) benzene TMS NPhMe (38b)

TMS O TMS Si 100 3 97 (95) benzene TMS NPh2 (38c) a In all entries shown in this table 2,3-dimethylbutadiene was used in excess. b The yield determined by integration of the 1H NMR spectrum of the crude reaction mixture.

In order to check how the reactivity of 39 is influenced by the nature of the substituents at the N atom, the carbamylpolysilanes 38b-g were also synthesized and their Diels-Alder reactions with 2,3-dimethyl-1,3-butadiene were studied. The results of these studies are summarized in Table 3. The substituents at the N atom significantly influence the rate of cycloadduct formation. The N,N-diphenylcarbamylpolysilane 38c was found to be the most reactive compound in these cycloaddition reactions. The reason for this observation could be either bulkiness of the phenyl group that forces the

28 amino group out of conjugation with the Si=C bond because of steric con- gestion, and/or delocalization of the lone pair electrons of the amino group to the phenyl ring. The latter implies that the nitrogen lone pair electrons would be less available for conjugation with Si=C bond and thereby making the silene molecule a less electron-rich dienophile, and less influenced by reversed polarization. Indeed, B3LYP/6-31G(d) calculations showed that the presence of Ph substitution at the amino group in silene 40c forces the amino group out of conjugation with the Si=C bond, thus making this silene a less electron-rich dienophile. Two factors can cause the rotation of the amino group: either the sheer bulk of the Ph groups forces the NPh2 groups to rotate, or the lp(N) is so involved in conjugation with Ph groups that it cannot conjugate with Si=C ʌ bond which can also force the NPh2 group to rotate. If the steric bulk of the phenyl ring is the reason for the enhanced reactivity of the intermediate silene 39c it should also be possible to achieve the same enhancement in reactivity of silenes 39 by any other sterically congested substituent at N. For i t instance, N( Pr)2 and N( Bu)2 substitution should have somewhat similar effects as the NPh2 group. On the other hand, if there is predominately an electronic effect of the phenyl ring, one would expect a similar reactivity for smaller substituents where the lp(N) is already involved in conjugation, as for instance in a N-pyrrolyl group where there will be no more possibility of the lp(N) to interact with the Si=C ʌ-bond. In order to test these two alternative explana- tions for the increased reactivity, we carried out thermolysis experiments on the silylamides 38d and 38i in the presence of 2,3-dimethyl-1,3-butadiene (Table 3). Since the silylamide 38e with the N-pyrrole functionality is more i reactive than the silylamide 38d with the relatively bulky N( Pr)2 group one can conclude that the enhanced reactivity of N,N-diphenylsilylamide is mainly due to lack of the conjugation of the lp(N) with the ʌ-electron system of the Si=C bond. Thus, silylamide 39c is the least influenced by the zwitterionic resonance structures, and it is therefore a less electron-rich dienophile than 39a or any other dialkylamino substituted silenes. Hence, the observed higher reactivity of 39c in the reaction with 2,3-dimethyl-1,3- butadiene. All products from these cycloaddition reactions are [4+2] adducts, obtained in nearly quantitative yields, which contrast those of naturally polarized silenes, compounds that also give large amounts of ene and/or [2+2] adducts.44,46 The structures of the cycloadducts have been confirmed by 1H, 13C, and 29Si NMR spectroscopy. For cycloadducts 45e and 47 (Scheme 19) the X-ray crystallographic structures are also available (Figure 5). One can observe that 45e and 47 have similar substituent patterns at the Si and quaternary C atoms of the silacyclohexene ring as 45c (Figure 1), supporting the view that the rearrangement proposed in Scheme 17 is common for these silacyclohexene systems.

29 45e 47 Figure 5. ORTEP drawings of silacyclohexenes 45e and 47.

3.3 Reactions of Zwitterionic Silenes with 1,3-Dienes The most reactive 2-amino-2-siloxysilene 39c was selected for further studies of the Diels-Alder reactions with different 1,3-dienes. Both electron-poor 1,3-dienes, e.g. 1,3-butadiene, and electron-rich 1,3-dienes, e.g. 2,3- dimethoxy-1,3-butadiene, were involved in these experiments (Table 4 and 5). Based on the results summarized in Table 4, we conclude that the cycloadditions between 39c and 1,3-dienes are inverse-electron-demand (IED) Diels-Alder reactions. The IED nature of these cycloadditions can be explained based upon the frontier MO interactions between the 1,3-diene and the dienophile. It is known that the most favorable interaction in IED Diels- Alder reactions involves the HOMO of the dienophile and the LUMO of the 1,3-diene. The HOMO of the dienophile is always lower in energy than the LUMO of the 1,3-diene (Figure 6). However, electron withdrawing substituents in the 2- and 3-positions of the 1,3-diene usually decrease its LUMO energy, reducing the gap between the silene HOMO and the 1,3-diene LUMO. This increases the favorable interactions between the two frontier molecular orbitals and explains why the electron-poor 2,3- dimethoxylcarbonyl-1,3-butadiene and 1,3-butadiene are the most reactive, while the electron-rich 2,3-dimethoxy-1,3-butadiene is the least reactive in these cycloaddition reactions (Table 4).

30 E 1,3-diene dienophile

LUMO LUMO

HOMO HOMO

Figure 6. Frontier MO interaction in IED Diels-Alder reaction.

R R R1 R2 1 2 R1 = R2 = H 47 TMS O R1 = R2 = Me 45c TMS Si R1 = Me, R2= H 48 Si NPh TMS NPh TMSO 2 1 2 2 C6D6, 100 ° C TMS TMS R = R = OMe 49 1 2 R = R = CO2Me 50

Scheme 19

Table 4. Reactivity of 1,3-dienes in Diels-Alder reaction with silene 39ca 1,3-diene Time (h) Yield,b % (isolated)

MeO2C CO2Me 1.4 88 ( - )

H H 1.5 98 (87)

Me 3.3 97 (89)

Me Me 5.5 97 (95)

MeO OMe 104 93 (83)

a Conditions: 100 ˚C, [38c] / [1,3-diene] = 1:1, benzene-d6. b The yield determined by integration of the 1H NMR spectrum of the crude mixture.

31 It seems rather surprising that the more electron-rich silene 39a showed the least reactivity in the reaction with 2,3-dimethyl-1,3-butadiene. To explain this observation we therefore studied the dependence of the rate of formation of cycloadducts 45a and 45c on the concentration of 2,3-dimethyl- 1,3-butadiene and 2,3-dimethoxy-1,3-butadiene because the overall rate of formation of cycloadducts 45 depends on the rate of silene 39 formation from the silylamide precursor 38 as well as on the rate of cycloaddition between in situ generated silene 39 and 1,3-diene (Scheme 15). The results from this study are summarized in Table 5. It has been found that increased concentration of 2,3-dimethyl-1,3-butadiene leads to decrease in the reaction time required for the formation of cycloadduct 45c from 39c. In the case of 39a we found that increasing the 1,3-diene concentration does not decrease the formation time for cycloadduct 45a. This finding indicates that the rate limiting step in the case of the formation of 45a is the silene 39a formation, while in the case of cycloadduct 45c the cycloaddition reaction between 39c and 2,3-dimethyl-1,3-butadiene is the rate determining step, preceded by fast equilibrium between 38c and 39c. This conclusion was also supported by DFT calculations on the rearrangement of carbamylpolysilanes 38a - c into the corresponding silenes 39a - c (Figure 7). The outcome of these calculations is that with increase of the degree of reverse polarization the energy gap between starting silylamide 38 and silene 39 becomes larger, which in turn means that isomerization of silylamide 38a into silene 39a becomes a rate limiting step in the reaction of 39a with 1,3-diene (Figure 7).

26.8 mol) / 15.4

kcal 18.6 ( not calculated

E 39a 13.4 10.8 39b 0.0 39c 0.0 0.0

38c 38b 38a

Increased importance of reverse polarization Figure 7. The B3LYP/6-31+G(d)//B3LYP/6-31G(d) energy profiles for the isomerization of silylamides 38a – c into the 2-amino-2-siloxysilenes 39a – c.

In order to study the regiochemistry of the IED Diels-Alder reaction between the 2-amino-2-siloxysilenes 39 and the 1,3-dienes, we investigated the reaction of 39c with isoprene as an unsymmetrical 1,3-diene. As expected, this reaction leads to two regioisomers, formed in a 3/2 ratio. The ratio was de-

32 termined by integration of the 1H NMR spectrum of the crude mixture and the structures of the major and minor regioisomers of 48 are shown in Figure 8.

Table 5. Influence of the 1,3-diene concentration on the reaction time required for complete conversion of 38c into cycloadducts 45c and 49. a [1,3-diene]/[38c] Reaction Time b (h) Reaction Time c (h) (45c) (49)

1:1 104 5.5

2:1 78 -

4:1 54 3

6:1 - 2

10:1 - 1.5 a Conditions: benzene-d6, 100 °C. bTime required for complete transformation of 38c into 45c (reaction with 2,3-dimethyl-1,3- butadiene). cTime required for complete transformation of 38c into 49 (reaction with 2,3-dimethoxy-1,3- butadiene).

6 4 6 4 H CH3 H3C H 5 3 5 3 7 2 7 2 Si TMSO Si NPh TMSO NPh 8 1 2 8 1 2 TMS TMS TMS TMS 48' 48'' Figure 8. The structure of the major (48') and minor (48'') regioisomers determined by 2D NMR experiments. The structure of the major regioisomer 48' was assigned through 2D-NMR experiments. In a 2D-NMR long range 1H-13C correlative spectra (g-HMBC experiment) the major regioisomer 48' showed no cross peaks between the vinyl H6 proton and the quaternary C1 atom of the silacyclohexene ring where the position of quaternary C1 atom in the 13C NMR spectrum was identified through a DEPT experiment. On the other hand, the minor regioisomer 48'' showed a cross peak between the vinyl proton H4 and the ring carbon C1. The preference for formation of 48' is consistent with frontier MO interactions. It is known from calculations on model systems related to

33 39 that the larger coefficient of both HOMO and LUMO is located on the Si atom of the Si=C bond.21a Hence, the observed regiochemistry of the [4+2] cycloaddition with isoprene, yielding 48' appears to be somewhat favored.

In conclusion the main features of the investigated IED Diels-Alder reactions between zwitterionic 2-amino-2-siloxysilenes 39 and 1,3-dienes are: x The [4+2] cycloadducts are formed in nearly quantitative yields as one diastereomer. x The [4+2] cycloaddition reactions between these zwitterionic silenes and 1,3-dienes are in accord with inverse-electron-demand Diels-Alder reactions. Therefore, the relative reactivity of various 1,3-dienes in IED Diels –Alder reaction with 39c is as follows: 2,3-dimethoxycarbonyl-1,3- butadiene ~ 1,3-butadiene > isoprene > 2,3-dimethyl-1,3-butadiene > 2,3- dimethoxy-1,3-butadiene. This findings are in contrast to cycloadditions between the naturally polarised silene Me2Si=C(SiMe3)2 and the same 1,3-dienes, where an opposite reactivity trend was observed.46 x The expected cycloadducts 46 arising from the reactions between 39 and 1,3-dienes rearrange to the more stable isomeric cycloadducts 45 via exchange of TMSO and TMS groups between the quaternary C and Si atoms of the silacyclohexene moiety. x The rate determining step for cycloadduct formation depends on the degree of Si=C bond polarization of the investigated silenes. In case of the more reverse polarized silene 39a the rate limiting step is the formation of silene from silylamide, while in the case of the less reverse polarized silene 39c the rate limiting step is the Diels-Alder reaction. x In reactions of 39c with isoprene as an unsymmetrical 1,3-diene, a mixture of the two possible regioisomers are formed.

34 3.4 Reactions of Zwitterionic Silenes with Alcohols As was mentioned in section 2.2, silenes can be trapped with alcohols, and these reactions also serve as evidence for transient silene formation. We therefore studied reactions of various alcohols ROH (R = Me, Et, iPr, Ph, Allyl, Bn) with 2-amino-2-siloxysilenes 39a and 39c (Scheme 20).

H OTMS

TMS Si NPh2 TMS OMe H X eO M h TMS O , 1 C6D6CD3 or C6D6 TMS OTMS o C TMS Si Si 0 12 TMS NR 100 -180 oC TMS NR2 2 1 R 2 ' 0 O 38 39 H -1 8 0 o O o C 38a / 39a R = Me 120 C, 128h H NMe2 38c /39c R = Ph +

TMS TMS Si OR' 40 - 96% TMS Si NMe2 TMS TMS OTMS R' = Me 52 R' = Et 53 R' = iPr 54 R' = Ph 55 R' = Allyl 56 Si TMSO NMe2 R' = Bn 57 TMS TMS 95% 51 Scheme 20

We found that the reactivity of 38 toward ROH was strongly influenced by the nature of the substituents at the N atom. Surprisingly, reaction of excess ROH with 38a proceeds anomalously leading to formation of alkoxysilanes 52 - 57 in nearly quantitative yields (Scheme 20, Table 6). As expected the reaction time is strongly dependent on the bulkiness of the R group in ROH. More bulky alcohols usually require longer reaction times. The structures of alkoxysilanes 52 - 57 were confirmed by 1H, 13C, and 29Si NMR spectroscopy, and MS analysis and the corresponding spectroscopic data are reported in the supporting information of paper III. Initially, reaction of 38a with MeOH was performed at 120 °C, i.e. under conditions when 38a gives no reaction with 2,3-dimethylbutadiene even after long reaction times, and therefore it may be concluded that at this temperature silene 39a cannot be formed from 38a. To clarify this point we investigated the reaction between 38a and MeOH at 180 ºC, i.e. under conditions when formation of 39a was

35 proved through trapping experiments with 2,3-dimethyl-1,3-butadiene (Scheme 15). The reaction of 38a with excess MeOH is completed in 30 min. at 180 °C producing methoxysilane 52 in quantitative yield (Table 6). Although isolated yields were always slightly lower than measured by 1H NMR spectroscopy, perhaps due to decomposition of part of the silylethers during purification (preparative TLC, pentane as an eluent).

Table 6. Formation of tris(trimethylsilyl)silylethers 52 – 57 from reactions of alcohols ROH with carbamylpolysilane 38a.a Yield b (%) R Temp (°C) Time (isolated) 4 h 52, Me 120 96 (92) 1.5 hc

52, Me 180 30 min 95 (92)

53, Et 120 8 h 94 (90)

53, Et 180 50 min 93 (89)

54, iPr 120 11 h 93 (87)

54, iPr 180 70 min 92 (85)

55, Ph 120 65 min - (40)

55, Ph 180 30 min - (60)

56, Allyl 120 1 h 76 (68)

56, Allyl 180 15 min 89 (84)

57, Bn 120 1 h 70 (63)

57, Bn 180 15 min 85 (78) a Conditions: Sealed NMR tubes, toluene-d8, [38a] / [ROH] = 1 / 4 b The yields of the silylethers were determined by 1H NMR spectroscopy. c [38a] / [ROH] = 1 / 10

36 In addition, the reaction between 1,3-butadiene and 38a at 120 °C leads to the sole formation of the cycloadduct 51 and confirms that indeed 38a equilibrates with silene 39a also at this temperature (Scheme 20). The preference of 38a for nucleophilic substitution at Si, rather than addition of MeOH to the Si=C bond of 39a, is not surprising. On the one hand, 39a will exist in very small concentration at the reaction temperatures, and on the other hand, 39a is strongly affected by reverse Si=C bond polarization and thus less reactive toward addition of MeOH. This is in line with Apeloig’s conclusion about the reactivity of reversed polarized Si=C bonded compounds in the reaction with alcohols.47b Thus, both these factors make the substitution reaction with 38a more favorable than the addition of ROH to the Si=C bond in 39a. Tris(trimethylsilyl)silylethers are interesting compounds, since it has been demonstrated that the tris(trimethylsilyl)silyl group is a fluoride resistant but photolabile protecting group for primary and secondary alcohols.65 It is noteworthy that the protection of various alcohols can be carried out in a 1.2/1 ratio of 38a and ROH, normally used for the protection of alcohols by chlorosilanes or silyl triflates,66 in reasonable reaction time (Table 7). A possible disadvantage of this protection is the use of the elevated reaction temperatures (120 - 180 °C), although this approach does not require usage of a base, which is usually required for the protection of alcohols with silyl groups.65,66 The yields of the protected alcohols are quantitative and in some cases slightly higher than reported previously.65

Table 7. Protection of alcohols ROH with tris(trimethylsilyl)silyl group.a

R Temp. (ºC) Time (h) Yield b (%)

CH3CH2- 180 6 96

(CH3)2CH- 180 23 86

C6H5CH2- 180 4 85

CH2=CH-CH2- 180 4 87 a Conditions: Sealed NMR tubes, toluene-d8, [38a] / [ROH] = 1.2 / 1.0. b The yields of the silylethers were determined by 1H NMR spectroscopy.

To explain the outcome of the reaction of 38a with alcohols the mechanism of the replacement of the C(O)NMe2 group by the OMe group in a model silylamide A (Figure 9), was investigated using DFT calculations at B3LYP/6-31+G(d)//B3LYP/6-31G(d) level. According to these calculations the reaction between the model silylamide and MeOH starts with formation

37 of a hydrogen-bonded complex B, where the lp(O) of the model silylamide A forms a hydrogen bond with the MeOH proton (Figure 9). This complex B is stabilized by 5.1 kcal/mol relative to the separated reactants. The transition state leading to methoxysilane C and carbene D is separated from the reactants by 33.6 kcal/mol. The optimized geometries of the hydrogen-bonded complex B and transition state TS1 are shown in Figure 10. ) l

mo TS / 2 54.6 l

E (kca 33.6 TS1

SiH3 O H3Si Si SiH NMe 3 2 17.8 A + SiH MeOH 3 C H3Si Si OMe + O N Me 0.00 SiH3 HMe C D - 5.1

Me O H - 22.9 SiH3 O H3Si Si NMe2 SiH3 SiH B 3 O H3Si Si OMe + HCNMe2 C SiH3 E

Figure 9. Relative energies calculated at B3LYP/6-31+G(d)//B3LYP/6-31G(d) level for formation of the model methoxysilane C from reaction of MeOH with the model silylamide A, related to 38a.

The final products of the reaction of the model silylamide A with MeOH are methoxysilane C and N,N-dimethylformamide E. The quantitative formation of N,N-dimethylformamide as a by-product of the reaction of 38a with alcohols was proved by 1H and 13C NMR spectra of the crude reaction mixture. Using the B3LYP/6-31G(d) method the addition of MeOH to the silene B related to the model silylamide A was also investigated (Figure 11). As can be seen from the calculated energy profile, formation of silene F and addition of MeOH to the Si=C bond requires at least 42.7 kcal/mol (Path AĺTS1ĺFĺTS2ĺH, Figure 11) which is higher than the energy for the

38 substitution reaction that leads to the formation of methoxysilane as shown in Figure 9.

0 .97 5 Å

1 0.972 Å .9 5 2.826 Å 5 Å Å 2.137 Å 7 2 4 1.350 Å . 3 Å 4 4 .2 Å 1 33 1.9606 Å 1.9

B TS1 Figure 10. Optimized B3LYP/6-31G(d) geometries of the hydrogen-bonded complex B and transition state TS1 (related to the reaction pathway shown in Figure 9).

TS2 50.1 ) l mo / TS3 l 42.7 kca ( 27.7 TS1 E

20.4

H Si OSiH SiH3 O 3 3 3.8 0.00 Si + MeOH H Si Si + MeOH 3 H Si NMe H OMe NMe 3 2 SiH3 2 H3Si Si OSiH3 A F H3Si NMe2

-27.6 G

MeO H

H3Si Si OSiH3

H3Si NMe2 H

Figure 11. Relative energies calculated at B3LYP/6-31+G(d)//B3LYP/6-31G(d) level for addition of MeOH to silene F, related to 39a.

39 Experimentally we also investigated the reaction of 38c with MeOH. In contrast to 38a, the reaction of 38c with MeOH does not lead to the formation of methoxysilane 52, instead a complicated product mixture was obtained and after separation of the crude mixture by flash chromatography none of the separated compounds match either methoxysilane 52 or the methanol-to-silene 39c addition product (Scheme 20). In conclusion we have investigated the reactivity of various alcohols in the reaction with carbamylpolysilane 38a. These reactions proceeds anomalously and lead to quantitative formation of tris(trimethylsilyl)silylethers. To the best of our knowledge this reaction represents a novel approach to pro- tect primary and secondary alcohols with the tris(trimethylsilyl)silyl group.

40 4. Novel Metal 2-Silenolates: Influence of the Counterion on Structure and Reactivity

2-Silenolates are heavy Group 14 element analogues of enolates (Scheme 21). Since the latter hold a central position in organic chemistry, one can consider the exploration of metal 2-silenolates of equal importance for organosilicon chemistry. The 2-silenolate is strongly influenced by reverse Si=C bond polarization because resonance structure II significantly contrib- utes to its electronic structure (Scheme 21). Different lithium 2-silenolates with bulky alkyl or aryl substituents at the 2-position had previously been synthesized and their spectroscopic and chemical properties studied.37,45,67- 68,27b However, the further use of lithium 2-silenolates is hampered by their low stability and propensity to dimerize. Perhaps for that reason they have not received much attention, and according to Sci-Finder Scholar, only ten papers were published on lithium 2-silenolates until October 2004, in contrast to thousands of papers on enolates.

O O Si Si R R R = Alk, Ar. I II Scheme 21

4.1. A Novel and Convenient Pathway to Stable Metal 2-Silenolates Normally, the reaction of a silyllithium reagent with acylpolysilanes 58 in THF is used for generation of lithium 2-silenolates 59 (Scheme 22). These reactions are reported to be sensitive to temperature and light, and the usual conditions for preparation of lithium 2-silenolates are low temperatures (-78 to -100 °C) and darkness.37 Although lithium 2-silenolates are stable in solution, provided they are kept at low temperatures and in dark, they usually degrade at ambient temperatures within hours presumably via dimerization (especially when in 59, R = 1-Ad or tBu),.37

41 Metal 2-silenolates are closely related to zwitterionic silenes because the SiC bond is significantly longer than a normal Si=C double bond (Table 1), and resonance structure II best explains their geometrical parameters and chemical properties (Scheme 21). Similarly to zwitterionic silenes, 2- silenolates are more selective in reactions with 1,3-dienes than are naturally polarised silenes because they give exclusively the [4+2] adducts.45 With unsymmetrical 1,3-dienes they give only one of two possible regioisomers.45 The most important chemical reactions of lithium 2-silenolates are summarized in Scheme 22.

Si TMS OH TMS R

, H2O

TMS Me O (TMS)3SiLi TMS O Li MeI O Si TMS Si TMS Si o THF, -78 C, dark TMS R TMS R R TMS 58 59 R = tBu, 1-Ad, Mes, o-Tol Et3SiCl

Si TMS OSiEt3 TMS R Scheme 22

As a part of our studies of Si=C bonded compounds strongly influenced by reverse Si=C bond polarization, we recently found a novel and convenient pathway to metal 2-silenolates. While dissolving equimolar quantities of acylpolysilane 60 and tBuOK in THF at room temperature (Scheme 23), and following the reaction by 1H NMR spectroscopy, we found that potassium 2- silenolate 61 is formed as a single product.27b

TMS O tBuOK, THF, rt TMS O K TMS Si t Si - BuOTMS t TMS tBu TMS Bu 60 61 Scheme 23

42 The novel potassium 2-silenolate 61 possesses nearly the same 1H, 13C and 29Si NMR spectroscopic characteristics as the lithium 2-silenolates (Table 8).To our surprise, the potassium 2-silenolate 61 is stable at room temperature for at least three months. The question why 61 is stable at ambient temperatures and under light, in contrast to lithium 2-silenolates, is still open. However, the enhanced stability of potassium 2-silenolates vs. the lithium analogues presumably relates to differences in the aggregation behavior of the two species in solution. In addition, B3LYP/6-31G(d) calculations also showed that lithium 2-silenolates are less strongly affected by reverse polarization than are potassium 2-silenolates.69 Therefore, potassium 2-silenolates should be more stable toward dimerization.

4.2 Trapping experiments and metal exchange reactions The reactivity of potassium 2-silenolate 61 (Scheme 24) toward 1,3-dienes and alkylhalides is similar to that of lithium 2-silenolates.

Si TMS OH 85 % TMS O 64 O O K O O , H2O -40 ° C, 14 h O

Me TMS O [18]Crown-6 TMS O K MeI O Si Si TMS Si TMS THF, rt TMS TMS 65 61 63 96 % 87 %

MX2 rt

TMS OMX Si TMS 98 % 62 MX2 = MgBr2 (62a), ZnCl2 (62b) Scheme 24.

Thus, potassium 2-silenolate 61 react with methyl iodide to yield the Si- methylated acylsilane 63 in nearly quantitative yield, and reaction of 61 with 2,3-dimethyl-1,3-butadiene affords only the [4+2] cycloadduct 64. Further- more, potassium 2-silenolate 61 undergoes metal exchange reactions with ZnCl2 and MgBr2 in THF at room temperature yielding the corresponding zinc and magnesium 2-silenolates 62 (Scheme 24).

43 One can observe that the counterion, as well as complexation of the counterion with [18]crown-6, influences only slightly the NMR chemical shifts of the 2-silenolates (Table 8).

Table 8. 13C and 29Si NMR chemical shifts for the central C and Si atoms of metal 2-silenolates. TMS O M Si TMS R R M į 13C (ppm) į 29Si (ppm) Solvent

t b b Bu (59) Li 274.3 -70.3 THF:THF-d8(7:3)

t Bu (61) K 274.1 -78.7 THF-d8

t a Bu (65) K 268.7 -93.8 THF-d8

t Bu (62a) MgBr 264.8 -69.4 THF-d8

t Bu (62b) ZnCl 245.9 -74.8 THF-d8 a 2-Silenolate complex with [18]Crown-6. b Reported in reference 37.

4.3 Formation and X-ray structural characterization of the first crystalline metal 2-silenolate [18]crown-6 salt The remarkable thermal stability of potassium 2-silenolate 61 encouraged us to crystallize this compound in order to characterize it by means of X-ray crystallography. To enhance its crystallization we used [18]crown-6, previously used to crystallize other polysilyl anions.70 Indeed, treatment of in situ generated 61 with [18]crown-6 in toluene at room temperature leads to an orange solution from which yellow crystals suitable for X-ray crystallography were grown in two days. The complex 65 is the first structurally characterized enolate with one C atom exchanged to a heavy Group 14 element, and the X-ray crystallographic studies revealed its remarkable geometric characteristics (Figure 12).

44 Figure 12. ORTEP drawing of 65. The hydrogen atoms have been omitted for clarity. The thermal ellipsoids are set at the 50 % probability level. Selected bond lengths [Å], bond angles [°], and dihedral angles [°]: Si1-Si2 2.350(1), Si2-Si3 2.333(1), Si2-C7 1.926(3), C7-O1 1.245(3), K1-O1 2.846(2), K1-Si2 3.714(1), Si1- Si2-Si3 103.21(3), Si1-Si2-C7 117.95(8), Si3-Si2-C7 96.61(8), C7-O1-K1 113.2(2), Si2-C7-O1 116.6(2), C8-C7-O1 117.2(2), C8-C7-Si2 125.9(2), O1-C7-Si2-Si1 15.1(3), O1-C7-Si2-Si3 56.7(2).

The potassium ion is closer to the O atom than to the Si atom, the formal Si=C double bond is even longer than a normal Si-C single bond (1.926 Å in 65 vs. 1.87 Å for reference Si-C single bonds71), and the C=O double bond distance is only slightly elongated (1.234 Å in 65 vs. ~1.20 Å for reference C=O double bonds 72). Finally, the unsaturated Si atom in 65 is markedly pyramidal ((6Si) = 317.8°) indicating that resonance structure II is dominant (Scheme 17). Normally in silyl anions the value of 6Si is smaller, for exam- t 73 ple 6Si = 304.8° in BuMe2(TMS)2SiLi(THF)3 as measured experimentally. To aid in understanding the structure of 65, quantum chemical calculations - + on the model complex [(H3Si)2Si=C(O)Me] K (Me2O)3 were performed at B3LYP/6-311G(d) level, and the calculated and experimentally determined geometries are in excellent agreement.27b It is noteworthy that the related silene, e.g. 4-silatriafulvene (14, Scheme 8) of Kira and co-workers,36 was predicted to have a fluxional nature of the skeleton which is not planar around the Si=C bond but trans-bent. The energy difference between trans-bent and planar structures for 14 is only 1 kcal/mol, estimated at MP2/6-311++G(d,p) + ZPE level of theory. However, Kira and co-workers found that the structure of 14 determined by X-ray crystallography was not consistent with the computed trans-bent form since the experimental structure of 14 is planar. It was considered that crystal packing enforces a planar structure. In this regard our potassium 2-silenolate is also the first experimental evidence that a compound that formally could be Si=C double bonded can a have pyramidal Si atom. Based on these results

45 we believe that one can generate 2-silenolates having chiral Si centers. The experiments in the latter matter are in progress and will be presented else- where. Furthermore, the enhanced stability of the potassium 2-silenolates will be useful in future exploration of the 2-silenolate chemistry.

4.4 Potassium 1-Dialkylamino-2,2-bis(trimethylsilyl)-2- Silenolates: A Class of Remarkably Stable 2-Silenolates

We have already demonstrated that the Si=C bond can be stabilized toward dimerization by adding two S-electron donating substituents at the C end. We have also shown that zwitterionic silenes are closely related to metal 2- silenolates, both by geometrical parameters and by chemoselectivity. There- fore, it was interesting to synthesize silylamide enolates 67 (Scheme 25) and explore if the presence of a second S-electron donating substituent can further stabilize the 2-silenolates and increase its chemoselectivity in various transformations. Also it motivated investigation on silylamide enolates in analogy to the interesting chemistry of the amide enolates in organic synthesis. Alkylation using metal amide enolates has recently become an important synthetic strategy for the construction of various complex carbon framework from simple precursors including development of asymmetric Į-alkylation,74 palladium-catalyzed intermolecular couplings of aryl and vinyl halides with amide enolates,75 and palladium-catalysed Ȗ- and į-lactam formation meth- odologies 76. Moreover, the amide bond in amide enolates has received some attention due to the importance of the amide moiety in peptide chemistry.77 The first potassium enolates of N,N-dialkylcarbamylpolysilanes 67a - c were synthesized by the same approach as used for the synthesis of potassium 2-silenolates (Scheme 24). Upon reacting an equimolar amount of N,N- dialkylcarbamylpolysilanes 38a, 38f and 38g with tBuOK at room temperature in THF or toluene, and monitoring the reaction by 1H and 13C NMR spectroscopy, it was found that the reactions are completed in less then 10 minutes, leading to the formation of potassium enolates of N,N- dialkylcarbamylpolysilanes as single products (Scheme 25). These new compounds were characterized by 1H, 13C and 29Si NMR spectroscopy. Al- kylation of the 1-dialkylamino-2-silenolates with MeI leads to quantitative formation of the Si methylated products 67a, 67f, and 67g (Scheme 25), whereas the reaction with 2,3-dimethyl-1,3-butadiene at room temperature leads to a complex product mixture. An attempt to trap 66a by 2,3-dimethyl- 1,3-butadiene in a Diels-Alder reaction at room temperature, and followed by O-silylation with TMSCl, failed because this reaction lead to formation of polymeric material. It can be noted that the carbon analogues, i.e. amide enolates, are often used as initiators of anionic polymerization of a wide

46 range of monomers.78 Lowering the reaction temperature to -45 °C significantly decreases polymer formation, but the cycloadduct 45a was formed in less than 5% yield (Scheme 25). The observed low rate of the Diels-Alder reaction should be due to the electron-rich characters of both the 2-silenolate 66a (as a dienophile) and 2,3-dimethyl-1,3-butadiene. The result is not surprising as the silene 39a which is closely related to 66a (both compounds are electron-rich dienophiles) undergoes reaction with 2,3-dimethyl-1,3- butadiene only at 180 ºC, and the reaction is completed in 48 h (Scheme 15).

TMS Si TMS TMSO NRR' l 45a C 5 % TMS O K S M Si I T , h TMS NRR' 2 1 , o C 5 Me TMS TMS O K 4 O t - O BuOK MeI, 15 min. TMS Si Si II TMS Si TMS NRR' NRR' TMS NRR' THF, rt, 10 min TMS 67 [18 88% 38 ]crown-6 TMS O K O III Si O O TMS NRR' K 66 O O O 38a / 66a / 45a / 67a / 68a R = R' = Me 96 % TMS O Si 38f / 66f / 67f / 68f NRR' = 1-piperidyl TMS NRR' 68 38g / 66g / 67g / 68g R = R' = nBu

Scheme 25

The reaction of 66a, 66f, and 66g with equimolar amount of [18]crown-6 leads to the formation of the corresponding crown ether complexes of the 2- silenolates (Scheme 25). These reactions were performed in order to enhance crystallization of 66a and 66f in a similar fashion as for potassium 2- silenolate 65. However, it was not possible to obtain suitable crystals of 68f for X-ray crystallography. Instead, suitable crystals of 66f for X-ray crystallography were grown directly in benzene solution during one day at room temperature without the use of any oxygen containing solvents. The X-ray crystal structure and selected geometrical parameters of 66f is shown in Fig- ure 13. The central Si atom in 66f is even more pyramidal than in 65 since the sum of the valence angles at Si is 309° vs. 317° in 65. The Si-C distance

47 was found at 1.933 Å i.e. slightly elongated by 0.007 Å when compared to 65.

Figure 13. ORTEP drawing of 66f. The hydrogen atoms have been omitted for clarity. The thermal ellipsoids are set at the 50% probability level. Selected bond lengths [Å], and bond angles [°]: Si1-Si2 2.348(3), Si2-Si3 2.341(3), Si2-C12 1.933(7), C12-O1 1.251(7), K1-O1 2.794(5), K1-Si2 3.574(2), C12-N13 1.345(8) Si1-Si2-Si3 104.03(10), Si1-Si2-C12 109.40(2), Si3-Si2-C13 96.4 (2), C12-N13- C11 121.4(8), C7-N13-C11 115.6(8), C12-O1-K1 164.2(5), N13-C12-O1 117.7(6), Si2-C12-N13 123.3(6).

The thermal stability of 66a, 66f, and 66g relative to that of 65 is also nota- ble. When heating benzene-d6 solutions of 65, 66a, 66f, and 66g in sealed NMR tubes at 90 °C and monitoring the stability of the 2-silenolates by 1H and 13C NMR spectroscopy, we found that 65 degrade completely within 5 hours, whereas 66a, 66f, and 66g are considerably more stable. Approxi- mately 40% of 66a and 66f have decomposed after 10 hours heating at 90 ºC. This finding is unexpected since the amide and ester enolates are less stable than ketone enolates. Whereas 1-dialkylamino-2-silenolates are remarkably stable, the opposite is true for the 1-amino-2-silenolate derived from N,N-diarylcarbamyl- polysilane 38c. Reaction of 38c with tBuOK under the same conditions used to prepare 66 leads to quantitative formation of potassium diphenylamide [18]crown-6 complex 69 (1H and 13C NMR spectroscopy and X-ray crystallographic data are in consistence with reported data79). The other products of this reaction is carbon monoxide, as verified by reduction of PdCl2 solution by the gaseous substance formed in course of the reaction,80 and a complicated mixture of silicon containing compounds. The pathway shown in Scheme 26 is proposed to explain formation of potassium diphenylamide 69, carbon monoxide, and silylene 70. To validate this reaction sequence, the 2- silenolate formation was carried out in the presence of excess 2,3-dimethyl- 1,3-butadiene, which also functions as a trapping reagent for silylene. In- deed, the latter experiment yielded the [4+1] adducts 71 in 17 % isolated

48 yield. The 1H, 13C NMR and MS spectroscopic data of 71 are in consistence with the literature data.81

TMS O tBuOK, [18]Crown-6 TMS TMS Si + CO+ Si TMS TMS NPh2 rt, 10 min. N 69K --- [18]Crown-6 70 38c 96%

17% Si TMS TMS 71

Scheme 26

In conclusion, we have reported the synthesis, chemical reactions, and X-ray structural characterization of the first Si analogues of amide enolates that are thermally more stable than metal 2-silenolates.

49 Summary and Outlook

Transient zwitterionic 2-amino-2-siloxysilenes and persistent potassium 2- silenolates have been formed, and their reactivity has been studied. Reac- tions of the 2-amino-2-siloxysilenes with 1,3-dienes give silacyclohexenes as one single diastereomer in nearly quantitative yields. Through variation of the dienes, it is shown that these reactions proceed in accord with inverse- electron-demand Diels-Alder reactions. In contrast to other silenes, it is found that the 2-amino-2-siloxysilenes do not react with alcohols, but that instead the precursor carbamylpolysilane reacts with the alcohol to form silylethers. To our knowledge, this should represent a new base-free protocol for protection or primary and secondary alcohols. With regard to the 2- silenolates it is revealed that these heavy enolates are structurally different to enolates. Similar as zwitterionic silenes, they react with 1,3-dienes to yield exclusively [4+2] cycloadducts. The first silicon analogues of amide enolates have also been prepared and their reactivity was studied. Future work could be directed toward trapping of the transient zwitterionic 2-amino-2-siloxysilenes with appropriate transition metal complexes, followed by NMR and possibly X-ray structural characterization. The cycloaddition reactions of the zwitterionic silenes with heterodienes could also be explored. For example, Si containing tetraquinoline derivatives arising from a proposed reaction of arylaldimines with zwitterionic silenes could be of interest for medicinal chemistry since the normal tetraqiunoline derivatives are used as pharmaceutical agents. With regard to stable metal 2-silenolates and potassium 1-dialkylamino-2- silenolates, further work could be directed in several areas: the exploration of Diels-Alder reactions with different 1,3-dienes, as well as possibility to use these species for construction of complex silacyclic systems, such as Si containing Ȗ- and į-lactam formation, via palladium mediated intramolecular coupling of aryl halides with the potassium 2-silenolates and 1- dialkylamino-2-silenolates. Research could also be conducted in order to investigate the remarkable stabilities of potassium 2-silenolates vs. the lithium analogues. Our method for the formation of stable metal 2-silenolates could presumably be employed for the formation of chiral 2-silenolates with Si as a chiral center. Finally, with regard to heavy enolates one could envi- sion that the presently reported findings in silenolate chemistry can be trans- ferred to the chemistry of germenolates.

50 Summary in Swedish

Till skillnad från andra naturvetenskapliga discipliner möjliggör syntetisk kemi framtagande av ämnen som inte är naturligt förekommande. Syntetisk kemi möjliggör design av kemiska föreningar med förbättrade egenskaper för att fungera inom vitt skilda tillämningar, exempelvis som läkemedel, som polymera material med funktionella egenskaper eller som laddningsbärare inom molekylärelektronik. Det är inte överraskande att utvecklande av nya kemiska reaktioner utgör en central men utmanande del av syntetisk kemi. Dessa nya reaktioner kan leda till att man mer effektivt kan bilda redan kän- da klasser av föreningar, men även att man kan framställa helt nya förening- ar med intressanta fysikaliska, kemiska, och/eller biologiska egenskaper som kan ges olika tillämpningar. Syntetisk organisk kemi innefattar i huvudsak föreningar baserade på grundämnet kol (C). Föreningar med enkel, dubbel, och trippelbindningar, så kallade alkaner (C-C), alkener (C=C), alkyner (CŁC), ketoner (C=O), iminer (C=N) m.fl., har en central position inom organisk kemi. Industriell produktion av nya läkemedel, polymera material, och parfymer involverar alla syntetisk organisk kemi under olika stadier av utvecklingsprocessen. Kisel (Si) är det grundämne som står närmast under kol i det periodiska systemet. Det har därför alltid varit en utmaning att framställa föreningar med dubbel och trippelbindningar till kisel (Schema 1) för att sedan studera deras struktur, reaktivitet och andra egenskaper. En av de tidiga industriellt viktiga upptäckterna inom kiselkemi var syntesen av silikoner, vilka är linjä- ra och cykliska s.k. polysiloxaner. Professor Kipping vid University of Not- tingham i England gjorde denna upptäckt under 1930–talet då han försökte framställa kisel-syre dubbelbundna föreningar (Si=O) vilka han kallade silikoner (silicon ketones, Sv. kiselketoner). Det visade sig dock att reaktionen mellan en bas och diklorosilaner istället för Si=O bundna föreningar gav linjära såväl som cykliska polysiloxaner, numera kallade silikoner, sannolikt via intermediära Si=O bundna föreningar. Dessa försök representerar därmed några av de första försöken att framställa kisel-analoger till alkener, ketoner, iminer, och alkyner, och numera har ett flertal av kiselanalogerna framställts och deras egenskaper har studerats. Under det gångna året har syntes av den första SiŁSi trippelbundna föreningen, en s.k. disilyn, rapporterats (se Scien- ce 2004, 305, 1755).

51 O C CC CC CC OC en alken ett enolat en alkyn en keton

O Si Si C Si C Si Si OSi en silen ett 2-silenolat en disilyn en "silikon"

Si Si

en disilen Schema 1

Det forskningsprojekt som presenteras i denna doktorsavhandling behandlar silener, d.v.s. föreningar med kisel-kol dubbelbindningar (Si=C). Speciellt fokus har lagts på Si=C bundna föreningar med en polaritet som är omvänd till den naturliga vilket innebär att de har en SiG-=CG+ polariserad bindning istället för SiG+=CG- polariserad (Schema 2). Den omvända polariseringen åstadkoms med substituenter på C atomen som donerar S-elektroner till Si=C bindningen. Detta leder till att man förflyttar negativ laddning till Si atomen och föreningen blir zwitterjonisk (Schema 2). Dessa silener, vilka vi benämner zwitterjoniska silener, saknar egentlig motsvarighet inom kolke- min. Nära ”kemiska släktingar” med de zwitterjoniska silenerna är 2- silenolater, och vi har även studerat sådana. De är dessutom kiselanaloger till de inom organisk kemi mycket användbara enolaterna (Schema 1).

G G G G Si C Si C

en naturligt polariserad silen en omvänt polariserad silen

X X X Si Si Si Y Y Y

X,Y=NR2,OR,SR. Schema 2

Tidigare har endast en handfull Si=C bundna föreningar som är stabila vid rumstemperatur framställts. Anledningen till att så få studerats beror på att de mycket gärna reagerar med luftens fukt och syre. Silenerna reagerar även i frånvaro av fukt och syre, främst genom att dimerisera. Att silener är insta- bila och svårhanterbara bör vara det främsta skälet till att de ännu inte funnit några tillämpningar inom organisk syntes. De i avhandlingen beskrivna stu- dierna visar att zwitterjoniska silener och 2-silenolater reagerar mer selektivt

52 än naturligt polariserade silener. Kvantkemiska datorberäkningar tidigare utförda inom gruppen har visat att de zwitterjoniska silenerna har en lägre tendens att dimerisera och en högre motståndskraft mot fukt och syre än naturligt polariserade silener. Vi fann att det mest effektiva sättet att fram- ställa transienta zwitterjoniska silener är att upphetta karbamylpolysilaner i närvaro av 1,3-diener (Schema 3), reaktioner som ger höga utbyten och mycket specifikt leder till att silacyklohexen-föreningar bildas i en enda s.k. diastereomer form. Så är inte fallet för naturligt polariserade silener vilka oftast reagerar mycket oselektivt med 1,3-diener och bildar ett flertal olika cykliska och acykliska produkter. Vi har vidare undersökt reaktionen mellan zwitterjoniska silener och 1,3-diener utförligt och funnit att dessa reaktioner tillhör en särskild grupp av Diels-Alder-reaktioner som benämns inverse- electron-demand Diels-Alder reaktioner. Därvid kan man förklara vilka diener som kommer att reagera snabbt med de zwitterjoniska silenerna och vilka som kommer att reagera långsamt. Enkla kinetiska studier visar vilket steg som är hastighetsbestämmande i reaktionen, beronde på silenens struktur och laddningsfördelningen i Si=C bindningen. I fallet med med zwitterjoniska silener har vi funnit att det hastighetsbestämmande steget utgör bil- dandet av själva silenen, medan vi för mindre elektronrika silener har funnit att reaktionen mellan silenen och 1,3-dienen är det hastighetsbestämmande steget. Vi har bevisat experimentellt att zwitterjoniska silener, framställda från karbamylpolysilaner, är måttligt motståndskraftiga mot fukt och syre. Man behöver därför inte utestänga luft eller använda vattenfria (torra) lös- ningsmedel när man utför reaktioner med dessa silener. I fallet med naturligt polariserade silener måste luft och vatten utestängas och reaktionerna utföras i inertgas atmosfär. De zwitterjoniska silener vi framställt reagerar ej heller med alkoholer utan silylamiden reagerar och bildar silyletrar.

SiMe3 O Toluen Me3Si OSiMe3 Si Me3Si Si Si Me3Si NMe2 Me3SiO NMe NMe2 180 ° C 2 SiMe3 Me3Si SiMe3 Schema 3

SiMe 3 O tBuOK Me3Si O Li Me3Si Si Si t Bu THF, rt Me3Si tBu SiMe3 Schema 4

53 Som nämndes ovan har även 2-silenolater studerats inom det projekt som presenteras i avhandlingen. 2-Silenolaternas kemi är ännu mindre utforskat jämfört med silenernas kemi, huvudsakligen på grund av deras tidigare låga stabilitet vid rumstemperatur och deras ljuskänslighet. Endast tio vetenskap- liga artiklar om 2-silenolater har publicerats fram till och med oktober 2004. I avhandlingen presenteras en mycket enkel och effektiv metod att vid rumstemperatur framställa kalium 2-silenolater genom att reagera olika acylpoly- silaner med kalium tert-butoxid (Schema 4). Till vår förvåning fann vi att dessa kalium 2-silenolater, till skillnad mot de litium 2-silenolater som tidigare studerats av andra, är stabila vid rumstemperatur under lång tid. De måste dock förvaras i en inertgas-atmosfär. Reaktionerna mellan kalium 2- silenolater och elektronfattiga alkylhalider och 1,3-diener är snarlika motsva- rande reaktioner för litium 2-silenolater. Fördelen med de kalium 2- silenolater vi framställt jämfört med tidigare litium 2-silenolater är framför- allt deras högre värmetålighet. Vi har även framställt två kristallina kalium 2-silenolater lämpliga för röntgenkristallografi och dessa är de första struk- turundersökningarna av grupp 14 enolater. Komplexens geometrier visade sig stämma väl överens med datorberäkningar vi utfört i gruppen. Strukturen är inte plan vid den centrala kiselatomen utan de tre bindningarna utgående från detta kisel utgör en pyramid. Vidare är bindningslängden för Si=C bindningen till och med längre än en normal Si-C enkelbindning. Ännu mer negativ laddning kan delokaliseras till Si atomen om alkylgruppen på kolatomen i Si=C bindningen byts ut mot en S-elektrondonerande dimetyla- minogrupp. Reaktionerna mellan denna nya 1-amino-2-silenolaten och alkylhalider samt 1,3-diener har studerats. Röntgenkristallografi visade en ännu mer pyramidaliserad central kiselatom och en förlängning av Si=C bindningen jämfört med andra 2-silenolater. Denna observation antyder att 1-amino-2-silenolater är lämpliga byggstenar för design och syntes av op- tiskt aktiva 2-silenolater.

Översatt från engelska till svenska av Anders Eklöf och Henrik Ottosson

54 Acknowledgement

I would like to express my sincere gratitude to my supervisor Docent Henrik Ottosson for accepting me in his newly established group, for introducing me to Silicon Chemistry and for his support during these years.

Past and present members of the HO group are gratefully acknowledged: Dr. Ibrahim El-Sayed, Dr. Khalil Chajara, Dr. Haruhisa Kato, Niclas Sandström, Dr. Hayato Tsuji, Dr. Ahmed El-Nahas, Dr. Arnaud Martel, Dr. Malgorzata Gajewska, Julius Tibbelin, and Anders Eklöf.

Dr. Haruhisa Kato, David Shanks, Anna Trifonova, Phil. Lic. Oleksiy Itsenko, Julius Tibbelin and Anders Eklöf are gratefully acknowledged for proof reading and criticism of this thesis.

Special thanks to the administrative and technical personnel of the Depart- ment of Organic Chemistry: Eva, Leif, Thomas and Gunnar.

Many thanks to all past and current members of the Organic Chemistry De- partment, especially: Anna, Oleksiy, Eskender, Senait, Antti, Magnus E. and Magnus B., Stefan M, Jarle, Klas, for a unforgettable times we spent together inside or outside of the Chemistry building.

Docent Adolf Gogoll is gratefully acknowledged for his invaluable help with 29Si NMR spectroscopy and for keeping NMR and MS machines in excellent working conditions.

Special thanks to my friends outside the Chemistry Department: Eka, George, Zura, Lyudmila, Elnara, Maral and Tibaut, David and Alexandra, Satkyn, Zhanna: without you it was not really possible for the south man to stay in the north for these couple of years.

Last but not least, thanks to my family Nino and Megi for their support during these years.

55 References and notes

1. a). Kipping, F.S. Proc. R. Soc. 1911, 27, 143. b). Kipping F. S.; Sands, J. E. J. Chem. Soc. (Trans). 1921, 119, 834. c). Kipping, F.S. J. Chem. Soc. 1924, 125, 2291. d). Kipping, F.S. Proc. R. Soc. 1937, 159, 139. e). Dasent, W.E. Nonexist- ing compounds. Marcel Dekker: New-York. 1965, chap. 4, p.61 f). Pitzer, K.S. J. Am. Chem. Soc. 1948, 70, 2140. 2. a). Jutzi, P. Angew. Chem. 1975, 87, 269. b). Jutzi, P. Angew. Chem. Int. Ed. 1975, 14, 232. 3. a). Gusel’nikov, L.E.; Flowers, M.C. J. Chem. Soc. Chem. Commun. 1967, 864. b). Nametkin, N.S.; Vdovin, V.M.; Gusel’nikov, L.E.; Zav’yalov, V.I. Isv. Akad. Nauk SSSR ser. Khim. 1966, 584. c). Gusel’nikov, L.E.; Nametkin, N.S. Chem. Rev. 1979, 79, 529. 4. a).Brook, A.; Abdesaken, G.; Gutekunst, B.; Kallury, R.K. J. Chem. Soc. Chem. Commun. 1981, 191. b). Brook, A.G.; Nyburg, S.C.; Abdesaken, F.; Gutekunst, B.; Gutekunst, G.; Krishna, R.; Kallury, M. R.; Poon, Y.; Chang, Y.;Winnie, W. J. Am. Chem. Soc. 1982, 104, 5667. 5. West, R.; Fink, M.; Michl, J. Science 1981, 214, 1343. 6. Yoshifuji, M.; Shima, I.; Inamoto, N.; Hirotsu, K.; Higuchi, T. J. Am. Chem. Soc. 1981, 103, 4587. 7. Wiberg, N.; Wagner, J.; Reide and Müller, G. Organometallics 1987, 6, 32. 8. Miracle, G.E.; Ball, J.L.; Powell, D.R and West, R. J. Am. Chem. Soc. 1993, 115, 11598. 9. a). Hemme, I.; Klingebiel, W. Adv. Organomet. Chem. 1996, 39, 193. b). Wiberg, N.; Schultz, K.; Fischer, G. Angew. Chem. Int. Ed. 1985, 24, 1053. c). Hesse, M.; Klingebiel, W. Angew. Chem. Int. 1986, 25, 649. 10. a). Arya, P.; Boyer, J.; Carre, F.; Corriu, R.; Lanneau, G.; Lapasset, J.; Perrot, M.; Priou, C. Angew. Chem. Int. Ed. 1989, 28, 1016. b). Suzuki, H.; Tokitoh, N.; Nagase, S.; Okazaki, R. J. Am. Chem. Soc. 1994, 116, 11578. 11. Tokitoh, N.; Wakita, K.; Okazaki, R.; Nagase, S.; Schleyer P.v.R.; and Jiao. H. J. Am. Chem. Soc. 1997, 119, 6951. 12. a). Jutzi, P.; Kanne, D.; Kruger, C. Angew. Chem. Int. Ed. 1986, 25, 164. b). Jutzi, P. Cyclopentadienyl Silicon Compounds, In The Chemistry of Organic Silicon Compounds, Rappoport, Z.; Apeloig, Y., Eds., Wiley, Chichester, 1998, vol. 2, Chap. 36, p. 2129 c). Denk, M.; Lennon, R.; Hayashi, R.; West, R.; Haaland, A.; Belyakov, H.; Verne, P.; Wagner, M.; Metzler, N. J. Am. Chem. Soc. 1994, 116, 2691. 13. Sekiguchi, A.; Kinjo, R.; Ichinohe, M. Science 2004, 305, 1755. 14. a). Michl, J. and Raabe, G. Chem. Rev. 1985, 85, 419. b). Raabe, G. and Michl, J. In The Chemistry of Organic Silicon Compounds, vol 1. Eds. Patai, S. and Rappoport, Z. Wiley, Chichester, 1989. c). Brook, A. and Brook, M.A. Adv. Or- ganomet. Chem. 1995, 39, 71. d). Müller, T.; Ziche, W.; Auner, N. In The Chemistry of Silicon Compounds, (Edited by Rappoport, Z. and Apeloig, Y. Wiley , New-York, 1998, Chap. 16, 859.

56 15. Sanganee, M. J.; Steel, P. G.; Whelligan, D. K. Org. Biomol. Chem. 2004, 2, 2393. 16. Kei, S. R.; Tilley, T.D.; Bergman R.G. Organometallics 2002, 21, 4648. 17. Yajima, S.; Nasegawa, Y; Okamura, K.; Natsuzowa, T. Nature 1978, 273, 525. 18. Procopio, L. J.; Berry, D.H. J. Am. Chem. Soc. 1991, 113, 4039. 19. Although masked disilenes have been polymerized anionically, monomeric disilenes (R2Si=SiR2) are not believed to be involved in this anionic ring- opening polymerization. See: a). Sakamoto K.; Obata, K.; Hirata, H.; Nakajima, M.; Sakurai, H. J. Am. Chem. Soc. 1989, 111, 7641. b). Yoshida, M.; Sakurai, H. Polymer 1994, 35, 4990. 20. Tsang, Chi-Wing.; Yam, Mandy.; Gates, D. P., J. Am. Chem. Soc. 2003, 125, 1444. 21. a). Apeloig, Y.; Karni, M. J. Am. Chem. Soc. 1984, 106, 6676. b). Ottosson, H. Chem. Eur. J. 2003, 9, 4144. 22. Schmidt, M.W.; Troung, P.N.; Gordon, M. S. J. Am. Chem. Soc. 1987, 109, 5217. 23. a). Gusel’nikov, L.E.; Nametkin, N. S.; Vdovin, V.M. Acc. Chem. Res. 1975, 8, 18. b). Walsh. R. J. Organomet. Chem. 1972, 38, 245. 24. a). Dobbs, K.D.; Hehre, W. J. Organometallics 1986, 5, 2057. b). Schmidt, M. W.; Gordon, M. S.; Dupuis, M. J. Am. Chem. Soc. 1985, 107, 2585. 25. Brook, M. Silicon in Organic, Organometallic and Polymer Chemistry, John Wiley & Sons, Inc., New York, 2000, p. 30. 26. a). Roelandt, F. F.; van de Vondel, D. F.; van der Kelen, G. P. J .Organomet. Chem. 1979, 165, 151. b). Lischka, H; Köhler, H. J. Chem. Phys. Lett. 1982, 85, 467. c). Fink, M. J.; Michalczyk, M. J.; Haller, K. J.; West, R., Michl, J. J. Chem. Soc. Chem. Commun. 1983, 1010. d). Masamune, S.; Urakami, S.; Snow, J. T.; Tobita, H.; Williams, D. J. Organometallics, 1984, 3, 333. 27. a). El-Sayed, I.; Guliashvili, T.; Hazell, R.; Gogol, A.; Ottosson, H. Org. Lett. 2002, 4, 1915. b). Guliashvili, T.; El-Sayed, I.; Fischer, A.; Ottosson, H. Angew. Chem. Int. Ed. 2003, 42, 1640. 28. a). Bailleux, S.; Bogey, M.; Breidung, J.; Burger, H.; Fajgar, R.; Liu, Y.; Pola, L.; Senzlober, M. and Thiel, W. Angew. Chem. Int. Ed. 1996, 35, 2513. b). Bail- leux, S.; Bogey, M.; Demaison, J.; Senzlober, M.; Breidung, J.; Thiel, W. Fa- jgar, R.; Pola, L. J. Chem. Phys. 1997, 196, 10016. 29. a). Volkova, V.V.; Avakyan, V.G.; Nametkin, N.S.; Gusel’nikov, L.E. J. Or- ganomet. Chem. 1980, 201, 137. b). Chen, J.; Hajduk, P.J.; Keen, J.D.; Chuang, C.; Emilsson, T.; Gutowsky, H.S. J. Am. Chem. Soc. 1991, 113, 4747. c). Mo- haffy, P.; Gutowsky, R.; Montgomery, L.K. J. Am. Chem. Soc. 1980, 102, 2854. 30. Boesveld, W. M.; Gehrhus, B.; Hitchcock, P.B.; Lappert, M. F.; Schlayer, P. v. R., J. Chem. Soc. Chem. Commun. 1999, 755 31. Blinka, T. A.; Helmer, B.J.; West, R. Adv. Organomet. Chem. 1984, 23, 193. 32. Koeing, J. L. Spectroscopy of Polymers American Chemical Society, Washing- ton, DC, 1992. 33. a). Brook, A. G.; Abdesaken, G.; Gutekunst, G.; Plavac, N. Organometallics 1982, 1, 994. b). Baines, K.M.; Brook, A. G.; Ford, R.R.; Lickiss, P.D.; Saxena, A.K.; Chatterton, W.J.; Sawyer, J. F.; Behnam, B.A. Organometallics, 1989, 8, 693. 34. Brook, A.; Wessely, H. J. Organometallics, 1985, 4, 1487. 35. Wiberg, N; Wagner, G.; Muller, G. Angew. Chem. Int. Ed. 1985, 24, 229. 36. Sakamoto, K.; Ogasawara, J.; Kon, Y.; Sunagawa, T.; Kabututo, M.; Kira, M. Angew. Chem. Int. Ed. 2002, 41, 1402.

57 37. Ohshita, J.; Masaoka, S.; Masaoka, Y. ; Hasebe, H.; Ishikawa, M.; Tachibana, A.; Yano, T.; Yamabe, T. Organometallics 1996, 15, 3136 38. Apeloig, Y.; Bendikov, M.; Yuzefovich, M.; Nakash, M.; Bravo-Zhivotovskii, D.; Bläser, D.; Boese, R. J. Am. Chem. Soc. 1996, 118, 12228. 39. Sekiguchi, A.; Ando, W. Organometallics 1987, 6, 1857. 40. a) Jensen, F. Introduction to Computational Chemistry, Wiley, Chichester, 1999. b) Christopher J. Cramer, Essentials of Computational Chemistry: Theo- ries and Models, John Wiley & Sons, New York, 2002. 41. a). Jones, P. R.; Cheng, A. H. B.; Albanesi, T.E., Organometallics, 1984, 3,78. b). Auner, N.; Gleixner, R. J. Organomet. Chem. 1990, 393, 33. c). Auner, N. J. Organomet. Chem. 1988, 353, 275. d). Auner, N.; Ziche, W.; Hardtweck, E. J. Organomet. Chem. 1992, 426, 1. e). Brook, A.; Nyburg, S. C.; Reynolds, W. F.; Poon, Y. C.; Chang, Y. M.; Lee, J. C.; Picard, J. P. J. Am. Chem. Soc. 1979, 101, 6750. f). Brook, A.; Baumegger, A.; Lough, A. J. Organometallics 1992, 11, 3088. g). Baines, K.M.; Brook, A.; Ford, R.R.; Lickiss, P.D.; Saxena, A.K., Chatterton, W.J.; Sawyer, J. F.; Behnam, B.A. Organometallics 1989, 8, 693. h). Ohshita, J.; Masaoka, Y.; Ishikawa, M. Organometallics 1991, 10, 3775. i).Ishikawa, M.; Kikuchi, M.; Kunai, A.; Takeuchi, T.; Tsukihara, T.; Kido, M. Organometallics 1993, 12, 3474. 42. Zhang, S.; Conlin, R. T.; McGarry, P. F.; Scaiano, J. C. Organometallics 1992, 11, 2317. 43. a). Brook, A.; Baines, K.M. Adv. Organomet. Chem. 1986, 25, 1. b). Chaubon, M. A.; Ranaivonjatovo, H.; Escudie, J.; Satge, J. Main Group Metal Chem. 1996, 19, 145. 44. Brook, A.; Vorspohl, R.R.; Ford R.R.; Hesse, M.; Chatterton, W.J. Or- ganometallics 1987, 6, 2128. 45. Ohshita, J.; Masaoka, S.; Morimoto, Y.; Sano, M.; Ishikawa, M. Organometal- lics 1997, 16, 1123. 46. Wiberg, N. J. Organomet. Chem. 1984, 273, 141. 47. a). Veszpremi, T.; Takahashi, M.; Hajgato, B.; Kira, M. J. Am. Chem. Soc. 2001, 123, 6629. b). Bendikov, M.; Quadt, A.R.; Rabin, O.; Apeloig, Y. Or- ganometallics 2002, 21, 3930. c). Nagase, S.; Kudo, T.; Ito, K. In Applied Quantum Chemistry, Eds.: Smith, V.H.; Schaefer, H.F.; Morokuma, K. Reidel- Dordrecht, The Netherlands, 1986, 249. 48. a). Morkin, T. L.; Leigh, W. J.; Tidwell, T.; Allen, A. D. Organometallics 2001, 20, 5707. b). Morkin, T. L.; Leigh, W. J. Acc. Chem. Res. 2001, 34, 129. 49. Sakurai, H. In The Chemistry of Silicon Compounds, Edited by (Rappoport, Z. and Apeloig, Y.) Wiley, New York, 1998, chap. 15. 859. 50. Lowry, T.H.; Richardson, K.S. Mechanism and Theory in Organic Chemistry, 2nd ed.; Harper & Row: New-York, 1981. 51. a). Barton, T.J.; Marquardt, G.; Kilgour, J. A. J. Organomet. Chem. 1975, 85, 317. b). Nametkin, N.S.; Dolgoporov, N.N.; Gusel’nikov, L.E. J. Organomet. Chem. 1979, 169, 165 c). Gordon, M.S.; Barton, T.J.; Nakano, H. J. Am. Chem. Soc. 1997, 119, 11966. 52. Conlin, R.T.; Namavari, M.; Chickos, J.S.; Walsh, R. Organometallics 1989, 8, 168. 53. Auner, N.; Davidson, S.; Ijadi-Maghsoodi; Lawrence, F.T. Organometallics 1986, 5, 431. 54. a).Wiberg, N.; Preiner, G. Angew. Chem. Int. Ed. 1977, 16, 328.b). Wiberg, N.; Preiner, G.; Scheida, O. Chem. Ber. 1981, 114, 2087. c). Wiberg, N.; Link, M.; Fischer, G. Chem. Ber. 1989, 122, 409. d). Wiberg, N.; Joo, K.S.; Polborn, K. Chem. Ber. 1993, 126, 67. d). Wiberg, N.; Wagner, G. Chem. Ber. 1986, 119,

58 1455. e). Wiberg, N.; Wagner, G. Chem. Ber. 1986, 119, 1467. f). Wiberg, N.; Wagner, G. Angew. Chem. Int. Ed. 1983, 22, 1005. 55. a). Krempner, C.; Reinke, H.; Oehme, H. Angew. Chem. Int. Ed. 1994, 33, 1625. b). Wustrack, R.; Oehme, H. J. Organomet. Chem. 1988, 352, 95. c). Oehme, H.; Wustrack, R.; Heiene, A.; Sheldrick, G.M.; Stalke, D. J. Organomet. Chem. 1993, 452, 33. 56. Eaborn, C.; Happer, D. A.; Hitchcock, P.B.; Hopper, S. P.; Safa, K. D.; Washburne, S. S.; Walton, D. R.M. J. Organomet. Chem. 1980, 186, 309. 57. a).Ohshita, J.; Masaoka, Y.; Ishikawa, M.; Takeuchi, T. Organometallics, 1993, 12, 876. b). Ohshita, J.; Masaoka, Y.; Masaoka, S.; Ishikawa, M.; Tachibana, A.; Yano, T.; Yamabe, T. J. Organomet. Chem. 1994, 473, 15. 58. Bravo-Zhivotovsky, D.; Braude, V.; Stanger, A.; Kapon, M.; Apeloig, Y. Organometallics 1992, 11, 2326. 59. Brook, A.G.; Harris, J. W.; Lennon, I.; El-Sheikh, M. J. Am. Chem. Soc. 1979, 101, 83. 60. Ishikawa, M.; Matsui, S.; Naka, A.; Ohshita, J. Organometallics 1996, 15, 3836. 61. a). Ishikawa, M.; Fuchikami, T.; Sugaya, T.; Kumada, M. J. Am. Chem. Soc. 1975, 97, 5923. b). Ishikawa, M.; Kumada, M. Adv. Organomet. Chem. 1981, 19, 51. 62. a). Ando, W.; Sekiguchi, A.; Sato, T. J. Am. Chem. Soc. 1982, 104, 6830. b) Ando, W.; Sekiguchi, A.; Sato, T. J. Am. Chem. Soc. 1981, 103, 5573. 63. Al-Juaid, S.; Derouiche, Y.; Hitchcock, P. B., Lickiss P. D.; Brook, A. G. J. Organomet. Chem. 1991, 403, 293. 64. Naka, A.; Ikada, J.; Motoike, S.; Yoshizawa, K.; Kondo, Y.; Kang, S-Y.; Ishi- kawa, M. Organometallics 2002, 21, 2033. 65. Brook, M.; Gottardo, C.; Balduzzi, S.; Mohamed, M. Tetrahedron Lett. 1997, 38, 6997. 66. a). Greene, T. W.; Protective Groups in Organic Synthesis, Wiley: New York, 1981. b). Greene, T. W.; Wuts, P. G. M. Protective Groups in Organic Synthe- sis, 2nd Ed., Wiley: New York, 1991. c). Kochienski, P. J. Protecting Groups, Theime Verlag: Stutgart, 1994. 67. a). Ohshita, J.; Masaoka, Y.; Masaoka, S.; Ishikawa, M.; Tachibana, A.; Yano, T.;Amabe, T. J. Organomet. Chem. 1994, 473, 15. b). Ohshita, J.; Masaoka, S.; Ishikawa, M. Organometallics 1996, 15, 2198. c). Ohshita, J.; Sakurai, H.; Tokunaga, Y.; Kunai, A. Organometallics 1999, 18, 4545. d). Ohshita, J.; Sa- kurai, H.; Masaoka, S.; Tamai, M.; Kunai, A.; Ishikawa, M. J. Organomet. Chem. 2001, 633, 131. e). Ohshita, J.; Tamai, M.; Sakurai, H.; Kunai, A. Or- ganometallics, 2001, 20, 1065. 68. a). Biltueva, I. S.; Bravo-Zhivotovskii, D. A.; Kalikhmann, I. D.; Vitkovski, V. Y.; Shevchenko, S. G.; Vyazankin, N. S.; Voronkov, M. G. J. Organomet. Chem. 1989, 368, 163. b). Bravo-Zhivotovskii, D. A.; Apeloig, Y.; Ovchinikov, V.; Igonin, V.; Struchkov, Y. T. J. Organomet. Chem. 1993, 446, 123. 69. Ottosson, H. unpublished results. 70. a). Tilley, T. D.; Yap, G. P. A.; Rheingold, A.; Freeman, W. P. Angew. Chem. Int. Ed. 1996, 35, 976. b). Blanton, J. R.; Diminnie, J. B.; Chen, T.; Wiltz, A. M.; Xue, Z. Organometallics 2001, 20, 5542. 71. Kaftory, M.; Kapon, M., Botoshansky, M. The Chemistry of Organic Silicon Compounds, Vol. 2 (Eds.: Z. Rappoport, Y. Apeloig), Wiley, Chichester, UK, 1998, pp. 181. 72. Berthier, G.; Serre J. The Chemistry of the Carbonyl Group, Vol. 1 (Eds.: S. Patai), Wiley, London, 1966, pp. 1.

59 73. Apeloig, Y.; Yuzefovich, M.; Bendikov, M.; Bravo-Zhivotovskii, D.; Klink- hammer, K. Organometallics, 1997, 16, 1265. 74. a). Kawabata, T.; Özturk, O.; Chen, J.; Fuji, K. Chem. Comm. 2003, 162. And references given therein. b). Futatsugi, K.; Yanagisawa, A., Yamamoto, H. Chem. Comm. 2003, 566. c). Matsuo, J. I.; Odashima, K.; Kobayashi, S. Org. Lett. 1999, 1, 345.d). Seyden-Penne, J. Chiral Auxiliaries and Ligands in Asymmetric Synthesis, John Wiley & Sons: New-York, 1995. 75. Culkin, D. A.; Hartwig, J. F. Acc. Chem. Res. 2003, 36, 234. 76. a). Lee S.; Hartwig, J. F. J. Org. Chem. 2001, 66, 3402. b). Honda, T.; Namiki, H.; Satoh, F. Org. Lett. 2001, 3, 631. and references given therein. 77. Pugh, J. K.; Streitwieser, A. J. Org. Chem. 2001, 66, 1334. 78. Shen, Y.; Zhu, S.; Pelton, R. Macromolecules 2001, 34, 376. 79. Hitchcock, P. B., Khvostov, A. V.; Lappert, M. F.; Protchenko, A. V. J. Orga- nomet. Chem. 2002, 647, 198. 80. There are various methods for the detection of CO, for example visit the web page: http://mattson.creighton.edu/CO/ (10/10/2004). 81. Ohshita, J.; Yoshitomi, T.; Ishikawa, M Organometallics 1994, 13, 3227. (In this work compound 71 was generated upon photolysis of polysilanes in the presence of 2,3-dimethylbutadiene, but the yield of 71 is only 7%).

Acta Universitatis Upsaliensis Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology Editor: The Dean of the Faculty of Science and Technology

A doctoral dissertation from the Faculty of Science and Technology, Uppsala University, is usually a summary of a number of papers. A few copies of the complete dissertation are kept at major Swedish research libraries, while the summary alone is distributed internationally through the series Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology. (Prior to October, 1993, the series was published under the title “Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science”.)

Distribution: Uppsala University Library Box 510, SE-751 20 Uppsala, Sweden www.uu.se, [email protected]

ISSN 1104-232X ISBN 91-554-6080-1