Silylenes Coordinated to Lewis Bases

Johannes Belzner    Institut für Organische Chemie der Georg-August-Universität Göttingen, D-37077 Göttingen, Germany

Heiko Ihmels    Institut für Organische Chemie der Universität Würzburg, D-97074 Würzburg, Germany

I INTRODUCTION

Silylenes are, in a certain sense, the younger but bigger siblings of carbenes, and their chemistry has been reviewed several times.1 In 1978, Gaspar asked the question: “Are we to conclude from our present knowledge that nature has designed silylenes to humbly mimic carbenes, or is the situation different?”1a Today, we know that silylenes and carbenes have a number of properties in common, but also show significant differences.

One of these differences is the fact that H2Si: has a singlet ground state with an empty p orbital and a doubly occupied σ orbital, which is in contrast to the well-known triplet ground state of H2C:. Moreover, calculations show that, with the exception of silylenes bearing electropositive substituents such as Li2(H)Si:, Li2Si:, and HBe(H)Si:, all silylenes are singlet ground state species and that the singlet–triplet gap often is bigger than that in the corresponding carbenes.2 Due to the availability of the unoccupied p-orbital, silylenes are prone to react with Lewis bases, which may donate their free lone pair into the empty orbital (Scheme 1). This donor–acceptor interaction results in the formation of a silylene complex. In the last two decades, much interest has been focused on this new class of complexes, and a considerable number of investigations, either of theoretical or of experimental nature, have been performed to obtain better insight into the chemistry of Lewis base–silylene complexes. This review gives a brief summary of the literature that documents the efforts to explore this new field of organosilicon chemistry.

Scanning the literature one notices that silylene–Lewis base complexes are drawn either using an arrow, which points from the heteroatom toward the silicon center, thus indicating a dative bond (structure A), or as a 1,2-dipole with a covalent bond (structure B). Because the theoretical results concerning the extent of charge separation in silylene–Lewis base complexes or silaylides are contradictory (Section II,A) we will use both grapical descriptions of these compounds alternatively and interchangeably.

II THEORETICAL STUDIES OF SILYLENE–LEWIS BASE COMPLEXES

Only few theoretical studies have been devoted exclusively to the coordination of Lewis bases to silylenes. Most calculational evidence for the formation of silylene–Lewis base complexes was obtained from the computational investigation of the insertion reaction of H2Si: into various H–X σ bonds, where X is an heteroatom center possessing one or more free electron pairs. As can be seen from Fig. 1 these exothermic reactions proceed via initial formation of a donor–acceptor complex and subsequent rearrangement through a three-membered cyclic transition state to yield the eventual insertion product. The kinetic stability of the initial silylene–Lewis base complex depends on the depth of this local minimum, i.e., on the complexation energy ∆Ecompl as well as on the height of the energy barrier for the rearrangement (∆Erearr)·

In the following, we will present the results of semiempirical and ab initio calculations addressing the formation and subsequent reactions of silylene–Lewis base complexes. The review is organized according to the nature of the coordinating base; i.e., we will start with Lewis bases of group 14 and go through the periodic table to group 17 Lewis bases.

A Complexes with Carbon Nucleophiles

The only carbon nucleophile whose coordination to silylenes was investigated theoretically is carbon monoxide. The main interest centers on the question whether this nucleophile forms a Lewis acid–Lewis base complex 1 or a silaketene 2 with silylenes. Using semiempirical methods such as MNDO or AMI, Arrington et al.3 localized a planar silaketene structure as minimum for the Me2SiCO system. The Si–C bond lengths of 162.5 pm (MNDO) and 164.4 pm (AMI) are clearly in the range of Si = C double bonds as expected for a silaketene. In contrast, at the HF/3-21G level a Me2Si: ← CO complex exhibiting a pyramidal silicon center was shown to be 84 kJ mol− 1 more stable than the planar silaketene structure.3 The Si–C bond length of 289.1 pm and the C–Si–CO bond angle of 89° are in good agreement with a complex in which a free electron pair at the carbon is donated into the empty p orbital at the silicon center. Ab initio calculations by Hamilton and Schaefer4 at considerably higher levels localized a weakly bound nonplanar structure 1 (Si–C: 193.8 pm, H–Si–C: 88.6°) as the equilibrium structure of H2SiCO. The planar ketene-like structure 2, which was calculated to be 77.5 kJ mol− 1 higher in energy using the CCSD method and a polarized triple-ζ basis set, represents at this level of theory the transition state for the inversion of the silicon center.4 Similar results were obtained for the Me2SiCO system.4 Finally, based on ab initio energies obtained at the MP2/3-21G*//3-21G* level of theory, an alternative structure 3, which features a CO molecule bridging two silylene units, was suggested by Zakharov and Zhidomirov5 for the adduct of R2Si: (R = H, Me) with carbon monoxide.

B Complexes with Group 15 Nucleophiles

The reaction of H2Si: with NH3 was investigated theoretically for the first time by Raghavachari et al.6a at the MP4SDTQ/6-31G**//6-31G* level. A staggered complex 4 is formed by interaction between the lone pair of NH3 and the empty p orbital of singlet H2Si:. The length of the dative Si ← N bond is 208.9 pm; i.e., it is considerably longer than the calculated Si–N bond length of silylamine (172 pm). Complex 4 is located in a fairly deep minimum as evidenced by the binding energy of 105.1 kJ mor− 1 as well as the energy barrier of 160.4 kJ mol− 1 for the rearrangement of 4, which would yield the insertion product 5 (Scheme 2). A more recent study by Conlin et al.7 estimated the bond dissociation enthalpy of 4 as 97 ± kJ mol− 1 at the G2 level of theory. The fact that the Mulliken charges of the H2Si: and NH3 units of 4 are equal within ± 0.20 was interpreted by the authors as evidence against an ylid structure of complex 4. In contrast to these results, Schoeller et al.8 found that, based on Löwdin populations, electron density is transferred significantly from the NH3 unit to the H2Si moiety in 4, and accordingly described 4 as a zwitterionic ylide species. CIS/6-31 + G* calculations suggested the S0 → Si transition to be shifted from 485 in free H2Si: to 301 nm in complex 4;7 this theoretical result is in good agreement with the hypsochromic shift, which is experimentally observed for matrix-isolated silylene–Lewis bases complexes (Section III).

The phosphine–silylene complex 6 is less stable than 4 as evidenced by a lower binding energy of 73.3 kJ mol− 1 as well as a lowered activation energy of 82.4 kJ mol− 1 for the rearrangement of 6 into 7 at the MP4SDTQ/6-31G**//6-31G* level (Scheme 3).6a A similar decrease in binding energy was found by Schoeller et al.8a comparing the ammonia complex 4 (∆Ecompl = 103.0 kJ mol− 1) with phosphine complex 6 (∆Ecompl = 87.5 kJ mol− 1), when an appreciably higher level of theory was used. In addition, these authors investigated the influence of substituents at phosphorus and silicon in 6 on the binding energy.8a It was found that the binding energy is increased by the introduction of electron-withdrawing substituents (Cl and F) at phosphorus in 6 as well as of electron-donating SiH3 groups at silicon. These results suggest, in line with the calculated Löwdin populations, that complex 6 has appreciable zwitterionic character.

The same authors also studied the coordination of heavier pnictogen–hydrogen compounds XH3 to H2Si:. The X–Si distances in these complexes correspond to slightly elongated single bonds. In general, the binding energy decreases going from NH3 to BiH3 as Lewis base (Table...