Silyl Cations Stabilized by Pincer Type Ligands with Adjustable Donor Atoms

Novel E,C,E′-pincer supported silyl cations (E, E′ = O, S, Se, Au) were prepared in three steps starting from 2,6-F2C6H3SiMe2H (1a) and 2,6-Br2C6H3SiMe2H (1b), which were first converted in two complementary ways into 2,6(Ph2P)2C6H3SiMe2H (2). The oxidation of 2 with H2O2·urea, S8, and Se8 afforded 2,6-(Ph2PE)2C6H3SiMe2H (3a, E = O; 3b, E = S; 3c, E = Se) and 2-(Ph2PE)-6-(Ph2P)-C6H3SiMe2H (4b, E = S; 4c, E = Se), which were reacted to the E,C,E-supported silyl cations [2,6-(Ph2PE)2C6H3SiMe2] (5a, E = O, counterion Br3; 5b, E = S, counterion B(C6F5)4; 5c, E = Se, counterion B(C6F5)4), the E,C-supported silyl cations [2-(Ph2PE)-6-(Ph2P)C6H3SiMe2] (6b, E = S, not isolated; 6c, E = Se, not isolated), the O,C,S-supported silyl cation [2-(Ph2PS)-6-(Ph2PO)C6H3SiMe2] (7, counterion B(C6F5)4) as well as the E,C,Au-supported silyl cations [2-

In the presence of (substituted) benzene (derivatives), Lewis pair complexes [(R 3 Si···D)] + [A] -(IV) such as [(Me 3 Si···arene)] + -[B(C 6 F 5 ) 4 ]were isolated in which the arenes serve as π-donors (arene = benzene, toluene etc.). [8] Some of these species may be even viewed as silyl-substituted arenium ions. The use of one bulky m-terphenyl substituent (occasionally decorated by halogen atoms at the flanking phenyl groups) gave rise to the formation of intramolecularly coordinated donor acceptor complexes, such as [(2,6-Mes 2 C 6 H 4 )Me 2 Si] + [B(C 6 F 5 ) 4 ] -. [9] In the ferrocenyl-substituted silyl cation [FcMe 2 Si] + 2 [B 12 Cl 12 ] 2-, the iron intramolecularly provides electron density to the electron deficient silicon atom (Fc = ferrocenyl). [10] The vast majority of silyl cations have been prepared by a variation of the Bartlett-Condon-Schneider reaction using trityl salts of weakly coordinating anions [Ph 3 C] + [A]for the hydride abstraction from neutral H-silanes R 3 SiH. Quite often, these reactions produced hydride-bridged silyl cations [R 3 Si(μ-H)SiR 3 ] + possessing three-center two electron (3c2e) bonds, rather than "free" silyl cations. [11] Despite having a somewhat reduced Lewis acidities some of the higher-coordinated silyl cations show remarkable catalytic activities, [12] including C-F bond activation/hydrodefluorination reactions [6,13] and Diels-Alder reactions. [14] The combination of silyl cations and bulky phosphanes has been considered as frustrated Lewis pairs (FLPs) for the fixation and activation of carbon dioxide and dihydrogen. [15] Pentacoordinate silyl cations [(R 3 Si···D 2 )] + [A] -(V) have also been prepared deliberately using intramolecularly coordinating substituents with donor atoms (so-called built in ligands). [16] A prominent and significant example involves the O,C,O-pincer supported silyl cation 6-{(iPrO) 2 P(O)} 2 C 6 H 3 (Ph 2 )Si] + (VI, counterion PF 6 -) containing two intramolecularly coordinating phosphonium oxide groups (Scheme 2). [17] It is nowadays understood that phosphonium oxides are best described by bipolar single bonds +P-Orather than P=O double bonds. [18] At present, there is a vivid debate how to present donor acceptor interactions within main group complexes and it has been pleaded to avoid extreme resonance formulas for marketing reasons. [19] In the centre of this debate are often lowcoordinate cations, which are stabilized by phosphine ligands or N-heterocyclic carbenes (NHCs). Although this work is not concerned with these compound classes, the question for the most significant resonance structure for the description of silyl cations, such as VI, also requires clarification and will be discussed below. Like the tricoordinate silyl cations [Mes 3 Si] + [HCB 11 Me 5 Br 6 ] -, [2] and [Pemp 3 Si] + 2 [B 12 Cl 12 ] 2-, [3] VI possesses a trigonal planar base consisting of three C atoms. [17] In addition to the tricoordinate silyl cations V comprises two axial O atoms that compensate the electron deficiency of the Si atom. For both tricoordinate silyl cations [R 3 Si + ][A -] (I) and pentacoordinate silyl cations [(R 3 Si···D 2 )] + [A] -(V), a simplistic "valence bond (VB) model" might attribute the bonding of the three equatorial substituents to three sp 2 -orbitals. It might further describe the axial substituents D of the pentacoordinate silyl cations [(R 3 Si···D 2 )] + [A] -(V) as three-center four electron (3c4e) bonds involving the p z -orbital of the Si atom. [16] In the tricoordinate silyl cations [R 3 Si] + [A] -(I) the p z -orbital remains vacant. Most recent computational work attributes the high affinity of Si for O to the high electronegativity differences and emphasizes the strongly polar or even ionic bond character of the Si-O bond, [20] which questions strong covalent contributions for the axial bonding in pentacoordinate silyl cations such as V. We became interested in the preparation of new E,C,E′pincer supported silyl cations with a number of different donor atoms E, E′ = O, S, Se, Au including also those with a smaller affinity and electronegativity difference between Si and O. The established synthetic route for the preparation of the O,C,Opincer supported silyl cation VI starts from the dibromo benzene VII, which was converted into the diphosphonate VIII by a transition metal catalyzed Arbuzov reaction, which restricts the donor atoms of this pincer ligand to O atoms. [17] We note in passing that efforts were undertaken to prepare related O,C,Spincer ligands, which, however, rely on the Arbuzov reaction. [21] The selective C-H metallation of VIII prior to the reaction with Ph 2 SiHCl gives rise to the formation of the H-silane IX. Hydride abstraction from IX via the Bartlett-Condon-Schneider reaction eventually provided the silyl cation VI. [17] In this work we present an alternative route that first introduces the Si moiety and avoids Arbuzov type reactions, which allows setting up more versatile pincer type ligands. [22] Using this route, a number of new E,C,E′-pincer supported silyl cations with various potential donor atoms (E, E′ = O, S, Se, Au) have been prepared and fully characterized. The nature of the bonds within these compounds was determined by analysis of a set of topological and integrated real-space bonding indicators (RSBIs) derived from the theoretically calculated electron densities (ED) and electron pair densities utilizing the atoms-in-molecules (AIM) [23] and electron-localizability-indicator (ELI-D) [24] space partitioning schemes which divide space in basins of atoms and paired electrons, respectively. A combination of these two methods provides quantitative information about the strength and nature of a bond and is very well suited to straightforwardly detect also weak atomic interactions, which is not the case for sole inspection of molecular orbitals (MOs) and/or natural bond orbitals (NBOs).
obtained (see ESI for details). The ratio between the products of this reaction remained constant for at least 12 h. The same reaction had a different outcome when it was carried out in fluorobenzene (see ESI for details). Shortly after the start of the reaction (its progress was monitored by 31 P NMR) a mixture of 6c, 5c and an unidentified product was obtained. The starting material was consumed in less than 13 minutes, and after this time the concentration of 5c and the unknown product decreased rapidly. Within 60 minutes the molar ratio between 6c and 5c became 1:0.02. While this solvent dependence is still not fully understood, the observation shows that the reaction mechanism might be quite complex. The E,C-supported silyl cations [2-(Ph 2 PE)-6-(Ph 2 P)C 6 H 3 SiMe 2 ] + (6b, E = S; 6c, E = Se) are fairly reactive and all attempts to isolate these species were impeded by decomposition. However, the clean formation of silyl cations 6b and 6c was unambiguously confirmed by 31 P, 29 Si and 77 Se NMR spectroscopy (see below). Freshly prepared samples of 6b and 6c were used for all subsequent reactions. Exposed to the air, 6b reacted rapidly with molecular oxygen to give the O,C,S-supported silyl cation [2-(Ph 2 PS)-6-(Ph 2 PO)-C 6 H 3 SiMe 2 ] + (7). The reaction of 6b with sulfur proceeded at a slower pace (6 d), but eventually produced the S,C,S-supported silyl cation [2,6-(Ph 2 PS) 2 C 6 H 3 SiMe 2 ] + (5b).

Solid-State and Solution Structure
The molecular structures of the precursors 2, 3a·H 2 O, 3b, 3c and 4b, the E,C,E′-pincer supported silyl cations 5a, 5c and 7 (E, E′ = O, S, Se) as well as the E,C,Au-pincer supported silyl  Table 1. The spatial arrangement of the Si atoms of the precursors 2, 3a·H 2 O, 3b, 3c and 4 is distorted tetrahedral. Only 3a shows an intramolecular Si···O contact (3.024(8) Å) significantly longer than that of VIII (2.738(2) Å), which is indicative to weak σ-hole bonding. [28] The tetrahedral geometry of the Si atoms is unaffected by this additional coordination as the sum of C-Si-C angles steadily increases in the series 2 (338.1(3)°), 3a (339.9(4)°), 3b (341.5(1)°) and 3c (343.6(4)°). The spatial arrangement of the Si atoms of the silyl cations 5a, 5c and 7 is distorted trigonal bipyramidal (type IV, Chart 1), whereby the three C atoms and chalcogen atoms E, E′ = O, S, Se adopt the equatorial and axial positions, Figure 3. ORTEP view of the of X-ray molecular structure of 5a, 5c and 7 showing the atom-numbering scheme. Displacement ellipsoids are drawn at the 30 % probability level and H atoms are shown as small spheres of arbitrary radii.
For the precursor IX (Scheme 2) the presence of hydrogen bridges of the Si-H···O-P type have been claimed. [17] A similar Si-H···O contact seems to be present in 3a, as evidenced by the H···O bond length (2.826(7) Å). The other precursors 3b, 3c and 4b show no evidence for contacts of the type Si-H···E interactions (E = S, Se). In the solid-state, the precursors 3a, 3b and 3c each adopt different conformations regarding the position of the chalcogen atoms relative to the central C 6 H 3 Si plane. In 3a·H 2 O, the O atoms are situated on the opposite side of the plane (transoid), whereas in 3b and 3c and the S and Se atoms lie on the same sides of the plane (cisoid). In 3a·H 2 O the transoid conformation might be related to the presence of PO···H-OH hydrogen bridges. The O···O donor acceptor distances (2.844(7) and 2.907(8) Å) are indicative of medium to weak hydrogen bonding. [30] While there appears to be no Si···Au contact in 8b and 8c, both compounds show two intramolecular hydrogen bonds of the type C-H···Au. The H···Au bond lengths of 8b (2.772(1) Å, 2.962(1) Å) and 8c (2.816(2) Å, 2.891(2) Å) and the C···Au distances of 8b (3.539(3) Å, 3.701(3) Å) and 8c (3.555(1) Å, 3.614(1) Å) fully consistent with parameters collected for a recent survey on H···Au contacts. [31] In the O,C,O-pincer supported silyl cation (5a), the O atoms are nearly coplanar with the central C 6 H 3 Si plane (largest deviation from the ideal plane 0.083(3) Å for O1). The same holds also for the O,C,S-pincer supported silyl cation 7 (largest deviation from the ideal plane 0.080(1) Å for S2). However, in the Se,C,Se-pincer supported silyl cation 5c, the dimethylsilyl  Table 1. Comparison of selected bond lengths (in Å) and angles (in deg) determined by X-ray single-crystal analyses of 2, 3a-3c, 4b, 5a, 5c, 7, 8b and 8c.

Gas-Phase and Electronic Structures
Using the coordinates of the X-ray structures as starting point the gas-phase structures of the E,C,E-pincer supported silyl cations 5a, 5b, and 5c (E = O, S, Se) and their precursors 2, 3a, 3b and 3c were fully optimized at the B3PW91/6-311+G(2df,p) level of theory. Due to the size and issues related to relativistic effects for the E,C,Au-pincer supported silyl cations 8b and 8c only single point calculations were carried out at the same level of theory. In general there is good agreement between experimental and calculated bond parameters (see ESI for details). For the precursor 3a, both the cisoid and the transoid conformers were preliminary calculated, however, as their relative stability differs only by about 6 kJ mol -1 only the transoid conformer of 3a resembling the solid-state structure most was taken into fur-ther considerations. For 3b and 3c the cisoid conformers were calculated. The bond topology according to the AIM theory and isosurface representations of the ELI-D localization domains are shown in Figure 5, Figure 6, and Figure 7. Selected bond lengths and AIM bond topological parameters are collected in Table 2, whereas topological and integrated ELI-D bond descriptors are listed in Table 3. Since similar bonds exhibit similar topological and integrated RSBIs only one example is given each (see ESI for all data). The considered RSBIs include: ρ(r) -the electron density at the bond critical point (bcp) -and its corresponding Laplacian (∇ 2 ρ bcp ), ε -the bond ellipticity (ε = (λ 1 /λ 2 )-1; λ 1 and λ 2 are the curvatures perpendicular to the bond axis), d 1 /d -the ratio of the atom-bcp distance over the atom-atom distance, G/ρ bcp and H/ρ bcp -the kinetic and total energy density over ρ bcp ratios, [33] V 001 ELI, ELI pop -the volume and electron population of the ELI-D basin, ELI max -the ELI-D value at the attractor position, Δ ELI -the distance of the attractor position perpendicular to the atom-atom axis and RJI -the Raub-Jansen index, [34] which provides the number of electrons (and also percentage contributions) of an bonding or lone pair ELI-D basin distributed between the adjacent AIM atoms which form a bond. The delocalization index [35] δ(A,B) is also calculated, which is a direct measure for electron sharing between two AIM atoms. In 3a-c no Si-E (E = O, S, Se) bond paths were observed in the AIM topology which rules out such types of interactions in these compounds, although the Si···S distances (< 3.7 Å) and Si···Se distances (< 3.8 Å) are shorter than the sum of the Van der Waals radii of 3.9 Å and 4.0 Å, respectively. This result is sup- Figure 5. Left column: molecular graphs of gas-phase structures of 3a, 3b and 3c (AIM2000 presentation). Right column: theoretical ELI-D localization domain representation of 3a, 3b and 3c. The colour code makes domains belonging to different basins distinguishable from each other, additionally the hydrogen basins are drawn in transparent mode for clarity; MolIso [32] graphics at isovalue Υ = 1.50. ported by the AIM atomic charges (see below). However, a Si-H···O bond path was observed in 3a indicating a weak interaction between the hydridic H atom and the also negatively Figure 6. Left column: molecular graphs of gas-phase structures of 5a, 5b and 5c. (AIM2000 presentation). Right column: theoretical ELI-D localization domain representation of 5a, 5b and 5c. The color code makes domains belonging to different basins distinguishable from each other, additionally the hydrogen basins are drawn in transparent mode for clarity; MolIso [32] graphics at isovalue Υ = 1.50.  [32] graphics at isovalue Υ = 1.50.

Eur
polarized O atom. In contrast, the molecular gas-phase structures of 5a-c show such Si···E bond paths and the electronic parameters determined at the Si···E bcps unravel the respective nature of the interaction. The positive (or very close to zero) values of Laplacian at the Si-E bcps indicate polarized interactions for all three types of contacts (  Table 2 and Table 3 (P-E, P/Si-C) show strong attributes of both covalent as well as ionic interactions. Generally, both types of Si-C bonds: Si-C sp 3 and Si-C sp 2 distances are polar covalent with quite strong polarization towards carbon atom, which is confirmed by the Raub-Jansen indexes (RJI) ranging between 83-89 % and the ratio H/ρ bcp close to -0.7 h/e. The partially ionic character is confirmed by a low value of electron density at the bond critical point, a low positive value of Laplacian and a low delocalization index. All these topological and integrated Table 2. Bond length (in Å) and AIM bond topological parameters [a] for selected bonds to phosphorus and silicon atoms of 2, 3a-3c, 5a-5c, 8b and 8c. parameters correspond well with literature data for C-Si linkages. [36] All C-P bonds are also polar covalent, however the polarization effect is lower in comparison to the C-Si bonds. Herein, the covalent contribution is seen from the electron density values close to 1 e/Å 3 , negative values of corresponding Laplacian, and a H/ρ bcp value close to -1 he -1 . The delocalization index unravels that more electron pairs are shared between C and P atoms (δ = 0.67-0.84, RIJ = 71-78 %) than for C and Si atoms (δ = 0.39-0.53, RIJ = 83-89 %). Aforementioned properties of C-P bond are in proper agreement with our recent results for peri-substituted (ace)naphthylphosphinoboranes where such type of bonds were also evaluated using both AIM and ELI-D approaches. [37] Similarly, as it was found for the solid-state structures, the E-P bond lengths are also longer in the gas-phase E,C,E-pincer supported silyl cations (5a-5c) in comparison to those of the precursors (2, 3a-3c), especially in case of the P-O bonds where the elongation is more pronounced due to the relatively small size of the O atom in comparison to other chalcogen atoms (S and Se). Such elongation is reflected in the topological parameters of these bonds; for example the electron density is lower in the pincer-type structures. As mentioned above, the P-O bonds are highly polarized covalent bonds with significant ionic contributions, which is consistent with findings from combined experimental and theoretical studies. [38] Herein, two parame- Table 3. Topological [a] and integrated [b] ELI-D bond descriptors for selected bonds to phosphorus and silicon atoms of 2, 3a-3c, 5a-5c, 8b and 8c. [a] ELI max -ELI-D value at the attractor position, Δ ELI the distance in Å of the attractor position perpendicular to the atom-atom axis. [b] δ -the delocalization index, V 001 ELIis the volume of the ELI-D basin in Å 3 cut at 0.001au, ELIpop -the electron population within the ELI-D basin in e, and RJI -the Raub-Jansen index in e and %.
ters: H/ρ bcp and the Raub-Jansen index, showing the degree of covalency, are very similar for the P-O bond type. The values 1.93 and 2.16 he -1 of the parameter G/ρ bcp , the highest observed for all structures analyzed, indicate the high degree of iconicity for such type of bond. Based on DFT studies combined with AIM and ELI-D approaches, P-S and P-Se bonds can be classified as shared-shell interactions with a negative value of the Laplacian and a negative H/ρ bcp ratio. The P-Se bonding shows only a slight polarization towards the phosphorus atom which is reflected by RJI = 60-65 %. In both cases a delocalization index higher than 1 indicates that the P and S/Se atoms share more than one but less than two electron pairs. The ionic contribution described by G/ρ bcp is very small and comparable to those found for P-C bonds. In the E,C,Au-pincer supported silyl cations 8b and 8c the nature of the C-Si/C-P/E-P bonds remains unchanged in comparison to the E,C,E supported silyl cations 5a-5c, which is confirmed by the very similar real-space bonding indicators characteristic for polar-covalent interactions. Interestingly, the Si-E (E = S, Se) bond lengths are significantly shorter in 8b and 8c (up to 0.3 Å) in comparison to the related gas-phase structures of the E,C,E-pincer supported silyl cations 5b and 5c. This discrepancy is not an effect of different computational techniques used (single point calculation vs. geometry optimization) since the X-ray analysis revealed the same tend-ency for experimental distances of 5c and 8c. Such a difference in bond lengths, however, has a considerable impact on topological (ρ bcp , G/ρ bcp , H/ρ bcp ) and integrated (δ) parameters, however the Laplacian values still remain close to zero. Considering the ELI-D bonding indicators of Si-Se contact in 8c, it is interesting to note that the electrons within its ELI-D basin are more localized (values of ELI max and ELI pop increased) than in structure 5c. It is reflected in the graphical representation of ELI-D isosurfaces of 8c and 5c. In the former case ( Figure 7) the Si1-Se2 basin is clearly visible in contrast to the latter case ( Figure 6) where the corresponding basin is not observed at Y = 1.5 (but becomes visible at smaller ELI-D values). For the related Si-S basins of 5b and 8b the enhancement of electron localization is not so clearly pronounced. Interestingly in all studied S, Se-pincer supported cations, the RJIs remain comparable (above 80 %) confirming the polar covalent character of studied Si-E (E = S, Se) contacts. The AIM analysis confirmed the existence of intramolecular C-H···Au interactions found in the crystal structures of 8b and 8c. The topological properties determined at the Au···H bond critical points clearly show that these interactions are weak with electron density close to zero (0.06-0.09 e Å -3 ), the associated Laplacian below 1e Å -5 , and the kinetic energy density dominates the potential energy density (|V|/G = 0.85-0.88). [33]

Analysis of AIM Charges
Selected atomic and fragmental charges obtained by integration of the AIM atomic basins are given in However, due to the obvious Si-E interactions in the E,C,Epincer supported silyl cations 5a-c distinct changes in the AIM atomic charges are observed in comparison to the corresponding precursors. Interestingly, in 5a the Si atom has 0.14 e less electrons compared to its precursor 3a, whereas an electron increase of 0.11 e and 0.20 e is found for the corresponding S-and Se-containing pincer molecules 5b and 5c.

Conclusions
A versatile route for the preparation of E,C,E′-pincer type supported silyl cations, such as [2,6-(Ph 2 PE) 2 C 6 H 3 SiMe 2 ] + (5a, E = O; 5b, E = S; 5c, E = Se), [2-(Ph 2 PS)-6-(Ph 2 PO)C 6 H 3 SiMe 2 ] + (7) and [2-(Ph 2 PAuC 6 F 5 )-6-(Ph 2 PE)C 6 H 3 SiMe 2 ] + (8b, E = S; 8c, E = Se) has been developed. The relative coordination number according to the 29 Si NMR chemical shifts decreases in the order 5a (-30.9 ppm) < 5b (-5.7 ppm) < 5c (-1.1 ppm) < 7 (5.9 ppm) < 8c (40.3 ppm) ≈ 8b (41.2 ppm). The nature of the intramolecularly coordinating chalcogen atoms E was analyzed by an electron density based set of RSBIs derived from the AIM and ELI-D space partitioning schemes Consistent with the aforementioned trend regarding the relative coordination number, the kinetic energy density over ρ bcp ratios reveal that the strength of the Si-E bonds decreases in the order 5a (1.19 he -1 ) < 5b (0.43 he -1 ) < 5c (0.31 he -1 ). Several other RSBIs including the electron density and its Laplacian at the bond critical point (bcp) as well as the Raub Jansen index (RJI) suggest that the Si-O bond is ionic in nature, which is in line with the large electronegativity difference of both elements. [20] In contrast, the Si-S and Si-Se bonds are slightly less ionic, which is also consistent with smaller electronegativity difference of In the E,C,E′-pincer type supported silyl cations, high positive charges are also situated on the P atoms of 5a (2.92 e), 5b (2.16 e) and 5c (1.89 e), which can be accounted for the by use of bipolar single bonds + P-Ein the Lewis formula representations (Scheme 2, Scheme 3, Scheme 4, Scheme 5, and Scheme 6). In this way, both Si and P atoms are assigned formal positive charges. In light of the recent controversial debate of the appropriate description of donor acceptor bonds within main group compounds, [19] we deliberately used "arrows" to emphasize the strongly polar or even ionic bond character of the Si-O bond in 5a. Resonance structures involving "formal charges" suggesting a higher degree of covalence, are rather inadequate to describe the bond situation of the Si-O bonds of 5a. The same holds true for the Si-S bonds of 5b and the Si-Se bonds of 5c, albeit the bond polarity of these bonds is less pronounced. In an effort to contribute to this debate [19] we propose to generally write donor-acceptor complexes with predominately covalent bond character using "lines and formal charges" and those with predominately ionic bond character using "arrows".

Synthesis of 2,6-(Ph 2 P) 2 C 6 H 3 SiMe 2 H (2). Method A.
A suspension of lithium granule (0.21 g, 30 mmol) in THF (20 mL) and diphenylchlorophosphine (3.6 g, 16.6 mmol) were stirred at room temperature until the lithium granules were consumed. The solution turned dark red shortly after the beginning of the reaction. Neat 1a (1.0 g, 5.8 mmol) was added and the mixture was stirred overnight at room temperature. The solvent was removed under reduced pressure and the residue was washed with acetonitrile (50 mL). To separate the lithium chloride from the product, dichloromethane (50 mL) was added and the suspension filtered. The solvent was removed in vacuo to give colorless crystals of 2 (2.6 g,5.3 mmol,91 %;Mp. 155°C). (27.2 mL, 68.0 mmol) in diethyl ether (250 mL) was cooled to -78°C. N,N,N′,N′-tetramethylethylenediamine (7.90 g, 68.0 mmol) and 1b (10.0 g, 34.0 mmol) were added and stirring was continued for 2 h at -78°C.

Method B. A 2.5-M solution of n-butyllithium
The solution was warmed up to room temperature and chlorodiphenylphosphine (15.0 g, 68.0 mmol) in diethyl ether (25 mL) was added drop wise. The mixture was stirred overnight. The solvent was removed under reduced pressure and the residue was washed with acetonitrile (50 mL). To separate the lithium chloride from the product, dichloromethane (50 mL) was added and the suspension filtered. The solvent was removed in vacuo to give colorless crystals of 2 (12.5 g,24.8 mmol,73 %;Mp 155°C).

Isolation of 8b·[BAr F ] -.
In the glovebox, a vial was charged with 4b (116 mg, 0.216 mmol) and trityl tetrakis [3,5-bis(trifluoromethyl)phenyl]borate (239 mg, 0.216 mmol). Anhydrous fluorobenzene (3 mL) was added. The mixture was stirred for 20 min then to the crude 6b·[BAr F ]was added (tht)AuC 6 F 5 (98 mg, 0.216 mmol) and stirring was continued for 10 additional minutes. The red solution was layered with n-hexane (approx. 7 mL) and the vial left standing for a week. After this time a red oily layer separated at the bottom of the vial. The vial was placed in a freezer at -30°C for several days causing the precipitation of a crystalline solid. The solution was removed and the solid washed with n-hexane (3×5 mL) then dried. Compound 8b·[BAr F ]was obtained as a tan solid (315 mg, 0.179 mmol, 83 %; . Crystals suitable for X-ray diffraction were obtained by diffusion of n-hexane vapors into a concentrated solution of 8b·[BAr F ]in CH 2 Cl 2 (approx. 0.3 mL).

Synthesis of [2-(Ph 2 PSe)-6-[Ph 2 P(AuC 6 F 5 )]C 6 H 3 SiMe 2 ] + [BAr F ] -(8c·[BAr F -). NMR monitored experiment.
In the glovebox, an NMR tube was charged with 4c (85 mg, 0.145 mmol) and [Ph 3 C][BAr F ] (160 mg, 0.145 mmol). Anhydrous C 6 D 6 (1 mL) was added to the solid mixture. The NMR tube was shaken occasionally for the next 3 h then was left standing for additional 9 h. Two layers formed. The dark yellow bottom layer was assessed by 31 P and 29 Si NMR; a complete conversion of 4c was observed. The formation of 6c·[BAr F ] -, 5c·[BAr F ] -, and of an unidentified side-product in a 1:0.25:0.11 molar ratio was indicated by 31 P-NMR. The NMR tube was brought in the glovebox and to the crude mixture containing 6c·[BAr F ]was added (tht)AuC 6 F 5 (65 mg, 0.145 mmol). The NMR tube was shaken for approximately 5 minutes. The upper layer removed and the bottom layer washed with C 6 D 6 (6×0.35 mL). Complete conversion of 6c·[BAr F ]and formation of the target product 8c·[BAr F ]was indicated by 31 P and 29 Si NMR. Compound 5c·[BAr F ]was present in approximately the same amount with respect to 8c·[BAr F ] -. The solvent was than evaporated in the glovebox and the remaining oil dissolved in fluorobenzene. Crystals suitable for X-ray diffraction were obtained by diffusion of n-hexane vapors into a concentrated solution of 8b·[BAr F ]in PhF. 31

Isolation of 8c·[BAr F ] -.
In the glovebox, a vial was charged with 4c (130 mg, 0.223 mmol) and trityl tetrakis [3,5-bis(trifluoromethyl)phenyl]borate (247 mg, 0.223 mmol). Anhydrous fluorobenzene (3 mL) was added. The mixture was stirred for 40 min. The color of the solution changed from brown-green to red. To the crude 6c·[BAr F ]was added (tht)AuC 6 F 5 (101 mg, 0.223 mmol) and stirring was continued for 10 additional minutes. The red solution was layered with n-hexane (approx. 7 mL) and the vial left standing for a week. The yellow crystals obtained were separated from the solution and then washed with n-hexane (3×5 mL) and dried. Compound 8c·[BAr F ]was obtained as a yellow crystalline solid ( Crystal structure determination. The X-ray data collection was carried out using Mo-K α radiation on a Siemens P4 diffractometer (2, 3a-3c, 4b, 5a), on Stoe IPDS I diffractometer (5c) and on Bruker Venture diffractometer (7, 8b, 8c, Ph 3 COOCPh 3 ) respectively. Singlecrystal structures were solved and refined on F 2 with SHELXL-2013 [41] (except of 8b-c, for which SHELXL-2014 [41] was used) including anisotropic displacement parameters for all non-hydrogen atoms. In structure 7, a group of atoms (S2, O1, P1, P2, C1 C2) were refined as disordered over two sets of sites with occupancy factors equal to 0.905(1):0.095 (1). Additionally, similarity restraints of the atomic displacements parameters of all disordered atoms were used in the refinement of 7. In all structures (C)-H-atoms were located in calculated positions and refined isotropically using a rigid body model.
In 2, 3a-c and 4b, the H atoms attached to Si were found on the difference map and their positional parameters were refined with U iso = -1.5 U eq (Si) (except for structures 3c and 4b, where this H-atom was constrained to Si). In structure (3a) a solvent water molecule was refined with hydrogen atoms constrained to oxygen atom. Further details of crystal data and measurement conditions are given in Tables S1-S3 in the Supporting Information (SI). The ORTEP drawings were made using Mercury. [42] Deposition Numbers 1024517 (2), 1024518 (3a), 1024519 (3b), 1024520 (3c), 1024521 (4b), 1024506 (5a), 1024507 (5c), 1024508 (7), 1424430 (8b), 1424431 (8c) and 1024509 (Ph 3 COOCPh 3 ). Copies of the data can be obtained free of charge on application to The Director, CCDC, 12 Union Road, Cambridge CB2 1 contain the supplementary crystallographic data for this paper. These data are provided free of charge by the joint Cambridge Crystallographic Data Centre and Fachinformationszentrum Karlsruhe Access Structures service www.ccdc.cam.ac.uk/structures.

Computational chemistry.
On the basis of experimental X-ray single-crystal coordinates of the precursors (2, 3a-c) and the E,C,E′pincer supported silyl cations (5a, E = O; 5c, E = Se ), a full geometry optimization of the gas-phase molecules was performed with the functional B3PW91 [43] and the 6-311++G(2df,p) [44] basis-set using the program package Gaussian09. [45] Since the X-ray structure of the S,C,S′-pincer cation (5b) was not determined, its gas-phase geometry was predicted and optimized using starting geometry of (5a) with S atoms instead of O atoms. The single-point calculations of the E,C,Au-pincer supported silyl cations (8b, E = S; 8c, E = Se) were performed at the B3PW91 level of theory using the effective core potential for Au [46] with the associated triple-basis-set [46] and the 6-311++G(2df,p) [44] basis-set for all other atoms.
Complete topological analysis of the electron density was performed according to Bader's quantum theory of atoms in molecules (QTAIM) using AIM2000 [47] and AIMAll. [48] The ELI-D bond properties were calculated from the Gaussian checkpoint file using the program DGrid-4.6 [49] with the grid step size 0.06 au.