S–H Bond Activation in Hydrogen Sulfide by NHC-Stabilized Silyliumylidene Ions

Abstract

Reactivity studies of silyliumylidenes remain scarce with only a handful of publications to date. Herein we report the activation of S–H bonds in hydrogen sulfide by mTer-silyliumylidene ion A (mTer = 2,6-Mes₂-C₆H₃, Mes = 2,4,6-Me₃-C₆H₂) to yield an NHC-stabilized thiosilaaldehyde B. The results of NBO and QTAIM analyses suggest a zwitterionic formulation of the product B as the most appropriate. Detailed mechanistic investigations are performed at the M06-L/6-311+G(d,p)(SMD: acetonitrile/benzene)//M06-L/6-311+G(d,p) level of density functional theory. Several pathways for the formation of thiosilaaldehyde B are examined. The energetically preferred route commences with a stepwise addition of H₂S to the nucleophilic silicon center. Subsequent NHC dissociation and proton abstraction yields the thiosilaaldehyde in a strongly exergonic reaction. Intermediacy of a chlorosilylene or a thiosilylene is kinetically precluded. With an overall activation barrier of 15 kcal/mol, the resulting mechanistic picture is fully in line with the experimental observation of an instantaneous reaction at sub-zero temperatures.

1. Introduction

Low-valent main group chemistry is a rapidly developing field and the wealth of new structural motifs, which have been isolated in the past two decades, have increasingly gained interest in using these species for the activation of small molecules and, potentially, for catalysis (for representative reviews see [1,2,3,4,5,6,7]). Key to these developments have been the usage of suitable synthetic methodologies in combination with thermodynamic and kinetic stabilization by appropriately chosen ligands. In particular, for the heavier carbon analogue silicon, a plethora of studies reported new low-valent compounds in recent years [8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25] and the chemistry of silylene base adducts has already been carefully developed [14,26,27,28,29,30,31,32,33,34,35,36]. Before these findings, silyliumylidene ions, cationic Si(II) species were found to be promising as similar versatile Lewis amphiphiles [37,38].

In 2004, Jutzi initiated the chemistry of silyliumylidene ions taking advantage of the stabilizing effects of the η⁵-coordinated pentamethyl-cyclopentadienyl ligand to prepare hypercoordinate silyliumylidene ion I (Figure 1) [39]. Driess and coworkers isolated the two coordinated silyliumylidene ion II, stabilized by aromatic 6π-electron delocalization as well as by intramolecular donation of the sterically encumbered β-diketiminate ligand [40]. N-heterocyclic carbenes (NHCs) represent another ligand class, widely used in modern main group chemistry. As NHCs are strong σ donors, their application in main group chemistry enabled the isolation of a large variety of low coordinate and low-valent main group compounds [41,42]. The first NHC-stabilized silyliumylidenes, III and IV were synthesized by Filippou and coworkers via a three-step protocol from SiI₄ [28], and by Driess and coworkers through the reaction of Roesky’s NHC-stabilized dichlorosilylene with their bridged bis-carbene ligand [23].

Sasamori, Matsuo, Tokitoh and coworkers obtained the bulky aryl-substituted silyliumylidenes Va-c by treatment of the corresponding diaryldibromodisilene with the carbenes ImMe₄ (1,3,4,5-Me₄-imidazol-2-ylidene) or ImⁱPr₂Me₂ (1,3-ⁱPr₂-4,5-Me₂-imidazol-2-ylidene) [43]. Around the same time our group reported the mTer- and Tipp-substituted silyliumylidenes A and A’ (mTer = 2,6-Mes₂-C₆H₃, Mes = 2,4,6-Me₃-C₆H₂, Tipp = 2,4,6-ⁱPr₃-C₆H₂) [44]. Different to all other known silyliumylidenes, A and A’ are accessible via an easy one-step synthesis: the addition of 3 equivalents of ImMe₄ to the corresponding Si(IV) aryldichlorosilanes to give the silyliumylidenes via HCl removal by ImMe₄ and nucleophilic substitution of chloride. The same approach was recently used by the group of Matsuo, obtaining Va via addition of ImMe₄ to a solution of (EMind)dichlorosilane (EMind = 1,1,7,7-tetraethyl-3,3,5,5-tetramethyl-s-hydrindacen-4-yl) [45]. It should be noted that the corresponding iodosilyliumylidene stabilized by one NHC and one cAAC (cyclic (alkyl)aminocarbene) moiety have been reported by So and coworkers [46], as well as the parent silyliumylidene [HSi⁺] stabilized by two ImMe₄ moieties [47].

Although a handful of silyliumylidenes have been reported in the last few years, reactivity studies are limited to the activation of elemental sulfur [24,48], the synthesis of a stable silylone from IV [23], and the catalytic application of I in the degradation of ethers [49]. For comparison, the neutral silylene derivative of silyliumylidene II (Figure 1) has been applied in the activation of several small molecules such as NH₃, H₂S, H₂O, AsH₃ and PH₃ [50,51,52]. A theoretical assessment of the observed divergent reactivity was provided by Szilvási and coworkers, revealing a unique insertion step to form the 1,4 adducts, followed by varying pathways towards the products [53]. The NHC-stabilized arylchlorosilylene corresponding to A has already been published by Filippou and coworkers in 2010, as well as the chlorosilylene with sterically more demanding mTer^iPr ligand (mTer^iPr = 2,6-Tipp₂C₆H₃) [54]. Subsequently, the mTer^iPr-chlorosilylene was employed as the precursor for the preparation of a silylidyne complex Cp(CO)₂Mo≡Si(mTer^iPr) [55]. The conversion of those chlorosilylenes with lithium diphenylphosphine and LiPH₂ to the corresponding phosphinosilylene and 1,2-dihydrophosphasilene reported by Driess and coworkers in 2015 [56,57] as well as the reaction towards diazoalkanes and azides presented by Filippou and coworkers [58] remain as the only reports regarding the reactivity of this species.

In any case, we consider silyliumylidene ions as promising candidates for small molecule activation as they possess two different reactive sites: an electron lone pair, and two empty p-orbitals at the silicon center. The electrophilicity of A and A’ is moderately mitigated by N-heterocyclic carbene coordination to the silicon center (Figure 2). Moreover, the zwitterionic representation of A/A’ (Figure 2) emphasizes the view of a silyl-anion, which appears useful further below.

We have already presented the silylene-like reactivity of A in the C–H activation of phenylacetylene to give the 1-alkenyl-1,1dialkynylsilane VI as the Z-isomer exclusively (Figure 2) [44]. We have also reported the application of A/A’ for the reduction of CO₂ yielding the first NHC-stabilized silaacylium ions (VII/VII’) [59]. In addition, we have demonstrated the importance of kinetic stabilization by the steric bulk of the aryl ligands. In contrast to VII, the less shielded compound VII’ is kinetically labile even at sub-zero temperatures and could only be characterized spectroscopically. Very recently, we reported the synthesis of the corresponding heavier silaacylium ions VIIIa-c obtained from the reactions with CS₂ or S₈, Se, and Te, respectively [60]. Also, we could demonstrate the recovery of silyliumylidene A from VIIIa-c by the treatment with AuI as well as chalcogen transfer reactions.

For the last 30 years, neutral congeners of VIIIa-c, silanechalcogenones R₂Si=E with E = S, Se, Te have been studied extensively [61]. In contrast, related compounds of type RHSi=E (E = S, Se, Te) are limited to the studies on intramolecularly stabilized silathioformamide by Driess and coworkers [50] and the NHC-stabilized heavier silaaldehydes by Müller and coworkers [62].

In this article, we further expand our series by the reaction of silyliumylidene A with hydrogen sulfide, yielding an NHC-stabilized thiosilaaldehyde, in a combined experimental and theoretical approach.

2. Results and Discussion

2.1. Reaction of Silyliumylidene A with H₂S

The reaction of NHC-stabilized silyliumylidene A, dissolved in acetonitrile, with 1 M H₂S solution in THF proceeded rapidly even at −20 °C, and the orange color of the starting material vanished within seconds. ¹H NMR spectroscopy indicated the formation of imidazolium salt, and one remaining ImMe₄ coordinated to the silicon center (3.46 ppm, 6H, NCH₃, ImMe₄). The splitting of the signals for the ortho-methyl groups and the benzylic protons in the mesityl moieties indicated reduced symmetry in the product. A new signal at 5.35 ppm with ²⁹Si satellites (¹J_SiH = 209.0 Hz) was assigned to the Si bound hydrogen atom by ¹H/²⁹Si-HMBC-NMR spectroscopy. The ²⁹Si NMR signal was shifted down-field from −69.03 ppm in the starting material to −39.59 ppm. Single crystals were obtained after storing the reaction solution at 8 °C overnight. X-ray crystallography confirmed the formation of the NHC-stabilized thiosilaaldehyde B (Figure 3). In earlier work by Müller and coworkers, they obtained the analogous species with the bulkier mTer^iPr ligand by the reaction of the NHC-stabilized hydridosilylene with elemental sulfur [62]. BmTer^iPr features the same ¹J_SiH coupling constant (209 Hz), which is smaller compared to the one of silathioformamide (255 Hz) reported by Driess and coworkers. [50]. Compound B is stable under the inert atmosphere and shows good solubility in acetonitrile, however, in contrast to thiosilaaldehyde BmTer^iPr, only a limited solubility in aromatic solvents is observed. Removal of the imidazolium byproduct from B was achieved by fractional crystallization from acetonitrile to obtain B as an analytically pure crystalline solid in 54% yield. Repetition of the experiments using A’ featured the same fast decoloring, but the attempts to isolate the corresponding B’ were not successful, most likely due to the kinetic lability of the formed product.

The tetracoordinate NHC-stabilized thiosilaaldehyde B (Figure 3) exhibits a distorted tetrahedral coordination around the silicon atom with a π-stacking of the NHC and a mesityl group of the terphenyl ligand. The Si–S bond length was 2.0227(9) Å, which is slightly longer than in compound VIIa (2.013(1) Å and 2.018(1) Å for the two independent molecules) [60], as well as the intramolecular, stabilized thiosilaaldehyde by Driess and coworkers (1.9854(9) Å) [50]. It is closer to the covalent double bond radii of sulphur and silicon (2.01 Å) than to the sum of single bond radii (2.19 Å) [63]. The Si–C^NHC bond (1.934(2) Å) and the Si–C^mTer bond (1.902(2) Å) are shortened compared to A (1.9481(19) Å and 1.9665(19) Å/1.9355(19) Å). The structural parameters are close to those of BmTer^iPr [62].

Further insight into the nature of the Si–S bond in B is provided by density functional theory (DFT) computations at the M06-L/6-311++G(2d,2p)//M06-L/6-31+G(d,p) level. For all bonding analyses, we chose a truncated molecular model replacing the mTer ligand by phenyl and ImMe₄ by ImMe₂H₂ (1,3-Me₂-imidazol-2-ylidene). The computed structural parameters of B^Model agreed well with the experimental molecular structure obtained from X-ray diffraction (see Table S3). Natural bond orbital (NBO) analysis reveal natural localized molecular orbitals (NLMOs, Figure 4b) corresponding to Si–H, Si–C^NHC, Si–C^mTer and Si–S single bonds as well as three NLMOs representing the electron lone pairs at sulfur. This zwitterionic representation of B^Model is also the dominant Lewis resonance structure within the natural resonance theory (NRT) formalism (Figure 4a). In line with analysis by Müller and co-workers [62], the short Si–S bond and the Wiberg bond index of 1.38 can be rationalized by negative hyperconjugation [64,65] of the sulphur lone pairs into the σ*(Si–R) orbitals: the occupancy of the LP(S) NBOs is significantly decreased (1.81 e, 1.76 e), while the NBOs for the anti-bonding σ*-orbitals are partly populated (Si–H: 0.11 e, Si–C^NHC: 0.14 e, Si–C^mTer: 0.12 e). Topological analysis of the computed electron density, by means of Bader’s quantum theory of atoms in molecules (QTAIM) [66,67], characterizes the Si–S bond as a strongly polar covalent interaction as indicated by a marked shift of the bond-critical point (bcp) towards the more electropositive Si site, a relatively large electron density ${\rho }_{\mathrm{bcp}}$, a positive Laplacian ∇²ρ_bcp as well as a negative total energy density H_bcp at the bcp (Figure 4c) [68,69].

For a classification of further characteristics at the Si–S bcp, we analyzed related species containing a Si–S single bond (1 and parent compound 3) or a Si=S double bond (2 and parent compound 4, Figure 4d). The results of the corresponding QTAIM analyses are summarized in

Table 1. Results of QTAIM analyses of B^Model and 1–4. Selected properties of the electron density distribution of the Si–S bond: Bond path lengths d_Si–S, and distances to bcps d_Si–bcp and d_bcp–S, the electron density ρ_bcp, the Laplacian of the electron density ∇²ρ_bcp, the total energy density H_bcp, the bond ellipticity ε_bcp = λ₁/λ₂ − 1 (derived from the two negative eigenvalues of the Hessian matrix of the electron density at the bcp with λ₁ ≥ λ₂), delocalization index δ_Si,S.

Compound	dSi–S [Å]	dSi–bcp [Å]	dbcp–S [Å]	ρbcp [eÅ−3]	∇2ρbcp [eÅ−5]	Hbcp [EhÅ−3]	εbcp	δSi,S
BModel	2.00	0.75	1.26	0.78	3.33	−0.52	0.01	0.78
1	2.13	0.77	1.36	0.66	1.36	−0.43	0.12	0.56
2	1.95	0.73	1.22	0.83	5.24	−0.55	0.21	1.15
3	2.14	0.78	1.37	0.64	1.36	−0.40	0.10	0.57
4	1.94	0.73	1.21	0.83	5.39	−0.56	0.23	1.25

. All molecular graphs display a characteristic shift of the Si–S bcp towards the more electropositive silicon site. In line with expectation, the values of ρ_bcp and ∇²ρ_bcp are higher for double bonded compounds 2 and 4 compared to single bonded compounds 1 and 3. The electron density and its Laplacian for the Si–S bond in B^Model are located in between, suggesting the presence of a partial double bond [70,71]. Also, the delocalization gradient δ_Si,S, i.e., the number of electron pairs shared between two atoms, lies between the values for the single and double bonded species. However, the bond ellipticity ε_bcp, a measure of ρ_bcp anisotropy indicating the π character of a bond, is rather small with a value of 0.01. Nevertheless, the Si–S bond shortening, as well as the decrease in ellipticity compared to ε_bcp for the Si–S single bonds in 1 and 3, agree with the presence of negative hyperconjugation in B^Model [72], which was already observed within the NBO framework.

These results support a zwitterionic nature of B with the partial double bond character of the Si–S bond due to negative hyperconjugation. This enhanced interaction between silicon and sulphur is reflected experimentally by the short bond length found in single crystal XRD analysis.

2.2. Mechanistic Investigations on the Reaction of Silyliumylidene A with H₂S

The reaction of H₂S and A proceeds instantaneously, preventing the NMR detection of intermediates to gain further information on this conversion. To rule out the formation of chlorosilylene C as reaction intermediate, we investigated the interconversion of C and A in a combined experimental and theoretical approach (Figure 5 and Figure 6).

Addition of one further equivalent of ImMe₄ to a solution of chlorosilylene C in benzene at RT lead to no change in color or ¹H-NMR. Heating of the reaction solution to 40 °C resulted in slow darkening of the solution to orange. After several hours the formation of orange crystals in the lower part of the Schlenk tube was observed, yielding silyliumylidene A in 58% isolated yield after a prolonged reaction time of 18 h. The reverse reaction could not be demonstrated experimentally due to the limited stability of chlorosilylene C and ImMe₄ in MeCN, the sole solvent in which A is soluble and stable.

DFT calculations on the interconversion of C and A were performed at the M06-L/6-311+G(d,p)(SMD = benzene)//M06-L/6-31+G(d,p) level of theory with a marginally reduced molecular model (mTer reduced to 2,6-diphenyl-C₆H₃ and ImMe₄ replaced by ImMe₂H₂). Silyliumylidene 7 is only slightly lower in energy than chlorosilylene 5 (Figure 6). This is in accordance with a report on the related silyliumylidene ions Vb and Vc, for which a substituent-dependent shift in relative stabilities was observed [43]. We further investigated the potential energy surface of the interconversion: NHC addition to chlorosilylene 5 via TS56 (Δ^‡G = 19.6 kcal/mol) leads to tetracoordinate 6, which is located as an unstable intermediate (18.8 kcal/mol). En route to silyliumylidene ion 7, a substantial effective barrier of 28.4 kcal/mol for the chloride dissociation in TS67 is found. This is in line with interconversion of chlorosilylene C to silyliumylidene A taking place at elevated temperatures but clearly incompatible with the H₂S activation that takes place at −20 °C. Based on our combined experimental and theoretical studies, we thus conclude that formation of thiosilaaldehyde B does not involve intermediacy of chlorosilylene C.

The reaction of 8 with hydrogen sulfide commences with a proton transfer via TS89 to give intermediate 9 as an ion pair in an exergonic step (Figure 7) (M06-L/6-311+G(d,p)(SMD:acetonitrile)//M06-L/6-31+G(d,p) level of theory). Subsequently, the SH moiety adds to the silicon center to yield 10, a pentacoordinate intermediate, with an effective barrier of 15.1 kcal/mol. The assistance of a second H₂S molecule in TS²89 is entropically disfavored compared to TS89. The NHC-stabilized silyliumylidene 8 acts as a nucleophile in the reaction with hydrogen sulfide, as its electrophilicity is saturated by the presence of two coordinating NHCs. Accordingly, the zwitterionic representation of 8 in the following best emphasizes its nucleophilic character.

Starting from 10, different pathways to NHC-stabilized thiosilaaldehyde 11 were examined (Figure 8). The concerted NHC dissociation and S–H proton abstraction in transition state TS1011 is connected with a barrier of 6.9 kcal/mol and directly yield 11 in a strongly exergonic reaction. Alternatively, dissociation of one NHC ligand from 10 to 12 was thermodynamically favored and proceeded barrierlessly, as indicated by relaxed potential energy surface scans along the Si–C^NHC bonds (see Figures S7 and S8). The NHC liberated subsequently abstracts, with clear kinetic preference, the S–H proton in 12 (TS1211: Δ^‡G = 8.7 kcal/mol). The alternative route for Si–H hydride abstraction via TS1213 is connected with a substantially higher activation barrier (Δ^‡G = 23.5 kcal/mol), which renders this path to 11 kinetically irrelevant. The atomic charges obtained by natural population analysis of 10 (H^Si: −0.17 e, H^S: 0.18 e) supported the view that the increased activation barrier goes back to the additional charge transfer occuring in the course of the hydride abstraction. Overall, the addition of H₂S to silyliumylidene 8 via TS910 is rate-limiting with an effective activation barrier of 15 kcal/mol. Subsequent isomerization to thiosilaaldehyde 11 is initiated by barrierless NHC dissociation and accomplished by abstraction of the S–H proton by the free carbene. Concerted proton abstraction and NHC dissociation (TS1011) is kinetically disfavored.

In conclusion, we have presented the activation of hydrogen sulfide by silyliumylidene ion A to give the thiosilaaldehyde B. Its nucleophilicity is best rationalized by assuming a zwitterionic character. Combined experimental and theoretical investigations reveal that the thiosilaaldehyde formation does not involve intermediacy of chlorosilylene C or thiosilylene 13. The NHC-stabilized silyliumylidene A adds H₂S in a stepwise reaction sequence followed by NHC dissociation. Proton abstraction by the latter yields thiosilaaldehyde in a strongly exergonic reaction. With an overall activation barrier of 15 kcal/mol, the resulting mechanistic picture is fully in line with the experimental observation of an instantaneous reaction at sub-zero temperatures.

3. Materials and Methods

3.1. General Methods and Instruments

All manipulations were carried out under the argon atmosphere using standard Schlenk or glovebox techniques. Glassware was heat-dried under vacuum prior to use. Unless otherwise stated, all chemicals were purchased from Sigma-Aldrich (Steinheim, Germany) and used as received. Benzene, n-hexane, and acetonitrile were refluxed over standard drying agents (benzene/hexane over sodium and benzophenone, acetonitrile over CaH₂), distilled and deoxygenated prior to use. Deuterated acetonitrile (CD₃CN) and benzene (C₆D₆) were dried by short refluxing over CaH₂ (CD₃CN) and/or storage over activated 3 Å molecular sieves (CD₃CN and C₆D₆). All NMR samples were prepared under argon in J. Young PTFE tubes. mTerSiHCl₂, chlorosilylene C and ImMe₄ were synthesized according to procedures described in literature [54,73,74]. NMR spectra were recorded on Bruker AV-400 spectrometer (Rheinstetten, Germany) at ambient temperature (300 K). ¹H, ¹³C, and ²⁹Si NMR spectroscopic chemical shifts δ are reported in ppm relative to tetramethylsilane. δ(¹H) and δ(¹³C) were referenced internally to the relevant residual solvent resonances. δ(²⁹Si) was referenced to the signal of tetramethylsilane (TMS) (δ = 0 ppm) as the external standard. Elemental analyses (EA) were conducted with a EURO EA (HEKA tech, Wegberg, Germany) instrument equipped with a CHNS combustion analyzer. Details on XRD data are given in the supplementary materials.

3.2. Improved and Upscaled Synthesis of Silyliumylidene A

mTerSiHCl₂ (1.00 g, 2.42 mmol, 1.0 eq.) and ImMe₄ (901 mg, 7.26 mmol, 3.0 eq.) were each dissolved in 17.5 mL of dry benzene in two different flasks. The ImMe₄ solution was added very slowly to the silane solution to generate a layer of immediately formed imidazolium hydrogenchloride salt separating both solutions without stirring. After complete addition/overlaying stirring was switched on, both solutions mixed thoroughly as fast as possible and the precipitated imidazolium hydrogenchloride salt was allowed to settle down for a short time. The supernatant dark red solution was filtered into a new flask, the residue was washed with 2 mL of dry benzene and the combined solutions were allowed to stand overnight for complete crystallization of the orange silyliumylidene. The yellow supernatant was separated from the orange crystalline solid, washed four times with 5 mL dry hexane to remove residues of white imidazolium hydrogenchloride salt and dried in vacuo. An orange crystalline product was obtained in 66% yield (1.00 g). Analytical data are the same as previously published [44].

3.3. Synthesis of Thiosilaaldehyde B

Silyliumylidene A (150 mg, 221 µmol, 1.0 eq.) was dissolved in MeCN (3.0 mL), cooled to −20 °C and an excess of H₂S solution (approx. 0.8 M) in THF was added. The solution quickly turned from orange to yellow to blue-green while a white precipitate was formed. The solution was allowed to warm to RT upon which the precipitate redissolved. The solution was concentrated to halve the volume and stored in the fridge for crystallization overnight. The supernatant was filtered off and the white residue was washed with MeCN (0.5 mL) at 0 °C. The solid was dried in vacuo. B was obtained as a white crystalline solid in 54% yield (59.0 mg, 118 μmol). Storage of a crude reaction mixture at 8 °C yield single crystals of B suitable for X-ray diffraction analysis.

¹H-NMR (400 MHz, 298 K, CD₃CN) δ 7.45 (t, J = 7.6 Hz, 1H, C⁴H, C⁶H³), 6.98 (s, 2H, C^3/5H, Mes), 6.91 (d, J = 7.6 Hz, 2H, C^3,5H, C₆H₃), 6.76 (s, 2H, C^3/5H, Mes), 5.36 (s, 1H, SiH, ¹J_SiH = 209.0 Hz), 3.46 (s, 6H, NCH₃, ImMe₄), 2.36 (s, 6H, C^1/3/5CH₃, Mes), 2.31 (s, 6H, C^1/3/5CH₃, Mes), 1.98 (s, 6H, CCH₃, ImMe₄), 1.93 (s, 6H, C^1/3/5CH₃, Mes); ¹³C-NMR (126 MHz, 298 K, CD₃CN) δ 149.84, 148.75, 141.06, 138.76, 137.45, 137.42, 137.05, 130.08, 129.63, 128.45, 34.12 (NCH₃, ImMe₄), 22.21 (C^2/4/6CH₃, Mes), 21.76 (C^2/4/6CH₃, Mes), 21.23 (C^2/4/6CH₃, Mes), 8.68 (CCH₃, ImMe₄); ²⁹Si-INEPT-NMR (99 MHz, 298 K, CD₃CN) δ −39.58; EA experimental (calculated): C 74.27 (74.65), H 7.66 (7.68), N 5.61 (5.62), S 6.25 (6.43) %.

3.4. Conversion of Chlorosilylene C to Silyliumylidene A

Chlorosilylene C (40.0 mg, 80 µmol, 1.0 eq.) and ImMe₄ (10.1 mg, 80 µmol, 1.0 eq.) were dissolved in 1.5 mL of dry benzene in a Schlenk tube. The tube was placed in an oil bath and heated to 40 °C for 18 h. After this time, a large amount of orange crystals was formed with some white precipitate (ImMe₄·HCl) and a slightly yellow supernatant, which was removed via the syringe. The orange crystals were washed two times with 2 mL benzene and three times with 2 mL hexane to remove the white precipitate. The crystalline material was dried in vacuo to give A in 58% yield (29.0 mg). Analytical data are the same as previously published [44].

3.5. DFT Calculations

Geometry optimizations and harmonic frequency calculations have been performed using Gaussian09 [75] employing the M06-L/6-31+G(d,p) [76,77,78] level of density functional theory. The SMD polarizable continuum model was used to account for solvent effects of acetonitrile and benzene [79]. The ‘ultrafine’ grid option was used for numerical integrations [80]. Stationary points were characterized as minima or transition states by analysis of computed Hessians. The connectivity between minima and transition states was validated by IRC calculations [81] or displacing the geometry along the transition mode, followed by unconstrained optimization. For improved energies, single point calculations were conducted at the SMD-M06-L/6-311+G(d,p) [82,83] level of theory; wave functions used for bonding analysis were obtained at the M06-L/6-311++G(2d,2p) [82,83] level. Natural bond orbital (NBO) and natural resonance theory (NRT) analyses were performed using the NBO 6.0 program [84], interfaced with Gaussian09 [85,86]. The AIMALL [87] program was used for QTAIM analyses [66,67]. Unscaled zero-point vibrational energies, as well as thermal and entropic correction terms, were obtained from Hessians computed at the M06-L/6-31+G(d,p) level using standard procedures. Pictures of molecular structures were generated with the ChemCraft [88] program.