Main

The debate on the existence of silylium ions (R3Si+) in the condensed phase, the so-called silylium ion problem1, was finally resolved with the isolation and characterization of the tricoordinate (free) trimesitylsilylium ion (Mes3Si+) (ref. 2). Despite these controversial beginnings3, silylium ions have nowadays emerged as potent reagents and catalysts in synthetic chemistry4,5. In catalytic transformations, tetracoordinate trialkyl-substituted silylium ions stabilized by a donor molecule such as the solvent6 or counteranion7 are typically used as initiators, and the catalytic cycle is then maintained by the self-regeneration of the silylium ion5. The traditional way to make these reactive intermediates is the Corey reaction, that is, the hydride abstraction from the corresponding hydrosilane with the trityl (Ph3C+) salt of a weakly coordinating anion8. Using halogenated 1-carba-closo-dodecaborates, Reed and co-workers succeeded in the synthesis and solid-state characterization of counteranion-stabilized trialkylsilylium ions such as [iPr3Si(HCB11H5Br6)] (refs. 7,9,10,11) (1c → 2c; Fig. 1a). Although the silyl cation is coordinated by one halogen atom of the carborate cluster, this interaction is energetically weak9,10, and the carborate salts show silylium-ion-like reactivity in solution5.

Fig. 1: Synthetic routes to silylium ions.
figure 1

a, Preparation of alkyl-substituted silylium carborates by hydride abstraction7. b, Preparation of hydrogen-substituted silylium carborates by dehydrogenative or dephenylative protolysis15. c, Chlorine-substituted silylium ions as key intermediates in the catalytic halodealkylation of alkyl(halo)silanes with relevance to the Müller–Rochow direct process16,17. d, Protolysis strategy to access halogen-substituted, counteranion-stabilized silylium ions. Alk, alkyl substituent; r.t., room temperature. White circles indicate B–H, and grey circles indicate B–Br.

Full size image

Our laboratory introduced a general approach for the rapid access of carborate-stabilized silylium ions based on the chemoselective protolysis of various tetraorganosilanes using Reed’s superacidic benzenium salt12 [H(C6H6)]+[HCB11H5Br6]− (refs. 13,14). Following this strategy, we achieved the synthesis of the full series of hydrogen-substituted silylium ions either by dehydrogenation (leaving group (LG) = H) or dephenylation (LG = Ph), as exemplarily shown for the generation of the secondary silylium ion [tBu2HSi(HCB11H5Br6)] (ref. 15) (3d or 4d → 5d; Fig. 1b).

Recently, we disclosed a protocol for the catalytic halodealkylation of fully alkylated silanes16 and proposed halogen-substituted silylium ions of the type [R2XSi(HCB11H5Br6)] as the key intermediates in the corresponding twofold halodealkylation process17 (Fig. 1c). This method enables the upcycling of Me4Si and Me3SiCl, accumulating as by-products from the Müller–Rochow direct process, into higher-value Me2SiCl2, which serves as feedstock monomer for the silicone production18,19. Within the catalytic cycle, an arenium ion is supposed to cleave the C(sp3)‒Si bond in trialkylhalosilanes, giving rise to hitherto unknown halosilylium ions17. Of note, Dolgov, Voronkov and Borisov have initially postulated the intermediacy of chlorosilylium ions in the substituent redistribution reaction of dialkylchlorosilanes promoted by catalytic amounts of AlCl3 (ref. 20). However, experimental evidence of these reactive intermediates is still lacking. While both inter- and intramolecular halogen stabilization of silylium ions have been investigated systematically21,22, work on halogen-substituted silylium ions has been limited to theoretical23 and gas-phase studies24.

The intermolecular interaction between silylium ions and halides is also the critical step of one of the most prominent applications of trialkylsilylium ions5, that is, the catalytic hydrodefluorination of aliphatic C(sp3)–F bonds in fluorocarbons25. In the search for even more fluoridophilic silylium ions, which even enable the hydrodefluorination of environmentally harmful per- and polyfluoroalkyl substances (PFAS), Gusev and Ozerov identified chlorine-substituted silylium ions as promising candidates by calculating the hydride ion affinity and fluoride ion affinity (FIA) of various silylium ions26. Compared with trialkyl- and hydrogen-substituted silylium ions, these cations are expected to be more electrophilic, as evident by the increasing FIAs in the order Me3Si+ < Me2HSi+ < Me2ClSi+.

Given the importance and relevance of halogen-substituted silylium ions, we envisaged their independent preparation and characterization in the condensed phase. Here, we report the synthesis and solid-state structures of halosilylium ions of type [Alk2XSi(HCB11H5Br6)] (X = F, Cl, Br or I; Alk = Me, Et, iPr or tBu). The key to accessing these long-sought reactive intermediates is the implementation of the protolysis approach by dehydrogenation (LG = H) or dephenylation (LG = Ph) of the corresponding hydro- or phenylsilanes using Reed’s superacidic benzenium ion [H(C6H6)]+[HCB11H5Br6]− (6 or 7 → 8; Fig. 1d).

Results and discussion

Synthesis

In our initial attempt to generate a halogen-substituted silylium ion, we tested the classical Corey hydride abstraction from a dialkyl-substituted halosilane and chose chlorodiisopropylsilane (6bc) as a test substrate, as we knew from our previous work that the bulky isopropyl group prevents any redistribution of substituents27,28 (Fig. 2a, top). When the trityl salt [Ph3C]+[HCB11H5Br6]− was treated with an excess of iPr2SiHCl (6bc) in benzene at room temperature, a slow reaction occurred, as indicated by a gradual fading of the yellow reaction mixture. However, nuclear magnetic resonance (NMR) spectroscopic and X-ray crystallographic analysis revealed the exclusive formation of hydrogen-substituted silylium carborate [iPr2HSi(HCB11H5Br6)] (5c), which is characterized by a 29Si NMR chemical shift of δ = 69.7 ppm and a diagnostic signal in the 1H NMR spectrum for the intact Si–H bond at δ = 5.27 ppm with a 1J(Si,H) coupling constant of 240 Hz. The same reaction outcome was also obtained starting from the corresponding iodosilane iPr2SiHI (6dc), and reducing the silane equivalents to stoichiometric quantities led only to a slower reaction rate. In all cases, no formation of the desired halogen-substituted silylium ion was detected. This result was unexpected, as the trityl cation should not be able to heterolytically cleave silicon‒halogen bonds owing to its substantially lower FIA compared with secondary silylium ions26. We assumed that initial hydride transfer generates a halosilylium intermediate that immediately abstracts the halide from the starting silane to eventually arrive at the observed hydrogen-substituted silylium carborate 5c. This hypothesis was later corroborated by an independent control experiment (see ‘Reactivity‘ section).

Fig. 2: Synthesis and NMR spectroscopic characterization of halogen-substituted silylium ions.
figure 2

a, Attempts to generate a chlorine-substituted silylium ion by classical hydride abstraction (top) and successful synthesis by our protolysis approach (bottom). b, Synthesis of the full series of halogen-substituted silylium carborates. For the preparation of the individual silanes as silylium ion precursors, see Supplementary Figs. 1–4 in Supplementary Information Section 3. c, Plots of the experimental (left) and calculated (right) 29Si NMR chemical shifts of the halosilylium carborates. The computed 29Si chemical shifts (in ppm relative to tetramethylsilane) were obtained at the X2C-PBE0/x2c-TZVPall-2c level of theory, comparing one-component (1c) calculations with only scalar-relativistic effects (values on the dashed lines) and two-component (2c) calculations including also spin–orbit effects (values on the solid lines). aAll reactions were performed at ca. 0.02 mmol scale in C6D6 (0.2 ml, 0.1 M) in a glovebox under an argon atmosphere. bIsolated yields. cExperimental 29Si NMR chemical shifts determined by 1H,29Si HMQC NMR spectroscopy (500/99 MHz, 298 K, optimized for J = 7 Hz) in 1,2-Cl2C6D4. 1J(Si,F) coupling constant in Hz in parentheses. dComputed 29Si NMR chemical shifts at the ‘exact two-component’ (X2C) spin–orbit PBE0/x2c-TZVPall-2c level.

Full size image

In another effort to access a halogen-substituted silylium ion, we tested our protolysis approach, which already enabled the synthesis of the hydrogen-substituted congeners15. Following this strategy, a pale yellow suspension of Reed’s benzenium salt [H(C6H6)]+[HCB11H5Br6]− in benzene was treated with equimolar amounts of iPr2SiHCl (6cb), resulting in immediate gas evolution and the formation of an off-white solid (Fig. 2a, bottom). NMR spectroscopic analysis now indicated successful Si–H bond heterolysis, as apparent by the disappearance of the signal for the Si–H bond in the 1H NMR spectrum. The measured 29Si NMR chemical shift of δ = 74.3 ppm was shifted downfield relative to the hydrogen-substituted silylium ion 5c (Δδ = 4.6 ppm). Because the silylium carborate salts are not very soluble in benzene, the NMR spectra were recorded in deuterated ortho-dichlorobenzene. Unambiguous evidence for the formation of the chlorine-substituted, counteranion-stabilized silylium ion [iPr2ClSi(HCB11H5Br6)] (8bc) was eventually provided by X-ray diffraction analysis (see ‘Crystallographic characterization’ section).

We next embarked on a systematic investigation of halogen-substituted silylium ions by variation of the halogen substituent (X = F, Cl, Br or I) and alkyl substituent (Alk = Me, Et, iPr or tBu) in [Alk2XSi(HCB11H5Br6)] (Fig. 2b). Using equimolar amounts of benzenium salt [H(C6H6)]+[HCB11H5Br6]−, iPr2SiHBr (6cc) and iPr2SiHI (6dc) were equally converted in high yields to the corresponding halosilylium ions 8cc and 8dc, respectively (table in Fig. 2b, entries 11 and 15). Conversely, the protolysis of iPr2SiHF (6ac) led to a 1:1 mixture of [iPr2HSi(HCB11H5Br6)] (5c) and [iPr2FSi(HCB11H5Br6)] (8ac). The observed reaction outcome was unclear at this stage, but we had already speculated that fluoride abstraction from iPr2SiHF (6ac) by the transiently formed fluorosilylium ion iPr2FSi+ (8ac) competes with the dehydrogenative protolysis. We therefore used phenyl as the leaving group (LG = Ph) because dearylation (protodesilylation) proved to be a more facile and faster process than the dehydrogenation (LG = H) in our previous study15. Indeed, selective formation of [iPr2FSi(HCB11H5Br6)] (8ac) was now achieved by dephenylation of iPr2PhSiF (7ac) under the same reaction set-up (table in Fig. 2b, entry 3). In addition to the isopropyl group, methyl, ethyl and tert-butyl substituents were also tolerated in this reaction (Fig. 2b). It is worth mentioning that, unlike hydrogen-substituted silylium ions 5, the halogen-substituted silylium ions 8 were obtained without any detectable substituent redistribution, even when decorated with small methyl or ethyl groups at the silicon atom that are prone to intermolecular exchange reactions27,28.

NMR spectroscopic characterization

The measured solution 29Si NMR chemical shifts of the halogen-substituted silylium carborates 8 span a relatively wide spectral range (δ(29Si) = 94–41 ppm; Fig. 2b) compared with the purely alkyl-substituted tertiary silylium carborates 2 (δ(29Si) = 113–93 ppm)5. Replacement of one alkyl group in [Alk3Si(HCB11H5Br6)] (2) by a halogen atom generally leads to a considerable high-field shift of the 29Si NMR resonance. Although the 29Si NMR chemical shift of classical trialkylsilylium ions is rather insensitive to the alkyl substituent9, an increasing low-field shift is seen in the order Me < Et < iPr < tBu (refs. 13,28). This trend is also evident for the halogen-substituted silylium ions 8 (except for X = F) and becomes increasingly pronounced in the order Cl < Br < I (Fig. 2c, left). While in the iodosilylium ion series, the 29Si NMR chemical shifts cover the broadest spectral range (δ(29Si) = 94–41 ppm), this alkyl effect is not reflected for the fluorine-substituted silylium ions, and all 29Si NMR chemical shifts are essentially identical (Δδ(29Si) < 4 ppm). However, a correlation is again observable for the 1J(Si,F) coupling constant, which increases steadily in the order Me < Et < iPr < tBu (table in Fig. 2b, entries 1–4). The large 1J(Si,F) values between 369 Hz and 411 Hz are markedly higher than those in neutral fluorosilanes, which typically fall in the range of 275–290 Hz (for example, 1J(Si,F) = 275 Hz for Me3SiF)29. The 1J(Si,F) coupling constant of zwitterionic [tBu2FSi(HCB11H5Br6)] (8ad) also exceeds that of tBu2FSiBr by 65 Hz (1J(Si,F) = 410.5 versus 345.6 Hz)30. The 19F NMR chemical shifts are also more shielded in the same order (δ(19F) = −116.7 ppm for 8aa, −129.0 ppm for 8ab, −137.0 ppm for 8ac and −142.0 ppm for 8ad).

The experimental 29Si NMR chemical shifts were further corroborated by density functional theory (DFT) calculations (table in Fig. 2b and Fig. 2c, right). To understand the larger dependence on the alkyl substituent for the heavier halogen-substituted silylium ions, the chemical shifts were computed at adequate, that is, relativistic quantum-chemical levels (see Supplementary Information for the computational details). As expected15, computations on the free gas-phase cations resulted in far too large shifts, and inclusion of the interaction with the carborate counteranion is crucial to obtain shifts in the right range31 (Extended Data Table 1). It should be mentioned here that the silylium ion can exhibit different binding modes with the carborate cluster28,32 (for an in-depth analysis, see Supplementary Information Section 7). Moreover, neglecting the effects of spin–orbit coupling (SOC) at scalar-relativistic one-component (1c) levels leads to an overestimation of the shifts. In particular, the 1c calculations fail to properly describe the trends observed when going to the heavier halogen substituents (Fig. 2c, right, and Extended Data Table 2). Moderate SOC effects for X = F and Cl can be attributed mostly to the interaction of the silicon centre with the bromine atom of the carborate anion, while the larger SOC effects for the heavier halogen substituents align closely with the known SOC shielding effects of covalently bound halogen substituents in many other systems, the so-called normal halogen dependence33,34.

Importantly, the increasing curvature of the experimental shift plot towards X = Br and I with decreasing electron-donating ability of the alkyl substituent is reproduced only when SOC effects are included (Fig. 2c). It is known that such SOC NMR shifts arise through a Fermi-contact mechanism that transfers SOC-induced spin polarization from the heavy substituent responsible for SOC to the respective NMR nucleus33,34. Obviously, the SOC effects increase with the atomic number (that is, nuclear charge) of the halogen. In addition, a better electron-donating ability of the alkyl substituent with Me < Et < iPr < tBu strengthens the covalency of the Si–C bonds and, thus, diminishes the covalency of the Si–X bond. This is clearly evident from the natural population analysis (NPA) charges and from the composition of the relevant natural localized molecular orbitals (Extended Data Table 3), where along this series of alkyl substituents a given Si–X bond becomes increasingly polarized towards the halogen atom. While this effect looks rather moderate, such small changes are known to be sufficient to reduce the effectiveness of the mentioned Fermi-contact mechanism of the SOC NMR shifts34. As a result, the shielding SOC effects for a given halogen tend to be largest for Alk = Me and smallest for Alk = tBu (for example, Δδ(29Si) = −70 ppm for Me2ISi+ and Δδ(29Si) = −40 ppm for tBu2ISi+; Fig. 2c, right, and Extended Data Table 2). The known dependence of the SOC effects on the bond covalency together with the interplay between Si–C and Si–X bond covalency thus largely explain the interesting experimental findings.

The monotonous increase of the computed 29Si chemical shifts in the absence of spin–orbit contributions (see values on the dashed lines in Fig. 2c, right) correlates with decreasing HOMO–LUMO gaps (the energy difference between the highest occupied and lowest unoccupied molecular orbitals; Extended Data Table 4) and may reflect, to some extent, decreasing halogen–silicon π-bonding along this series. The more electron-donating alkyl substituents decrease the gaps and therefore contribute to larger shifts at the scalar relativistic level, that is, without spin–orbit effects.

Crystallographic characterization

For the isopropyl substituent, we succeeded in the crystallographic characterization of [iPr2XSi(HCB11H5Br6)] for all four halogen atoms (X = F, Cl, Br and I). Single crystals suitable for X-ray diffraction analysis were obtained from a solution of the corresponding silylium salt in ortho-difluorobenzene (for X = F, Br and I) or ortho-dichlorobenzene (for X = Cl) by vapour diffusion with n-pentane and n-hexane, respectively. All molecular structures show the same general features as the reported tertiary trialkyl-7,9,28 and secondary hydrogen-substituted15 silylium carborates (Fig. 3). The positively charged silicon atom is coordinated by one of the bromine atoms from the pentagonal belt of the icosahedral carborate cluster. This interaction leads to a distorted tetrahedral coordination environment at the silicon atom. The key bond lengths and angles of the halosilylium ions 8ac–dc are summarized in Table 1. For comparison, the geometric parameters for [iPr2HSi(HCB11H5Br6)] (5c) and [iPr3Si(HCB11H5Br6)] (2c)7,9 are also listed. As expected, the Si–X bond length increases in the order F < Cl < Br < I. The Si–Br bond distance to the coordinating carborate anion decreases from 8ac (2.439 Å) to 8bc (2.417 Å) and 8cc (2.393 Å), but is longer again for 8dc (2.436 Å). Although all Si–Br bonds are shorter than in trialkyl-substituted silylium carborate 2c, they are still substantially elongated compared with neutral bromosilane Me3SiBr (2.235 Å)35 and dibromosilane (F5C2)2SiBr2 (2.154 Å)36. The bromosilylium salt [iPr2BrSi(HCB11H5Br6)] (8cc) serves as a good reference for assessing the nature of the different Si–Br bonds, as the silicon atom combines both a coordinating and a bound bromine atom. While the former Si–Br bond length is 7% longer, the latter is 3% shorter than in Me3SiBr. This reflects the weakly coordinating nature of the anion and the cationic character of the silylium ion. In addition to the shortened Si–Br bond to the counteranion, the halosilylium ions 8ac–dc exhibit an increasing degree of pyramidalization compared with trialkylsilylium carborate 2c, which is indicative of a higher bromonium ion character arising from the electron-withdrawing effect of the halogen substituent. It should be noted, however, that the Si–Br bond lengths to the carborate anion (2.393–2.439 Å) as well as the sum of the three angles around the silicon atom as a measure of pyramidalization (∑R–Si–R = 344–346°) fall within a relatively narrow range and cannot be correlated with the Lewis acidity of the individual halosilylium ion (see ‘Lewis acidity’ section). A trend is seen only for each Si–C bond to the isopropyl substituent, which lengthens steadily from fluorine to iodine within the halosilylium ion series.

Fig. 3: Solid-state molecular structures of halogen-substituted silylium carborates.
figure 3

Thermal ellipsoids are shown at the 50% probability level. All hydrogen atoms are omitted for clarity.

Full size image
Table 1 Comparison of selected bond lengths and angles of silylium carborates [iPr2XSi(HCB11H5Br6)]
Full size table

Lewis acidity

In an attempt to experimentally correlate the relative Lewis acidities of the halogen-substituted silylium ions, we initially chose the established Gutmann–Beckett method using triethylphosphine oxide37. However, it turned out that this analysis is not compatible with the exceptional Lewis acidity of the halosilylium ions, and multiple signals were observed in the 31P{1H} NMR spectrum. We therefore turned to the Lewis acidity scale introduced by Müller and co-workers using para-fluorobenzonitrile (FBN) as Lewis base38. This NMR probe indeed proved to be suitable, and immediate Lewis adduct formation with the silylium ion led to clean generation of the corresponding silylnitrilium carborate (Fig. 4a). The addition of substoichiometric amounts of FBN ensured that the formation of any pentacoordinate silicon complexes (silanium ions) by double nitrile coordination was prevented. All silylnitrilium ions were fully characterized by multinuclear NMR spectroscopy (Supplementary Table 1 in Supplementary Information Section 5). The 19F NMR chemical shifts are summarized in the table of Fig. 4a. The degree of deshielding of the 19F NMR chemical shift relative to free FBN (δ(19F) = −103.9 ppm in C6D6) is a measure for scaling the relative Lewis acidities of the corresponding silylium ions. The results depicted in Fig. 4b rank the halosilylium ions 8ac–dc among the strongest isolable silicon Lewis superacids so far39. They are stronger Lewis acids than the corresponding hydrogen-, trialkyl- and triaryl-substituted congeners. The decreasing Lewis acidity from the fluorine to iodine substituent in iPr2XSi+ is in line with the observed trend of intramolecularly halo-stabilized silyl cations reported by Müller and co-workers22.

Fig. 4: Lewis acidity and reactivity of halogen-substituted silylium ions.
figure 4

a, Experimental and theoretical evaluation of the Lewis acidity based on Müller’s FBN scale and calculated fluoride ion affinities. b, The FBN acidity scale for different silylium ions in benzene. The 19F NMR chemical shift for Ar3Si+ (Ar = pentamethylphenyl) is taken from ref. 22. c, Chloride abstraction from a chlorosilane by a halogen-substituted silylium ion provides access to hydrogen-substituted silylium ions. d, Decomposition pathways of the halosilylium carborates. a19F NMR chemical shifts of the silylnitrilium carborates [iPr2XSi(FBN)]+[HCB11H5Br6]− determined from the 19F NMR spectrum (471 MHz, 298 K) in C6D6. bDeshielding of the 19F NMR chemical shift relative to free FBN (δ(19F) = −103.9 ppm). cDirectly computed gas-phase fluoride ion affinities at the LH20t-D4/def2-TZVPD//BP86-D4/def2-TZVP level. For a comprehensive list and a comparison with the anchor-point approach using OCF2, see Extended Data Table 5.

Full size image

The Lewis acidity ranking of the different silylium ions established with the FBN scale was further supported by DFT calculations of the FIAs, a generally accepted more direct thermodynamic index for Lewis acidity39. The FIA is defined as the negative of the enthalpy of the reaction LA + F− → LA–F−. Some of us have recently shown that the widely used anchor-point methods, which are based on isodesmic reactions for Lewis acids with a known FIA (for example, of F2CO) to avoid the calculation of free fluoride, are an improvement only when inadequate density functionals and/or insufficient basis sets lacking diffuse functions are used40. This is also corroborated for the present silylium ions, as supported by a comparison of directly computed and anchor-based FIAs for the full set of silylium ions (Extended Data Table 5). The directly computed FIAs for the different isopropyl-substituted silylium ions are summarized within the table of Fig. 4a. As expected, substitution of hydrogen by fluorine or chlorine enhances the Lewis acidity for a given silylium ion, while bromine or iodine substitution has a much smaller effect. In general, the electron-donating ability of the hydrogen and alkyl substituents increases along the series H < Me < Et < iPr < tBu, and this trend is also reflected in the computed FIAs (as well as in the NPA charges; Extended Data Table 3). Strikingly, Me2FSi+ has the largest FIA only just behind H3Si+ (Extended Data Table 5). Overall, the data clearly confirm that such silyl cations are among the strongest known Lewis acids.

Reactivity

The relative order of Lewis acidity was further tested experimentally by a competition experiment (Fig. 4c). Treatment of chlorosilylium carborate 8bc with equimolar amounts of iPr2SiHCl (6bc) resulted in immediate chloride abstraction from 6bc and the formation of hydrogen-substituted silylium salt [iPr2HSi(HCB11H5Br6)] (5c) along with iPr2SiCl2 (9bc), as monitored by NMR spectroscopy. The reaction outcome can be rationalized by the higher chloride ion affinity of iPr2ClSi+ compared with iPr2HSi+. At the same time, this result also provides an explanation why the established Corey reaction cannot be used to prepare halogen-substituted silylium carborates (Fig. 2a, top). As hydride abstraction from iPr2SiHCl (6bc) by the trityl cation is slow, the incipiently generated chlorosilylium ion iPr2ClSi+ immediately abstracts a chloride from the starting silane 6bc, eventually leading to the selective formation of [iPr2HSi(HCB11H5Br6)] (5c). All halosilylium carborates 8 are crystalline salts stable under inert atmosphere for several weeks at −30 °C, but slow decomposition of any halogen-substituted silylium ion is observed in ortho-dichlorobenzene solution (Fig. 4d, top; for an analysis of this decomposition pathway, see Supplementary Fig. 5). Exposure to sunlight led to even faster decomposition of the halosilylium ions in solution, now resulting in the formation of the corresponding dichlorosilane (Fig. 4d, bottom).

Conclusions

Previously elusive halogen-substituted silylium ions are not only exceptionally strong Lewis acids of academic interest, but also highly relevant in synthetic chemistry. Examples include their intermediacy in the halodealkylation of waste material from the Müller–Rochow process16,17 and their long-sought role as reagents in the hydrodefluorination of PFAS, exploiting their enormous fluoridophilicity25,26. The present work discloses a reliable synthetic access to the whole family of these superreactive cations in the form of their counteranion-stabilized carborate salts [Alk2XSi(HCB11H5Br6)] (X = F, Cl, Br or I). This was achieved by our dehydrogenative (LG = H) or dephenylative (LG = Ph) protolysis13,15 of the corresponding halosilanes R2XSi–LG using Reed’s superacidic benzenium salt12 [H(C6H6)]+[HCB11H5Br6]−. The molecular structures have been characterized by X-ray diffraction analysis with the fluorine congener being particularly fragile and delicate to handle. DFT calculations lend support to the understanding of the stabilization of these superacids, and the agreement between the measured and the computed 29Si NMR chemical shifts is excellent when SOC effects are taken into consideration. An assessment of the Lewis acidity using Müller’s FBN method38 and calculated FIAs40 confirm that halogen-substituted silylium ions are among the strongest Lewis acids ever isolated. This work opens the door to unexplored reactivity in catalysis with superelectrophilic silicon reagents.

Methods

General procedure for the generation of halogen-substituted silylium carborates

In a high-quality glovebox (O2, H2O < 1.0 ppm), the indicated hydrosilane 6 or phenylsilane 7 (1.0 equiv.) was added to a suspension of [H(C6H6)]+[HCB11H5Br6]− (1.0 equiv.) in C6D6 (0.2 ml). After stirring the reaction mixture for 2–3 h at room temperature, n-hexane (0.4 ml) was added. The suspension was filtered, and the residue was washed with additional n-hexane (3 × 0.2 ml) and dried for 10 min under high vacuum (~10−3 mbar). The halogen-substituted silylium carborates 8 were obtained as white solids, which can be stored for several weeks at −30 °C in the glovebox.