Silylated Sulfuric Acid: Preparation of a Tris(trimethylsilyl)oxosulfonium [(Me3Si−O)3SO]+ Salt

Abstract The chemistry of silylated sulfuric acid, O2S(OSiMe3)2 (T2SO4, T=Me3Si; also known as bis(trimethylsilyl) sulfate), has been studied in detail with the aim of synthesizing the formal autosilylation products of silylated sulfuric acid, [T3SO4]+ and [TSO4]−, in analogy to the known protonated species, [H3SO4]+ and [HSO4]−. The synthesis of the [TSO4]− ion only succeeds when a base, such as OPMe3 that forms a weakly coordinating cation upon silylation, is reacted with T2SO4, resulting in the formation of [Me3POT]+[TSO4]−. [T3SO4]+ salts could be isolated starting from T2SO4 in the reaction with [T−H−T]+[B(C6F5)4]− or T+[CHB11Br6H5]− when a weakly coordinating anion is used as counterion. All silylated compounds could be crystallized and structurally characterized.

Almost 50 years of silylium ion chemistry have shown that many applications for silylium ions in the field of catalysis have emerged from the pure basic research of the first decades. [1][2][3][4][5][6][7][8][9] The development of silylium ion chemistry is closely related to carbenium ion chemistry, and it is no coincidence that silicon is also called the "kissing cousin" of carbon. [2] And while we are on the subject of relationships: The [Me 3 Si] + ion (T + ) can also be understood as the "big brother" of the proton (Scheme 1). [10][11][12][13][14] Replacing H + with T + has several advantages. Substitution usually results in a thermodynamic (e.g. through hyperconjugation) and kinetic stabilization (through a larger steric demand) of the species under consideration. In the case of the pseudohalogen acids HX vs. TX (X = pseudohalogen), [15][16][17] for example, this leads to a significant stabilization, as can be seen in the increased melting and boiling points as well as the reluctance to oligomerize (e.g. X = CN, SCN, OCN). While HN 3 is a highly explosive substance, TN 3 can be handled safely even in large quantities at higher temperatures. [18][19][20][21] Nevertheless, the chemistry of a protonated species is often similar to that of a silylated species (Scheme 1). For example, a classical neutralization reaction can also be formulated for the silylated species. Furthermore, the neutral dimers H 2 and hexamethyldisilane T 2 , (Me 3 Si) 2 , show a similar reactivity towards dihalogens (X 2 ), that is, they form HX and TX, respectively, in the reaction with X 2 , even with X = I. [22,23] Like a free proton that does not exist in the condensed phase, also the T + ion is always coordinated either to a neutral solvent, anion or any other Lewis basic site in a molecular system. Therefore, it is not surprising that in analogy to the protonated species, such as [HÀXÀH] + (X = halogen, [24,25] pseudohalogen), [26,27] [H n+1 E] + ] (E= element of group 15 [28,29] for n = 3 and 16 [28,[30][31][32] for n = 2) or arenium ions [12] in aromatic systems, also the silylated species [13,14,[33][34][35][36] can be isolated in the presence of a weakly coordinating anion (Scheme 1). [4,6] Like the protonated species, all these silylated cations should be regarded as strong Lewis acids that can be utilized as T + transfer reagents.
Interestingly, while the chemistry of T 3 PO 4 and its silylated cationic species [T 4 PO 4 ] + has been explored, [37][38][39] nothing has been reported about a silylated cation of the type [T 3 SO 4 ] + to the best of our knowledge. However, protonated sulfuric acid, [H 3 SO 4 ] + , was isolated by Minkwitz et al. in a super acidic system (HF/SbF 5 ) as [SbF 6 ] À salt. [40] As early as 1945, Patnode and Schmid reported on the synthesis of bis(trimethylsilyl)sulfate, T 2 SO 4 (1), which they obtained in the reaction of TCl with H 2 SO 4 (Scheme 2, Eq. 1). [41] Since then, T 2 SO 4 has often been used as a silylation reagent. [42][43][44][45][46][47][48][49][50] Following our interest in [Me 3 Si] + chemistry, we studied the similarities between sulfuric acid, H 2 SO 4 , and its silylated congener T 2 SO 4 . Especially, we were intrigued by the idea to synthesize the formal autosilylation products of 2 We started this project with the synthesis of crystalline T 2 SO 4 (1, T= Me 3 Si) from TCl and 95 % H 2 SO 4 (Scheme 2, Eq. 1, Figure 1), which we obtained in 33 % yield after vacuum distillation at 100 8C (10 À3 mbar, see SI). With T 2 SO 4 in hand, we reacted it with various bases, such as DMAP (4-(dimethylamino)pyridine), KOtBu and OPMe 3 to "neutralize" exactly one T + ion in order to generate [TSO 4 ] À (Scheme 2, Eq. 2-5). With KOtBu as base (independent of the stoichiometry), we always isolated K 2 SO 4 and observed in solution the formation of the ether TÀOÀtBu as evidenced by 1 H, 13 C and 29 Si NMR studies. Also, the reaction of Cs 2 (SO 4 ) with T 2 SO 4 in toluene in the presence of [18]crown-6 (to increase the solubility) did not lead to the formation of a [TSO 4 ] À salt, but the pyrosulfate [Cs [18]crown-6] 2 S 2 O 7 (Xray, see SI) and T À O À T (NMR) were produced in a condensation reaction. The reaction with DMAP was carried out in 2:1, 1:1 and 1:2 ratios in CH 2 Cl 2 and followed by 14 N, 29 Si and 17 O NMR spectroscopy (Figure S1a-c). In the 14 N spectra, a strong broadening and shift of the two DMAP resonances (d[ 14 N] = À324.9 and À104.7 ppm) were observed, increasing with increasing amount of T 2 SO 4 . The two resonances (153 and 174 ppm) in the 17 O NMR spectra are shifted to higher field with increasing amount of T 2 SO 4 and the broad resonance at 153 ppm even vanishes. Interestingly, in the 29 Si NMR studies, we always observed only one resonance strongly shifted and not resolved compared to that of pure T 2 SO 4 (pure T 2 SO 4 : d[ 29 Si] = 33.6 ppm, cf. T 2 SO 4 /DMAP ratio: 2:1 29.7, 1:1 27.7 and 1:2 24.4 ppm, Figure S1c). Therefore, we assume a highly dynamic DMAP/T 2 SO 4 system from which we could only isolate crystalline [DMAPÀT] 2 SO 4 (X-ray, see SI). To avoid the problems as discussed before, we tried the slightly weaker base OPMe 3 ,   Table 1. which then, indeed, led to success. When exactly one equivalent of OPMe 3 is reacted with one equivalent of pure T 2 SO 4 in toluene, a crystalline trimethylsilylsulfate salt, 4 ], is obtained after concentration of the solution in 77 % yield (2, Scheme 2, Eq. 5, Figure 1). Only on one occasion could we isolate from such a reaction mixture one crystal of a side product, which was found to be the doubly desilylated pyrosulfate, [Me 3 POÀT] 2 [S 2 O 7 ] (X-ray, see SI).
The synthesis of a tris(trimethylsilyl)oxosulfonium [T 3 SO 4 ] + salt is achieved by reacting [T À H À T][B(C 6 F 5 ) 4 ] with silylated sulfuric acid in an 1:1 ratio in toluene. Attempts to crystallize the salt [T 3 SO 4 ][B(C 6 F 5 ) 4 ] failed both at room temperature and at lower temperatures such as 5 8C and À20 8C. Attempts to remove the entire solvent in vacuum (1-10 À3 mbar) at 60 8C resulted in the decomposition of the salt, which can be observed by the formation of a black insoluble residue. [12] The addition of non-polar solvents such as nhexane to precipitate the salt also failed. Changing the solvent from toluene to 1,2-dichlorobenzene was also unsuccessful. For this reason, we changed the counterion, as we assumed that the decomposition was initiated by a C À F activation at the borate anion. It is known that carborates are much more chemically robust compared to the [B(C 6 F 5 ) 4 ] À anion. [4,12,51] Indeed, when [Me 3 Si][CHB 11 Br 6 H 5 ] is reacted with T 2 SO 4 in toluene, colorless crystals of the desired [T 3 SO 4 ] + -salt are obtained in 68 % yield after 30 min ultrasound treatment at 60 8C and recrystallization (Scheme 2, Eq. 7). The formation of the [T 3 SO 4 ] + -ion with [CHB 11 Br 6 H 5 ] À as counterion was unequivocally proven by single-crystal X-ray studies (Figure 1, bottom). It should be noted that although we were able to generate the formal autosilylation products of T 2 SO 4 by separate synthesis routes, dissociation into [T 3 SO 4 ] + and [TSO 4 ] À was not observed for T 2 SO 4 , but [T 3 SO 4 ] + and [TSO 4 ] À react to give two T 2 SO 4 molecules immediately.
All three silylated sulfuric acid species [TSO 4 ] À , T 2 SO 4 and [T 3 SO 4 ] + were studied by different 13 C, 17 O, 29 Si, and 31 P NMR techniques in solution (see SI) as well as IR/Raman spectroscopy. As expected, the 29 [37] and computed 385 ppm for naked [Me 3 Si] + (g) , [52,53] see SI, Table S3). As the 29 Si NMR chemical shifts can be used as an indicator for the silylium ion character (and the deviation from planarity, see below), [52][53][54] that is, for the strength of the [Me 3 Si] + interaction with the solvent T 2 SO 4 , it can be assumed that T 2 SO 4 is a rather strong coordinating solvent utilizing the scale by Cremer et al. (À50 to 90 ppm, cf. 90-190 weakly coordinating, 200-370 weakly interacting, 370-385 noncoordinating solvents and 385 ppm gas phase). [52] Crystals of all three silylated sulfuric acid species are moisture sensitive but thermally considerably stable with defined melting points (  (Figure S3), for T 2 SO 4 correspondingly with two neighboring T 2 SO 4 molecules ( Figure S2) and in [T 3 SO 4 ] + exactly one such interaction ( Figure S4), however, with one adjacent cation. In the latter case, interestingly, weak Br anion ···H À C cation interactions are added. Likewise, weak Br anion ···H À C anion interactions are found between the H atom attached to the C atom of one carborate anion and the Br atom in para-position to the CÀH bond atom of an adjacent second carborate anion ( Figure S4). This leads to a zig-zag chain of carborate anions in the solid. The [T 3 SO 4 ] + cations coordinate with this chain via the abovementioned weak Br anion ···H À C cation interactions.
As depicted in Figure 1, the central SO 4 core always adopts a highly distorted tetrahedral geometry, with two different SÀO bond lengths (   [8], [55]  To get some insight into the charge transfer upon silylation and desilylation, respectively, we computed the partial net charges of the elements and the [Me 3 Si] as well as [SO 4 ] moieties within all three silylated species (Table 1