Pseudohalogen Chemistry in Ionic Liquids with Non‐innocent Cations and Anions

Abstract Within the second funding period of the SPP 1708 “Material Synthesis near Room Temperature”,which started in 2017, we were able to synthesize novel anionic species utilizing Ionic Liquids (ILs) both, as reaction media and reactant. ILs, bearing the decomposable and non‐innocent methyl carbonate anion [CO3Me]−, served as starting material and enabled facile access to pseudohalide salts by reaction with Me3Si−X (X=CN, N3, OCN, SCN). Starting with the synthesized Room temperature Ionic Liquid (RT‐IL) [nBu3MeN][B(OMe)3(CN)], we were able to crystallize the double salt [nBu3MeN]2[B(OMe)3(CN)](CN). Furthermore, we studied the reaction of [WCC]SCN and [WCC]CN (WCC=weakly coordinating cation) with their corresponding protic acids HX (X=SCN, CN), which resulted in formation of [H(NCS)2]− and the temperature labile solvate anions [CN(HCN)n]− (n=2, 3). In addition, the highly labile anionic HCN solvates were obtained from [PPN]X ([PPN]=μ‐nitridobis(triphenylphosphonium), X=N3, OCN, SCN and OCP) and HCN. Crystals of [PPN][X(HCN)3] (X=N3, OCN) and [PPN][SCN(HCN)2] were obtained when the crystallization was carried out at low temperatures. Interestingly, reaction of [PPN]OCP with HCN was noticed, which led to the formation of [P(CN)2]−, crystallizing as HCN disolvate [PPN][P(CN⋅HCN)2]. Furthermore, we were able to isolate the novel cyanido(halido) silicate dianions of the type [SiCl0.78(CN)5.22]2− and [SiF(CN)5]2− and the hexa‐substituted [Si(CN)6]2− by temperature controlled halide/cyanide exchange reactions. By facile neutralization reactions with the non‐innocent cation of [Et3HN]2[Si(CN)6] with MOH (M=Li, K), Li2[Si(CN)6] ⋅ 2 H2O and K2[Si(CN)6] were obtained, which form three dimensional coordination polymers. From salt metathesis processes of M2[Si(CN)6] with different imidazolium bromides, we were able to isolate new imidazolium salts and the ionic liquid [BMIm]2[Si(CN)6]. When reacting [Mes(nBu)Im]2[Si(CN)6] with an excess of the strong Lewis acid B(C6F5)3, the voluminous adduct anion {Si[CN⋅B(C6F5)3]6}2− was obtained.


Synthesis of Pure Pseudohalide Containing ILs from ILs with Decomposable Anions
In addition to the synthesis of the pseudohalide salts, we aimed to isolate possible intermediates such as [CO 2 X] À (X = pseudohalogen), which are presumably formed during the decomposition reaction (Scheme 2). The fluoro-and cyanoformates [CO 2 F] À [131] or [CO 2 CN] À [132] are already literature known, of which the latter was synthesized as the [Ph 4 P] + salt by exposure of CO 2 to a concentrated solution of [Ph 4 P]CN in acetonitrile.
First, we tried to isolate pseudohaloformate anions from concentrated and cooled reaction solutions of [R 3 MeN][CO 3 Me] and Me 3 SiÀ X in acetonitrile. [133] But even at À 40°C the release of gaseous CO 2 was observed and no new signals for any [CO 2 X] À species could be observed by means of 13 C{ 1 H} NMR spectroscopy. Subsequently, a solvent free process was used, from which the [CO 2 CN] À anion was synthesized with [Et 3 MeN] + as counter cation, when heating pure [Et 3 MeN]CN to 150°C in a CO 2 atmosphere at a pressure of 3 bar. Initially, the salt remained solid at high temperature but then liquefied when it was cooled down to room temperature. Raman spectroscopy revealed that at room temperature the RT-IL [Et 3 MeN][CO 2 CN] with ν CN = 2196 cm À 1 is present, while at elevated temperature (100°C) CÀ C bond cleavage (activation barrier~40 kJ/mol) [132] and formation of the starting materials [Et 3 MeN]CN with ν CN = 2049 cm À 1 and CO 2 is observed (Scheme 3). In a closed system this process is reversible, whereby [Et 3 MeN][CO 2 CN] is formed again upon cooling. However, when the system is opened, CO 2 is irreversibly released and [Et 3 MeN]CN is formed.
Nevertheless, a reaction of other ammonium pseudohalides of the type [nBu 3 MeN]X (X = N 3 , OCN und SCN) with CO 2 could not be observed under similar reaction conditions.

Reactions of Pseudohalides Containing ILs with P 4 and P 4 S 10 [133]
As Schmidtpeter et al. could show, white phosphorus is degraded by cyanide salts to dicyanophosphide and different polyphosphides in solution (Scheme 4). [134] We wondered whether new phosphorus pseudohalide compounds could be synthesized, using the pure ILs without further solvents. [133] To prevent the white phosphorus from sublimating, a mixture of P 4 and the ILs [nBu 3 MeN]X (X = CN or N 3 ) was filled into an ampoule. This was sealed and then heated to 65°C (X = N 3 ) and 105°C (X = CN), respectively, causing the solids to liquefy. When heating the mixture for 24 h, a shiny violet-black solid was formed in both reaction vessels and a red (X = N 3 ) or colorless (X = CN) liquid was formed as well. Subsequently, the reaction products were extracted with Scheme 1. Synthesis of methylcarbonates with trialkylmethylammonium (E = N, R = Et, nPr, nBu) or methyltriphenylphosphonium (E = P, R = Ph) counter ions in methanol with dimethyl carbonate and a tertiary alkylated amine or triphenylphosphine, respectively.   ). [134] acetonitrile and benzene, leaving an insoluble black solid. By means of 1 H and 13 C{ 1 H} NMR spectroscopic analysis of the liquid phase, the decomposition of the cation to nBu 3 N could be determined in both reactions. The reaction of P 4 with cyanide also resulted in the formation of traces of [P(CN) 2 ] À as was observed by Schmidtpeter et al. [134] The solid residues could not be examined further by Raman spectroscopy due to fluorescence, but elemental analyses of the substances indicated a high phosphorus content of the compound with 61 % for X = N 3 and 33.7 % for X = CN. The formation of further unknown pseudohalide phosphorus compounds could not be observed by means of 31 P NMR spectroscopy. Due to the instability of the cation, no further experiments with [nBu 3 MeN] X (X = OCN and SCN) were performed.
Likewise, the aim was to investigate whether the pseudohalide salts undergo reaction with phosphorus pentasulfide, hopefully leading to the formation of new compounds via a solvent-free synthesis process. Roesky et al. already observed the formation of [(NCPS 2 ) 2 S] À , [(N 3 ) 2 PS 2 ] À and [(N 3 PS 2 ) 2 S] À or [(SCN) 2 PS 2 ] À , which were isolated as [nPr 4 N] salts by reaction of MX (M = Na or K; X = CN, N 3 or SCN) with P 4 S 10 in acetonitrile. [134] We heated a mixture of two equivalents of [nBu 3 MeN]CN with one equivalent of P 4 S 10 at 105°C for 24 h, thus resulting in liquefaction of both reactants and formation of a homogeneous phase. In solvents such as Et 2 O, benzene, thf or n-hexane, no formed products could be extracted, whereas in acetonitrile the complete residue was dissolved. From the complex reaction mixture, only two formed products could be assigned by 31 P NMR and (ESI-TOF)-MS. One was the adamantane-like [P 4 S 9 N] À anion [135,136] (δ[ 31 P] = 67 and 33 ppm; m/z = 426) and the second was the [(SCN) 2 PS 2 ] À anion [134] (δ[ 31 P] = 53 ppm, m/z = 211), which are already known from literature and are shown in Scheme 5. Interestingly, Roesky et al. prepared the [P 4 S 9 N] À anion by the reaction of [(N 3 ) 2 PS 2 ] À with P 4 S 10 , while the [(SCN) 2 PS 2 ] À was formed from P 4 S 10 and KSCN.
As the products could neither be extracted from the mixture nor crystallized in different mixtures of solvents, e. g. acetonitrile/benzene or acetonitrile/diethyl ether, no further investigations with other ILs [nBu 3 MeN]X (X = N 3 , OCN, SCN) were carried out. [ [137] A direct method for the synthesis of this compound was achieved when [PPN]SCN was treated with in situ formed HNCS, generated from MeOH and Me 3 SiÀ NCS (Scheme 6).

Reactions of Pseudohalides containing ILs with Acids of Pseudohalides (HA)
According to X-ray analysis, a slightly bent anion is formed, with one N···H···N hydrogen bridge (Figure 1, contacts a and [139,140] Already in 1978, Salthouse and Waddington observed that, when HCN was added to [nPr 4 N]CN, the solution became increasingly viscous due to polymerization of the HCN and also noticed that the color of the solution quickly intensified. Crystals, which formed in the black oil, were examined by IR spectroscopy. A single band for ν CN = 2060 cm À 1 indicated the formation of hydrogen dicyanide [H(CN) 2 ] À , but a solid state structure was not determined. [141] Thus motivated, we tried to synthesize salts containing the [CN(HCN) n ] À (n = 1, 2, 3 …) ions by reaction of pure  HCN [125] In an initial study the salts were suspended in Me 3 Si-CN and then MeOH was added to generate HCN in situ, but no crystallization could be achieved when cooling the solutions. Therefore, the [WCC]CN were dissolved directly in an excess of 15 equivalents of HCN, which was cooled to 0°C. The concentrated, highly ionic solution with the characteristics of an IL turned from yellow to brown within an hour and became increasingly viscous, which can be explained by the formation of polymeric HCN species. [142] Since in all cases no crystallization of a product could be achieved, different salts of the type   1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57 to temperature as well as moisture and decompose over time, observed by Raman spectroscopy. X-ray structure analysis proved the formation of molecular dihydrogen tricyanide anions [CN(HCN) 2 ] À (Figure 2). The anions form distorted chains which are formed by hydrogen bonds (Figure 2, contacts a and  b). The chains are connected to each other by weak Van der Waals interaction (Figure 2, contact c) Figure 3, contacts a-f) and interionic Van der Waals interactions ( Figure 3, contacts g-j) also lead to chains along the b-axis in the crystal. Y-shaped molecular anions are built, due to the larger cavities formed by the bulky [PPN] cation, which allows the aggregation of a further HCN molecule. This Y-shaped isomer, which contains only hydrogen cyanide molecules, also represents the energetically preferred structure of all calculated isomers of the trisolvates.
In addition, we were curious, whether it is possible to stabilize and isolate HCN aggregates with other pseudohalides, such as azide, cyanate, thiocyanate and the phosphaethynolate anion in form of their [PPN] + salts. [138] Again, the pseudohalides were dissolved in cooled HCN, but unfortunately no crystals could be obtained from the concentrated and cooled phase. Hence, a rather unusual crystallization method was used. Fomblin YR-1800 perflouropolyether was placed in a lowtemperature stage for crystal picking, which was cooled to À 60°C. The HCN salt mixtures were cooled until the liquid layer of the IL-like mixture slowly solidified and subsequently, small portions of the solid were transferred into the ether. Even at À 60°C, the release of gaseous HCN could be observed which resulted in formation of X-ray suitable crystals. The azide and cyanate salts crystallized as trisolvates [PPN][X(HCN) 3 ] (X = N 3 , OCN) (Figure 4 and Figure 5). Both molecular anions exhibit a distorted Y-shaped geometry. Head-to-tail contacts of adjacent anions can be observed but with quite long interionic distances. Hence, the formation of anionic strands, which are located in the cavities formed by the bulky [PPN] + ions, are better explained by packing effects in the crystal. Due to disorder of the central cyanate anion in [OCN(HCN) 3 ] À , two different isomers have to be considered. On the one hand, an isomer (52(1) %) in which two CÀ H···O and one CÀ H···N H-bridges are observed ( Figure 5, left), and on the other hand an isomer (48(1) %) in which two CÀ H···N and one CÀ H···O H-bridges are found ( Figure 5, right). The thiocyanate salt crystallized as HCN     (Figure 6, left). Computations suggest that the energetically favored isomers exhibit a Y-shaped structure with two CÀ H···N Hbridges. However, the observed L-shaped geometry was found to be only 0.58 kcal/mol higher in energy. Interestingly, if the pseudohalide PCO À , [143] which was synthesized in form of [PPN] PCO, is dissolved in HCN, the degradation of the phosphaethynolate anion was observed. This led to the formation of the dicyanophosphide anion [P(CN) 2 ] À , which crystallized as HCN disolvate [P(CN·HCN) 2 ] À (Figure 6, right).

Synthesis of Pseudohalogen Borate and Phosphate Species in Pseudohalide Containing
ILs [123] The Lewis acidities of trimethyl phosphate OP(OMe) 3 and trimethyl borate B(OMe) 3 were investigated towards the pure ILs [nBu 3 MeN]X (X = CN, N 3 , OCN, SCN), hopefully leading to new pseudohalide phosphates or borates. However, a reaction of [nBu 3 MeN]X with OP(OMe) 3 did not lead to the formation of new pseudohalide phosphates as hoped (Scheme 8, route A), but in all cases to methylation of the pseudohalide and to formation of [nBu 3 MeN][O 2 P(OMe) 2 ], which could be shown clearly by means of NMR spectroscopy (Scheme 8, route B). [123] When heating [nBu 3 MeN]X and B(OMe) 3 in a flask, a reaction was observed exclusively with X = CN. When heating an equimolar mixture slowly to 75°C, the RT-IL [nBu 3 MeN][B (OMe) 3 (CN)], with a melting point of À 51°C, was formed. The cyanidotrimethoxyborate anion is thermally unstable and can be transferred to the starting materials in vacuum, which is accelerated when raising the temperature (Scheme 9). After cooling the RT-IL to room temperature, the formation of crystals could be observed. X-ray elucidation of the single crystals revealed the formation of the double salt [nBu 3 4 ] À and cyanotrimethylsilane. [144] 6. Reaction of Pseudohalogen Borate ILs with Persilylated Compounds of Group 15 [133] Since it is already known that [B(OMe) 4 ] À undergoes reaction with cyanotrimethylsilane to form [B(OMe) 4 À n (CN) n ] À (n = 2, 3) under elimination of Me 3 SiÀ OMe, the reactivity of the synthesized [B(OMe) 3 (CN)] À anion towards tris(trimethylsilyl)phosphane and tris(trimethylsilyl)amine was investigated. [133] The aim was to synthesize new cyanide-containing compounds by the same silylether elimination reaction under formation of a boron-element bond (Scheme 10).
An equimolar mixture of [nBu 3 MeN][B(OMe) 3 (CN)] and E (SiMe 3 ) 3 (E = N, P) was placed in a flask. In both cases, a twophase system was formed in which the starting materials remained unchanged at room temperature, showing no reaction. When heating the mixtures to 55°C, no reaction with N (SiMe 3 ) 3 was observed according NMR analysis. However, with P  (SiMe 3 ) 3 partial decomposition of the ammonium cation to nBu 3 N and formation of Me 3 SiÀ OMe and B(OMe) 3 was found, accompanied with the formation of [B(OMe) 2 (CN) 2 ] À , which could be observed by means of (ESI-TOF)-MS. According to 11 B and 31 P NMR, a desired product synthesis with BÀ P bond formation did not occur.

Synthesis of Coordination Polymers Utilizing
Cyanide Silicate Containing ILs with Decomposable Cations [124,145] The synthesis of hexacyanidosilicate dianions [Si(CN) 6 ] 2À was motivated by the fact that this anion, besides the already synthesized pseudohalide analogues [SiX 6 ] 2À (X = N 3 , [146,147] OCN, [148,149] SCN, [150,151] SeCN, [152] NCCrCo 5 [153] ), was not yet known. The first attempt to synthesize hexacyanido silicates was performed via Cl/CN substitution reactions. [145] Figure 3). Portius et al. faced the same problem when synthesizing hexacyanidosilicates as they could show in their publication, which appeared at the same time as ours. [154] Hence, we thought, that utilization of IL-like reaction mixtures could help to solve this problem.
Therefore, we changed the synthesis strategy. We started with the synthesis of [R 3 HN] 2 [SiF 6 ] (R = Et, nPr), by protonation of tertiary amines with H 2 SiF 6 (Scheme 12, route A), and [R 3 MeN] 2 [SiF 6 ] (R = nPr, nBu), which were formed by decom-position reactions of H 2 SiF 6 with the previously synthesized ILs, the ammonium methylcarbonates (Scheme 12, route B). Subsequently, the fluorido silicates were suspended in 20 equivalents of Me 3 SiÀ CN and selectively converted to [SiF(CN) 5 ] À and [Si(CN) 6 ] À by temperature-controlled F/CN substitution reactions (Scheme 12, Figure 7), which could be isolated on a preparative scale. Catalytic amounts of GaCl 3 as Lewis acid shortened the reaction time when synthesizing the hexacyanidosilicate dianions.
In addition, the reaction of the hexacyanidosilcate dianion towards the Lewis acid tris(pentafluorophenyl)borane B(C 6 F 5 ) 3 was investigated. [145] But since all synthesized ammonium and alkali metal salts were only soluble in Lewis basic solvents such as MeCN, MeOH or H 2 O, which react themselves with the borane, a cation was needed, which allowed the salt to dissolve in non-Lewis basic solvents such as CH 2   is used. Unfortunately, only different hexacyanidosilicate-borane substitution patterns could be observed by means of 11 B{ 1 H} NMR spectroscopy, but no compound could be isolated from the reaction mixture. But when the borane is used in excess (n > 6), a complete functionalization of all cyanide ligands with the borane was found and the voluminous {Si[CN·B(C 6 F 5 ) 3 ] 6 } 2À adduct anion (V anion~2 .77 nm 3 ) is formed (Scheme 13, bottom; Figure 8).

Conclusions
In conclusion, we have shown that salts containing the methyl carbonate anion like [R 3 MeE][CO 3 Me] (E = N, R = Et, nPr or nBu; E = P, R = Ph) are suitable starting materials for the quantitative synthesis of the corresponding ammonium and phosphonium pseudohalides [R 3 MeE]X (X À = CN, N 3 , OCN, SCN) 6 ] and novel imidazolium salts, respectively. Further, derivatization of the hexacyanidosilicate dianion with the Lewis acid B(C 6 F 5 ) 3 led to formation of the bulky adduct anion {Si[CN·B(C 6 F 5 ) 3 ] 6 } 2À , which can be regarded as a very bulky WCA (weakly coordinating anion).