V. S. V. S. N.
Swamy
a,
Shiv
Pal
b,
Shabana
Khan
*b and
Sakya S.
Sen
*a
aInorganic Chemistry and Catalysis Division, CSIR-National Chemical Laboratory, Dr Homi Bhabha Road, Pashan, Pune 411008, India. E-mail: ss.sen@ncl.res.in; Web: http://academic.ncl.res.in/ss.sen Fax: +9120-2590-2633; Tel: +9120-2590-2052
bDepartment of Chemistry, Indian Institute of Science, Education, Research, Pune, Dr Homi Bhabha Road, Pashan, Pune 411008, India. E-mail: shabana@iiserpune.ac.in; Web: http://www.iiserpune.ac.in/~shabana Fax: +9120-2589-9790; Tel: +9120-2590-8137
First published on 1st June 2015
Cations and dications of heavier group 14 elements in their low oxidation state have received widespread attention in recent years. The journey started with the isolation of a series of cations of the composition [(C5Me5)E:]+ [E = Si–Pb], followed by the more recent isolation of a Ge(II) dication encapsulated within a cryptand, a carbodiphosphorane stabilized [GeCl]+ monocation with a two coordinate Ge atom, Si(II) cations and dications stabilized by N-heterocyclic carbenes (NHCs), which highlights the ongoing growth and interest in the chemistry of tetrel(II) cations. This is presumably because the central atom (E) in these compounds contains two or three unoccupied valence orbitals as well as holds a lone pair of electrons. Such an electronic description represents ambiphilicity, which is of great interest for catalysis. The successful synthesis of divalent group 14 cations requires new synthetic strategies based on the sterically demanding neutral or monoanionic ligands, utilization of counter anions, and solvents with low nucleophilicity in order to minimize the degree of interactions with the cations. An alternative approach for the realization of divalent cations of group 14 elements is their coordination to the transition metals. This synthetic approach was successfully applied for the isolation of a range of transition metal coordinated divalent cations of group 14 elements. Apart from arousing academic interest some of these cations have found application as activators in the Ziegler–Natta polymerization of alkenes.
Considering the rich chemistry displayed by carbenium ions, the realization of their heavier analogs i.e. [R3E+ (E = Si–Pb)] has become one of the important areas of modern inorganic chemistry and is still an area of continuous investigation. However, in contrast to the thoroughly studied chemistry of carbenium ions, only a little is known about the chemistry of heavier group 14 cations. The cations of heavier group 14 elements can be classified into two distinct groups, based on the oxidation state of the atom (Scheme 1). The most common class of the cations is tetrylium ions (I) (group 14 is called the tetrel group), which possess six valence electrons and are considered as heavier analogues of carbenium ions (R3C+). The generation of such a cation is easier in the gas phase, which provides an environment devoid of possible interfering solvents. However, their generation in the condensed phase is very difficult due to the highly electrophilic nature of R3E+ and their propensity to react with any donor moiety, even with arenes. Therefore, bulky donor substituents are required to quench the electrophilicity of the cation as well as to shield the cationic center from any nucleophilic attack. With these new synthetic strategies as well as taking advantages of the low nucleophilicity of borates and carboranes as counter anions and toluene as a solvent, Lambert and Reed et al. independently isolated the first silylium ions [Et3Si(toluene)]+[B(C6F5)4]− and [iPr3Si]+[CB11H6Br6]−, respectively.4 However, the Si centers in these silylium cations adopted a markedly pyramidal geometry instead of the expected planar geometry and resonated at δ = 92.3 and 102.8 ppm in their respective 29Si NMR spectrum. The calculated gas phase chemical shift of Et3Si+ was 354.6 ppm,5 which was considerably downfield than the experimentally observed values. Consequently, the cationic character of Si atoms in these compounds was heavily questioned and eventually both Lambert and Reed agreed that neither [Et3Si(toluene)]+ nor [iPr3Si]+ were “bona fide” silylium cations.6 The state of affairs was similar to those of germanium and tin. The groups of Lambert and Kira independently reported the isolation of [nBu3Sn]+X− [X = B(C6F5)3H and B{(3,5-CF3)2C6H3}4] featuring a tri-coordinate Sn atom with 119Sn NMR of δ = 360 and 356 ppm.7 However, Edlund et al. proposed that the NMR data of the aforementioned tin cations corresponded to the arene bound Sn atoms as three-coordinate trigonal-planar R3Sn+ was computationally predicted to resonate at δ = 1500–2000 ppm.8 Thus, a lot of controversies were sparked and simultaneously a huge effort was dedicated for the isolation of three-coordinate trigonal planar tetrylium ions. This was finally accomplished in 2000s with the isolation of “bona fide” three coordinate trigonal planar silylium [{Mes3Si}+{CB11HMe5Br6}−, Mes = 2,4,6-Me3-C6H2],9 germylium [{(tBu2MeSi)3Ge+}{B(C6F5)4}−], and stannylium [{(tBu2MeSi)3Sn+}{B(C6F5)4}−],10 [{Tipp3Sn+}{B(C6F5)4}−], (Tipp = 2,4,6-iPr3-C6H2) cations (Chart 1).11 Subsequently, the advances in the syntheses of novel condensed-phase cations as well as the reactivity studies have become important research areas with mainly two fold objectives: (i) to study these exotic compounds in the solid state due to the paucity of previous reports of structurally characterized heavier analogues of carbenium ions and (ii) to explore these species as efficient Lewis acid catalysts. Hallmark studies by Lambert, Reed, Müller, Corriu, Belzner, Oestreich, and others not only led to the isolation of many more such derivatives but revealed them as potential Lewis acids in homogeneous catalysis.12 Their enhanced electrophilicity has already been exploited in Diels–Alder reactions,13 C–C bond formation,14 and C–F activation reactions,15 as well as in small molecule activation e.g. dihydrogen.16
Another distinctly different class of group 14 cations is tetrylium-ylidenes (II) (Scheme 1), which features a group 14 element with a lone pair. It can be imagined that a species of composition (RE:+), which combines the nucleophilic nature of carbene as well as the high electrophilicity of cations is of high synthetic importance as a potential noninnocent ligand in catalysis. However, the number of valence electrons in type II cations is only four, while there are six valence electrons in the type I cations. Thus, further loss of two electrons from the closed shell might imply even greater reactivity and more difficulty in isolation. It was obvious that like R3Si+, the laboratory realization of RSi+ (a derivative of HSi+) could only be accomplished through kinetic stabilization, but neither steric protection nor electronic stabilization may be adequate if only one substituent is attached to the Si center.
Increasing the coordination number of Si in RSi+ without perturbing the formal oxidation state can be achieved mainly by two ways: (i) taking advantage of the cyclopentadienyl type of ligands, which can undergo haptotropic shift depending on electronic requirements of the system; (ii) attach a neutral donor ligand to the cation. The donor moiety occupies the empty coordination site of the cation and thereby reduces its electrophilicity. Both these synthetic tricks have been successfully implemented. The elegant use of the pentamethylcyclopentadienyl (Cp*) ligand by Jutzi et al. allowed the isolation and characterization of the first Si(II) cation, [(η5-C5Me5)Si:]+.17 The concept of donor-stabilization has recently been used for the isolation of a series of highly reactive compounds which were otherwise inaccessible. For example, the groups of Roesky and Filippou showed that the reactivity of silylenes like SiCl2 or SiBr2 can be tamed when they are attached to an N-heterocyclic carbene.18 The N-donor stabilized amidinato chlorosilylene [PhC(NtBu)2SiCl] from the Roesky group has also enjoyed substantial attention in recent years.19 Similar N-donor stabilization afforded a unique cationic silyliumylidene ([LSi][B(C6F5)4] (L = CH[C(MeNAr)2], Ar = 2,6-iPr2-C6H3)) through protonation of the Si(II) center of the corresponding silylene, L1Si (L1 = CH[C(Me)(C = CH2)(NAr)2]).20 These results kick-started the chemistry of low-valent Si(II) cations although the existence of mono-coordinated Si(II) cations is yet to be established.
In contrast to carbon and silicon atoms, germanium, tin, and lead are more stable in their MII oxidation state as the stability increases with the increase of the principal quantum number. Going down the periodic table, the s–p separation increases with increasing nuclear charge. As a result s/p hybridization becomes more and more difficult, leading to the “inert s-pair effect”, in which only the p electrons are used in the bonding. Therefore, the isolation of LE:+ (E = Ge–Pb) is relatively simpler. To the best of our knowledge, the first mention of LE:+ (E = Ge–Pb) came from the group of Jutzi, who reported a series of nido-cluster type cations of the composition [(C5Me5)E:]+.21 Since then a number of germanium(II), tin(II), and lead(II) mono- and dications have been reported. A few of these compounds have been cited in recent review articles22,23 as well as in an excellent book by Lee and Sekiguchi, which primarily deals with “carbenium derivatives of heavier group 14 elements (R3E+)”.24 However, to our knowledge no efforts have been dedicated to compile only the cations of composition RE:+, where E is a heavier group 14 element in the +2 oxidation state. Moreover, several important accomplishments like [Cl–Si:]+, Si(II) dication, polyether ligated Ge(II) and Sn(II) mono- and dications etc. that deserve mentioning have been recently reported. However, the chemistry of cations and dications of heavier group 14 elements in low oxidation states is very far from being complete; there are still many challenges left. The practical application of these classes of compounds is still awaited. Recent studies showed that heavier carbenes have the potential to compete with transition metal complexes for single site small molecule activation, which is of significant importance in order to find cheap and green alternatives to transition metal complexes for this important class of reactions.25 Therefore, the utilization of the cations and dications of heavier group 14 elements in various catalytic transformations and small molecule activations seems to be attractive and promising. These arguments justify the requirement of a review at this juncture dedicated solely to tetrel(II) cations, which will extensively kindle further interest, directed towards its development. This present review will deal with the available synthetic routes for the preparation of cations of heavier group 14 elements in low oxidation states, starting from the compounds of the composition Cp*E+, which were the first to be prepared. We shall mainly concentrate on the cations stabilized by using bulky ligands. We shall also cover the transition metal supported EII (E = Si–Pb) cations. The literature coverage of this review is up to 2014. Because it is difficult to definitively conclude the formation of the cations by spectroscopy alone, we have chosen to mainly cover compounds which were structurally characterized by single crystal X-ray diffraction studies.
Chart 2 Selective ligands used for stabilizing cations of heavier group 14 elements in low oxidation states. |
In an effort to explore the feasibility of isolating dications of heavier group 14 elements in the +2 oxidation state, synthetic chemists moved from the formally anionic ligands and focused on neutral ligands like N-heterocyclic carbenes (NHCs). Baines and coworkers introduced NHC for the isolation of the first Ge(II) dication coordinated by three IiPr (1,3-iPr2-imidazol-2-ylidene) groups (vide infra).27 Similar to β-diketiminato ligands, NHCs also offer high synthetic flexibility by varying the “wingtip” substituents (substituent at the nitrogen atom), which makes them one of the most sought after ligands in main group chemistry.28 Another leading motif over the last few years has been the linking of two NHCs by several bridging spacers to generate discrete multitopic bis-NHCs, which exhibit high affinities toward a broad range of metals.29 Complexes featuring such ligands are stabilized by the chelate effect and offer various possibilities for tuning their geometric and electronic properties.
The enormous success of NHCs prompted the investigation of other carbon-based neutral ligands. The groups of Alcarazo and Vidović used a divalent C(0) compound, commonly known as carbodiphosphorane [(PPh3)2C], for the realization of B,30a Ge,30b and P31 centered cations. The main advantage of these C(0) ligands over NHCs is that they are capable of acting as σ- as well as π-donors. A recent computational paper by Toner and Frenking proposed that the substitution of an NHC by a carbodiphosphorane in the Grubbs’ catalyst for alkene metathesis might lead to an enhanced reactivity.32
Macrocyclic ligands like cryptands, crown ethers, azamacrocycles etc. have been noted for their remarkable metal complexation properties through numerous weak donor sites. There are many examples of cryptand and crown ether coordinated s-block and d-block elements in the literature.33 Schmidbaur et al. found that such a macrocyclic ligand (cyclophane) can also stabilize germanium and tin cations.34 However, the use of macrocyclic ligands for realizing p-block cations has extensively increased since Baines’ report of a germanium(II) dication encapsulated in [2.2.2]cryptand.35 In the following years, a range of macrocyclic ligands like crown ethers, cryptands, azamacrocycles etc. were used to give rise to an array of cations and dications of germanium(II) and tin(II), which revealed an interrelation between the cavity size of the macrocycles and the stability of the E(II) dication or the [EX]+ (E = Ge and Sn) monocation (vide infra section 4.1.4).
It is only recently that the various research groups have started to exploit the coordination sphere of the electron rich coordinatively unsaturated transition metals such as W, Pt, Pd to stabilize germanium(II), tin(II), and lead(II) cations (vide infra). The precedent for this approach came from the isolation of an array of complexes featuring BE (E = N and O),36 BC,37 SiMo,38 GeRe,39 SiO40 bonds, which had thus far been otherwise inaccessible. However, interpretation of the bonding situation in transition metal supported tetrel(II) cationic complexes is not very straightforward taking into consideration other possible canonical forms.
Following the synthesis of the [Cp*Si]+ cation, Jutzi, Scheschkewitz, and others reported many reactions taking advantage of 1 as a starting material, which are outlined in Scheme 4. The reaction of 1 with LiN(SiMe3)2 afforded silylene (Me5C5)SiN(SiMe3)2 (2) as a reactive intermediate, which dimerizes to give the disilene E-{(η1-Me5C5)[N(SiMe3)2]Si}2 (3).17 Later Jutzi et al. found that a rare reversible phase dependent dynamic equilibrium exists between silylene 2 and disilene 3.46 During crystallization colorless silylene 2 dimerized to yellow colored disilene 3. Again when 3 was dissolved in solvents, a colorless solution of 2 was obtained. This unusual behavior could be rationalized by steric strain in the solid disilene, flexibility in bonding modes of Cp* groups, low activation energy for the equilibrium process, and a small energy difference between 2 and 3. It should be noted here that dehydrochlorination of [(η1-Me5C5)SiHCl2] with KN(SiMe3)2 also led to disilene 3 in a higher yield.471 was further reacted with various lithium containing anions like Li(2,6-Tipp2-C6H3) (Tipp = 2,4,6-iPr3-C6H2) and Li[NC{N(Dipp)CH}2] (Dipp = 2,6-iPr2-C6H3) to obtain [(η3-C5Me5)(2,6-Tipp2-C6H3)]Si: (4) and [(C5Me5)(NC{N(Dipp)CH}2)]Si: (5), respectively, in salt elimination reactions.48 The former features a silylene with one σ-donor and one π-donor substituent attached to the Si(II) center. In contrast, the reaction of 1 with Li[HC(CMeNDipp)2] did not result in the putative Cp*[HC(CMeNDipp)2]Si: but a constitutional isomer 6.491 was also utilized as a stoichiometric source of silicon in the reaction with Na[Fe(η5-C5Me5)(CO)2], which gave rise to a ferrio-substituted silylene [Fe(η5-C5Me5)(CO)2{Si(η3-C5Me5)}] (7) at low temperatures (Scheme 5).50 However, under ambient conditions, the silylene fragment was inserted into one of the C–H bonds of the Cp* ring leading to a rearranged product 7′. The substitution reaction with the lithium disilenide [Tipp2SiSi(Tipp)(Li{dme}2)] led to straightforward access to the first cyclotrisilene with only carbon-based substituents (8) (Scheme 4).51
The study of compounds featuring low valent main group elements continues to be a worthwhile subject due to their anticipated application in metal free catalysis. Recently, 1 has been found to catalytically convert 1,2-dimethoxyethane (DME) to 1,4-dioxane and dimethyl ether (Scheme 6), which is a rare example of a metal free catalytic transformation.52Ab initio calculation suggests that the O→Si dative bond in the DME→1 complex is electrostatic in nature and the subsequent enhancement of positive charge at the silicon center facilitates the attack of another equivalent of DME. This catalytic process is found to be useful for a range of oligo(ethylene glycol)diethers, leading to 1,4-dioxane and dimethyl ether in each case.
Presented in Table 1 are selected NMR data for the related silicon(II) compounds. In general, more cationic charge density on the silicon center results in a more downfield shifted 29Si resonance. However, both 10 (δ = 40.5 ppm) and 11 (δ = 69.3 ppm) exhibit upfield shifted resonances compared to that of 9 (δ = 88.4 ppm) although the Si centers in 10 and 11 are more Lewis acidic in nature. This was presumably a consequence of a pπ–pπ interaction between the π-electron system of the β-diketiminate ligand and the Si atom. Moreover, the γ-H resonances of 10 (δ = 6.79 ppm) and 11 (δ = 6.92 ppm) also indicated the existence of aromatic ring current in the systems. X-ray studies on 10 and 11 showed that the six-membered SiC3N2 rings were planar and the endocyclic N–C bond lengths were shortened by 0.05 Å compared to those in 9. Such features were in accord with the aromatic nature of these systems, which was further confirmed by nuclear independent chemical shift (NICS) calculations (NICS (1): −3.9 ppm).
Compound | Solvent | 29Si NMR (ppm) | Ref. |
---|---|---|---|
1 | CD2Cl2 | −400.2 | 17 |
10 | THF-d8 | 40.5 | 20 |
11 | CD2Cl2 | 69.3 | 20 |
12 | CD2Cl2 | −3.3 | 54 |
13 | CD2Cl2 | −58.4 | 55 |
14 | CD3CN | −89.9 | 56 |
15 | CD3CN | −55.3 | 56 |
16a | C6D6 | 8.3 (d, 1JP–Si 7.4 Hz; SiMe3), 234.5 (dt, 2JP–Si 219.4 Hz, 48.5 Hz, SiPt2) | 59 |
16b | THF-d8 | 0.01 (d, 1JP–Si 6.0 Hz, SiMe3), 187.8 (dt, 2JP–Si 43.4 Hz, 184.3 Hz, SiPd2) | 59 |
A chelating bis(iminophosphorane)ligand has been recently used by Driess et al. to obtain a chlorosilyliumylidene complex 12 through ligand exchange of the SiCl2 unit.54 NHC·SiCl218a was used as a stoichiometric source of SiCl2 and its reaction with the bis(iminophosphorane) ligand resulted in the ion pair with concomitant liberation of one equivalent of NHC (Scheme 8). Utilizing the same synthetic strategy, Driess et al. prepared another remarkable chlorosilyliumylidene derivative 13 by the reaction of the bidentate bis-NHC ligand with NHC·SiCl2 (Scheme 9).55 The 29Si NMR resonances due to the three-coordinate Si nuclei in 12+ and 13+ were observed at δ = −3.3 and −58.4 ppm, respectively. The upfield shift can also be attributed to the stronger electron donation from bis-NHC and iminophosphorane ligands. Alternatively, such shifts to a higher field may also indicate the decrease of the cationic character. Inspection of the frontier MOs revealed that the HOMO in 12+ is associated with the ten π-electrons from the naphthyl moiety and the nitrogen lone pair of the P–N ylide bonds and the lone pair on the silicon atom is depicted by the HOMO−1 orbital. This is in contrast to 13+ where the lone–pair of silicon constitutes the HOMO and the difference may be attributed to the stabilization of the Si lone pair by the π-system in 12+.
Unlike NHC stabilized SiCl2 and SiBr2, which were reported back in 2009,18 the synthesis of the first stable NHC stabilized diiodosilylene (IDipp·SiI2)56 (IDipp = 1,3-(2,6-iPr2-C6H3)2-imidazol-2-ylidene) has recently been accomplished by following the same synthetic protocol used for the isolation of IDipp·SiBr2.18b Following this, Filippou et al. have found that the reaction of IDipp·SiI2 with IiPr2Me2 (1,3-iPr2-4,5-Me2-imidazol-2-ylidene) resulted in the displacement of one iodide ligand leading to [(IDipp)(IiPr2Me2)·SiI]+I− (14) (Scheme 10), the first formal derivative of [I–Si:]+. The structure of 14 revealed a fixed orientation of one isopropyl C–H group pointing towards the Si(II) atom to maximize a Si⋯H–C anagostic interaction, as also supported by the 29Si–1H coupling constant (J = 10.4 Hz) and subsequent DFT calculations.
Addition of a less bulky NHC (IMe4 = 1,3,4,5-Me4-imidazol-2-ylidene) to IDipp·SiI2 led to the first Si(II) dication [(IMe4)3→Si:]2+I2− (15) (Scheme 11).56 The removal of the iodide anions from the Si center can be attributed to the steric bulk exerted by three NHCs. The Si center adopts a pyramidal propeller like conformation (Fig. 1), akin to the analogous Ge(II) dication (492+) published before by Baines and coworkers (vide infra).27 However, interpretation of the ionicity of 152+ as well as other donor stabilized Si(II) cations (12+–14+) is complicated by the fact that such stabilization takes place at the cost of the cationic character of the Si atom. Three possible canonical forms of 12+–14+ are outlined in Scheme 12. Two of the resonating structures oppose and diminish the cationic nature of the Si atom and indicate that the positive charge is distributed over the ligands. This was further manifested in the appearance of the 13C NMR signals of the carbene C in the relatively high field in these cations (13: δ = 161.6; 14: δ = 151.5 and 158.3; 15: δ = 150.7 ppm), which are closer to that of the imidazolium salt (IMe4H)Cl (δ = 136.9 ppm) than that of IMe4 (δ = 212.7 ppm). However, various charge calculations indicate the accumulation of positive charge on silicon atom. Therefore, perhaps it is safe to comment that all three resonating structures contribute to the overall bonding of these cations.
Scheme 14 Ge(II) and Sn(II) cations stabilized by pentamethylcyclopentadienyl, cyclophane, and pyrazole ligands. |
Reger and Coan reported the synthesis and structural elucidation of [HB(3,5-Me2pz)3Ge]+ (20+) with an iodide counter ion (Scheme 14).61 The shortest Ge⋯I distance is over 4 Å, clearly indicating the lack of a substantial covalent interaction between these atoms. The Ge(II) center adopted a pyramidal geometry with three neighboring nitrogen atoms coordinated to the Ge center. This result kick-started the use of N-donor ligands for the isolation of p-block cations. Following this, Dias et al. used bi-dentate, monoanionic amino-troponiminate (ATI) with a distinct 10π electron backbone for the isolation of Ge(II) and Sn(II) monocations. Substitution of the corresponding chloro precursor [(iPr2ATI)GeCl] with AgOSO2CF3 resulted in [(iPr2ATI)Ge][OSO2CF3] (21) (Scheme 15).62a The germanium atom was weakly bound to the oxygen atom of the triflate anion (Ge–O: 2.255(2) Å). In order to prepare a “free” Ge(II) cation, (iPr2ATI)GeCl was reacted with (η5-C5H5)ZrCl3, which serves as a chloride scavenger to result in [(iPr2ATI)Ge][(η5-C5H5)ZrCl2(μ-Cl)3ZrCl2(η5-C5H5)] (22a).62a A related Sn(II) cation, [(iPr2ATI)Sn][(η5-C5H5)ZrCl2(μ-Cl)3ZrCl2(η5-C5H5)] (22b) was earlier reported by the same group.62b However, the weak interactions between the cations and the terminal chloride atoms were found in the solid state structures of 22a,b.
Following the synthesis of germylenemonochloride [LGeCl] [L = CH(CMeNAr)2; Ar = 2,6-iPr2C6H3] using the versatile mono-anionic bidentate β-diketiminate ligands,63 Power's group prepared a cyclic Ge(II) mono-cation [LGe][(HO)B(C6F5)3] (23) by reacting LGeCl and B(C6F5)3 in the presence of water (Scheme 16).64 The geometrical parameters of 23+ such as the decrease of C–N bond lengths (1.34av Å) with the concomitant increase of the Ge–N bond length (1.91av Å) and NICS(1) values (−2.4 ppm) were evocative of those observed for its silicon analogue 11, indicating the presence of a slight aromatic character in 23+.65 However, one must note here that the γ-H signal of 23+ appeared at a significantly higher field (δ = 4.23 ppm), than that of 11 (δ = 6.92 ppm).
The tin versions of 23+ with the [AlCl4]− (24), [B(C6F5)4]− (25), and [B(C6F5)3(Me)]− (26) counter-anions were later reported by Fulton and coworkers (Scheme 16).66 The 119Sn NMR chemical shifts for these cations are −626.7, 197.0, and −139.5 ppm, respectively (Table 2). Among them, only 25 displays a low-field 119Sn NMR shift from that of its precursor (−224 ppm). Unfortunately, 24 and 25 were not characterized by X-ray crystallography. The single crystals of 26 were only grown from a diethyl ether solution and in the solid-state structure of 26 one ether molecule was bound to the Sn atom, which explained the high-field chemical shift in the 119Sn NMR.
Compound | Solvent | 119Sn NMR (ppm) | Ref. |
---|---|---|---|
18 | — | 21a | |
19 | — | 21a | |
22b | CDCl3 | 734 | 62b |
24 | CD2Cl2 | −626.7 | 66 |
25 | CD2Cl2 | 197.0 | 66 |
26 | CD2Cl2 | −139.5 | 66 |
32 | CD2Cl2 | 46.3 (solid-state: 68) | 70 |
34 | CD2Cl2 | −30 | 70 |
39 | CD2Cl2 | 249.71 | 30b |
40 | CD2Cl2 | 17.28(br) | 30b |
44 | THF-d8 | −60.27 (SnCl3−) & −435.07 | 75 |
46 | CDCl3 | −73.2 (SnCl3−) and −330.4 | 76 |
54 | CD3CN | −1468 | 80 |
64a–c | Solid-state: −980 (Cl), −920 (Br), −810 (I) | 86b | |
65 | Solid-state: −1533 | 86b | |
66 | Solid-state: −1436 | 86a | |
67 | Solid-state: −1457 | 86a | |
68 | Solid-state: −840, −58 | 86a | |
69 | Solid-state: −1578 | 86a | |
70 | Solid-state: −1405 | 86a | |
71 | Solid-state: −1721, −1706 | 86a | |
72 | Not mentioned | 87 | |
74 | THF-d8 | −8.4 | 89a |
75 | THF-d8 | 224.0 | 89b |
76 | THF-d8 | 249.9 | 89b |
77 | CDCl3 | 70.6 | 90 |
78 & 79 | Not observed | 92 |
Mochida and coworkers introduced a chiral ligand, (1,1-bis[(4S)-4-phenyl-1,3-oxazolin-2-yl]ethane), popularly known as ({(S)-box-Ph}H) to isolate a germanium(II) cation.67 The motivation for this work presumably came from the use of the enantiomeric carbene as a ligand for metal complexes to catalyze asymmetric reactions.68 Abstraction of the chloride atom from the DME solution of the corresponding germylene with Ag[SbF6] in the presence of excess THF afforded the Ge(II) cation, [Ge((S)-box-Ph)(THF)](SbF6) (27) (Scheme 17). Substitution of the THF ligand with other Lewis bases like PMe2Ph and pyridine yielded [Ge((S)-box-Ph)(PMe2Ph)](SbF6) (28) and [Ge((S)-box-Ph)(py)](SbF6) (29), respectively. Inspection of the Ge–N bond lengths in these cations revealed that the average bond length increases on going from 27 to 29 [27: 1.928 (2) Å, 28: 1.933(3) Å, and 29: 1.939(4) Å]. This lengthening was assumed to be due to the increased D→Ge π donation from 27 to 29, resulting in less σ as well as π interactions, which in turn increases the Ge–N bond lengths.
By adopting the same synthetic protocol that has been used for the synthesis of 11, Müller et al. isolated two more Ge(II) cations, 30 and 31 upon protonating 1,3-di(2,6-iPr2C6H3)-germaimidazol-2-ylidene with [H(OEt2)2]+[B(C6F5)4]− and [Et3Si(C6H6)]+[B(C6F5)4]− (Scheme 18).69 The protonation took place at the C-4 position instead of germanium and as a result the ligand underwent a shift from di-anionic to mono-anionic by rupturing the symmetry of the GeN2 moiety. The molecular structure of 30 revealed that one ether molecule is strongly coordinated to the Ge atom. Although exhibiting no interaction with the benzene solvent, 31 was weakly bound to the borate anion through the fluoride atoms. A detailed theoretical calculation was carried out by Müller et al. to understand why the protonation took place in the C-4 position instead of Ge. It was observed that the C-protonated product was more stable than the hypothetical Ge-protonated product by 68.3 kJ mol−1 (B3LYP/6-311G) and 44.8 kJ mol−1 (MP2), respectively. Another underlying factor responsible for this regioselective protonation is the formation of a C–H bond instead of a highly polarized Ge–H bond.
Scheme 19 Quasi-monocoordinate Ge and Sn cations, 32+ and 33+ and their adduct formation with 4-DMAP. |
The electrophilic nature of 32+ and 33+ was observed upon addition of the N-donor ligand, 4-DMAP, which resulted in 4-DMAP coordinated Sn(II) (34+) and Ge(II) (35+) cations. The 4-DMAP coordination led to the displacement of C⋯E interactions in the latter, which is reflected in the longer C⋯E distances and higher 119Sn NMR resonance (δ = −30 ppm).
Interestingly, the P–C–Ge–Cl torsion angle in 36+ (Fig. 2) was only 8.18°, a feature that facilitated good overlap between the filled p orbital of C and the vacant orbital of Ge, leading to the shortening of the C–Ge distance (1.954(2) Å) than that in (Ph3P)2C·GeCl2 (2.063(2) Å). The WBI of the C–Ge bond in 36 (0.84) was considerably higher than that in (Ph3P)2C·GeCl2 (0.54), revealing further a significant strong π-donation from the central carbon of (Ph3P)2C to the germanium atom. The HOMO of 36+ represents mainly the σ-lone-pair while the HOMO−1 corresponds to a C–Ge π-bonding (Fig. 3). The LUMO of 36+ which was associated with a π* (C–Ge) orbital became populated upon addition of 4-DMAP leading to an increase of the C–Ge bond length (2.053(3) Å) and WBI value (0.64) in the corresponding Lewis adduct, 4-DMAP·[(Ph3P)2C(GeCl)]+ (38+).
Fig. 3 Frontier orbitals of 36+ (left: HOMO, right: HOMO−1). Adapted from ref. 30b. |
In contrast, the reaction of (Ph3P)2C·SnCl2 with an equivalent of AlCl3 did not yield the anticipated monomeric [(Ph3P)2C(SnCl)]+ cation and instead led to the formation of a chlorine bridged dimeric [(Ph3P)2C(SnCl)]22+ (39+) species (Scheme 21).30b This result indicates that the π(C→Sn) bond in 39+ is very weak, presumably due to a less-efficient overlap between the C(2p) and Sn(5p) orbitals. The dimeric structure was stabilized by the formation of chloride bridges between the Sn atoms to provide the extra electron density to the Sn center. The reaction of 4-DMAP with 39 resulted in a monomeric 4-DMAP·[(Ph3P)2C(SnCl)]+ adduct (40+).
The germanium variants of 12 and 13, were reported by Driess and coworkers.77 Apparently, the reaction of GeCl2 with iminophosphorane and bis-NHC ligands led to the cleavage of one of the Ge–Cl bonds and thus afforded 4777a and 48 (Scheme 24).77b The Ge centers were well separated from the chloride anions (47: 6.83 Å and 48: 6.53 Å). The geometric features and the frontier molecular orbitals of 47+ and 48+ were almost identical to those mentioned earlier for their corresponding silicon analogs. The dative nature of the N→Ge and C→Ge bonds was further apparent from their respective WBI values (47+: 0.436 and 0.430; 48+: 0.608 and 0.611).
Fig. 5 Germanium κ-edge XANES of 552+ (green line) and 492+ (red line). Adapted from ref. 78. |
Another elegant variation of the NHCs is the linking of two NHCs using a borate spacer, which gives the ligand a certain degree of flexibility. This ligand has been recently used for the isolation of a compound featuring a Ge–Ge bond where one Ge atom is monocationic and another Ge atom is dicationic.79 The preparation of 53 is unique (the reaction sequence is shown in Scheme 26) and represents a tour de force in organometallic synthesis. A key step in this sequence was the generation of the zwitter-ionic Ge(II) cation (50) from a salt metathesis route using potassium bis(NHC)-borate and GeCl2·dioxane. The isolated Ge(II) cation was further derivatized to yield the zwitterionic Ge(II) cation with a [H–Ge:]+ moiety, 51, which upon reaction with [Ph3C]+[B(C6F5)4]− generated 52. Spontaneous elimination of the Ph3CH from the latter apparently led to a highly reactive Ge(II) dication, which underwent donor–acceptor stabilization with unreacted 51 to form an unprecedented compound with a [HGe:+→Ge:2+] motif (53). A deliberate synthetic route that involved the 1:1 reaction of 51 and 52 also led to compound 53 with concomitant elimination of Ph3CH.
Müller and coworkers obtained a borate salt of tris toluene Sn(II) dication [Sn(C7H8)3]2+ (542+) from the serendipitous decomposition of a stannylium cation during its recrystallization in toluene.80 It was reported that the reaction of Tipp2Sn (Tipp = 2,4,6-iPr3-C6H2) with the silylarenium ion [iPr3Si(tol)]+[B(C6F5)4]− presumably generated the corresponding stannylium ion (Tipp2Sn(SiiPr3))+ initially, which in due course decomposed to furnish [Sn(C7H8)3][B(C6F5)4]2 (Scheme 27). No deliberate synthetic route was reported for 542+. A single crystal X-ray study of 542+ revealed that two toluene molecules were significantly closer to the Sn atom than the remaining third toluene molecule leading to an unsymmetrical coordination around the Sn atom. The large isomeric shift (4.14(1) mm s−1) confirmed the +2 oxidation state of the Sn center and indicated that the Sn(II) valence electrons reside in orbitals that are almost exclusively of 5s character. Computations showed significant charge transfer from the arene ring to the empty p orbitals of the Sn atom leading to accumulation of high positive charge (+1.28) on the Sn atom.
In a subsequent systematic study it was demonstrated that reacting differently sized crown ethers, such as [12]-crown-4, [15]-crown-5, and [18]-crown-6 with GeCl2·dioxane led to a range of Ge(II) mono- and dications (56–61) (Scheme 29).83 The structural properties of these cations were governed by the size of the crown ether employed. Ge2+ fits into the cavity of [15]crown-5 and [18]crown-6; while it formed a sandwich complex with two [12]crown-4 ligands. The structural properties can also be influenced by the substituent on the Ge atom. [15]-Crown-5 adopted a folded structure with the [GeCl]+ fragment and a planar conformation with [GeOTf]+. Related compounds with tacn (1,4,7-triazacyclononane) and cyclam (1,4,8,11-tetramethyl-1,4,8,11-tetraazacyclotetradecane) ligands have also been reported.84 Like in the cases of crown ethers, the size and the denticity of the azamacrocycles strongly influence the coordination geometry of the Ge2+ dications and the nature of the counter anion.
The reactivity of the Ge(II) crown ether cation towards H2O and NH3 has been recently demonstrated. Instead of oxidative addition, the formation of stable adducts like [([15]-crown-5) Ge·OH2][OTf]2 (62) and [([15]-crown-5)Ge·NH3][OTf]2 (63) was observed (Scheme 30).85 Subsequent computational studies showed that the purported oxidative addition products of water and ammonia are less stable than the adducts by 13 kJ mol−1 and 57 kJ mol−1, respectively, which can be attributed to the preferential formation of the O–H bond rather than the Ge–H bond. The hydrogen abstraction from 62 with a suitable base like pyridine, ammonia, NHC etc. was found to be feasible, as the formation of the corresponding conjugate acids was detected in the 1H NMR spectra. Such dehydrogenation would lead to multiply bound Ge–X [X = O or N] bonds, the formation of which was computationally supported with the increase of the electron density of the putative dehydrogenated variants at the bond critical points.
The groups of Baines and Macdonald recently published the cryptand and crown ether complexes of tin (Scheme 31).86 Unlike germanium, tin forms both mono-[cryptand[2.2.2]SnX][SnX3] (X = Cl; 64a, X = Br; 64b, X = I; 64c) and dicationic compounds [cryptand[2.2.2]Sn][OTf]2 (65) with [2.2.2]-cryptand depending on the Sn(II) precursor used in the reaction. The more flexible glyme ligands are also found to be suitable in isolating Sn(II) cations, as manifested in the isolation of [triglymeSn(OTf)]+ (66+) and [tetraglymeSn]2+ (672+) with triflate as the counter-anion. [18]-Crown-6 tends to form 1:1 adducts with Sn(II) precursors leading to 68 and 69, whereas the smaller crown ethers, [15]-crown-5 and [12]-crown-4, were not able to accommodate the tin atom within the crown ether cavity, and thereby resulted in 2:1 sandwich complexes with tin (70 and 71).
It was observed that crown ethers and glymes form Sn(II) dications with triflate precursors whereas [Sn–X]+ derivatives with the halide precursors. The difference in the reaction pattern can be attributed to the greater s-character of the Sn valence electrons in triflates than in halides. 119Sn solid state NMR data revealed that 119Sn nuclei are highly shielded in the triflate salt whereas considerably deshielded in the chloride salt, and thus support the formation of the resulting complexes. The +2 oxidation state of Sn(II) atoms in these polyether ligated cations was further reflected in Mössbauer spectroscopy, which revealed the pure 5s lone pair as the Sn valence orbital. The ligation of [15]crown-5 and [12]crown-4 to the Sn atom hardly perturbed its electronic configuration owing to the almost symmetrical Sn bonding environment in 70 and 71. In contrast, the coordination geometry around the Sn(II) centers in the [18]crown-6, triglyme, and tetraglyme complexes of Sn(II) is less symmetrical, leading to more perturbation in the valence electron as indicated by the increase of the quadrupolar splitting of the resonance in the respective Mössbauer spectrum. The effect of the poly-ether ligand and the substituent on Sn was also illustrated through cyclic voltammetry, where complexes in which the valence electrons of tin have more s-character need more energy to become oxidized. Besides, it also depends on the steric properties of the ligands around the tin atom as cyclic voltammetry data of 70 and 71 show no oxidation current.86a
The oxidative addition of the Sn–Cl bond in 2,6-Mes2-C6H3SnCl to trans-[dppe]2W(N2)2, followed by salt elimination afforded the cation, 72 (Scheme 32). Replacing the 2,6-Mes2-C6H3 substituent by a Cp* moiety, the same group isolated the germanium variant, 73.88 Both Sn and Ge atoms adopt a nearly linear geometry with WE triple bond distances of 2.4641(7) and 2.303(1) Å, respectively. Subsequent theoretical calculations suggested that the π-back donation from W to germanium/tin may indeed contribute to the stability of these cations. One may raise the question here that the formal oxidation state of Sn in 72 and 73 is not +2, so they are not “true” germylium and stannylium ylidenes. Although we do not completely disagree with this tenet, but such a classification for transition metal supported p-block cations is always questionable taking into account other resonance contributors manifesting a germylium or stannylium character. Therefore, for the sake of completion we incorporated these cations in the review.
The metathetical reaction between bulky pincer based ligands with (THF)W(CO)5·SnCl2 or (THF)Cr(CO)5·SnCl2, followed by the abstraction of the halides resulted in several Sn(II) cations, such as [2,6-(Me2-NCH2)2C6H3(H2O)Sn{W(CO)5}]+ (74+), [{(2,6-MeOCH2)2C6H3}Sn(OH2){Cr(CO)5}]+ (75+) and [{(2,6-MeOCH2)2C6H3}Sn(THF)2{Cr(CO)5}]+ (76+) with various counter anions (Scheme 33).89 Another pincer based stannylene, 77 has recently been reported by Jambor et al. featuring a Sn–Pt bond.90 Although one might be tempted to assign the tin atom the +2 oxidation state, NBO analysis revealed that there was no lone pair at the tin atom. A plausible explanation is that the lone pair was shared by both Sn and Pt atoms and used up for the formation of a Sn–Pt bond, leading to some, by no means complete, stannylium ion character. Low-field 119Sn NMR (70.6 ppm) resonance indicated accumulation of negative charge on the Pt atom. DFT calculations on this complex indicated that there is significant charge donation from the [N→Sn]+ fragment to the [Pt(pyt)2Cl]− moiety. This finding was in accord with the NPA analysis which disclosed a significant fraction of positive charge at the Sn atom (+1.709) and negative charge on the Pt (−0.486). The decrease of the Sn–Pt bond length (2.466(1) Å) with respect to the other reported Sn–Pt bond lengths (vide infra) reflects an increased contribution of backbonding from the metal.
Very recently, Braunschweig et al. showed the latent σ-donor ability of the Pt(PCy3)2 to coordinate with Ge(II), Sn(II), and Pb(II) dihalides, which resulted in the formation of the respective Metal Only Lewis Pairs (MOLPs).91 The mono stannylene complex, [(Cy3P)2Pt–SnBr2] was found to be an excellent starting material for realizing the low-coordinate Sn(II) mono- and dications via halide abstraction reaction, as manifested by the formation of [(Cy3P)2Pt–SnBr]22+ (78+) and [(Cy3P)2Pt(Sn)] (792+).92 The Sn–Pt bond lengths in these cations [78+: 2.524(1) and 792+: 2.502(1) Å] are significantly shorter than the Sn–Pt bond present in their precursor complex [(Cy3P)2Pt–SnBr2] (2.605(2) Å) (Scheme 34). Although exhibiting no interaction with the solvent molecule, these cations were not totally free, being weakly bound to the counter anion through their bromide atoms. No 119Sn NMR signals were detected presumably due to the broadening of resonance to an undetectable level caused by the increased electric field gradient and large chemical shift anisotropies induced by the tin environment.
The next milestone contribution to the low coordinate Pb(II) cations was achieved by Power et al., who synthesized [Ar*Pb·η2-toluene][MeB(C6F5)3] (Ar* = 2,6-(2,4,6-iPr3C6H2)2-C6H3) (81) (Scheme 36) from the reaction of Ar*PbMe with B(C6F5)3 in toluene.94 The low coordination of the Pb center in 81 was manifested in its 207Pb NMR resonances observed at δ = 8974 ppm, shifted 13500 ppm downfield of the [(η5-C5Me5)Pb]+ salts. This very large difference is consistent with a lower effective coordination number of Pb in 81+ and weak interactions with toluene. The solid-state structure of 81 was determined by single crystal X-ray analysis (Fig. 7). The analysis revealed that there are no close interactions (3.963(6) Å) between the lead atom and the anion in 81. However, the Pb center interacted with the solvent toluene and the Pb–toluene interaction can be best described as being of η2 type. The weakly coordinating toluene molecule can be easily displaced by two pyridine molecules to yield [Ar*Pb(py)2][MeB(C6F5)3] (82). The upfield shift of the 207Pb NMR resonance (δ = 4764 ppm) of 82 with respect to that of 81 (δ = 8974 ppm) and the pyramidalization at the Pb(II) cationic center clearly reflected the decrease of the cationic character in line with the increase of the coordination number of the Pb atom.
A series of tetrel(II) cations supported by the β-diketiminato ligand has recently been completed by Fulton et al. with the isolation of lead(II) cations (83+ and 84+) with the B(C6F5)4 and MeB(C6F5)3 counter-anions (Scheme 37). The Pb center of these compounds was well separated from its anion (the closest approach from Pb to the nearest F atom was 3.319 (4) Å), reminiscing the structural properties of their tin homologue. However, in the solid-state structure of 83+ a dichloromethane molecule was found with 64% occupancy to give a long-range lead–chlorine interaction (Pb–Cl(1) = 3.213(4) Å). Calculation points to the increased stability of the solvated complexes LPb+·CH2Cl2 over LPb+ by 3 kcal mol−1 (L = CH(CMeNDipp)2). No 207Pb NMR resonances for 83 and 84 were observed, which according to the authors, may be attributed to the fast relaxation of the lead nucleus.
The success of stabilizing Sn(II) mono- and dications in the coordination sphere of [Pt(PCy3)2] led Braunschweig and co-workers to isolate the lead variants. Adopting the same synthetic protocol that was employed for 78+ and 792+, they reported two Pb(II) mono-cations with AlCl4 (85a) and [BArCl4] [ArCl = 3,5-Cl2-C6H3] (85b) and a rare example of Pb(II) dication [{(Cy3P)2Pt(Pb)}{AlCl4}2 (86)] (Scheme 34, vide supra).92 The Pb(II) cations and dications were structurally similar to the analogous Sn(II) cations upon comparison of bonding motifs and metrical parameters. The Pb centers in 85a and 85b revealed weak bonding interactions (3.420(1) Å) to the chloride ion in the counter-anions. In the solid-state form, 86 also exhibited a weak interaction of the Pb atom with the two chloride atoms of two aluminate counter-anions. Detailed theoretical investigations were additionally carried out to further characterize the bonding situation in these complexes. It was proposed before that the Pt(0) and Pb(II) fragments in the starting material (Pt(PCy3)2→PbCl2) donate σ-electron density to each other, leading to the concept of “synergic σ-donation”,91b which was further proposed by Su et al. in the transition metal complexes featuring multiple bonds between group 10 and group 14 elements.95 A recent EDA-NOCV study by Braunschweig et al. revealed that the Pt–Pb bond in (Pt(PCy3)2→PbCl2) is a Pt→Pb dative bond whereas in 85 and 86, it is an electron sharing bond. So, the real bonding phenomena of these complexes are not very well defined and clearly require further theoretical investigation.
The selection of ligands has turned out to be often decisive in the synthesis of these compounds because the stabilization requires the usage of sterically encumbering ligands with donor substituents. Therefore, the design of many more new ligands for stabilizing low oxidation species continues to be a central theme in this area and will remain so for many more years. For example, the emergence of carbodiphosphorane, which is a simultaneous σ- and π-donor and its utilization for realizing two coordinate [Ge–Cl]+ (36+) will rival NHC for the stabilization of low coordinate group 14 cations in the coming years. In a similar way, bis-carbene has already obtained a foothold in silicon (13+) and germanium (48+ and 53) chemistry and many more such fascinating compounds supported by bis-carbene ligands are expected. The use of macrocycles to protect and stabilize germanium(II) and tin(II) cations and dications rendered an alternative route to the stabilization of low valent species.
Bearing in mind the infancy of cations of heavier group 14 elements in low oxidation states, there are plenty of objectives to achieve and problems to solve to develop the fundamental aspects of this field. For example, a monocoordinate silicon(II) cation, which is truly a higher homologue of HSi+ is elusive. The scenario is the same for other heavier group 14 elements. Similarly, a Si(II) dication encapsulated in crown ethers or cryptands is yet to be realized and remains as a sought after goal. One would also notice that unlike silicon and germanium, no monomeric tin(II) cation was reported with neutral ligands like NHC or carbodiphosphorane which is due to the poor overlap between 2p (C) and 5p (Sn). The examples of Pb(II) cations and dications are still very scant and isolation of many more Pb(II) cations is highly desirable.
The reactivity of cations of low valent heavier group 14 elements has not been very well studied but the seminal [Cp*Si]+ (1) was found to be a potential workhorse for a wider range of transformations. In fact, 1 led to novel neutral silicon(II) compounds as well as a cyclotrisilene simply by the addition of appropriate anionic nucleophiles. Most of the other cations were only reacted with some Lewis bases like 4-DMAP to establish the Lewis acidic nature of the cations. However, Driess et al. recently demonstrated that bis-NHC stabilized Si(II) cation, 13 and Ge(II) cations, 48 are excellent precursors to give rise to the hitherto elusive cyclic silylone55 and germylone,77b (heavier analogues of carbodiphosphorane) respectively (Scheme 38). These results ensure that the search for new chemistry involving heavier group 14 elements will not abate anytime soon. However, more systematic studies of the reactivities of such cations are highly desirable. The main target in this chemistry is metal-free catalysis which is yet to be realized. Nevertheless, catalytic conversion of 1,2-dimethoxyethane to 1,4-dioxane by 1 is a very promising step towards metal free catalysis and it is only a matter of time before this goal will be accomplished. Besides, Rausch et al. reported [(η5-Me5C5)Sn]+[B(C6F5)4]− to be an effective co-catalyst for Ziegler–Natta olefin polymerization, which manifests that low-valent group 14 cations are not only of academic interest but also have potential for applications in industrial processes.
Footnote |
† This review is dedicated to Prof. Herbert W. Roesky on the occasion of his 80th birthday. |
This journal is © The Royal Society of Chemistry 2015 |