Article pubs.acs.org/JACS
A Stable Silylene with a σ2, π- Butadiene Ligand Zhaowen Dong, Crispin R. W. Reinhold, Marc Schmidtmann, and Thomas Müller* Institute of Chemistry, Carl von Ossietzky University of Oldenburg, Carl von Ossietzky-Str. 9-11, D-26129 Oldenburg, Federal Republic of Germany, European Union S Supporting Information *
ABSTRACT: The synthesis of a new type of silylene 1 is reported. It adopts a bicyclo[2.1.1]hexene structure in which a hafnocene group is incorporated. The silylene is stabilized by homoconjugation with the remote CC double bond. This is indicated by its highly shielded 29Si NMR chemical shift (δ29Si = −155) and is firmly established by its experimental molecular structure from XRD analysis. The results of a detailed bonding analysis based on DFT calculations suggest for model compounds of silylene 1 and for its heavier germanium, tin, and lead homologues uniformly electronic structures of carbene analogues that are stabilized by homoconjugation. This stabilization mode is equivalent to a σ2, π-coordination of the butadiene ligand to the element atom as it is typical for zirconocene or hafnocene butadiene complexes.
1. INTRODUCTION One of the mainsprings in contemporary main group chemistry is the possibility to use cheap and prevalent main group elements in bond activation reactions as mimic and, finally, as replacement of more conventional transition metal based compounds.1 The prerequisite for this reactivity is the presence of donor and acceptor type orbitals in a defined spatial arrangement and with a small energetic separation. These requirements are met for example in frustrated Lewis pairs (FLPs) and in some multiple bonded systems involving heavy main group elements.2 Also carbenes and their heavier analogues are perfect candidates, due to the existence of small HOMO/LUMO gaps. The challenge is however to tame the resulting inherent amphiphilic reactivity while maintaining the potential for bond activation reactions.3 Recently, we described the synthesis of a new type of germylene 2 that is stabilized by homoconjugation with the remote C2C3 double bond (Chart 1).4 Preliminary studies disclosed a mostly nucleophilic reactivity for germylene 2. As important key properties, such as HOMO energy and the energy difference between singlet and triplet states ΔE(ST) of carbene analogues, strongly depend on the dicoordinated element,1b,c,5 we extended our methodology to the synthesis of the higher congener, silylene 1 (Chart 1). Parallel to our work on germylene 2, Saito and co-workers disclosed the preparation and characterization of tin compounds 3 and 3(Cl).6 According to their analysis, these compounds are transition metal-like complexes between Sn(0) atoms and neutral butadiene. This bonding description is substantially different from our model of germylene 2, a carbene analogue with the germanium atom in the formal oxidation state II. The controversy has its roots in the well-established dichotomy of the butadiene ligand in transition metal chemistry, which on the one hand, allows the formation of metallacyclopentene © 2017 American Chemical Society
Chart 1. Silylene 1 and Related Compounds (Cp = Cyclopentadienyl)
compounds (σ2, π-coordination) with early transition metal fragments such as MCp2 (M = Zr, Hf).7 On the other hand, η4, π-complexes are formed with electron rich late transition metal fragments such as Fe(CO)3, (Scheme 1).8 Driven by this controversy, we conducted a comparative computational investigation of the bonding situation for the silicon, germanium and tin compounds 1a−3a which are close models to the experimentally investigated species 1−3 and we extend this series to the lead compound 4a (Scheme 1).
2. RESULTS AND DISCUSSION Silylene Synthesis and Characterization. Silylene 1 was synthesized by addition of a THF solution of hafnocene Received: April 9, 2017 Published: May 3, 2017 7117
DOI: 10.1021/jacs.7b03566 J. Am. Chem. Soc. 2017, 139, 7117−7123
Article
Journal of the American Chemical Society Scheme 1. σ2, π- and η4 π-Coordination of the Butadiene Ligand in Transition Metal and in Main Group Chemistry7,8
Figure 1. 29Si NMR chemical shifts of silylenes and related compounds (Ter: 2,4-bis(2,4,6-triisopropylphenyl)phenyl).
absorption at λmax = 388 nm. Spectroscopic indications for the intermediate silylene complex 6 or its THF complex were provided by low temperature NMR measurements (Scheme 2). An aliquot of the reaction mixture was transferred into a NMR tube at T = −90 °C, and 29Si NMR spectroscopy at T = −70 °C displayed the formation of one major product, which is characterized by two resonances at δ29Si = 164.5 and −14.5. At higher temperatures these signals gradually disappeared and at ambient temperature the signals of silylene 1 dominated the 29 Si NMR spectrum of the residual reaction mixture. The low field signal of the intermediate at δ29Si = 164.5 is comparable to 29 Si NMR resonances reported for related hafnocene silylene complexes.10 Additional NMR spectroscopic data supported its structural assignment as hafnocene silylene complex 6 (see SI material). Orange crystals suitable for X-ray diffraction analysis (XRD) of compound 1 were obtained from hexane/THF solution at T = −30 °C. The structure solution revealed one independent molecule of compound 1 along with disordered THF molecules. The shortest distance between an oxygen atom and the divalent silicon atom is 449 pm, which indicates no interaction between the THF molecules and the silylene 1 in the solid state. In the crystal, silylene 1 adopts a bicyclo[2.1.1]hexene-like molecular structure (Figure 2). The Cp2Hf fragment and the four carbon atoms form a five-membered
Scheme 2. Synthesis of Silylene 1
dichloride to a freshly prepared solution of dipotassium silacyclopentadiendiide K2[5] in THF at T = −105 °C (Scheme 2).9 A logic intermediate in this reaction is the hafnocene silylene complex 6 which was formed after double salt metathesis. Complexation of THF to the hafnium center is known to stabilize similar hafnium silylene complexes.10 At higher temperatures, the hafnocene complex 6 underwent a rearrangement reaction to give silylene 1. The stable but highly reactive silylene 1 was isolated at ambient conditions. The overall isolated yield after recrystallization from pentane/THF (3:1) was 28%. The identity of compound 1 was initially verified by NMR spectroscopy. 1H NMR spectroscopy suggested the formation of a symmetric product with two nonequivalent cyclopentadienyl (Cp) substituents (δ1H = 5.84, 5.73). The high field shift of the 13C NMR resonance of carbon atoms C1 and C4 of silole dianion 5 to silylene 1 indicated product formation (δ13C(C1/4) = 145.8 (5), 110.0 (1)). Most characteristic for the formation of silylene 1 are however the obtained 29Si NMR spectroscopic results. The formally dicoordinated silicon atom of silylene 1 shows a resonance at very high field (δ29Si = −155.2) which is shielded by Δδ29Si = −303.7 compared to the silole dianion 5 (δ29Si = 148.5)11 and which is outside the expectation for dicoordinated silylenes.12 Comparison with 29 Si NMR chemical shift data of silylenes and related Si(II) compounds (Figure 1) suggests a coordination number clearly larger than two for the divalent silicon atom in silylene 1.13 A hexane solution of the orange crystals showed a low energy
Figure 2. Molecular structure of silylene 1 in the crystal. Hydrogen atoms are omitted for clarity. Thermal ellipsoids are shown at the 50% probability level. Selected atom distances [pm] and angles [deg]: Si1−C1 199.0(3), Si1−C2 214.5(3), Si1−C3 217.1(4), Si1−C4 198.8(3), Si1−Hf1 293.9(9), C1−C2 148.8(4), C2−C3 143.6(4), C3−C4 148.4(4), C1−Hf1 221.7(3), C4−Hf1 221.0(3), α (Si) 94.9, α (Hf) 128.3. For the definition of the flap angles α, see Table 1. 7118
DOI: 10.1021/jacs.7b03566 J. Am. Chem. Soc. 2017, 139, 7117−7123
Article
Journal of the American Chemical Society ring with a regular envelope conformation (flap angle α(Hf) = 128.3°, Figure 2). The bonds between the hafnium atom and the carbon atoms C1 and C4 are slightly shorter than expected for Hf−C single bonds (221.7, 221.0 pm (1) vs regular Hf−C = 227 pm).14 The former silole ring adopts an extreme envelope conformation, with the silicon atom located above the plane spanned by the four carbon atoms (flap angle α(Si) of 94.9°, Figure 2). A remarkable feature of this molecular arrangement is that the interaction between the silicon atom and the double bond is more significant than that of the hafnium atom and the double bond. The Si−C distances in compound 1 (Si−C1/4 = 199.0, 198.8 pm; Si−C2/3 = 214.5, 217.1 pm) are all longer than regular Si−C single bond lengths involving tetracoordinated silicon and carbon atoms (Si−C = 186−188 pm),15 or than the Si−C bonds in silylenes I and II involving dicoordinated silicon atoms and tetracoordinated carbon atoms (Si−C = 190−191 pm).12b,m The Si−C1/4 distances are however only slightly longer than the Si−CTer bond in silylene V (Si−C = 197.3 pm).13h In addition, we note that for structurally related 7-silanorbornadienes elongated bonds between the tetracoordinated silicon atom and the bridgehead carbon atoms due to π*−σ-hyperconjugation were reported (C−Si = 191−197 pm).16 The C−C distances in the central C4 unit are alternating with relative short C1−C2/C3−C4 single bonds (148.8, 148.4 pm) and long C2C3 double bond (143.6 pm). The close approach of the silicon atom to the carbon atoms C2 and C3 and the long formal CC double bond suggest electron donation from the π-orbital of the CC double bond into the empty 3p orbital of the silicon atom. The clear separation of the silicon atom from the hafnium center (Si−Hf 293.9 pm) indicates only small bonding interaction. Overall, based on the experimental molecular structure, compound 1 can be described as a dicoordinated silylene with significant homoconjugative interaction between the silicon atom and the remote C2C3 double bond that increases its coordination number. It is noteworthy, that the Tilley group has previously studied the reactivity of silolyl monoanions versus pentamethyl cyclopentadienyl hafnium chlorides. In these cases structurally very different η5-silolyl complexes have been isolated.17 As expected, the structure of
silylene 1 is very similar to that of its germanium congener 2, as shown by comparison of the corresponding structural data (Table 1).4 We note however that also the structure of the tin compound 3 shows a close resemblance.6 In compound 3, the bonds between the tin and the bridgehead carbon atoms are long (Sn−C1/4 = 234.1, 230.0 pm) but they are in the expected range for bonds between dicoordinated tin and carbon atoms (Sn−C = 222−232 pm).18 In addition, the bending of the tin atom toward the C2−C3 bond results in short Sn/C2/3 separations (Sn−C2/3 = 244.7, 241.9 pm). Finally, the metrics of the carbon skeleton C1−C4 are almost identically for all three compounds. In particular, the C1−C2 and C3−C4 bonds are in each case longer than the central C2−C3 bond (Table 1). The question that emerges from this comparative presentation of the experimental structural data is whether two different bonding description for the three in principal homologues compounds 1−3 is justified or a unified concept should be applied. We will address this question in following by applying computational methods. Computational Investigation: Silylene Formation and Properties. For the computations of all compounds the M06-2X density functional method in combination with a triple-ζ basis set and relativistic pseudo potentials for the elements Sn, Pb and Hf were applied (M06-2X/def2-tzvp).19 The molecular structures obtained by this method for compounds 1−3 are very close to that in the solid state. The largest deviation between the computed atomic distances and the experimental data is 4.6 pm (2% deviation, for C1−Sn in compound 3). In addition, we calculated the molecular structures of model compounds 1a−4a which have no substituents at the carbon atoms (Scheme 1). For the silicon, germanium and tin containing model compounds 1a−3a the deviations of their computed structures relative to the experimental structures of compound 1−3 are also small (maximum 3% for bond lengths). At first, we note that the silylene hafnocene complex 6 is substantially less stable than silylene 1 (ΔE = −179 kJ mol−1). Even when stabilization of compound 6 by complexation with THF is taken into account, the formation of silylene 1 is exothermic by ΔE = −42 kJ mol−1 (Chart 2). For compound 1,
Table 1. Comparison of Structural Parameter of Carbene Analogues 2−4 and of Model Compounds 1a−4a; Experimental Data from XRDab (Bold) Computed Data at M06-2X/def2-tzvp
C1−C2 C2−C3 C3−C4 C1−E C2−E C3−E C4−E C4−Hf C1−Hf E−Hf C4−E−C1 α(Hf)c α(E)d
1
2
3
1a
2a
3a
4a
E = Si
E = Ge
E = Sn
E = Si
E = Ge
E = Sn
E = Pb
148.8(4), 147.6 143.6(4), 142.1 148.4(4), 147.7 199.0(3), 200.1 214.5(3), 211.4 217.1(4), 213.2 198.8(3), 199.8 221.0(3), 220.8 221.7(3), 222.1 293.9(9), 297.2 88.0(12), 87.1 128.3, 129.7 94.9, 92.5
148.3(8), 148.0 142.5(8), 142.5 148.2(8), 147.2 210.0(7), 209.3 226.7(6), 224.0 227.0(6), 225.9 208.8(6), 209.3 220.1(6), 220.4 220.3(7), 220.8 301.0(9), 307.4 85.0(2), 84.2 124.8, 122.3 97.9, 96.2
145.8 141.2 145.8 201.8 209.5 209.5 201.8 221.1 221.1 301.1 83.0 137.8 88.3
145.6 141.8 145.6 213.6 219.4 219.4 213.6 221.2 221.1 310.3 79.0 139.4 88.6
145.7 141.8 145.7 232.3 240.4 240.4 232.3 221.1 221.2 327.3 73.5 136.3 91.8
145.2 142.1 145.2 242.9 250.0 250.0 242.9 220.8 220.9 336.3 70.4 137.1 91.8
150.5, 142.3, 147.5, 234.1, 244.7, 241.9, 230.0, 220.9, 222.9, 319.4, 77.1, 132.9, 95.3,
147.8 143.0 148.4 229.5 242.4 243.7 233.1 221.6 223.6 322.5 76.7 131.8 95.2
Reference 4. bReference 6. cFlap angle α(Hf) is defined as angle between the Hf atom, the midpoint of C1/C4 and midpoint of C2/C3 distances. Flap angle α(E) is defined as angle between the E atom, the midpoint of C1/C4 and the midpoint of C2/C3 distances. eSum of the covalent radii expected for Hf−E single bonds: E = C: 227, E = Si: 268; E = Ge: 273; E = Sn: 292; E = Pb: 296 pm.14 fSum of the covalent radii expected for C−E single bonds: E = C: 150, E = Si: 191; E = Ge: 196; E = Sn: 215; E = Pb: 219 pm.14 a
d
7119
DOI: 10.1021/jacs.7b03566 J. Am. Chem. Soc. 2017, 139, 7117−7123
Article
Journal of the American Chemical Society
It is noteworthy that the frontier orbitals of the related germanium 2 and tin 3 compounds are, as reported earlier, of very similar shape.4,6 This suggests a comparable electronic situation in all three compounds. Computational Investigation: Bonding Situation. Based on the close structural resemblance of our models 1a - 3a to the experimental structures of compounds 1−3 (Table 1), we conducted a comparative computational investigation of their bonding situation. The calculated C−C bond lengths in the carbon skeleton of compounds 1a−4a are intermediate between typical C−C single and CC double bonds (Table 1) and show only small differences. Although this suggests a delocalized bonding situation it is noteworthy that for all model compounds a clear long−short−long sequence is found, indicating a higher CC bond order for the central C2C3 bond. This is supported by natural bond orbital (NBO) analysis for compounds 1a−4a that predicts a higher Wiberg bond index, WBI, for the central bond (WBI (C2C3) = 1.38−1.39 vs WBI (C1C2) = 1.18−1.24, Table 3).20 Based on the calculated structural features there is
Chart 2
29 Si NMR chemical shifts of δ29Si = −5 (SiMe3) and −176 (Si) (GIAO/M06L/def2-tzvp//M06-2X/def2-tzvp) were calculated, which are in fair agreement with the experimental data (δ29Si = −3 and −155 (Si)). The calculated frontier orbitals of silylene 1 are dominated by the interaction of the π-orbital of the C2C3 double bond and the Si 3p orbital (Figure 3). The HOMO indicates
Table 3. Results of NBO Analysis of Compounds 1a−4a
C1−C2 [pm] WBI (C1−C2) C2−C3 [pm] WBI (C2−C3) E−C1/4 [pm] WBI (E−C1/4) E−C2/3 [pm] WBI (E−C2/3) E−Hf [pm] WBI (E−Hf) occupation (σ(E−C1)) occupation (πα(C2−C3)) occupation (π(C2−C3)) occupation (np(E)) NRT (Scheme 4) weight of A sum of weight of B/C
Figure 3. Surface diagrams of the frontier orbitals of silylene 1 (M06-2X/def2-tzvp, at an isodensity value of 0.04). Color code: Yellow, silicon; light blue, hafnium; light gray, carbon; hydrogen atoms are omitted.
delocalization of π-electron density from the π(C2C3) orbital into the Si(3p) orbital. The LUMO features large contributions from the 3p(Si) orbital. In addition, the interaction between the π*(C2C3) orbital and the attached phenyl rings contribute to the shape of the LUMO. A TD/M06-2X calculation predicts a low energy absorption in the gas phase at λmax = 347 nm, which is at slightly higher energy than the experimental value obtained in hexane (λmax = 388 nm). The calculated HOMO energy level of silylene 1 is by 0.74 eV higher than predicted for the N-heterocyclic silylene, NHSi 7 (Chart 2) and the energy separation of singlet and triplet state ΔE(ST) is smaller by 1.06 eV, indicating both a higher nucleophilic character and larger electrophilicity for silylene 1 (Table 2).
E(HOMO) ΔE(ST) ΔER(1) ΔER(2) ΔER(3)
−6.03 2.65
7
−6.77 3.71
2b
1a
2a
3a
4a
E= Ge
E= Si
E= Ge
E= Sn
E= Pb
−5.98 2.69
−5.99 2.55 −15 −35 −144
−5.91 2.49 −32 −48 −135
−5.66 2.21 −20 −38 −124
−5.40 1.88 −22 −24 -c
2a
3a
4a
E = Ge
E = Sn
E = Pb
145.8 1.18 141.2 1.38 201.8 0.58 209.5 0.30 301.1 0.18 1.71 e 0.40 e 1.66 e 0.30 e
145.6 1.21 141.8 1.38 213.6 0.54 219.4 0.27 310.3 0.19 1.68 e 0.42 e 1.68 e 0.27 e
145.7 1.22 141.8 1.39 232.3 0.50 240.4 0.27 327.3 0.18 1.67 e 0.42 e 1.70 e 0.24 e
145.2 1.24 142.1 1.38 242.9 0.49 250.0 0.24 336.3 0.20 1.65 e 0.43 e 1.69 e 0.23 e
35% 18%
34% 14%
35% 11%
34% 8%
only little interaction between the Hf atom and the C2C3 bond in compounds 1a−4a although the Hf atom is electron deficient. The Hf−C2/C3 distance decreases slightly from the silicon compound 1a to the lead compound 4a (295.7 (1a), 287.3 (4a) pm) but all are significantly larger than the Hf−C2/C3 distance reported for the hafnocene butadiene complex 8 (264.1 pm, Chart 2).7b Likewise, there is no structural indication for an Hf/E interaction. All Hf/E distances are longer than the sum of the single bond covalent radii14 and the flap angles α(Hf) are large (136−139°, Table 1). In agreement, the calculated WBI for the Hf/E pairs are rather small (WBI(Hf/E) = 0.18−0.20). In contrast, for all four model compounds, 1a−4a, strong intramolecular interactions between the element atom E and the C2C3 double bond are suggested by short C2/3−E separations and by flap angles α(E) around 90°. The isodesmic reactions shown in Scheme 3 provide a thermodynamic estimate of the different intramolecular interactions. In agreement with the discussed structural parameter, the interaction between the hafnocene group and the C2C3 double bond (eq 1 in Scheme 3) and between the hafnocene
Table 2. Calculated HOMO Energies E(HOMO) and Vertical Singlet Triplet Separations, ΔE(ST) for compounds 1−4 and NHSi, 7a 1
1a E = Si
Calculated reaction energies ΔER(1) - ΔER(3) of the isodesmic reactions eq 1- 3 (Scheme 3) (kJ mol−1, at M06-2x/def2-tzvp). bFrom ref 4. cThe reaction energy for eq 3 (Scheme 3) for the lead compound could not be calculated since 5-plumba-6-germabicyclo[2.1.1]pentene is not a stationary point on the potential energy surface. a
7120
DOI: 10.1021/jacs.7b03566 J. Am. Chem. Soc. 2017, 139, 7117−7123
Article
Journal of the American Chemical Society Scheme 3. Isodesmic Reactionsa
significant contributions of canonical structure C as predicted by the NRT calculations. An analysis based on the quantum theory of atoms in molecules (QTAIM) revealed for all four model compounds very similar molecular graphs.22 The results for silylene 1a are shown in Figure 4 and those obtained for model compounds
a
The intramolecularly interacting groups on the product side of the equations are marked in red.
group and the element atom (eq 2 in Scheme 3) are small (−15 to −48 kJ mol−1, Table 2). In contrast, the interaction energies between the element atom and the C2C3 double bond are large (eq 3 in Scheme 3, −144 to −124 kJ mol−1, Table 2) and contribute significantly to the thermodynamic stability of model compounds 1a−3a. The frontier orbitals for the models 1a−4a are very similar to those calculated for compound 1 (see Figure 3). The predicted increase of the HOMO energy level and concomitant decrease of ΔE(ST) for model compounds 1a−4a implies a growth of the reactivity along this series (see Table 2). NBO analysis of the model compounds 1a−4a confirmed their highly delocalized structures. In the framework of natural resonance theory (NRT), the dominating Lewis structure for all four compounds is the dicoordinated carbene analogue A (Scheme 4, Table 3).21 Canonical structures that describe the homoconjugative interaction between the element atom and the remote C2C3 double bond (B and C, Scheme 4) are in
Figure 4. (a) Molecular graph of silylene 1a based on QTAIM analysis (M06-2X/ADZP(Hf),def2-tzvppd(Si,C,H)//M06-2X/def2-tzvp). Black lines indicate bond paths, green circles represent bond critical points, and small red circles are ring critical points. (b) Part of the molecular graph of silylene 1a projected on a contour plot of the Laplacian of the electron density in the plane spanned by the midpoint of the C2−C3 bond, the hafnium, and the silicon atom. Regions of local charge accumulation are marked in red contours, regions of local charge depletion are marked in blue contours.
2a−4a are given in the Supporting Information. The calculated graph of compound 1a reveals the structure of a dicoordinated silylene which is consistent with the results of the NBO analysis. An indication for the homoconjugative interaction between the silicon atom and the C2C3 double bond is the location of the ring critical point of the folded five-membered SiC1−C4 ring in the plane spanned by the Si atom and the carbon atoms C2 and C3 (Figure 4b). In addition, the calculated Laplacian of the electron density reveals a distinct valence shell charge concentration (VSCC) at the silicon atom that is reminiscent to the silylene lone pair (Figure 4b). The elliptical electron concentration around the bond critical point of the C2−C3 bond indicates its π-bond character and a silacyclopentene structure for compound 1a.23 At this point of the discussion, several conclusions are drawn. Although in the model compounds 1a−4a two electron deficient centers are present, the 14e hafnocene group and the 6e main group element atom, their structure and their electronic situation is dominated by the interaction of the element atom and the remote double bond. This can be interpreted as homoconjugative interaction as in classical examples such as the boranorbornadiene 924 or related carbocations i.e. cation 10.25 For silylenes, this stabilization mode was previously suggested by Maier and co-workers for a transient compound.26 The alternative description of an intramolecular σ2, π-coordination of the element atom to the unsaturated C4 ligand is equivalent (Figure 5). Our bonding analysis revealed for all four model compounds 1a−4a essentially uniform structural and electronic parameter. We found only small gradual differences between the compounds that do not imply a significantly different bonding situation. Based on our analysis, the model compounds 1a−4a and thereby also the experimentally investigated compounds 1−3 are carbene analogues, classical E(II) compounds that are stabilized by homoconjugation. Identifying the C4 perimeter as a butadiene ligand, the description as σ2, π-coordination as in zirconocene or hafnocene complexes is appropriate. We found, however, no indication for the alternative 4π-coordination of the
Scheme 4. Leading Resonance Structures for Compounds 1a−4a According to Natural Resonance theory
particular for the silicon 1a and germanium compound 2a of importance (Table 3). The natural population analysis (NPA) of this leading resonance structure reveals significantly depleted σ-(E−C1/4) bonds (occupation 1.71 e to 1.65 e) compared to the ideal value of 2e. This depletion results from σ−π* hyperconjugation between the E−C1/4 bonds and the π*(C2C3) orbitals, that leads to significant occupations of the π*-orbitals (0.40−0.43 e). Consequently, this interaction weakens both participating bonds while increasing the C1−C2 and C3−C4 bond orders. Further reduction of the C2C3 bond order results from homoconjugative interaction between the π(C2C3) (occupation 1.66 e to 1.70 e) and the formally empty np(E) orbitals (occupation 0.23−0.30 e). These delocalization modes are reflected by the calculated WBIs for compounds 1a−4a. Strongly reduced bond orders were predicted for the C2C3 (WBI 1.38−1.39) and the E−C1/4 bonds (WBI 0.49−0.58). In contrast, increased bond orders were computed for the C1−C2 bonds (WBI 1.18−1.24, Table 3, and, for comparison with standard compounds, see the Supporting Information). Notably, WBIs ranking from 0.24 to 0.30 between the element atom and the C3 atom hint to a substantial covalent bonding and are in agreement with 7121
DOI: 10.1021/jacs.7b03566 J. Am. Chem. Soc. 2017, 139, 7117−7123
Article
Journal of the American Chemical Society
Fellowship to Z.D. The simulations were performed at the HPC Clusters HERO and CARL, located at the University of Oldenburg (Germany) and funded by the DFG through its Major Research Instrumentation Program (INST 184/108-1 FUGG/INST 184/157-1 FUGG) and the Ministry of Science and Culture (MWK) of the Lower Saxony State.
■
(1) (a) Power, P. P. Nature 2010, 463, 171−177. (b) Asay, M.; Jones, C.; Driess, M. Chem. Rev. 2011, 111, 354−396. (c) Driess, M. Nat. Chem. 2012, 4, 525−526. (d) Fischer, R. C.; Power, P. P. Chem. Rev. 2010, 110, 3877−3923. (2) (a) Spikes, G. H.; Fettinger, J. C.; Power, P. P. J. Am. Chem. Soc. 2005, 127, 12232−12233. (b) Peng, Y.; Ellis, B. D.; Wang, X.; Fettinger, J. C.; Power, P. P. Science 2009, 325, 1668−1670. (c) Longobardi, L. E.; Russell, C. A.; Green, M.; Townsend, N. S.; Wang, K.; Holmes, A. J.; Duckett, S. B.; McGrady, J. E.; Stephan, D. W. J. Am. Chem. Soc. 2014, 136, 13453−13457. (d) Erker, G.; Stephan, D. W.; et al. Frustrated Lewis Pairs I & II. In Topics in Current Chemistry; Erker, G., Stephan, D. W., Eds.; Springer-Verlag: Berlin, Germany, 2013. (3) (a) Bourissou, D.; Guerret, O.; Gabbai, F. P.; Bertrand, G. Chem. Rev. 2000, 100, 39−92. (b) Hahn, F. E.; Jahnke, M. C. Angew. Chem., Int. Ed. 2008, 47, 3122−3172. (c) Hopkinson, M. N.; Richter, C.; Schedler, M.; Glorius, F. Nature 2014, 510, 485−496. (d) Vignolle, J.; Cattoen, X.; Bourissou, D. Chem. Rev. 2009, 109, 3333−3384. (e) Melaimi, M.; Soleilhavoup, M.; Bertrand, G. Angew. Chem., Int. Ed. 2010, 49, 8810−8849. (f) Soleilhavoup, M. l.; Bertrand, G. Acc. Chem. Res. 2015, 48, 256−266. (4) Dong, Z.; Reinhold, C. R. W.; Schmidtmann, M.; Müller, T. Angew. Chem., Int. Ed. 2016, 55, 15899−15904. (5) (a) Tokitoh, N.; Okazaki, R. Coord. Chem. Rev. 2000, 210, 251− 277. (b) Kü hl, O. Coord. Chem. Rev. 2004, 248, 411−427. (c) Nagendran, S.; Roesky, H. W. Organometallics 2008, 27, 457− 492. (d) Mizuhata, Y.; Sasamori, T.; Tokitoh, N. Chem. Rev. 2009, 109, 3479−3511. (e) Mandal, S. K.; Roesky, H. W. Chem. Commun. 2010, 46, 6016−6041. (f) Kühl, O. Mini-Rev. Org. Chem. 2010, 7, 324−334. (g) Marschner, C. Eur. J. Inorg. Chem. 2015, 3805−3820. (h) Haaf, M.; Schmedake, T. A.; West, R. Acc. Chem. Res. 2000, 33, 704−714. (i) Gehrhus, B.; Lappert, M. F. J. Organomet. Chem. 2001, 617, 209− 223. (j) Kira, M. Chem. Commun. 2010, 46, 2893−2903. (k) Yao, S.; Xiong, Y.; Driess, M. Organometallics 2011, 30, 1748−1767. (l) Sen, S. S.; Khan, S.; Samuel, P. P.; Roesky, H. W. Chem. Sci. 2012, 3, 659− 682. (m) Sen, S. S.; Khan, S.; Nagendran, S.; Roesky, H. W. Acc. Chem. Res. 2012, 45, 578−587. (n) Bag, P.; Ahmad, S. U.; Inoue, S. Bull. Chem. Soc. Jpn. 2017, 90, 255−271. (6) Kuwabara, T.; Nakada, M.; Hamada, J.; Guo, J. D.; Nagase, S.; Saito, M. J. Am. Chem. Soc. 2016, 138, 11378−11382. (7) (a) Krueger, C.; Mueller, G.; Erker, G.; Dorf, U.; Engel, K. Organometallics 1985, 4, 215−223. (b) Erker, G.; Krüger, C.; Müller, G. Adv. Organomet. Chem. 1985, 24, 1−39. (8) Deeming, A. J. In Comprehensive Organometallic Chemistry; Wilkinson, G., Stone, F. G. A., Abel, E. W.; Eds.; Pergamon Press: Oxford, 1982; Vol. 4, p 377. (9) Fekete, C.; Nyulászi, L.; Kovács, I. Phosphorus Sulfur Silicon Relat. Elem. 2014, 189, 1076−1083. (10) Nakata, N.; Fujita, T.; Sekiguchi, A. J. Am. Chem. Soc. 2006, 128, 16024−16025. (11) Measured in THF. We note that ref 9 reports for K2[5] in benzene/THF a 29Si NMR chemical shift of δ29Si = 49.3. (12) (a) Denk, M.; Lennon, R.; Hayashi, R.; West, R.; Belyakov, A. V.; Verne, H. P.; Haaland, A.; Wagner, M.; Metzler, N. J. Am. Chem. Soc. 1994, 116, 2691−2692. (b) Kira, M.; Ishida, S.; Iwamoto, T.; Kabuto, C. J. Am. Chem. Soc. 1999, 121, 9722−9723. (c) West, R.; Denk, M. Pure Appl. Chem. 1996, 68, 785−788. (d) Driess, M.; Yao, S.; Brym, M.; van Wüllen, C.; Lentz, D. J. Am. Chem. Soc. 2006, 128, 9628−9629. (e) Gehrhus, B.; Lappert, M. F.; Heinicke, J.; Boese, R.; Bläser, D. J. Chem. Soc., Chem. Commun. 1995, 1931−1932. (f) Asay,
Figure 5. Equivalence between σ2, π-coordination and homoconjugation in bicyclic compounds.
butadiene ligand to an element in its E(0) valence state as it is realized, for example, in iron carbonyl butadiene complexes and suggested by Saito and co-workers for the tin compound 3.
3. CONCLUSIONS We found that salt metathesis reaction of dipotassium silacyclopentadiendiide K2[5] with hafnocene dichloride provides access to a new type of silylene 1. Low temperature NMR spectroscopy provides evidence that the reaction proceeds via hafnocene silylene complex 6, which is stabilized by the solvent THF. Silylene 1 is stable at ambient conditions and its reactivity is currently explored in our laboratories. The analysis of its molecular structure and the computational investigation of its electronic properties supported our interpretation that compound 1 should be regarded as a silylene which is stabilized by homoconjugation with the remote CC double bond. Hence, the close relationship to the germanium compound 2 is shown. The description of compounds 1 and 2 as carbene analogues with the tetrel element in the formal oxidation state II is in contrast to the formulation of the corresponding tin compound 3 as Sn(0) butadiene complex. The results of a comparative computational analysis of the silicon, germanium, tin and lead containing model compounds 1a−4a revealed very similar bonding situations for all four compounds. Furthermore, the analysis favors the structure of a carbene analogue which is stabilized by homoconjugation for all four compounds. Adapting the well-established terminology for the bonding modes of the butadiene ligand, this description is analogous to a σ2, π-coordination of a butadiene ligand to the element atom.
■
ASSOCIATED CONTENT
* Supporting Information S
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.7b03566. Experimental and computational details, NMR spectra, and details of the structure solution of compound 1 (PDF) Crystal data for silylene 1 (CIF)
■
REFERENCES
AUTHOR INFORMATION
Corresponding Author
*
[email protected] ORCID
Thomas Müller: 0000-0002-4247-3776 Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was supported the Carl von Ossietzky University Oldenburg and by the Lower Saxony State by a Lichtenberg 7122
DOI: 10.1021/jacs.7b03566 J. Am. Chem. Soc. 2017, 139, 7117−7123
Article
Journal of the American Chemical Society M.; Inoue, S.; Driess, M. Angew. Chem., Int. Ed. 2011, 50, 9589−9592. (g) Protchenko, A. V.; Schwarz, A. D.; Blake, M. P.; Jones, C.; Kaltsoyannis, N.; Mountford, P.; Aldridge, S. Angew. Chem., Int. Ed. 2013, 52, 568−571. (h) Protchenko, A. V.; Birjkumar, K. H.; Dange, D.; Schwarz, A. D.; Vidovic, D.; Jones, C.; Kaltsoyannis, N.; Mountford, P.; Aldridge, S. J. Am. Chem. Soc. 2012, 134, 6500− 6503. (i) Rekken, B.; Brown, T.; Fettinger, J.; Tuononen, H.; Power, P. J. Am. Chem. Soc. 2012, 134, 6504−6507. (j) Abe, T.; Tanaka, R.; Ishida, S.; Kira, M.; Iwamoto, T. J. Am. Chem. Soc. 2012, 134, 20029− 20032. (k) Lee, G.-H.; West, R.; Müller, T. J. Am. Chem. Soc. 2003, 125, 8114−8115. (l) Hadlington, T. J.; Abdalla, J. A.; Tirfoin, R.; Aldridge, S.; Jones, C. Chem. Commun. 2016, 52, 1717−1720. (m) Kosai, T.; Ishida, S.; Iwamoto, T. Angew. Chem., Int. Ed. 2016, 55, 15554−15558. (13) (a) Jutzi, P.; Kanne, D.; Krüger, C. Angew. Chem., Int. Ed. Engl. 1986, 25, 164−164. (b) Takeda, N.; Suzuki, H.; Tokitoh, N.; Okazaki, R.; Nagase, S. J. Am. Chem. Soc. 1997, 119, 1456−1457. (c) So, C.-W.; Roesky, H. W.; Magull, J.; Oswald, R. B. Angew. Chem., Int. Ed. 2006, 45, 3948−3950. (d) Yao, S.; Brym, M.; van Wüllen, C.; Driess, M. Angew. Chem., Int. Ed. 2007, 46, 4159−4162. (e) Rodriguez, R.; Gau, D.; Contie, Y.; Kato, T.; Saffon-Merceron, N.; Baceiredo, A. Angew. Chem., Int. Ed. 2011, 50, 11492−11495. (f) Junold, K.; Baus, J. A.; Burschka, C.; Tacke, R. Angew. Chem., Int. Ed. 2012, 51, 7020−7023. (g) Rodriguez, R.; Contie, Y.; Gau, D.; Saffon-Merceron, N.; Miqueu, K.; Sotiropoulos, J.-M.; Baceiredo, A.; Kato, T. Angew. Chem., Int. Ed. 2013, 52, 8437−8440. (h) Jutzi, P.; Leszczyńska, K.; Neumann, B.; Schoeller, W. W.; Stammler, H. G. Angew. Chem., Int. Ed. 2009, 48, 2596−2599. (i) Azhakar, R.; Ghadwal, R. S.; Roesky, H. W.; Wolf, H.; Stalke, D. Chem. Commun. 2012, 48, 4561−4563. (j) So, C.-W.; Roesky, H. W.; Gurubasavaraj, P. M.; Oswald, R. B.; Gamer, M. T.; Jones, P. G.; Blaurock, S. J. Am. Chem. Soc. 2007, 129, 12049−12054. (k) Sen, S. S.; Hey, J.; Herbst-Irmer, R.; Roesky, H. W.; Stalke, D. J. Am. Chem. Soc. 2011, 133, 12311−12316. (l) Gao, Y.; Zhang, J.; Hu, H.; Cui, C. Organometallics 2010, 29, 3063−3065. (14) Pyykkö, P.; Atsumi, M. Chem. - Eur. J. 2009, 15, 12770−12779. (15) Kaftory, M.; Kapon, M.; Botoshansky, M. In The Chemistry of Organic Silicon Compounds; Rappoport, Z., Apeloig, Y.; Eds.; John Wiley & Sons: Chichester, 1998; Vol. 2, p 181. (16) (a) Gerdes, C.; Saak, W.; Haase, D.; Müller, T. J. Am. Chem. Soc. 2013, 135, 10353−10361. (b) Lutters, D.; Severin, C.; Schmidtmann, M.; Müller, T. J. Am. Chem. Soc. 2016, 138, 6061−6067. (17) (a) Dysard, J. M.; Tilley, T. D. J. Am. Chem. Soc. 1998, 120, 8245−8246. (b) Dysard, J. M.; Tilley, T. D. J. Am. Chem. Soc. 2000, 122, 3097−3105. (18) Klinkhammer, K. W. In The chemistry of organic germanium, tin, and lead compounds; Rappoport, Z.; Ed.; John Wiley & Sons: Chichester, 2002; Vol. 2, p 283. (19) (a) The Gaussian 09 program was used. Frisch, M.; et al. Gaussian 09, Revision B.01/D.01; Gaussian, Inc.: Wallingford, CT, 2010. (b) For detailed description, see the Supporting Information. (20) Wiberg, K. B. Tetrahedron 1968, 24, 1083−1096. (21) (a) Glendening, E. D.; Weinhold, F. J. Comput. Chem. 1998, 19, 610−627. (b) Glendening, E. D.; Weinhold, F. J. Comput. Chem. 1998, 19, 593−609. (c) Glendening, E. D.; Badenhoop, J.; Weinhold, F. J. Comput. Chem. 1998, 19, 628−646. (22) (a) Bader, R. F. W. Atoms in Molecules: A Quantum Theory, Clarendon Press, Oxford, U. K., 1990;. (b) The QTAIM analysis was performed with the AIMALL program: Keith, T. A., AIMAll, Version 11.05.16, 2011. (23) Further discussion of the QTAIM results and additional numerical data are provided in the Supporting Information. (24) Fagan, P. J.; Burns, E. G.; Calabrese, J. C. J. Am. Chem. Soc. 1988, 110, 2979−2981. (25) Laube, T.; Lohse, C. J. Am. Chem. Soc. 1994, 116, 9001−9008. (26) Maier, G.; Reisenauer, H. P. Eur. J. Org. Chem. 2003, 479−487.
7123
DOI: 10.1021/jacs.7b03566 J. Am. Chem. Soc. 2017, 139, 7117−7123