Neutral “Cp-Free” Silyl-Lanthanide(II) Complexes: Synthesis, Structure, and Bonding Analysis

Complexes featuring lanthanide silicon bonds represent a research area still in its infancy. Herein, we report a series of Cp-free lanthanide (+II) complexes bearing σ-bonded silyl ligands. By reactions of LnI2 (Ln = Yb, Eu, Sm) either with a 1,4-oligosilanyl dianion [K-Si(SiMe3)2SiMe2SiMe2Si(SiMe3)2-K)] (1) or with 2 (Me3Si)3SiK (3) the corresponding neutral metallacyclopentasilanes ({Me2Si(Me3Si)2Si}2)Ln·(THF)4 (Ln = Yb (2a), Eu (2b), Sm (2c)), or the disilylated complexes ({Me3Si}3Si)2Ln·(THF)3 (Ln = Yb (4a), Eu (4b), Sm (4c)), were selectively obtained. Complexes 2b, 2c, 4b, and 4c represent the first examples of structurally characterized Cp-free Eu and Sm complexes with silyl ligands. In both series, a linear correlation was observed between the Ln–Si bond lengths and the covalent radii of the corresponding lanthanide metals. Density functional theory calculations were also carried out for complexes 2a–c and 4a–c to elucidate the bonding situation between the Ln(+II) centers and Si.


■ INTRODUCTION
Complexes with the metal in the oxidation state 3+ bearing Cp or substituted Cp ligands are dominating the organometallic chemistry of the rare-earth elements. With respect to the divalent oxidation state, 1,2 for a long time only three lanthanide elements, samarium, europium, and ytterbium, were readily accessible. However, in the meantime the number of elements has increased substantially, and quite recently Evans et al. were successful in completing the series of crystalline examples of divalent molecular complexes of all lanthanides. 3−6 This enabled for the first time a comparison of all lanthanides in a single uniform coordination environment and postulation of their electronic ground states. 6 The results show conclusively that ligands can change the electronic ground state of Ln(II) complexes whereas this has not been observed for the Ln(III) complexes. Due to the high energy of the 5d orbitals and the limited radial extensions of the 4f orbitals for Ln(III) complexes, interactions with the ligands are weaker. 6 Although numerous examples of d-block metal-silyl compounds exist, examples of rare-earth metal-silyl compounds are still scarce or entirely unknown for some elements. The few reported samarium silyl compounds, which have also been characterized by X-ray diffraction analysis, feature Sm with Cp* ligands almost exclusively in oxidation state 3+ 7−9 with the exception of some divalent silylene complexes reported by Evans and co-workers 10 and our group. 11 The reaction of Cp* 2 Yb·(Et 2 O) with (Me 3 Si) 3 SiLi was shown to result in the formation of Cp*YbSi(SiMe 3 ) 3 ·(THF) 2 accompanied by elimination of Cp*Li. 12 Reacting Cp*YbSi(SiMe 3 ) 3 ·(THF) 2 with excess (Me 3 Si) 3 SiLi led to the formation of some 30% of [(Me 3 Si) 3 Si] 2 Yb·(THF) x , which could not be isolated. 12 The reaction of Cp* 2 Ln·(THF) 2 (Ln = Sm, Eu, Yb) with PhSiH 3 / KH, which yields metalate complexes of the type K-[Cp* 2 Ln II (SiH 3 )·(THF)], 13 also seems to involve silyl anions. The respective Eu and Yb complexes could be successfully characterized using single crystal X-ray diffraction analysis. In addition Bochkarev and co-workers reported the synthesis of neutral Cp-free (Ph 3 Si) 2 Yb II (THF) 4 by direct treatment of elemental ytterbium with Ph 3 SiCl in THF. 14 Divalent rare-earth amides, particularly silylamides [Ln II -N(SiMe 3 ) 2 ], have also received some attention. 15−18 A few of these compounds do not contain cyclopentadienyl ligands on the metal and are therefore interesting for the exploration of catalytic processes such as hydrosilylation 19 or enantioselective hydroamination 20 and polymerization of polar monomers like methyl methacrylate and lactones. 21 Recently, Evans et al. could show reactions of N 2 , CO, and CO 2 with [(Me 3 Si) 2 N] 3 Ln complexes, where N 2 is reduced to (NN) 2− and (N 2 ) 3− and CO and CO 2 are reduced to the rare (CO) 1− and (CO 2 ) 1− radicals. 4,22 Despite the lack of direct Ln−Si bonds, β-agostic interactions 23,24 play an important role in lanthanide complexes that contain the N(SiMe 2 H) 2 group. 18 Several types of agostic lanthanide hydridosilylamido complexes 25 3 led to methyl deprotonation of a trimethylsilyl group rather than to Si−Yb III bond formation. 28,29 β-Agostic interactions are important not only in complexes with N(SiMe 2 H) 2 groups but also in complexes like Yb[C-(SiMe 2 H) 3 ] 2 ·THF 2 30 or Yb[C(SiMe 2 H) 3 ] 2 ·TMEDA 31 containing the C(SiMe 2 H) 3 ligand. The short Yb−Si distances as well as the small Yb−C−Si angles in the crystal structure are in perfect concordance with the β-agostic Si−H−Yb interactions. 30,31 We recently reported synthesis and characterization of some σ-bonded Cp 2 -lanthanide-silyl complexes (Ln = Tm, Ho, Gd, Tb, Ce) in the oxidation state 3+. 11 Herein, the synthesis of Cp-free samarium-, europium-, and ytterbium-silyl complexes in the oxidation state 2+ are described in addition to a thorough investigation into their spectroscopic properties, X-ray diffraction studies, and the first systematic density functional theory (DFT) investigation for this class of complexes.

■ RESULTS AND DISCUSSION
Synthesis. Ytterbium(II), europium(II), and samarium(II) diiodides were investigated with regard to their reactivity with oligosilanyl anions. The THF complexes LnI 2 ·(THF) 2 [Ln = Yb (yellow), Eu (dark yellow), Sm (dark blue)] were prepared by reaction of small pieces of metal with 1,2-diiodoethane in THF. 32,33 The dark blue SmI 2 ·(THF) 2 was either dissolved in THF or suspended in toluene and treated with silanyl mono-or dianions. Using the less reactive silyl magnesium compounds [(Me 3 Si) 3 Si] 2 Mg 34 or [Me 2 Si(Me 3 Si) 2 Si] 2 Mg 34 did not cause any reaction at ambient temperature within 18 h as judged by in situ 1 H and 29 Si NMR spectroscopy, and the solutions remained dark blue. In addition, heating the reaction mixture to 80°C for 3 h did not lead to any observable changes. With the more reactive system (Me 3 Si) 3 SiK·18-crown-6 35 in toluene only the formation of (Me 3 Si) 4 Si could be observed. The formation of (Me 3 Si) 2 Sm·(THF) x had been reported to occur in the reaction of hexamethyldisilane with samarium amalgam leading to an inseparable mixture of the desired Sm(II) compound and Sm(III) derivatives. 36 It was further proposed that reaction at low temperature and large excess of samarium would lead to a more selective reaction to the Sm(II) complex. 36 Eventually, we tried the treatment of SmI 2 ·(THF) 2 in THF with [Me 2 Si-(Me 3 Si) 2 Si] 2 K (1) or 2 equiv of (Me 3 Si) 3 SiK (3) and obtained dark violet reaction mixtures. Reaction monitoring by 1 H and 29 Si NMR spectroscopy showed the formation of [Me 2 Si-( M e 3 S i ) 2 S i ] 2 S m · ( T H F ) 4 ( 2 c ) (Scheme 1) and [(Me 3 Si) 3 Si] 2 Sm·(THF) 3 (4c) (Scheme 2), respectively. Both compounds could be isolated as very air and moisture sensitive but stable crystalline compounds.
With this success, we reasoned that the synthesis of analogous ytterbium complexes might be accomplished in a similar manner. Monitoring the reaction of 3 with YbI 2 ·(THF) 2 by NMR spectroscopy showed after only 15 min complete consumption of starting material 3 and the formation of 4a along with substantial amounts of tetrakis(trimethylsilyl)silane and tris(trimethylsilyl)silane. In an analogous manner, reaction of YbI 2 ·(THF) 2 with 1 showed formation of 2a, but again the generation of several side products was observed. These results are similar to what was previously described for titanium 37 and yttrium 38 complexes bearing tris(trimethylsilyl)silyl groups. The problem could be overcome by strict exclusion of light during the formation and storage of 2a and 4a. Exposure of 2a to ambient light over the duration of a week caused decomposition to 1,1,2,2-tetrakis(trimethylsilyl)tetramethylcyclotetrasilane 39 as the main product. The yields of 2a and 4a could be increased by changing the solvent from THF to toluene or DME (2a·DME). Both 2a and 4a are not only light sensitive but also extremely air and moisture sensitive. Exposing 2a to vacuum over several hours resulted in complete decomposition of the complex and in generation of 1,1,4,4-tetrakis(trimethylsily)tetrasilane and uncharacterizable insoluble metal species.
The synthesis of the europium compounds 2b and 4b turned out to be similar to the ytterbium complexes. Again, selective reactions with short reaction time were observed in toluene, and air and moisture sensitive crystalline compounds were obtained.
Reaction of LnI 2 (Ln = Yb, Eu, or Sm) according to Scheme 2 with just 1 equiv of silyl anion 3 led in all our attempts only to products 4a−c, and no evidence for the formation of a monosilylated complex (Me 3 Si) 3 SiLnI was found.
The determination of yields of all compounds was somewhat hampered by some loss of THF during the isolation process due to the use of reduced pressure. For the diamagnetic Yb

Inorganic Chemistry
Article compound 2a the yield could however be determined NMR spectroscopically by using a defined amount of toluene as internal standard.
To get an estimate of the reactivity of the Si−Yb bond of 2a a reaction with Cp 2 ZrCl 2 was carried out. The analogous reactions with the potassium or magnesium 1,4-disilanides 34,40,41 are known to give the respective zirconacyclopentasilane (5). 40,41 The same course was also observed in the reaction of 2a with Cp 2 ZrCl 2 (Scheme 3). This outcome certainly marks compound 2a as possessing a strong disilanide character.
NMR Spectroscopy. With respect to NMR spectroscopy, the diamagnetic ytterbium complexes 2a and 4a are the most interesting ones with respect to insight into the bonding situation. While for the europium complexes 2b and 4b no meaningful NMR spectra could be obtained at all, the samarium complex 2c and 4c exhibited spectra with paramagnetically shifted signals.
So far two examples of tris(trimethylsilyl)silylated ytterbium-(II) complexes are known. The neutral complex Cp*YbSi-(SiMe 3 ) 3 ·(THF) 2 , 12 reported by Lawless and co-workers, displays 29 Si NMR signals at −158.3 ppm (Yb−Si) and −2. Very similar spectroscopic data were also found for the analogous DME complex 2a·DME with signals at −158.4 ppm ( 1 J Yb−Si = 656 Hz, Si(SiMe 3 ) 3 ), −29.8 ppm (SiMe 2 ), and −2.9 ppm ( 2 J Yb−Si = 20.4 Hz, SiMe 3 ). When the formation of 2a was carried out directly in C 6 D 6 instead of toluene with an approximate THF concentration of 9 equiv per Yb atom, the 29 Si NMR spectrum of the reaction solution indicated a higher degree of shielding of the metalated silicon atom with the resonance shifted to −163.9 ppm and a 1 J Yb−Si coupling constant of 633 Hz. The substantial shift from −154.0 to −163.0 ppm suggests a higher degree of solvatization of the Yb atom in the presence of THF.
A comparison of the 29 Si NMR spectroscopic data with structurally related compounds such as analogous magnesium, zinc, zirconocene, and hafnocene complexes helps to obtain some insight into the nature of the Si−Yb interaction. The oligosilanyl magnesium compounds [(Me 3 Si) 3 Si] 2 Mg· (THF) 2 42 [δ 29 Si = −171.9 ppm (SiMg), −6.4 ppm (SiMe 3 )], [(Me 3 Si) 3 Si] 2 Mg·(TMEDA) 34 3 Si] 2 Zn·(bipy). 45 For the Si−Yb−Si complexes 4a and 2a a similar behavior can be observed, where the shift of the undistorted complex 4a of δ = −144.8 ppm is raised to δ = −154.0, −158.5 ppm for the constrained geometry of complexes 2a and 2a·DME with much smaller Si−Yb−Si angles than found for 4a. From these NMR data it seems likely that the polarization of the Si−Yb bond should be somewhat more pronounced than that of a Si−Zn bond but less than a Si−Mg bond. This conclusion seems to be consistent with the reaction of 2a with Cp 2 ZrCl 2 to 5 which also hints at a high degree of polarization.
The number of silylated samarium complexes with elucidated crystal structures is scant, 7−10,47 and only for one compound, Cp* 2 SmSiH(SiMe 3 ) 2 , are 29 Si NMR spectroscopic data known (δ 29 Si = −23.5 ppm (SiMe 3 ), no signal detected for Si−Sm). 7,8 None of the known complexes for which NMR spectroscopic data are given features Sm in the oxidation state 2+. However, a comparison of the NMR spectroscopic properties of Cp* 3 Sm 48 and Cp* 2 Sm·(THF) 2 49 provides some ideas as to what to expect for the NMR spectra of 2c and 4c. The 13 C NMR resonances for the Cp* unit were found for Cp* 3 Sm at δ = 113.2 ppm (C 5 Me 5 ) and 28.3 ppm (C 5 Me 5 ) while for Cp* 2 Sm· (THF) 2 these signals were found at δ = −73.7 ppm (C 5 Me 5 ) and 94.6 ppm (C 5 Me 5 ). The THF resonances for Cp* 2 Sm· (THF) 2 were found at δ = 149.5 ppm (OCH 2 ) and 33.4 ppm (OCH 2 CH 2 ). It seems thus reasonable to expect 29    Some key metrical parameters from the crystal structures of the isostructural complexes 4a−4c. Thermal ellipsoids at the 30% probability level and hydrogen atoms omitted for clarity. Color code: green = lanthanide metal (Yb (4a), Eu (4b), Sm (4c)), pink = silicon, light gray = carbon.

Inorganic Chemistry
Article Fluorescence Spectroscopy. Among the divalent rareearth ions, Eu 2+ has attracted increasing attention over the past decades for its unique fluorescence properties. 50,51 In Eu 3+ (4f 6 ) the emission after short wavelength excitation consists of relatively sharp f−f bands found always at the same wavelengths. In contrast the corresponding emission of Eu 2+ (4f 7 ) is much broader, and depending on the material, it can occur at wavelengths in the blue, green, or red 52 part of the visible spectrum. This is due to the 4f 6 5d 1 → 4f 7 nature of the Eu 2+ emission. It is parity allowed and originates from the first excited state configuration, which involves 5d orbitals. Since the energy of the d-orbitals is strongly influenced by the surrounding crystal field, the transition can occur at different wavelengths in different Eu 2+ containing materials.
The fluorescence spectrum of solid compound 4b upon excitation at 366 nm shows a broad emission with a maximum intensity at 490 nm in the blue/green wavelength range (Supporting Information Figure S24). The fluorescence is strong enough to be seen with the naked eye. After a few minutes in the spectrometer's sample holder, however, the yellow sample in the central region of the sample holder, where it is hit by the excitation UV light, turns black indicating its decay. As expected a comparison with the fluorescence spectrum of solid EuI 2 (THF) 2 53 (Supporting Information Figure S24) reveals that the silyl groups are stronger crystal field ligands than iodide (Supporting Information Figure S25).
X-ray Crystallography. Molecular structures of all compounds 2a−c ( Figure 1) and 4a−c (Figure 2) in the solid state could be determined by means of single crystal X-ray diffraction. Interestingly, neutral metallacyclopentasilanes 2a and 2c were found to crystallize in the monoclinic space group P2 1 /n whereas 2b prefers the trigonal space group R3̅ . The difference in the space group is caused by a different packing pattern of 2b. Within the crystal of complex 2b, channels with a diameter of 13.62 Å are formed by an alternative arrangement of CH 3 units from the SiMe 3 groups and the CH 2 groups of the coordinate THF molecules. Compounds 2a−c all contain four THF molecules coordinating to the lanthanide center with complex 2c crystallizing with an additional THF molecule in the asymmetric unit. The five-membered rings of 2a−c all engage in envelope conformations with one of the SiMe 2 units on the flap with distances of 0.73 Å for 2a, 0.56 Å for 2b, and 0.78 Å for 2c from the ring plane.
Compounds 4a and 4c ( Figure 2) crystallize in the orthorhombic space group Pbcn with three THF molecules coordinating to the lanthanide atoms. The asymmetric units consist of half a molecule, with the Yb or Sm atoms and one oxygen atom of a THF molecule residing on a symmetry plane. The europium compound 4b crystallizes in the space group Pna2 1 but with the whole molecule in the asymmetric unit. The packing patterns of 4a and 4c are identical whereas that of the europium compound 4b is different.
The Si−Yb bond distances in 2a (3.106 and 3.171 Å) and 4a (3.0644 Å) are in good agreement with published Cp-free Yb(II) complexes where the distances range from 3.039 to 3.191 Å. 14,27,29−31 The bond distances of the Eu−Si bonds in 2b (3.2052, 3.2162 Å) and 4b (3.1497 Å) are comparable to that found in Cp* 2 Eu II SiH 3 K(THF) 2 (3.239 Å). 13 For 2c This methodology yields an adequate description of the electronic structure of f-block-silyl complexes as was shown in a previous study. 11 Table 1 lists the calculated average Ln−Si bond lengths (Å), Mayer-bond order (MBO) of Ln−Si bonds, Natural Population Analysis charges, and HOMO energies (eV) for 2a−c and 4a−c. These data are in excellent agreement with experimental bond lengths (see Table 1), again confirming the appropriateness of our method of choice. The electronic structures of complexes containing the same lanthanide centers are similar to each other in the sense that they have the same number of unpaired f electrons (0, 7, and 6 for structures a, b, and c, respectively). Natural Population Analysis (NPA) reveals strong ionic character of Ln−Si bonds, with the silicon atoms possessing almost a clear extra electron (Table 1)

■ CONCLUSION AND SUMMARY
The chemistry of silyl lanthanides is still a poorly investigated field of research. In the current study we reported a number of lanthanide (+II) complexes with oligosilanyl ligands. The formation of these compounds occurs in a surprisingly facile way by reactions of LnI 2 (Ln = Yb, Eu, Sm) either with a 1,4oligosilanyl dianion [K-Si(SiMe 3 ) 2 SiMe 2 SiMe 2 Si(SiMe 3 ) 2 -K]  Sm (4c)). While the NMR spectra of the samarium complexes 2c and 4c display the extreme chemical shifts typical for paramagnetic compounds, the spectra of the diamagnetic Yb complexes 2a and 4a provide useful insight into the electronic situation, revealing strongly shielded resonances consistent with rather anionic silyl units. This picture is also supported by DFT calculations, which were carried for complexes 2a−c and 4a−c to elucidate the bonding situation between the Ln(+II) centers and Si. The calculated HOMOs of the complexes resemble very much electron lone pairs at the silicon atoms attached to the metal. The ionic character of the compounds is also exemplified by a reaction of the ytterbacyclopentasilane 2a with zirconocene dichloride, which smoothly proceeded to the respective zirconacyclopentasilane.
■ EXPERIMENTAL SECTION General Remarks. All reactions involving air sensitive compounds were carried out under an atmosphere of dry nitrogen or argon using either Schlenk techniques or a glovebox. All solvents were dried using a column based solvent purification system. 55 Chemicals were obtained from different suppliers and used without further purification.
[Me 2 Si(Me 3 Si) 2 SiK] 2 (1) 40,41 and (Me 3 Si) 3 SiK (3) 35 were prepared following reported procedures. SmI 2 ·(THF) 2 , YbI 2 ·(THF) 2 , and EuI 2 · (THF) 2 were prepared by treatment of the metals in THF with 1,2diiodoethane. 32 Fluorescence Spectroscopy. The fluorescence of solid compounds 4b and EuI 2 (THF) 2 was recorded with a PerkinElmer LS55 fluorescence spectrometer equipped with a Xe flash lamp as the light source. The solid compound was transferred to the instrument's solid sample holder in the glovebox and kept under nitrogen atmosphere until the sample holder was introduced to the sample chamber of the spectrometer.