Using Functionalized Silyl Ligands To Suppress Solvent Coordination to Silyl Lanthanide(II) Complexes.

The reaction of the potassium 1,3-trisilanediide Me2Si[Si(Me3Si)2K]2 with SmI2 and YbI2 was found to give the respective disilylated complexes Me2Si[Si(Me3Si)2]2Sm·2THF and Me2Si[Si(Me3Si)2]2Yb·2THF. Desolvation of coordinated solvent molecules in these complexes made their handling difficult. However, using a number of functionalized silanide ligands, complexes with a diminished number or even no coordinated solvent molecules were obtained ((R3Si)2Ln(THF)x (x = 0-3)). The structures of all new lanthanide compounds were determined by X-ray single-crystal structure analysis. NMR spectroscopic analysis of some Yb-silyl complexes pointed at highly ionic interactions between the silyl ligands and the lanthanides. This bonding picture was supported by DFT calculations at the B3PW91/Basis1 level of theory. Detailed theoretical analysis of a disilylated Eu(II) complex suggests that its singly occupied molecular orbitals (SOMOs) are very close in energy to the ligand silicon lone pairs (HOMO), and SQUID magnetometry measurements of the complex showed a deviation from the expected behavior for a free Eu(II) ion, which might be due to a ligand-metal interaction.

Introduction

Lanthanide complexes with the metal in the +3 oxidation state continue to dominate rare-earth chemistry.[1] The divalent oxidation state for lanthanides was first discovered for their dihalides MX2 in the solid state,[2] and for a long time only three lanthanide elements, samarium, europium, and ytterbium, were readily accessible in the M2+ oxidation state. However, quite recently Evans at al. succeeded in preparing examples of all stable lanthanides in the oxidation state +2.[3,4] An explanation for this unexpected success was given by Evans, who pointed out convincingly that the reduction process Ln(III) → Ln(II) is not necessarily associated with a change in electron configuration from 4f → 4f but can rather be regarded as 4f → 4f5d1.[5] The respective 5d orbital is energetically available because of strong ligand field splitting caused by the C5H4SiMe3 ligand. Our interest in other strong-field ligands led us to study the quite neglected area of silyl lanthanides. Recently we reported a number of lanthanide(+II) complexes bearing oligosilanyl ligands (Figure ).[6]

Figure 1

Oligosilanylated Ln(II) complexes.

Our first excursion into this chemistry led us to study the influence of bidentate versus monodentate silanides in their reactions with ytterbium, europium, and samarium diiodides.[6] The formation of the very air, moisture, and light sensitive compounds occurred in a facile way, and all of the neutral compounds formed shown in Figure were characterized by multinuclear NMR spectroscopy and single-crystal X-ray diffraction analysis. The cyclic silyl lanthanide compounds 2a,b and 3a contain four THF molecules, whereas the acyclic examples 2c,d and 3b bear only three THF molecules.[6] These coordinated solvent units do not only prevent the use of vacuum during the workup procedure but also cause a restriction to THF as a solvent in subsequent reactions in order to ensure a homogeneous product distribution with only one type of coordinated base. To overcome this limitation, we set out to find silyl frameworks that diminish solvent coordination possibilities, thus facilitating further handling and derivatization.

Results and Discussion

Synthesis

Continuing the line of our previous study led us to examine reactions of the 1,3-trisilanediide Me2Si[Si(Me3Si)2K]2 (1)[7,8] with SmI2·2THF and YbI2·2THF, which yielded the cyclic complexes Me2Si[Si(Me3Si)2]2Sm·2THF (2) and Me2Si[Si(Me3Si)2]2Yb·2THF (3) (Scheme ). Like the previous reactions with 1,4-tetrasilanediides, which led to the formation of five-membered rings with the lanthanide atom coordinated by four THF molecules, complexes 2 and 3, which exist as four-membered rings, also contain the same number of coordinated THF molecules. A first attempt to crystallize samaracyclotetrasilane 2 in pentane failed but provided a few crystals of the dimeric samarium complex 4, which were subjected to single-crystal X-ray diffraction analysis (Scheme ). The formation of 4, which bears only two THF units, constitutes a dimerization of 2 caused by the loss of three THF molecules from each molecule of 2. All further efforts to selectively abstract THF molecules from 2 under defined and reproducible conditions to repeat the formation of 4 were unsuccessful.

Scheme 1

Synthesis of Sm(II) (2) and Yb(II) (3) Complexes with a Bidentate 1,3-Trisilanediide Ligand and Formation of Dinuclear Complex 4 by THF Loss of 2

This difficulty encouraged us to find ways of avoiding or at least diminishing coordination of THF molecules to the lanthanide. The simple idea of replacing THF by a more strongly coordinating crown ether was quickly discarded. NMR spectroscopic analysis of the addition of 18-crown-6 to the bis[tris(trimethylsilyl)silyl]ytterbium·3THF[6] complex 3b showed just two detectable products: namely, tetrakis(trimethylsilyl)silane and tris(trimethylsilyl)silane (Scheme ). Therefore, new silyl ligands were designed that would permit the avoidance of donor solvents, such as THF and DME, or the use of crown ethers for the preparation of the required silyl anions.

Scheme 2

Decomposition of Bis[tris(trimethylsilyl)silyl]ytterbium Complex 3b by Addition of 18-crown-6

For the design of such a ligand, two possible strategies are self-evident. The first is to enhance the steric demand of the ligand to such an extent that no further space for coordinating solvent molecules is available. The second option strives to offer additional intramolecular coordination sites to the lanthanide center by incorporating donor atoms into the ligand backbone. The chosen example for a more bulky silyl ligand was pentakis(trimethylsilyl)disilanylpotassium (5).[9] Unfortunately, reaction of 5 with YbI2·2THF did not proceed as expected but led to a palette of products. Upon crystallization, one of the main products, identified by X-ray crystallography, was alkoxysilylytterbium complex 6 (Scheme ). Similar reactions involving ring-opening reactions of THF are known[10] and occur upon coordination of THF to a strongly Lewis acidic site, which renders the α-carbon electron deficient and prone to nucleophilic ring opening. In the solid state the alkoxysilylytterbium species thus formed undergoes a further dimerization step (Scheme ). Since the steric bulk of the substituted butoxy ligand in the vicinity of the metal atom is not high enough, additional THF molecules coordinate to each of the ytterbium atoms.

Scheme 3

Attempted Synthesis of Bis[pentakis(trimethylsilyl)silanyl]ytterbium Yielding Silylalkoxyytterbium Complex 6 and Other Products

The disiloxane [(Me3Si)3SiSiMe2]2O (7) (Scheme ) is easily accessible and was chosen to test the concept of incorporation of an oxygen atom into the silicon backbone to provide additional donor sites to the lanthanide atom. Reaction of 7 with 2 equiv of t-BuOK led to the formation of the respective α,ω-disilanide [K(Me3Si)2SiSiMe2]2O (8), and in subsequent reaction steps with LnI2·2THF (Ln = Yb, Sm) the clean formation of 9 and 10 (Scheme ) was observed. In both compounds the lanthanide is coordinated by the two silanide atoms, three THF molecules, and the oxygen atom within the oligosilane ligand backbone.

Scheme 4

Synthesis of Yb(II) (9) and Sm(II) (10) Complexes with a Tridentate Bis-Silyl Ligand

Several other starting materials (11, 14, 16, and 17) (Scheme ) that might offer possibilities for intramolecular stabilization were further taken into account. α-Alkoxysilyl anions feature silylenoid behavior, making self-condensation a typical reaction.[11,12] The latter can be avoided by addition of crown ether, whereupon α-alkoxysilyl anions can be isolated. Since the presence of crown ether was found to be incompatible with silyl lanthanides, the question arose which alkoxy-substituted silyl anions could be formed without the use of crown ether.

Scheme 5

Attempted Synthesis of an Yb(II) Complex with Two Bidentate 1,4-Silylalkoxy Ligands (Top Reaction) and Failed Synthetic Approaches to Other Alkoxy-Containing Silanides (Bottom Reactions)

Reaction of cyclotetrasilane 12 with t-BuOK was reported to give silanide 11 by a ring-opening attack on a dimethylsilylene unit.[13] Unfortunately, treatment of 11 with YbI2·(THF)2 did not result in formation of the expected Yb complex; instead, back-reaction to 12 and formation of Yb(OBu)2 took place (Scheme ). To find out whether the bulkiness of the tert-butoxy group impedes reaction of 11 with YbI2·(THF)2 to the desired lanthanide complex, compounds 14, 16, and 17 were chosen as starting materials bearing methoxy or methoxymethyl substituents (Scheme ). Unfortunately, reactions of 14 and 16 with t-BuOK did not yield the respective tetrasilanediides but led to complete decomposition of both starting materials in either THF or toluene. Treatment of the methoxy compound 17 with t-BuOK in DME gave an inseparable mixture of the desired silyl anion together with tetrakis(trimethylsilyl)silane and tris(trimethylsilyl)silylpotassium (Scheme ).

These problems caused us to reconsider our approach, and we decided to combine the requirements of bulkiness and a stabilizing atom in the skeleton within one silyl ligand and to put up with the disadvantage that the generation of the silyl anion has to be done in THF. In two reaction steps, ytterbium complex 19 was obtained from tris(trimethylsilyl)methoxysilane 18.[11] The eventual ligand, which is formed by condensation of 2 equiv of (Me3Si)2Si(OMe)K,[11] provides two coordination sites. Despite their bulkiness, the two attached ligands in 19 leave enough space at the Yb atom for the addition of one THF molecule (Scheme ).

Scheme 6

Synthesis of Yb(II) Complex 19 with Two Bidentate 1,3-Silylalkoxy Ligands

Encouraged by this result, we decided to use a simple silocane as our next ligand. Silocanes, 1,3-dioxa-6-aza-2-silacyclooctanes, contain a Si–(O–C–C−)2N fragment, which we hoped would offer further coordination possibilities for the lanthanide atom. Chlorosilocane 20 (Scheme ) was obtained by reaction of MeSiCl3 with MeN(CH2CH2OSiMe3)2. Treatment of 20 with tris(trimethylsilyl)silylpotassium·DME afforded 21 in a clean reaction (Scheme ). The formation of the respective oligosilocanylsilylpotassium 21a was achieved in DME. For the next reaction steps with LnI2·(THF)2 (Ln = Yb, Eu) the solvent was exchanged for toluene. In both the Yb complex 22 and the Eu complex 23 two silocanyl units serve as ligands. The lanthanides are coordinated by the two silanides, two nitrogen atoms, and two oxygen atoms without any further solvent molecule present.

Scheme 7

Synthesis of a Bis(trimethylsilyl)silocanylsilanide and Its Subsequent Use as a Tridentate Ligand in Complexes 22 and 23

NMR Spectroscopy

In addition to the previously described disilylated Ln(II) complexes 2a,c and 3a,b,[6] in the current study NMR spectroscopic data of 2, 3, 9, 10, 19, and 22 were obtained. The majority of these complexes are Yb compounds (3a,b, 3, 9, 19, and 22), which are diamagnetic and therefore well suited for NMR spectroscopic analysis.

1H and 13C NMR spectra of the Yb complexes display the expected ligand signals. All 1H spectra display broadened lines indicating dynamic coordination–decoordination behavior. For complex 3 the spectra clearly show decomposition upon removal of THF.

With respect to the silyl ligand–Yb(II) interaction 29Si NMR spectroscopy provides the most intimate insight. Complexes 3a,b were found to exhibit 29Si NMR signals at −158.4 and −144.8 ppm, respectively.[6] These chemical shifts indicate a silanide character of the silyl ligand. Related magnesium silanides exhibit somewhat more shielded signals,[14] whereas analogous oligosilanyl zinc compounds[15] resonate at slightly lower field. This NMR behavior is not without precedence. The tris(trimethylsilyl)silylytterbium complex Cp*YbSi(SiMe3)3·(THF)2, described by Lawless and co-workers,[16] showed a 29Si NMR signal at −158 ppm. A signal with a similar chemical shift at −148.6 ppm was reported by Niemeyer for the ate complex K[(Me3Si)3SiYb{N(SiMe3)2}2].[17] Most recently Krempner and co-workers[18] described the ytterbium silanide [Si(SiMe2OMe)3-κ3]2Yb, where the interaction of the ligand with the Yb atom is only occurring via the OMe groups and no interaction between the silanide atom and the lanthanide is observed. The respective 29Si NMR signal for this compound was found at −202.9 ppm. The latter value likely might be regarded as typical for a free silanide.

When the compounds of the current study are added to this analysis, the two values of −136.9 ppm and even −134.7 ppm for compounds 3 and 19, respectively, seem to indicate a less ionic bonding situation for these cases. However, in both of these compounds the silicon atom in question is incorporated into a four-membered ring. Such an arrangement is known to cause a downfield shift of the resonances of the ring silicon atoms. The chemical shift of the magnesium analogue of 3 is at −137.3 ppm,[14] indicating a very similar polarity of the respective Si–Yb and Si–Mg bonds. The 29Si chemical shift of compound 9 (−153.8 ppm) is comparable to that of 3a, whereas the −182.0 ppm value of the silocan complex 22 are definitely rather strongly shielded into a spectroscopic range where usually potassium or lithium silanides are found.[19] This unusual shift seems to be caused by the geometry of the ligand, which by its polydentate nature forces the silanide atom to coordinate in a distorted way with diminished directionality of the silanide lone pair and the Yb atom.

For the aforementioned neutral complex Cp*YbSi(SiMe3)3·(THF)2 a 1JSi–Yb coupling constant of 829 Hz[16] was found, and for K[(Me3Si)3SiYb{N(SiMe3)2}2] a value of 716 Hz[17] was reported. For the acyclic complex 3b a similar value of 732 Hz was found, and for a DME complex of 3a a coupling of 656 Hz was reported.[6] Formation of 3a in C6D6 with an approximate THF concentration of 9 equiv per Yb atom gave a 29Si NMR spectrum with a diminished 1JYb–Si coupling constant of 633 Hz.[6] For reasons not yet completely clear a 1JYb–Si coupling constant cannot always be observed. In the current study we were only able to detect it for complex 9, where it amounted to a value of 531 Hz. This decreased value certainly hints at a diminished bonding interaction.

Two-dimensional NMR correlation spectroscopy, in particular 1H29Si-gHMBC, proved to be a good tool for the assignment of signals. Especially for weakly paramagnetic complexes such as the samarium compounds 2 and 10, this allows obtaining meaningful 29Si NMR data. Using 1H29Si-gHMBC spectroscopy, resonances for the SiMe3 groups of 2 and 10 were found at −119.3 and −118.1 ppm, respectively. The respective silanide resonances were detected at 117.9 and 117.2 ppm. The spectra did not show signals for the SiMe2 units of either 2 or 10.

As it is strongly paramagnetic, for the Eu complex 23 no meaningful NMR spectra could be obtained at all. Nevertheless, on the basis of structural data and computational results it is reasonable to state very similar bonding situations for isostructural Yb, Sm, and Eu complexes. Therefore, the results derived from NMR spectroscopic analysis of Yb complexes, which indicate a highly ionic bonding interaction between the silanide ligands and the respective Ln(II) ions, should be valid also for Sm and Eu complexes.

Crystal Structure Analysis

Samarium complex 2 crystallizes in the monoclinic space group P21/c, with four THF molecules coordinating to the Sm atom (Figure ). Unexpectedly, the related Yb compound 3 (Figure S32 in the Supporting Information) crystallizes in the orthorhombic space group Pna21. The reason for this can be seen in Figure and is a result of the different arrangement of the molecules within the crystal. In both structures, no unusual short inter- or intramolecular contacts are found. The four-membered ring in 2 is folded by 12.65° between the plane fixed by the three Si atoms and the plane fixed by Si–Sm–Si atoms. In 3 this angle is diminished to 10.75°.

Figure 2

Molecular structure of 2 (thermal ellipsoid plot drawn at the 30% probability level). All hydrogen atoms are omitted for clarity. Selected bond lengths (Å) and angles (deg): Sm(1)–O(2) 2.561(3), Sm(1)–O(4) 2.567(3), Sm(1)–O(1) 2.587(3), Sm(1)–O(3) 2.627(2), Sm(1)–Si(3) 3.1746(11), Sm(1)–Si(1) 3.1813(11), Si(1)–Si(5) 2.3382(16), Si(1)–Si(4) 2.3397(15), Si(1)–Si(2) 2.3833(14); Si(3)–Sm(1)–Si(1) 74.66(3).

Figure 3

(left) Packing of Sm compound 2 along the a axis. (right) Packing of Yb compound 3 along the b axis.

The dinuclear samarium complex 4 (Figure ) consists of three annulated four-membered rings with the mirror plane through the central ring. This central ring is planar; the others are folded by 41.90° between the plane fixed by the three Si atoms and that fixed by Si–Sm–Si atoms.

Figure 4

Molecular structure of 4 (thermal ellipsoid plot drawn at the 30% probability level). All hydrogen atoms are omitted for clarity. Selected bond lengths (Å) and angles (deg): Sm(1)–O(1) 2.5044(15), Sm(1)–Si(1) 3.0864(8), Sm(1)–Si(3) 3.2010(8), Sm(1)–Si(3a) 3.2931(7), Sm(1)–Sm(1a) 3.8898(9), Si(1)–Si(4) 2.3306(8), Si(1)–Si(5) 2.3413(8), Si(1)–Si(2) 2.3604(10), Si(2)–C(2) 1.891(2), Si(2)–C(1) 1.906(2), Si(2)–Si(3) 2.3963(8), Si(3)–Si(6) 2.3555(10), Si(3)–Si(7) 2.3601(9); Si(1)–Sm(1)–Si(3) 76.428(18), Si(1)–Sm(1)–Sm(1a) 104.330(12), Si(3)–Sm(1)–Sm(1a) 54.297(16), Si(3a)–Sm(1)–Sm(1a) 52.126(17), Si(4)–Si(1)–Sm(1) 107.25(2), Si(5)–Si(1)–Sm(1) 143.14(3), Si(2)–Si(1)–Sm(1) 81.72(2), Sm(1)–Si(3)–Sm(1a) 73.58(2).

Alkoxysilylytterbium complex 6 (Figure ) crystallizes in the orthorhombic space group Pbca with an additional THF molecule in the asymmetric unit. The latter molecule is so seriously disordered that after refinement no meaningful geometry and no unambiguous atom assignment were possible; therefore, the SQUEEZE procedure[20] was applied. Several independent parts of complex 6 are disordered: namely, one trimethylsilyl group, two carbon atoms of the butylene units, and the ytterbium atoms. The four-membered ring consisting of alternate ytterbium and oxygen atoms is planar with a Yb–O bond distance of 2.246 Å (Table ). This short Yb–O bond shows explicitly the difference between coordinated oxygen atoms of THF or the ligands and ionically bonded atoms such as in 6 (Table ). The distance between the two Yb atoms amounts to 3.482 Å. A related dinuclear Yb(III) complex [Cp2YbOCH2CH=CHCH3]2[21] features a comparable distance of 3.536 Å.

Figure 5

Molecular structure of 6 (thermal ellipsoid plot drawn at the 30% probability level). All hydrogen atoms are omitted for clarity. Selected bond lengths (Å) and angles (deg): Yb(1)–O(1) 2.371(8), Yb(1)–O(2) 2.248(5), Yb(1)–Si(1) 3.015(3), Yb(1)-Yb(1A) 3.4825(9), Si(1)–Si(2) 2.322(3), Si(1)–Si(4) 2.381(3), Si(2)–C(1) 1.848(8); O(2)–Yb(1)–Si(1) 118.67(18), O(1)–Yb(1)–C(36) 80.6(3).

Table 1

Compilation of Structural Data Derived by Single-Crystal XRD Analysis of Complexes 2–4, 6, 9, 10, 19, 22, and 23 and the Related Compounds 2a–d and 3a,b

Si–Sm (Å)	Sm–O (Å)	Si–Yb (Å)	Yb–O (Å)	Yb–N (Å)	Si–Eu (Å)	Eu–O (Å)	Eu–N (Å)
2: 3.175, 3.181	2: THF, 2.561– 2.627	3: 3.057, 3.066	3: THF, 2.437–2.492
2a:[6] 3.213, 3.244	2a:[6] THF, 2.544– 2.613	3a:[6] 3.106, 3.171	3a:[6] THF, 2.433–2.485		2b:[6] 3.205, 3.216	2b:[6] THF, 2.546–2.613
2c:[6] 3.172	2c:[6] THF, 2.551– 2.528	3b:[6] 3.064	3b:[6] THF, 2.421–2.435		2d:[6] 3.150	2d:[6] THF, 2.515- 2.524
4: 3.086, 3.201 3.293	4: THF, 2.504	6: 3.015	6: THF, 2.371; ionic, 2.248
10: 3.183, 3.147	10: THF, 2.558–2.568; intra, 2.588	9: 3.053, 3.086	9: THF, 2.434–2.460; intra, 2.471
		19: 3.047, 3.063	19: THF, 2.364; intra, 2.365, 2.377
		22: 3.050	22: intra, 2.491	22: 2.705	23: 3.157	23: intra, 2.602	23: 2.765

Samarium complex 10 (Figure S34 in the Supporting Information) and the related Yb compound 9 (Figure ) both crystallize in the monoclinic space group P21/c with nearly identical cell parameters. The two annulated four-membered rings in 9 and 10 are only slightly bent toward each other, engaging an angle of 5.2° in 10 and 6.0° in 9. The four-membered rings are almost planar, with one atom out of planarity by 0.334 and 0.819 Å in 10 and 0.354 and 0.849 Å in 9.

Figure 6

Molecular structure of 9 (thermal ellipsoid plot drawn at the 30% probability level). All hydrogen atoms are omitted for clarity. Selected bond lengths (Å) and angles (deg): Yb(1)–O(2) 2.4350(16), Yb(1)–O(4) 2.4354(17), Yb(1)–O(3) 2.4598(18), Yb(1)–O(1) 2.4707(15), Yb(1)–Si(2) 3.0528(9), Yb(1)–Si(6) 3.0860(8), Si(1)–C(1) 1.889(3), Si(1)–Si(2) 2.3415(10); O(2)–Yb(1)–O(1) 86.87(6), O(4)–Yb(1)–O(1) 124.26(6), O(3)–Yb(1)–O(1) 156.64(6), O(1)–Yb(1)–Si(2) 68.98(4), O(1)–Yb(1)–Si(6) 68.60(4), Si(2)–Yb(1)–Si(6) 133.95(2), O(1)–Yb(1)–Si(5) 28.34(4), Si(2)–Yb(1)–Si(5) 97.318(18), Si(6)–Yb(1)–Si(5) 41.987(16), Si(2)–Yb(1)–Si(4) 41.760(16), Si(6)–Yb(1)–Si(4) 95.64(2), Si(5)–Yb(1)–Si(4) 55.770(19).

The disiloxane 7 (Figure S33 in the Supporting Information), which serves as the starting material for 9 and 10, crystallizes in the monoclinic space group P21. The SiMe2–O–SiMe2 unit and two SiMe3 groups show disorder, which can be seen in the number of restraints used. The Si–O distances in 7 are 1.628 and 1.645 Å, which under the influence of the coordinating lanthanide in 9 and 10 are elongated to 1.690 and 1.697 Å.

There are more than 350 compounds in the CCDC containing at least one Sm–THF interaction, and the average Sm–O bond length is 2.51 Å. With distances between 2.50 Å for 4 and 2.63 Å for 2 (Table ) compounds 2, 4, and 10 fit well into the range of observed values. The distance of the intramolecular interaction of Sm with the oxygen atom of the ligand in 10 is, at 2.588 Å, in the same range as the Sm–THF distances. Differences can be seen in the Sm–Si distances: 3.1746 and 3.1813 Å for 2, 3.0864, 3.2010, and 3.2931 Å for 4, and 3.183 and 3.147 Å for 10 (Table ). These values are in good agreements with distances in our previously reported Sm(II)–silyl complexes 2a,c[6] (Figure and Table ) and other reported examples.[22,23] Nevertheless, for all these compounds the bond lengths are slightly longer than the sum of the covalent radii (Sm–Si 3.09 Å).[24] Only in 4 is the distance of the tetracoordinated Si(1) to Sm(1), at 3.0864 Å, within the range of the covalent radii and is considerably shorter than those of pentacoordinated Si(3) to Sm(1a) at 3.2010 Å and Si(3a) to Sm(1) at 3.2931 Å. The complexes known in the literature with divalent samarium are mainly mononuclear, and only a few dinuclear complexes have been reported so far.[25] The Sm–Sm distance varies from within the bonding range of 3.3159 Å[25] or 3.552 Å[26] up to 4.195 Å,[27−30] far from a bonding distance considering the Sm–Sm atom distances in the metal being 3.588 and 3.620 Å.[31] In complex 4 a Sm–Sm distance of 3.8898 Å was found and some interaction might be possible.

Spirocyclic ytterbium compound 19 crystallizes in the triclinic space group P1̅ (Figure ). The two four-membered rings are almost planar with Si(7) 0.311 Å and Si(1) 0.334 Å out of planarity. Further, the two rings are engaging an angle of 142.7° and thus provide space for the additional THF molecule.

Figure 7

Molecular structure of 19 (thermal ellipsoid plot drawn at the 30% probability level). All hydrogen atoms are omitted for clarity. Selected bond lengths (Å) and angles (deg): Yb(1)–O(3) 2.364(4), Yb(1)–O(1) 2.365(4), Yb(1)–O(2) 2.377(4), Yb(1)–Si(8) 3.0469(17), Yb(1)–Si(2) 3.0627(17), Si(1)–O(1) 1.750(4), Si(1)–Si(2) 2.361(2), Si(3)–C(5) 1.863(7), Si(7)–O(2) 1.749(4); O(1)–Yb(1)–O(2) 143.23(14), O(1)–Yb(1)–Si(8) 104.42(10), O(2)–Yb(1)–Si(8) 68.59(10), O(1)–Yb(1)–Si(2) 69.05(9), O(2)–Yb(1)–Si(2) 105.62(10), Si(8)–Yb(1)–Si(2) 161.35(5), O(1)–Yb(1)–Si(1) 27.85(9), O(2)–Yb(1)–Si(1) 139.25(10), Si(8)–Yb(1)–Si(1) 132.05(4), Si(2)–Yb(1)–Si(1) 42.01(4), O(1)–Yb(1)–Si(7) 137.95(10), O(2)–Yb(1)–Si(7) 27.55(9), Si(8)–Yb(1)–Si(7) 41.75(4), Si(2)–Yb(1)–Si(7) 132.86(4), Si(1)–Yb(1)–Si(7) 155.22(4).

Silocanyl chloride 20 (Figure S35 in the Supporting Information) crystallizes in the monoclinic space group P21/n. The Si–N distance in 20 is 2.117 Å and thus is essentially longer than the 2.023 Å in a related chlorosilatrane.[32] Nevertheless, this distance is within the range (2.0–2.2 Å) of a transannular interaction between Si and N.[33] No interaction between Si and N is observed in 22 (Figure S36 in the Supporting Information) and 23 (Figure ), where the Si–N distance is 3.711 Å for 22 and 3.709 Å for 23 due to the influence of the Ln–N coordination. Both Ln–Si complexes crystallize in monoclinic space groups, 22 in C2 and 23 in C2/c. Here again, as was mentioned for 2 and 3, the differences in the cell parameters arise from the arrangement within the crystal packing (Figure S37 in the Supporting Information). Only a couple of structures bearing Yb(II)–Si bonds can be found in the literature,[6,16,17,34−37] but these distances are in good agreement with those found for 3, 9, 19, and 22 (Table ). For Eu(II)–Si only three examples can be found and again distances of compound 23 compare well to these.[6,37]

Figure 8

Molecular structure of 23 (thermal ellipsoid plot drawn at the 30% probability level). All hydrogen atoms are omitted for clarity. Selected bond lengths (Å) and angles (deg): Eu(1)–O(2) 2.602(4), Eu(1)–N(1) 2.765(5), Eu(1)–Si(1) 3.157(2), Si(1)–Si(2) 2.342(3), Si(4)–O(1) 1.659(5), Si(4)–O(2) 1.721(5), Si(2)–C(1) 1.881(7); O(2)–Eu(1)–O(2A) 143.80(19), O(2)–Eu(1)–N(1) 148.91(14), O(2)–Eu(1)–Si(1) 65.43(10), N(1)–Eu(1)–Si(1) 99.05(12), Si(1)–Eu(1)–Si(4A) 129.10(5), O(2)–Eu(1)–Si(4) 28.28(10), N(1)–Eu(1)–Si(4) 124.82(11).

Theoretical Studies

The electronic structures of 2 and 4 were calculated at the B3PW91/Basis1 level of theory. Previously we proved that this methodology yielded an adequate description of the electronic structure of f-block-silyl complexes.[38]Table contains the calculated average Sm–Si bond lengths (Å), Mayer bond order (MBO) of Sm–Si bonds, Natural Population Analysis charges, and HOMO energies (eV) for 2 and 4. We found excellent agreement between computed and measured bond lengths (see Tables and 2), which confirmed again that our chosen method is appropriate. The electronic structures of 2 and 4 are similar to each other in the sense that all samarium centers have six–six unpaired f electrons (S = 3) in all complexes while the silyl groups bear an additional electron, best described as silyl anions. Compound 4, the dimerization product of 2, contains two samarium centers 3.90 Å from each other. We found that the ground state is a high-spin system (S = 6); no lower energy broken-symmetry states were found. This suggests that the samarium centers have direct exchange interaction in spite of the relatively large distance.

Table 2

Calculated Average Ln–Si Bond Lengths (Å), Mayer Bond Order (MBO) of Ln–Si Bonds, Natural Population Analysis Charges, and HOMO Energies (eV) for 2, 4, and 23 at the B3PW91/Basis1 Level of theory

compound	Sm–Si bond distance (Å)	MBO of Sm–Si bond	NPA charge (Sm/Si)	HOMO energy (eV)
2	3.180	0.36	+1.37/–0.74	–3.11
4	3.080	0.61	+1.25/–0.80
	3.204	0.35	–0.72	–3.82
	3.297	0.21
23	3.157	0.25	+1.64/–0.88	–3.44
	3.493	0.04

The Si–Sm bond lengths in 2 and 4 (3.17–3.20 Å) are longer than those of our previously characterized relatively covalent f-block-silyl structures[6,38] that also suggest strong ionic character. Natural population analysis (NPA) also revealed strong ionic character of Sm–Si bonds; silicon atoms possess almost a clear extra electron (Table ), as NPA charges ranged between −0.72 and −0.80, while samarium centers have large positive charge. Mayer bond orders (MBO) are relatively low (∼0.3), similar to the case for our previous Sm–silylene complexes,[38] which also indicates negligible covalent character. HOMO orbitals underline this analysis (Figures and 10), as they resemble lone pairs situated on silicon centers with no extension toward the samarium centers.

Figure 9

HOMO orbital of 2 (−3.48 eV), calculated at the B3PW91/Basis1 level of theory (isovalue 0.02). White, gray, blue, red, and teal colors refer to hydrogen, carbon, oxygen, samarium, and silicon atoms, respectively.

Figure 10

HOMO (left, −3.82 eV) and HOMO-2 (right, −4.65 eV) orbitals of 4 calculated at the B3PW91/Basis1 level of theory (isovalue 0.02). White, gray, blue, red, and teal colors refer to hydrogen, carbon, oxygen, samarium, and silicon atoms, respectively.

We also calculated the electronic structure of 23 at the B3PW91/Basis1 level of theory as for the previous f-block complexes. Complex 23 was found to possess seven unpaired electrons (S = 7/2), all occupying f orbitals. These singly occupied molecular orbitals (SOMOs) are very close in energy to each other in the range of −2.60 to −3.31 eV. Interestingly, very close to the unpaired f electrons in energy there are two doubly occupied molecular orbitals (formally denoted as the HOMO of 23, Table ) of the silicon lone pairs, −3.75 and −3.44 eV, respectively. This small energy difference between the lowest lying SOMO and HOMO might be one reason for the peculiar magnetism of 23. Finite temperature effects or crystal-packing effects can overcome the only 0.12 eV difference, and the error of DFT can be very large in strong correlation problems. Nevertheless, none of our efforts using other DFT functionals resulted in a ground state solution that could explain the magnetic properties of 23.

HOMO orbitals show very clear silyl anion character (Figure ), which is in agreement with the very large NPA negative charge (−0.88) on the silicon centers that is even larger than those of 2 and 4. Together with very low Mayer bond order values (0.25), these results all indicate the highest ionic character that we have observed in our f-block silyl complexes that can explain the very high energy of the silyl anion orbitals close to the f orbitals.

Figure 11

HOMO (left, −3.44 eV) and HOMO-2 (right, −3.75 eV) orbitals of 23 calculated at the B3PW91/Basis1 level of theory (isovalue 0.02). White, gray, blue, red, and teal colors refer to hydrogen, carbon, oxygen, samarium, and silicon atoms, respectively.

SQUID Magnetometry on Silocane–Eu Complex 23

Among the rare-earth elements, europium in the valence state Eu(II) plays an exceptional role due to the compensation of orbital momentum (L = 0) in the free-ion state 8S7/2 providing the utmost spin momentum of seven unpaired spins (S = 7/2). The absence of orbital momentum makes Eu(II) unsusceptible to spin–orbit coupling and diminishes ligand field effects. The first excited state (6P7/2) is located at some 30.000 cm–1 above the ground state energy.[39] Therefore, at ambient temperature only the ground state becomes mainly populated, and its zero-field splitting is very weak, on the order of 10–2 cm–1, only detectable in the temperature range of sub-Kelvin. From these findings of the free-ion Eu(II) state the magnetic susceptibility is almost perfectly isotropic and follows the paramagnetic Curie law for S = 7/2 (Landé g factor equal to 2) (eq ).

The fundamental constants are Avogadro’s number NA = 6.022 × 1023/mol, Bohr magneton μB = 9.274 × 10–21 emu, and Boltzmann’s constant kB = 1.38 × 10–23 J/K. The Curie law holds for small magnetic fields and high temperatures. We measured the Curie constant χmolT [cm3 K mol–1] of the silocane–Eu compound 23 at very high magnetic fields (B = 1, 7 T) down to low temperatures T → 2.7 K, for which the Curie law has to be modified, as shown in eq where B(x) is the Brillouin-function for spin S as depicted in eq .

The top blue solid line in Figure a displays the calculated χmolT according to eq as a function of temperature for B = 1 T. Clearly, this curve approaches the high-temperature value 7.88 cm3K mol–1 as calculated by eq . It shows the characteristic turn-down at low temperature due to saturation. Interestingly, the measured molar Curie constant of the Eu complex 23 (blue crosses of Figure ) exhibits a deviation from the curvature at low temperature T < 100 K. This deviation from the free-ion magnetic state becomes less pronounced for higher magnetic field (7 T), as shown by the red graphs of Figure a. Figure b shows an enlarged view in the temperature range 20–100 K.

Figure 12

Calculated molar Curie constants χmolT of the “free ion” Eu(II), in comparison with SQUID measurements of 23 (marked as crosses) represented by the blue graphs (for B = μ0H = 1 T) and by the red graphs (for B = μ0H = 7 T): (a) the common Curie constant of 7.88 cm3 K mol–1 in the limit at high temperature (300 K); (b) detailed view in the reduced temperature range of 20–100 K, in which the measurement deviates from the free-ion S = 7/2 spin state.

As mentioned already, the ground state level of isolated Eu(II) is not split. However, in the silocane–Eu complex 23 the influence of ligands and of crystallinity could change the situation. It is out of the scope of this paper to recover these effects or the influence of a possible charge transfer.

Figure shows the field-dependent molar magnetization for four temperatures (T = 2.7, 6, 10, 20 K). The fits (solid lines) approximate the measured data (dots) only if 88% of Eu is in the “pure” S = 7/2 spin state, what again implicates that not all Eu ions can be represented by “free” ions.

Figure 13

Isothermal magnetization curves of Eu complex 23 for various temperatures. The measured data (dots) are best fitted (solid lines) by 88% of the “pure” 7/2 spin magnetization of Eu(II).

As a result of the magnetic measurements we conclude that an anticipated (S = 7/2) state of seven unpaired and unsplit (zero-field) spin states in Eu(II) is superficial but does not explain in detail the experimental data in complex 23 at reduced temperatures. DFT data (vide supra) suggest a moderate splitting of 0.12 eV between SOMO and HOMO molecular orbitals with a corresponding influence on the Eu spin state. In any case, all the Eu atoms cannot be in the S = 7/2 pure spin states, as demonstrated.

Conclusion

The current work continues the study of silylated lanthanides. Reaction of a potassium 1,3-trisilandiide with SmI2 and YbI2 was found to give disilylated complexes of samarium (2) and ytterbium (3). As the bite angle of the oligosilanediyl ligand is small, four THF molecules coordinate to the Ln(II) metal atoms. Dissociation of THF is a facile process, complicating isolation of homogeneous substances. In the case of samarium complex 3 it was possible to isolate the defined dimerization product 4, which is a dinuclear complex where dissociated THF molecules are replaced by the silanide units of the neighboring complex.

To facilitate isolation and handling of silyl lanthanide complexes, we set out to modify the silyl ligands’ coordinative properties by incorporating additional donor atoms which serve as Lewis basic sites. Using this approach, it was possible to diminish the coordinating THF molecules to one or even none. While the obtained complexes are still very sensitive toward light and ambient conditions, the ligand modifications diminish the chance of dissociation of coordinated solvent molecules and thus increase the stability of the complex.

Due to the paramagnetism of unpaired f electrons, NMR spectroscopy of lanthanide complexes is frequently difficult. Diamagnetic Yb(II) complexes with a completely filled f shell are an exception to this and provide valuable insight into the electronic properties. 29Si NMR spectroscopic analysis of the Yb–silyl complexes studied in this work reemphasizes the already previously gained insight that the nature of the Si–Ln bond in these complexes is highly ionic. The strongly downfield shifted resonances of the metal-attached silicon atoms indicate high silanide character. This is particularly true for the Yb complex with the bis(trimethylsilyl)silocanylsilyl ligand, for which a chemical shift of −182.0 ppm was observed.

The NMR analysis was supported by DFT calculations, which show that the HOMOs of the complexes consist mainly of silicon lone pairs without extension toward the lanthanide centers. According to natural population analysis (NPA) the silicon atoms possess almost a clear extra electron while the lanthanide centers have a large positive charge. In addition, the Mayer bond orders (MBOs) are relatively low (∼0.3). Calculation of the electronic structure of europium complex 23 confirmed the expected half-filled shell of seven unpaired electrons (S = 3/2). It is interesting to note that these singly occupied molecular orbitals (SOMOs), which are very close in energy to each other in the range of −2.60 to −3.31 eV, are also very close to the two doubly occupied molecular orbitals (HOMOs) of the silicon lone pairs, −3.75 and −3.44 eV, respectively.

With its seven unpaired electrons europium(II) complex 23 is also interesting with respect to its magnetic properties. Using SQUID magnetometry measurements the Curie constant χmolT [cm3 K mol–1] of 23 at very high magnetic fields (B = 1, 7 T) down to low temperatures T → 2.7 K was determined. Interestingly, molar Curie constants χmol of the “free” Eu(II) ion, in comparison with the SQUID measurements of 23 show that in a plot of χmol vs T the observed curve approaches the expected high temperature value of 7.88 cm3 K mol–1. However, the measured molar Curie constant of 23 shows a deviation from the expected curvature at low temperature T < 100 K which becomes less pronounced for higher magnetic field (7 T). The origin of this behavior is not yet clear but may be caused by the close proximity of the lowest SOMO and the HOMO of 23.

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.[40] 1,3-Dichloro-1,1,3,3-tetramethyldisiloxane, MeSiCl3, and other chemicals used were obtained from different suppliers and used without further purification. K-(Me3Si)2Si-SiMe2-Si(SiMe3)2-K (1),[7,8] (Me3Si)3Si-(Me3Si)2Si-K (5),[9]BuO(Me2Si)2-(Me3Si)2Si-(Me3Si)2Si-K (11),[13] (Me3Si)2SiH-(SiMe2)2-SiH(SiMe3)2 (13),[8] K-(Me3Si)2Si-(SiMe2)2-Si(SiMe3)2-K (15),[8] (Me3Si)3SiOMe (18),[11] MeN(CH2CH2OSiMe3)2,[41] (Me3Si)3Si-Me2Si-SiMe2Cl,[42] and (Me3Si)3SiK[9] were prepared following reported procedures. SmI2·2THF, YbI2·2THF, and EuI2·2THF were prepared by treatment of the metals in THF with 1,2-diiodoethane.[43−45]

1H (300 MHz), 13C (75.4 MHz), and 29Si (59.3 MHz) NMR spectra were recorded on a Varian INOVA 300 spectrometer. If not noted otherwise, all samples were measured in C6D6. To compensate for the low isotopic abundance of 29Si, the INEPT pulse sequence was used for the amplification of the signal.[46,47] Elementary analyses were carried out using a Heraeus VARIO ELEMENTAR instrument.

X-ray Structure Determination

For X-ray structure analyses the crystals were mounted onto the tip of glass fibers, and data collection was performed with a BRUKER-AXS SMART APEX CCD diffractometer using graphite-monochromated Mo Kα radiation (0.71073 Å). The data were reduced to Fo2 and corrected for absorption effects with SAINT[48] and SADABS,[49,50] respectively. The structures were solved by direct methods and refined by full-matrix least-squares methods (SHELXL97).[51] If not noted otherwise, all non-hydrogen atoms were refined with anisotropic displacement parameters. All hydrogen atoms were located in calculated positions to correspond to standard bond lengths and angles. Crystallographic data (excluding structure factors) for the structures of compounds 2–4, 6, 7, 9, 10, 19, 20, 22, and 23 reported in this paper have been deposited with the Cambridge Crystallographic Data Center as supplementary publication nos. CCDC 1026160 (2), 1509714 (3), 1026161 (4), 1526976 (6), 1509720 (7), 1509719 (9), 1509718 (10), 1509716 (19), 1509721 (20), 1509715 (22), and 1509717 (23). Copies of data can be obtained free of charge at http://www.ccdc.cam.ac.uk/products/csd/request/. Figures of solid-state molecular structures were generated using Ortep-3 as implemented in WINGX[52] and rendered using POV-Ray 3.6.[53]

Me2Si[Si(Me3Si)2]2Sm·2THF (2)

To a solution of SmI2·2THF (288 mg, 0.525 mmol) in THF (10 mL) was added freshly prepared silanide 1 (0.500 mmol) in THF (8 mL) dropwise. The addition went along with a color change to dark green and formation of a gray precipitate. The mixture was stirred for 3 h followed by concentration of the volume to 1 mL under reduced pressure. The dark green residue was extracted with toluene (5 × 4 mL). The extract was concentrated to 15 mL and stored at −35 °C for 3 days, affording green crystals of 2 (192 mg, 45%). Mp: 167–169 °C. 1H NMR (δ in ppm): 3.17 (s, 16H, THF), 1.34 (bs, 36H, SiMe3), 0.30 (s, 16H, THF), −1.16 (s, 4H, SiMe2). 13C NMR (δ in ppm): 119.4 (THF), 25.0 (THF), SiMe2 and SiMe3 signals not detected. 29Si NMR (δ in ppm): no signals detected. 1H NMR (δ in ppm, d8-THF): 3.58 (s, 16H, THF), 1.69 (s, 16H, THF), −0.12 (s, 36H, SiMe3), −1.07 (s, 6H, SiMe2). 13C NMR (δ in ppm, d8-THF): 70.6 (THF), 26.8 (THF), 23.4 (SiMe2), −18.1 (SiMe3). 29Si NMR (δ in ppm, d8-THF): 117.9 (Siq), SiMe2 not detected, −119.3 (SiMe3). Anal. Calcd for C30H74O4Si7Sm [846.32]: C, 42.60; H, 8.82. Found: C, 42.38; H, 8.37. UV absorption: λ1 242 nm (ε1 = 3.5 × 104 M–1 cm–1).

Me2Si[Si(Me3Si)2]2Yb·2THF (3)

To a solution of YbI2·(THF)2 (83 mg, 0.14 mmol) in toluene (4 mL) was added freshly prepared silanide 1 (0.14 mmol) in toluene (8 mL) dropwise under strict exclusion of light. The addition went along with a color change to orange. The mixture was stirred for 15 min, after which the solvent was removed. The residue was extracted with pentane (3 × 2 mL) and the extract filtered over Celite. The solvent was reduced to a volume of 2 mL, and a gray precipitate emerged which was dissolved by adding a few drops of THF. This solution was stored for 1 week at −37 °C, affording red crystals of 3 (47 mg, 37%). Mp: 139–140 °C. 1H NMR (δ in ppm): 3.73 (bs, THF), 1.58 (bs, THF), 1.17 0.59 (s, 6H, SiMe2), 0.74 (s, 36H, SiMe3). 13C NMR (δ in ppm): 68.2 (THF), 25.6 (THF), 7.9 (SiMe3), 3.2 (SiMe2). 29Si NMR (δ in ppm): −4.6 (2JYb–Si = 19.4 Hz, SiMe3), −32.7 (SiMe2), −136.9 (Siq).

Dinuclear Samarium Complex 4

Crystalline 2 (10 mg) was dissolved in pentane. Crystals suitable for X-ray analysis formed after several weeks at room temperature. Mp: 170–173 °C.

Reaction of Pentakis(trimethylsilyl)disilanylpotassium (5) with YbI2·(THF)2

To a solution of YbI2·(THF)2 (98 mg, 0.17 mmol) in toluene (2 mL) was added freshly prepared silanide 5 in toluene (4 mL) dropwise under strict exclusion of light. (5 was obtained from hexakis(trimethylsilyl)silane (170 mg, 0.34 mmol) and KOBu (40 mg, 0.36 mmol) in THF (1 mL). After complete conversion THF was removed and replaced by toluene.) Further treatment was as reported for 3. After crystallization for 2 weeks at −37 °C out of pentane a few crystals of 6 (mp: 233–234 °C) were obtained. NMR measurements showed a number of different unidentified products.

1,3-Bis[tris(trimethylsilyl)silyl]-1,1,3,3-tetramethyldisiloxane (7)

To a solution of 1,3-dichlorotetramethyldisiloxane (10.5 g, 51.7 mmol) in toluene (250 mL) were added tris(trimethylsilyl)silylpotassium (obtained from tetrakis(trimethylsilyl)silane (30.0 g, 93.5 mmol) and KOBu (11.0 g, 98.0 mmol)) in toluene (100 mL) dropwise. After it was stirred for 1 h, the reaction mixture was transferred to 0.5 M H2SO4/ice (300 mL) and extracted with diethyl ether (3 × 150 mL). The organic layers were dried over Na2SO4, and the solvent was removed. After recrystallization from diethyl ether colorless 7 (28.5 g, 97%) was obtained. Mp: 183–185 °C. 1H NMR (δ in ppm): 0.53 (s, 12H, SiMe2), 0.33 (s, 54H, SiMe3). 13C NMR (δ in ppm): 8.7 (SiMe2), 3.1 (SiMe3). 29Si NMR (δ in ppm): 13.4 (SiMe2), −10.5 (SiMe3), −132.8 (Siq).

1,3-Bis[potassiobis(trimethylsilyl)silyl]-1,1,3,3-tetramethyldisiloxane (8)

A solution of 7 (1 equiv) and KOBu (2.05 equiv) in THF was stirred at 60 °C for 2 h. After complete conversion, THF was removed and replaced by toluene. 1H NMR (δ in ppm): 0.69 (s, 12H, SiMe2), 0.44 (s, 54H, SiMe3). 13C NMR (δ in ppm): 11.1 (SiMe2), 7.2 (SiMe3). 29Si NMR (δ in ppm): 27.6 (SiMe2), −7.0 (SiMe3), −185.7 (SiK).

Synthesis of 9 by reaction of [K(Me3Si)2Si]2O (8) with YbI2·(THF)2

To a suspension of YbI2·(THF)2 (69 mg, 0.13 mmol) in toluene (2 mL) was added a solution of 8 (0.12 mmol) in toluene (2 mL) dropwise under strict exclusion of light. After 1 h the solvent volume was reduced, and the residue was treated with pentane (3 × 6 mL) and filtered over Celite. The solution volume was again reduced to 4 mL and stored at −37 °C for 24 h, after which yellow-orange crystalline 9 (87 mg, 86%) was obtained. Mp: 130–132 °C. 1H NMR (δ in ppm): 3.68 (bs, 12H, THF), 1.32 (bs, 12H, THF), 0.67 (s, 12H, SiMe2), 0.50 (s, 36H, SiMe3). 13C NMR (δ in ppm): 68.7 (THF), 25.2 (THF), 9.7 (SiMe2), 6.4 (SiMe3). 29Si NMR (δ in ppm): 32.4 (2JYb–Si = 47 Hz, SiMe2), −5.0 (2JYb–Si = 25 Hz, SiMe3), −153.8 (1JYb–Si = 531 Hz, SiYb).

Synthesis of 10 by Reaction of [K(Me3Si)2Si]2O (8) with SmI2·(THF)2

Compound 10 was obtained following the procedure for 9 (using 7 (100 mg, 0.16 mmol), KOBu (37 mg, 0.33 mmol), and SmI2·(THF)2 (92 mg, 0.17 mmol)) to yield 10 as dark green crystals (93 mg, 69%). Mp: 147–148 °C. 1H NMR (δ in ppm): 3.65 (m, 12H, THF), 2.30 (m, 12H, THF), 0.23 (s, 36H, SiMe3), −3.04 (s, 12H, SiMe2). 13C NMR (δ in ppm): 67.4 (THF), 28.4 (SiMe2), 24.1 (THF), −16.1 (SiMe3). 29Si NMR (δ in ppm): 117.2 (SiSm), −118.1 (SiMe3), n.d. (SiMe2).

2,5-Dimethoxy-2,5-bis(trimethylsilyl)decamethylhexasilane (14)

A solution of 2,2,5,5-tetrakis(trimethylsilyl)decamethylhexasilane (889 mg, 2.53 mmol) in CCl4 (3 mL) and pentane (3 mL) was stirred for 48 h. After completion of conversion was checked by 29Si NMR spectroscopy, triethylamine (1.0 mL) and DME (10 mL) were added followed by methanol (0.5 mL). After 15 min the solvent was removed and the residue was extracted with pentane (3 × 10 mL). After removal of the pentane, compound 14 was obtained as a colorless oil (843 mg, 84%). 1H NMR (δ in ppm): 3.30 (s, 6H, OCH3), 0.48 (s, 12H, SiMe2), 0.29 (s, 36H, SiMe3). 13C NMR (δ in ppm): 55.8 (CH3O), 1.2 (SiMe3), −2.2 (SiMe2). 29Si NMR (δ in ppm): 9.8 (SiO), −15.1 (SiMe3), −36.7 (SiMe2).

2,5-Bis(methoxymethyl)-2,5-bis(trimethylsilyl)decamethylhexasilane (16)

To a solution of methoxymethyl chloride (67 mg, 0.83 mmol) in DME (2 mL) was added silanide 15 (obtained from 2,2,5,5-tetrakis(trimethylsilyl)decamethylhexasilane (250 mg, 0.41 mmol) and KOtBu (94 mg, 0.84 mmol)) dropwise, whereupon formation of a white precipitate occurred. After 15 min the solvent was removed and the residue was extracted with pentane (3 × 5 mL). After removal of pentane, compound 16 was obtained as a colorless oil (208 mg, 92%). 1H NMR (δ in ppm): 3.47 (s, 4H, SiCH2), 3.15 (s, 6H, OCH3), 0.48 (s, 12H, SiMe2), 0.31 (s, 36H, SiMe3). 13C NMR (δ in ppm): 63.0 (CH3OCH2), 1.8 (SiMe3), −1.2 (SiMe2). 29Si NMR (δ in ppm): −12.2 (SiMe3), −34.8 (SiMe2), −77.5 (SiCH2OCH3).

1-Methoxy-3,3-bis(trimethylsilyl)heptamethyltetrasilane (17)

To a solution of (Me3Si)3Si-Me2Si-SiMe2Cl (300 mg, 0.75 mmol) and NEt3 (1 mL) in DME (4 mL) was added methanol (1 mL) dropwise. After 2 h the solvent was removed and the residue extracted with pentane (3 × 5 mL). The solvent was removed, and the colorless semisolid 17 (296 mg, 99%) was obtained. 1H NMR (δ in ppm): 3.22 (s, 3H, OCH3), 0.33 (s, 6H, SiMe2), 0.31 (s, 27H, SiMe3), 0.24 (s, 6H, OSiMe2). 13C NMR (δ in ppm): 50.0 (CH3O), 3.2 (SiMe3), −0.5 (OSiMe2), −1.2 (SiMe2). 29Si NMR (δ in ppm): 17.7 (Me2SiO), −9.6 (SiMe3), −42.2 (SiMe2), −133.0 (Siq).

Bis[2-methoxy-1,1,2,2-tetrakis(trimethylsilyl)disilanyl]ytterbium–THF (19)

The same procedure as for 3 was used, employing YbI2·(THF)2 (97 mg, 0.17 mmol), 18 (100 mg, 0.36 mmol), and KOBu (40 mg, 0.36 mmol) in toluene. After recrystallization from pentane at −37 °C yellow crystalline 19 (71 mg, 79%) was obtained. Mp: 177–181 °C. 1H NMR (δ in ppm, C6D6): 3.74 (bs, 4H, THF), 3.47 (s, 6H, OMe), 1.41 (bs, 4H, THF), 0.51 (s, 36H, YbSiSiMe3), 0.39 (s, 36H, OSiSiMe3). 13C NMR (δ in ppm, C6D6): 69.0 (THF), 56.7 (OMe), 25.2 (THF), 7.2 (YbSiSiMe3), 2.2 (OSiSiMe3). 29Si NMR (δ in ppm, C6D6): 34.1 (MeOSi), −6.7 (YbSiSiMe3), −16.1 (OSiSiMe3), −134.7 (SiYb).

Synthesis of MeN(CH2CH2O)2SiMeCl (20)

MeN(CH2CH2OSiMe3)2 (1.16 g, 4.32 mmol) and MeSiCl3 (679 mg, 4.54 mmol) were dissolved in toluene (6 mL) and heated to reflux for 20 h. Colorless crystalline 20 (563 mg, 69%) was obtained by decantation. Mp: 128–131 °C. 1H NMR (δ in ppm, CDCl3): 3.89 (m, 4H, OCH2), 2.86 (quintet, JH–H = 6.1 Hz, 2H, NCH2), 2.66 (quintet, JH–H = 6.1 Hz, 2H, NCH2), 2.46 (s, 3H, NCH3), 0.55 (s, 3H, SiCH3). 13C NMR (δ in ppm, CDCl3): 59.7 (OCH2), 54.6 (NCH2), 43.6 (MeN), 3.9 (MeSi). 29Si NMR (δ in ppm, CDCl3): −56.3. Anal. Calcd for C6H14ClNO2Si [195.72]: C, 36.82; H, 7.21; N, 7.16. Found: C, 37.16; H, 7.12; N, 7.25.

1,1-Methylaminobis(ethyloxy)-2,2-bis(trimethylsilyl)tetramethyltrisilane (21)

To a solution of 20 (315 mg, 1.605 mmol) in toluene (1 mL) was added tris(trimethylsilyl)silylpotassium (obtained from tetrakis(trimethylsilyl)silane (500 mg, 1.558 mmol) and KOBu (180 mg, 1.605 mmol) in DME; after complete conversion, DME was removed and replaced by toluene) in toluene (5 mL). After it was stirred for 5 h, the solution mixture was filtered, the precipitate washed with pentane, and the solvent from the combined organic layers removed. After recrystallization from pentane at −57 °C colorless crystalline 21 (373 mg, 61%) was obtained. Mp: 45–50 °C. 1H NMR (δ in ppm, CDCl3): 3.71 (m, 4H, OCH2), 2.56 (t, JH–H = 4.5 Hz, 4H, NCH2), 2.41 (s, 3H, CH3N), 0.31 (s, 3H, SiCH3), 0.19 (s, 27H, (CH3)3Si). 13C NMR (δ in ppm, CDCl3): 62.3 (OCH2), 58.3 (NCH2), 44.3 (MeN), 3.7 (MeSi), 2.4 [(Me3Si)3-Si]. 29Si NMR (δ in ppm, CDCl3): 3.5 (SiO2), −10.1 (Me3Si), −134.8 [(Me3Si)3Si]. Anal. Calcd for C15H41NO2Si5 [407.92]: C, 44.17; H, 10.13; N, 3.43. Found: C, 44.24; H, 9.63; N, 3.41.

Silocane–Yb Complex 22

The same procedure was used as for the preparation of 3 with YbI2·(THF)2 (56 mg, 0.10 mmol), 21 (80 mg, 0.20 mmol), and KOBu (23 mg, 0.21 mmol) (silanide 21a: 29Si NMR (δ in ppm, D2O capillary) 35.9 (MeSiO2), −5.4 (SiMe3), −210.2 (SiYb)). The formation of the silanide was carried out in DME, which after complete conversion was replaced by toluene. After recrystallization from pentane at −37 °C yellow crystalline 22 (36 mg, 43%) was obtained. Mp: 143–144 °C dec. 1H NMR (δ in ppm, C6D6): 3.70 (m, 4H, OCH2), 2.12 (s, 3H, NCH3), 1.79 (bs, 4H, NCH2), 0.54 (s, 21H, SiMe2 and SiMe3). 13C NMR (δ in ppm, C6D6): 80.4 (OCH2), 60.9 (bs, NCH2), 47.9 (NCH3), 6.3 (SiMe3), 4.0 (SiMe). 29Si NMR (δ in ppm, C6D6): 38.8 (MeSiO2), −5.0 (SiMe3), −182.0 (SiYb).

Silocane–Eu Complex 23

The same procedure as for 3 was used with EuI2·(THF)2 (169 mg, 0.31 mmol), 21 (228 mg, 0.56 mmol), and KOBu (66 mg, 0.58 mmol). The reaction to the anion was done in DME, which was after complete conversion replaced by toluene. After recrystallization with pentane at −37 °C green crystalline 23 (48 mg, 21%) was obtained. Mp: 185–190 °C. NMR: not detectable. Anal. Calcd for C24H64EuN2O4Si8 [821.43]: C, 35.09; H, 7.85; N, 3.41. Found: C, 34.92; H, 7.81; N, 3.43.