Stabilization of reactive rare earth alkyl complexes through mechanistic studies

Abstract

Rare earth tris(alkyl) complexes such as M(CH₂SiMe₃)₃(sol)_n are widely used as precursors for many compounds and as homogeneous catalysts for alkene polymerization and alkane functionalization. However, the thermal instability of those most conveniently made from the commercially available lithium salt of the neosilyl anion, LiCH₂SiMe₃, Li(r), restricts their utility. We present a new range of synthetically useful, more kinetically stable rare earth neosilyl solvates, derived from a full kinetic study of the various possible decomposition mechanisms of 7 known and 12 new solvated rare earth neosilyl complexes M(CH₂SiMe₃)₃(sol)_n M = Sc(III), Y(III), Lu(III), Sm(III), and sol = THF; TMEDA; DMPE; diglyme ((CH₃)₂(OCH₂CH₂)₂O, G₂), triglyme ((CH₃)₂(OCH₂CH₂)₃O, G₃). Surprisingly, simply using higher-denticity donors to sterically disfavor neosilyl γ-H elimination is not effective. While Sc(r)₃((CH₃)₂(OCH₂CH₂)₂O) has a half-life, t_1/2, of 258.1 h, six times longer than for Sc(r)₃(C₄H₈O)₂ (t_1/2 = 43 h), Lu(r)₃((CH₃)₂(OCH₂CH₂)₂O) and Y(r)₃((CH₃)₂(OCH₂CH₂)₂O) do not show the expected, analogous increased t_1/2. This is because new decomposition pathways appear for poorly fitting donors. Finally, kinetic studies demonstrate the impact of small, and increasing amounts of LiCl on the kinetics of the reactivity of the smaller alkyls Y(r)₃(THF)₂ and Lu(r)₃(THF)₂; molecules used in hydrocarbon chemistry and catalysis for fifty years. A new route to pure Y(r)₃(THF)₂, which avoids the traditional use of Li(r), is presented.

---

Introduction

Since their discovery in 1973,¹ neutral rare-earth metal complexes of the type RE(CH₂SiMe₃)₃(THF)_x, (RE = Sc, Y, Sm–Lu; x = 2, 3) featuring trimethylsilylmethyl ligands^2,3 have been valuable reagents for (pre)catalyst design, in particular for olefin- and diene polymerization,^4–7 and hydrocarbon C–H bond functionalization catalysis.^8,9 More recently, RE(CH₂SiMe₃)₃(donor)_x (RE = Y, Dy, Er, Ho; donor = THF, quinuclidine, lutidine and OPCy₃) have shown promise as single molecule magnets.¹⁰ Given that the Li and Mg salts of the trimethylsilylmethyl anion are commercially available, it is frustrating that all neosilyl complexes exhibit varying degrees of thermal instability, both in solution and in the solid state.² The larger –CH(SiMe₃)₂ alkyl¹¹ generates complexes with improved thermal stability, but its salts are not commercially available.

The employment of donor-ligands to enhance kinetic stability, but yet can be easily substituted by protic reagents,¹² should pave the way for novel RE-alkyl reactivity in these simple compounds.

Scheme 1 illustrates the three potential mechanisms for the elimination of SiMe₄ from RE(CH₂SiMe₃)₃(THF)₂. The first mechanism, Scheme 1a, involves the attack by a coordinated neosilyl on an α-CH₂ group. This results in the formation of a RE-alkylidene. However, this product has not been observed and is expected to be unstable. An early study of Er(CH₂SiMe₃)₃(THF)₃ suggested this decomposition mechanism, via silylalkylidene [Er(CH₂SiMe₃)(CHSiMe₃)]_n, but no evidence was provided to substantiate this, although this reactivity is well-documented for Ti.^13,14 The second mechanism, denoted as β-H in Scheme 1b, involves the attack by a neosilyl on a β-CH₂ group of a coordinated THF molecule, forming a metal vinyl alkoxide and ethylene, also anticipated to have low kinetic stability.^15,16 The third mechanism, Scheme 1c, involves the attack of the CH₂ carbon on a SiCH₃, forming a metallacycle. Deuterolysis experiments with Lu(CH₂SiMe₃)₃(THF)₂ suggested a γ-H elimination pathway to a metalacyclic intermediate Me₂Si(μ-CH₂)₂LuCH₂SiMe₃ although this was not observed directly.¹⁵

Scheme 1 The three possible mechanisms for elimination of SiMe₄ from RE(r)₃(THF)₂ (r = CH₂SiMe₃): (a) α-H elimination; (b) β-H elimination, and (c) γ-H elimination.

The addition of diglyme, (CH₃)₂(OCH₂CH₂)₂O, G₂, and glyme (CH₃OCH₂)₂, G₁, to Lu(CH₂SiMe₃)₃(THF)₂ generated the unanticipated mixed solvates Lu(CH₂SiMe₃)₃(κ²-G₂)(THF) and Lu(CH₂SiMe₃)₃(κ²-G₁)(THF) which the authors reported to be thermally robust but without providing data.¹⁵ Most recently, N,N tetramethyl ethylene diamine (TMEDA) or bis-dimethylphosphino ethane (DMPE) have been reported to enhance the kinetic inertness of [Y(CH₂SiMe₃)₃], with no further data.¹⁷

Herein, we report a comprehensive study of the relative thermal stabilities of nineteen rare earth complexes under conditions widely applied for protonolysis reactions, identifying the decomposition mechanisms, characterizing the important intermediates, and highlighting the importance of compound purity. We have prepared a library of seven known and twelve new solvated rare earth neosilyl complexes M(r)₃(donor)_x for r = CH₂SiMe₃ (Fig. S5–S30, S55, and S56†) that cover a range of rare earth radii, between Sc (r_{cov, 6-coord} = 0.745 Å) and Sm (r_{cov, 6-coord} = 0.958 Å);¹⁸ tris(CH₂SiMe₃) adducts with measurable stability at room temperature have not been made yet for any RE larger than Sm.^2,19 Previously reported complexes are: M(r)₃(THF)₂; M = Sm(III),¹⁹ Y(III),¹⁰ Lu(III),¹⁵ Sc(III);²⁰ Y(r)₃(donor);¹⁷ donor = DMPE/THF, TMEDA. The new complexes added here are: M(r)₃(G₂); M = Sm(III), Y(III), Lu(III), Sc(III); M(r)₃(G₃); M = Sm(III), Y(III), Lu(III); M(r)₃(TMEDA); M = Sm(III), Lu(III), Sc(III); M(r)₃(DMPE)(THF); M = Lu(III), Sc(III); G₂ = diglyme, (CH₃)₂(OCH₂CH₂)₂O, G₃ = triglyme (CH₃)₂(OCH₂CH₂)₃O (Table 1). For simplicity we have abbreviated them here to a label that describes the metal and donor solvent, i.e. M(r)₃(donor) = M-donor. Single crystal X-ray data for all new compounds (except Sm-G₂ and Sm-G₃) are discussed in the ESI (Tables S1–S10 and Fig. S31–S40).†

Table 1 Half-lives t_1/2 [h] of selected rare earth tris(neosilyl) solvates RE(CH₂SiMe₃)₃(donor)_x at 30 °C in C₆D₆. Donor = (THF)₂, (DMPE)(THF), TMEDA, G₂ = diglyme, (CH₃)₂(OCH₂CH₂)₂O, G₃ = triglyme (CH₃)₂(OCH₂CH₂)₃O, DMPE = Me₂PCH₂CH₂PMe₂, TMEDA = Me₂NCH₂CH₂NMe₂

RE(r)3(donor)x	Label	t 1/2 [h]	Ref.
Sm(CH2SiMe3)3(triglyme)	Sm-triglyme	0.7	This work
Sm(CH2SiMe3)3(THF)3	Sm-THF	2.8	19
Sm(CH2SiMe3)3(diglyme)	Sm-diglyme	3.7	This work
Y(CH2SiMe3)3(DMPE)(THF)	Y-DMPE	5.8	17
Y(CH2SiMe3)3(triglyme)	Y-triglyme	7.8	This work
Sc(CH2SiMe3)3(DMPE)	Sc-DMPE	9	This work
Sm(CH2SiMe3)3(TMEDA)	Sm-TMEDA	11.7	This work
Y(CH2SiMe3)3(diglyme)	Y-diglyme	24.4	This work
Y(CH2SiMe3)3(TMEDA)	Y-TMEDA	36.5	17
Sc(CH2SiMe3)3(THF)2	Sc-THF	43.9	20
Sc(CH2SiMe3)3(TMEDA)	Sc-TMEDA	73.4	This work
Lu(CH2SiMe3)3(DMPE)(THF)	Lu-DMPE	74.1	This work
Y(CH2SiMe3)3(THF)2	Y-THF	119.6	This work
Li[YCl(CH2SiMe3)3(THF)2]	LiClY-THF	213	10
Lu(CH2SiMe3)3(triglyme)	Lu-triglyme	190.7	This work
Sc(CH2SiMe3)3(diglyme)	Sc-diglyme	258.1	This work
Lu(CH2SiMe3)3(diglyme)	Lu-diglyme	278.2	This work
Lu(CH2SiMe3)3(THF)2	Lu-THF	482.6	15
Lu(CH2SiMe3)3(TMEDA)	Lu-TMEDA	487.8	This work

---

Results and discussion

Thermal stability and mechanistic studies

In a typical thermolysis experiment, 9.1 mM of a M complex is dissolved in C₆D₆ with hexamethylbenzene as an internal standard, in a J. Young NMR tube and the solution monitored by ¹H NMR spectroscopy for at least two half-lives during incubation at 30 °C.

Each compound's thermal stability is expressed as a half-life t_1/2 [h] of the complex in solution. Since all three mechanisms result in the formation of SiMe₄ (Scheme 1), the half-life (time for 50% of the compound to be converted into SiMe₄) is determined from linearized plots of the relative integration of the SiMe₄ resonance (Table 1) against hexamethylbenzene (Fig. S41–S45†). We first examined the half-life as a function of the metal center's ionic radius in M-THF and found a stability trend in the order Sm³⁺ < Sc³⁺ < Y³⁺ < Lu³⁺ (Table 1). The low thermal stability of Sm-THF is in line with its large size, and metals with ionic radii larger than Sm do not afford isolable neosilyl complexes.^17,21

Notably, the stability trend Sc < Y < Lu is in line with the metals' Lewis-acidity rather than size.²²

We examined the thermal decomposition mechanism of diamagnetic Lu-THF in greater depth. The compound decomposes with a half-life of t_1/2 = 482.6 h. Kinetic studies using initial rates methods are first-order in [Lu-THF] suggesting an intramolecular decomposition process. We prepared the d₂-r, d₈-r, and d₉-THF labelled isotopomers of Lu-THF (Fig. S48†).

First, α-H elimination can be ruled out based on the decomposition studies on Lu(d₂-r)₃(THF)₂ and Lu(d₉-r)₃(THF)₂ since NMR spectra show resonances corresponding to d₂-and d₁₀-labelled SiMe₄ as decomposition products, respectively (Fig. 1, and S48, inset†). This strongly suggests γ-H elimination as the primary decomposition mechanism, Scheme 1c. We measured a primary KIE = 2.09 for Lu(d₉-r)₃(THF)₂ (Fig. S53†) and no discernible KIEs for Lu(r)₃(d₈-THF)₃ or Lu(d₂-r)₃(THF)₂.

Fig. 1 Proposed γ-H elimination route converting Lu(d₂-r)₃(THF)₂ to the metallacycle that is visible in NMR spectra of the d₂-congener, with sections of the ¹H NMR spectra below measured at t = 0 h (red), 78 h (orange), 103 h (purple), 127 h (green), 165 h (violet) in C₆D₆, 30 °C. The resonance circled in green and red show the increase of SiMe₃(CD₂H) and the –CH₂ group of a d₂-metallacyclic species, respectively.

The thermolysis of Lu-THF, determined from the disappearance of the CH₂ neosilyl resonance, proceeds with a rate (9 × 10⁻⁴ h⁻¹) that is similar to that of SiMe₄ formation (11 × 10⁻⁴ h⁻¹), in agreement with decomposition via the liberation of 1 equiv. of SiMe₄ per Lu-THF. The slight discrepancy between the rates of Lu-THF consumption and SiMe₄ formation can be explained by an increasing resonance at δ = −0.894 in ¹H NMR spectrum that overlaps with the CH₂ – neosilyl resonance at δ = −0.891 (Fig. 1, and S49†). We assigned the growing resonance to the μ-CH₂ of the formed metallacycle Me₂Si(μ-CH₂)₂LuCH₂SiMe₃, (circled red in Fig. 1 and S49†), since the resonance is almost coincident with the LuCH₂SiMe₃ resonance, and can be observed in the ¹H NMR spectrum of Lu(d₂-r)₃(THF)₂.

Donor – dependence of RE-alkyl thermal stability

We initially synthesized Lu-G₂, which is fac-pseudo octahedral at Lu in solution and the solid-state (Fig. S9, S10, and S35†) to target an increased thermal stability of Lu-THF by sterically disfavoring γ-H elimination. However, a kinetic study reveals that decomposition occurs via γ-H elimination from the glyme rather than from the neosilyl, which is reasonable considering the greater acidity of the glyme CH₂ groups, decreasing the compound's half-life to t_1/2 = 278.2 h (Table 1, and Scheme 2). The same destabilizing effect is observed for Y-G₂. The size-fit correlation trend is supported by the observed increase in thermal stability for Sc-G₂, which showed a negligible activation of the glyme CH₂ groups (∼1.48% glyme decomposition by release of CH₂=CHOMe²³ into solution, Fig. S46, and S47†) granting the compound with an approximately six-fold increased half-life (t_1/2 = 258.1 h) than Sc-THF (t_1/2 = 43.9 h). Somewhat surprisingly to us, the destabilization is even more pronounced for Sm, Y, and Lu adducts of the larger glyme M-G₃ (Table 1). To target stable complexes of larger Ln that are desirable synthesis precursors, we made M-donor for M = Sm, Y, Lu, Sc and donor = TMEDA, DMPE, recognizing both have less acidic CH₂ groups than G₂ and G₃. This increases the half-life by a factor of up to four for Sm-TMEDA (t_1/2 = 11.7 h). Notably, the kinetic study shows that Sm-TMEDA decomposes cleanly with release of one equiv. of SiMe₄ (Fig. 2). Therefore we expect this will be an excellent precursor for protonolysis reactions. In contrast to M-TMEDA, all the M-DMPE congeners showed shorter half-lives compared to their THF counterparts (Table 1).

Scheme 2 Suggested route to the hypothesized ethoxyether-stabilised Lu silametallacycle from Lu-diglyme.

Fig. 2 ¹H NMR spectrum (600 MHz) of Sm-TMEDA in d₆-benzene (t = 28 h) showing SiMe₄ and TMEDA release.

¹H NMR studies on the d₉-labelled isotopomers of Lu(r)₃(sol) (sol = diglyme, TMEDA, DMPE) (Fig. S50–S53†) indicate thermal decomposition via γ-CH activation as shown by the emergence of a 1 : 1 : 1 triplet resonance that can be assigned to SiMe₃(CH₂D).

Impact of LiCl on the thermal stability and reactivity of the RE alkyls

While M(r)₃(THF)_x (M = Sc, Lu; x = 2 and Sm; x = 3) were shown to decompose according to a first-order rate law in complex, our initial experiments indicated a second-order rate-law for decomposition of the Y^III congener. This was supported by crossover studies on equimolar reaction mixtures of the d₂-r and d₉-r labelled complexes (Scheme S1, and Fig. S54†). All the reported routes to Y-THF originate from Li(r)^6,10 (Table S11†). We note that crystal structures of LiCl – free, neutral homoleptic M(r)₃(THF)_x (M = Y, Lu, Yb, Er, Ho) have been reported.^10,14,24 While the previous reports did not discuss whether ⁷Li NMR spectra were measured. However, we observed no resonance under normal conditions in ⁷Li NMR spectroscopic studies of solutions.

The addition of 12-crown-4 ether (12-c-4) to a C₆D₆ solution of Y-THF results in a species we assign as Li(12-c-4)YCl(r)₃(THF)·(12-c-4)_1.75, which shows a ⁷Li NMR resonance at δ = −2.09 (Fig. S57†)²³ and an ¹H NMR spectrum that is different to the known Y(r)₃(12-c-4) (Fig. S58, and S59†).²⁵ In contrast, 12-c-4 addition to Lu(r)₃(THF)₂ solutions shows no change in the (silent) ⁷Li NMR spectrum.

ICP – OES analyses combined with titration for chloride content (Mohr's method) indicate that Y-THF prepared from the reagent Li(r) is contaminated with 7% LiCl. This content explains why samples isolated in previous reports were able to ‘pass’ elemental analyses for carbon and hydrogen content (Table S11†). The use of elemental analysis to confirm bulk purity of a sample of a new complex has been the subject of recent scrutiny.^26,27

We examined the effect of increasing the amount of incorporated LiCl from zero to 1 equiv. on the properties of Lu-THF in benzene solution. We observe a factor of 2.5 increase in the half-life (t_1/2) of the complex from t_1/2 = 482 h for Lu-THF to t_1/2 = 1122 h (Fig. S63†). It is notable that the chemical shifts of the THF CH₂ groups are significantly shifted to higher frequencies in both the ¹H – & ¹³C NMR spectra compared to those of the alkyl groups. We suggest that in benzene solution at higher concentrations the THF is preferentially binding to Li and the Cl to the RE center, (Fig. S61, and S62†). Although the LiCl impurity improves the complexes' half-lives,^28,29 we expect it to hamper reactivity.

The inclusion of LiCl in s- and p-block alkyl complexes has been shown to significantly influence their stability, speciation, and reaction rates.²⁸ Knochel and co-workers have demonstrated that organomagnesium and organozinc compounds, in the presence of alkali metals, exhibit increased basicity, allowing them to deprotonate otherwise inert aromatic and heteroaromatic C–H bonds.³⁰ The combination of Grignard reagents with LiCl, known as turbo Grignard reagents, marks a significant advancement in main-group organometallic chemistry, leading to more accessible Grignard-like compounds that enable more powerful and controlled reactions. Recent ab initio molecular dynamics simulations suggest that LiCl aggregates form through μ₂-Cl bridging ligands, creating Li₄Cl₄ cubane-like structures in THF solutions.²⁹

The observation of second-order kinetics described above for the decomposition of our Li(r)-prepared Y-THF suggests the facile formation of a chloride bridged intermediate shown in Scheme 3, as a potential deviation from the mechanism found for the pure alkyls.

Scheme 3 Suggested route for the decomposition of LiCl-incorporated Y-THF that accounts for the observed kinetics.

Finally, a new route to pure Y-THF from Na(r)³¹ is described in this work. As anticipated, ICP analyses and Mohr's method rule out LiCl or NaCl contamination of the complex, which decomposes following first order kinetics, with a t_1/2 = 119.6 h, compared with t_1/2 = 213 h for Y-THF prepared from Li(r) (Fig. S60, S64–S66,† and Table 1).

Conclusions

A new range of synthetically useful, more kinetically inert rare earth neosilyl solvates has been developed, based on the study of known and new solvates M(CH₂SiMe₃)₃(sol)_n, where M = Sc(III), Y(III), Lu(III), Sm(III), and sol = THF, TMEDA, DMPE, diglyme (G₂), and triglyme (G₃). The donor adduct with the longest half-life has been identified for each metal and is: Sc-G₂, 258 h; Y-THF, 120 h, Lu-TMEDA, 488 h; Sm-TMEDA, 12 h at room temperature. Samarium is the largest rare earth cation that we have been able to stabilise as a tris(r) adduct by adding simple donors.

Notably, simply increasing the denticity of donors to sterically hinder neosilyl γ-H elimination proves ineffective, as poorly fitting donors lead to new decomposition pathways. A precise size match with polydentate ligands can dramatically improve half-lives, as observed with Sc-G₂. However, when the fit is less optimal, alternative decomposition mechanisms, such as γ-H activation at a glyme CH₂ group, become more prominent. Thus, Sc-donor shows a significant increase in half-life when switching donor from THF (44 h) to diglyme (258 h), but Lu-donor shows a significant decrease in half-life when switching donor from THF (483 h) to diglyme (278 h). The stability trend of Sc < Y < Lu within the M-THF solvate series aligns more with Lewis acidity than size.

The incorporation of LiCl into the system results in an increase in half-life, particularly for Lu-THF by a factor of 2.5, although the effect on reactivity, particularly protonolysis, has not yet been evaluated.

Furthermore, while neutral complexes such as Lu(r)₃(sol) (sol = THF, TMEDA, DMPE, G₂) exhibit first-order kinetics involving γ-C–H activation, as confirmed by labelling experiments, ate complexes follow second-order kinetics, likely due to μ-Cl incorporation. Finally, a new route to pure Y-THF from Na(r) is described. While Y-THF exhibits a lower thermal stability compared to LiClY-THF, we expect it to open new perspectives for clean and fast protonolysis relevant to rare earth synthetic chemistry and catalysis.

Experimental

Materials and methods

All moisture and air sensitive materials were manipulated using standard high-vacuum Schlenk-line techniques and MBraun gloveboxes and stored under an atmosphere of dried and deoxygenated argon. All glassware items, cannulae and Fisherbrand 1.2 μm retention glass microfiber filters were dried in a 160 °C oven overnight before use. M(CH₂SiMe₃)₃(THF)₂; M = Y(III),¹⁰ Lu(III),¹⁵ Sc(III)²⁰, Sm(CH₂SiMe₃)₃(THF)₃,¹⁹ and Y(CH₂SiMe₃)₃(donor);¹⁷ donor = (DMPE)(THF), TMEDA and NaCH₂SiMe₃ (ref. 31) were prepared according to published literature procedures and characterized with ¹H and ⁷Li NMR spectroscopy.

n-Hexane, tetrahydrofuran (THF), diethyl ether (Et₂O) and toluene for use with moisture and air sensitive compounds were dried using an MBRAUN SPS 800 Manual solvent purification system and stored over activated 3 Å molecular sieves. Benzene-d₆ was purchased from Cambridge Isotope Laboratories and refluxed over potassium metal for 24 hours, freeze–pump–thaw degassed and purified by trap-to-trap distillation prior to use. THF-d₈ was purchased from Cambridge Isotope Laboratories and dried over sodium/benzophenone before being freeze–pump–thaw degassed and purified by trap-to-trap distillation prior to use. Diglyme ((CH₃)₂(OCH₂CH₂)₂O, G₂) was dried over sodium metal before freeze–pump–thaw degassed and purified by trap-to-trap distillation prior to use. Triglyme (CH₃)₂(OCH₂CH₂)₃O, G₃) was dried over sodium metal before being purified by dynamic vacuum distillation prior to use. Tetramethylethylenediamine (TMEDA) was dried over sodium before being freeze–pump–thaw degassed and purified by trap-to-trap distillation prior to use. 1,2-Bis(dimethylphosphino)ethane (DMPE) was purified by trap-to-trap distillation and stored over 3 Å molecular sieves.

All solvents were purchased from Sigma-Aldrich or Fisher Scientific and stored over 3 Å molecular sieves for at least 12 hours before being used.

Selected synthetic procedures (all other data in the ESI†)

Synthesis of LiCl-free Y(CH₂SiMe₃)₃(C₄H₈O)₂ (Y-THF). In an Ar-filled glovebox, anhydrous YCl₃ (26 mg, 0.133 mmol) was suspended in 260 mL THF and 2.6 mL 1 : 1 mixture of Et₂O and pentane. The resulting white suspension was cooled to −78 °C and NaCH₂SiMe₃ (ref. 31) (44 mg, 0.4 mmol, 3 eq.) was added portion wise as a solid. After stirring the resulting white suspension for 1 h at −78 °C, it was warmed to room temperature for 15 min. Filtration and removal of the solvent at −40 °C yielded the title compound as a white crystalline solid. Yield: 28 mg, 44% based on Y.

¹H NMR (600 MHz, C₆D₆) δ 3.88 (s, 8H), 1.34–1.32 (m, 8H), 0.30 (s, 27H), −0.68 (d, J = 2.7 Hz, 6H). ¹³C NMR (151 MHz, C₆D₆) δ 70.60, 33.88, 33.65, 25.13, 4.63.

Anal. calcd for YC₂₀H₄₉Si₃O₂: C, 48.55; H, 9.98. Found: C, 48.43; H, 9.81.

Determination of chloride content. In an Ar-filled glovebox, Y-THF (6.7 mg) was weighed out in a 4 mL glass vial. The solid was then dissolved in water (4 mL) suitable for trace metal analyses, followed by the addition of a potassium chromate indicator (9.1 mg). A 25 mM solution of silver nitrate was titrated against the solution, resulting in the formation of silver chloride as white precipitates. The titration continued until the endpoint was reached, indicated by the formation of a dark orange solid (silver chromate) (Fig. S63†). The chloride content was determined to be ∼7.6%, which matched the Li content found by ICP analysis.

Synthesis of Y(CH₂SiMe₃)₃(C₆H₁₄O₃) (Y-G₂). In an Ar-filled glovebox, anhydrous YCl₃ (100 mg, 0.512 mmol) was suspended in 3 mL THF at 60 °C overnight. Following evaporation of the solvent, the residue was re-suspended in pentane (4 mL). The white suspension was cooled down to −78 °C and a solution of LiCH₂SiMe₃ (146 mg, 1.55 mmol, 3 eq.) in pentane (2 mL) was added dropwise. The resulting reaction mixture was stirred for 1 h at room temperature, filtered and treated with diglyme ((CH₃)₂(OCH₂CH₂)₂O, G₂) (68.7 mg, 0.51 mmol), yielding Y(CH₂SiMe₃)₃(C₆H₁₄O₃) (Y-G₂) as colorless crystalline solids. Diffraction quality crystals of Y(CH₂SiMe₃)₃(C₆H₁₄O₃) (Y-G₂) were grown over three days from a saturated hexane solution at −40 °C. Yield: 51%, based on Y.

¹H NMR (600 MHz, C₆D₆) δ 3.07 (s, 6H), 2.97 (s, 4H), 2.68 (s, 4H), 0.44 (s, 27H), −0.42 (d, J = 2.9 Hz, 6H). ¹³C NMR (151 MHz, C₆D₆) δ 69.18, 67.97, 60.68, 35.42 (d, J = 36.5 Hz), 4.78.

Anal. calcd for YC₁₈H₄₇Si₃O₃: C, 44.60, H, 9.77. Found: C, 42.37; H, 9.16.

Synthesis of Sc(CH₂SiMe₃)₃(C₆H₁₄O₃) (Sc-G₂). In an Ar-filled glovebox, Sc(r)₃(THF)₂ (40.35 mg, 0.0895 mmol) was dissolved in 2.5 mL hexane. The yellowish solution was cooled down to −78 °C followed by addition of diglyme ((CH₃)₂(OCH₂CH₂)₂O, G₂) (100 μL, 0.716 mmol, 8 eq.) resulting in yellow precipitates which were extracted with hexane (3 × 2 mL). The combined hexane extracts were concentrated under reduced pressure and cooled down to −40 °C, yielding diffraction quality crystals of Sc(CH₂SiMe₃)₃(C₆H₁₄O₃) (Sc-G₂) after six days. Yield: 54%, based on Sc(r)₃(THF)₂.

¹H NMR (600 MHz, C₆D₆) δ 3.23 (t, J = 5.2 Hz, 3H), 3.02 (s, 6H), 2.76 (t, J = 5.3 Hz, 5H), 0.41 (s, 36H), 0.09 (s, 6H). ¹³C NMR (151 MHz, C₆D₆) δ 70.05, 69.71, 60.73, 4.33.

Anal. calcd for ScC₁₈H₄₇Si₃O₃: C, 49.05, H, 10.75. Found: C, 48.97; H, 10.66.

Synthesis of Sm(CH₂SiMe₃)₃(C₆H₁₄O₃) (Sm-G₂). In an Ar-filled glovebox, anhydrous SmCl₃ (45.62 mg, 0.178 mmol) was suspended in 2.5 mL THF at room temperature for 20 min. Following the evaporation of the solvent, the residue was re-suspended in a 1 : 1 mixture of pentane and Et₂O (2.8 mL). The white suspension was cooled down to −78 °C and a solution of LiCH₂SiMe₃ (50.2 mg, 0.533 mmol, 3 eq.) in pentane (1.5 mL) was added dropwise. The resulting reaction mixture was stirred for 2 h at room temperature. The yellow filtrate was evacuated cold (−40 °C), the residues extracted with pentane (3 × 2 mL) and diglyme ((CH₃OCH₂CH₂)₂O, G₂) (22.7 mg, 0.16 mmol) added to the filtrate to yield Sm(CH₂SiMe₃)₃(C₆H₁₄O₃) (Sm-G₂) as a yellow crystalline solid after storage overnight at −40 °C. Yield: 41%, based on Sm.

¹H NMR (600 MHz, C₆D₆) δ 5.46 (s, 6H), 5.29 (s, 6H), 0.44 (s, 27H), −0.14 (s, 4H), −1.53 (s, 4H). ¹³C NMR (151 MHz, C₆D₆) δ 131.79, 31.97, 23.05, 14.34, 2.84.

Anal. calcd for SmC₁₈H₄₇Si₃O₃: C, 39.58, H, 8.67. Found: C, 39.29; H, 8.30.

Synthesis of Lu(CH₂SiMe₃)₃(C₆H₁₄O₃) (Lu-G₂). In an Ar-filled glovebox, anhydrous LuCl₃ (49.5 mg, 0.176 mmol) was suspended in 2.5 mL of tetrahydrofuran (THF) and heated to 50 °C for 40 min. After the solvent was evaporated, the residue was re-suspended in 2 mL of pentane. This white suspension was then cooled to −78 °C. Subsequently, a solution of LiCH₂SiMe₃ (50.2 mg, 0.533 mmol, 3 eq.) in 1.5 mL of pentane was added dropwise. The resulting reaction mixture was stirred for 2 hours at −78 °C. After being filtered while still cold, diglyme ((CH₃)₂(OCH₂CH₂)₂O, G₂) (23.6 mg, 0.176 mmol, 1 eq.) was added to the filtrate. This process yielded white crystalline solids with a yield of 60%, based on Lu. Finally, diffraction-quality crystals were grown from a concentrated solution of diethyl ether (Et₂O) and hexane at −40 °C.

¹H NMR (600 MHz, C₆D₆) δ 3.04 (s, 6H), 2.96 (t, J = 5.3 Hz, 4H), 2.61 (t, J = 5.3 Hz, 4H), 0.44 (s, 27H), −0.66 (s, 6H). ¹³C NMR (151 MHz, C₆D₆) δ 69.41, 68.43, 60.70, 41.60, 4.91.

Anal. calcd for LuC₁₈H₄₇Si₃O₃: C, 37.88, H, 8.30. Found: C, 37.54; H, 8.13.

Synthesis of Y(CH₂SiMe₃)₃((CH₃)₂(OCH₂CH₂)₃O) (Y-G₃). In an Ar-filled glovebox, anhydrous YCl₃ (586 mg, 3 mmol) was suspended in 10 mL of THF at 60 °C overnight. Following evaporation of the solvent, the residue was re-suspended in pentane (20 mL). The white suspension was cooled down to −78 °C and a solution of LiCH₂SiMe₃ (856 mg, 9.09 mmol, 3 eq.) in pentane (5 mL) was added dropwise. The reaction mixture was stirred at room temperature for 1 h. After stirring, it was cooled down to −78 °C, at which temperature it was filtered. The resulting mixture was treated with triglyme ((CH₃)₂(OCH₂CH₂)₃O, G₃) (534.7 mg, 3 mmol) at −78 °C. The result of this process was a white solid. This solid was then extracted using cold hexane (4 × 5 mL). From the combined saturated hexane fractions, diffraction quality crystals of Y(CH₂SiMe₃)₃((CH₃)₂(OCH₂CH₂)₃O) were grown at −40 °C. The yield of this entire process was 56%, a calculation based on Y.

¹H NMR (400 MHz, C₆D₆) δ 3.47 (s, 6H), 3.09 (t, J = 4.8 Hz, 4H), 3.03–2.93 (m, 8H), 0.39 (s, 27H), −1.00 to −1.15 (m, 6H). ¹³C NMR (151 MHz, C₆D₆) δ 72.32, 71.98, 69.51, 68.37, 5.27, 4.92.

Anal. calcd for YC₂₀H₅₁Si₃O₄: C, 45.43, H, 9.72. Found: C, 45.38; H, 9.76.

Synthesis of Sm(CH₂SiMe₃)₃(CH₃)₂(OCH₂CH₂)₃O) (Sm-G₃). In an Ar-filled glovebox, anhydrous SmCl₃ (140 mg, 0.545 mmol) was suspended in 1.1 mL THF at room temperature for 20 min. Following the evaporation of the solvent, the residue was re-suspended in a 1 : 1 mixture of pentane and Et₂O (8.4 mL). The white suspension was cooled down to −78 °C and a solution of LiCH₂SiMe₃ (146.5 mg, 1.556 mmol) in pentane (1.6 mL) was added dropwise. The resulting yellow reaction mixture was stirred for 2 h at room temperature. After stirring, it was cooled down to −78 °C at which temperature it was filtered. The yellow filtrate was evacuated cold (−40 °C), the residues were extracted with pentane (3 × 5 mL), and triglyme ((C₈H₁₈O₄), G₃) (92.4 mg, 0.52 mmol) was added to the combined fractions yielding Sm(CH₂SiMe₃)₃((CH₃)₂(OCH₂CH₂)₃O) (Sm-G₃). Yield: 48%, based on Sm.

¹H NMR (600 MHz, C₆D₆) δ 10.25 (s, 1H), 6.00 (s, 1H), 1.39 (s, 6H), −1.80 (s, 1H), −3.79 (s, 1H). ¹³C NMR (151 MHz, C₆D₆) δ 152.17, 86.66, 72.19, 70.77, 69.74, 67.62, 58.73, 3.61.

Anal. calcd for SmC₂₀H₅₁Si₃O₄: C, 40.7, H, 8.71. Found: C, 40.32; H, 8.43.

Synthesis of Lu(CH₂SiMe₃)₃((CH₃)₂(OCH₂CH₂)₃O) (Lu-G₃). In an Ar-filled glovebox, anhydrous LuCl₃ (154.8 mg, 0.550 mmol, 1 eq.) was suspended in 1.1 mL THF at 50 °C for 2 h. Following evaporation of the solvent, the residue was re-suspended in pentane (4.4 mL) and Et₂O (4.4 mL). The white suspension was cooled down to −78 °C and a solution of LiCH₂SiMe₃ (155.5 mg, 1.651 mmol, 3 eq.) in pentane (1.5 mL) was added dropwise. The cold well was lowered and the resulting reaction mixture was allowed to gradually warm with stirring over 1 h, then cooled down to −40 °C at which temperature it was filtered. Solvent was removed under vacuum. Following extraction with pentane at −40 °C (3 × 5.5 mL), triglyme ((C₈H₁₈O₄), G₃) (98.1 mg, 0.550 mmol, 1 eq.) was added to the filtrate, yielding white crystalline solids. The resulting pentane supernatant was stored at −40 °C overnight, yielding diffraction quality crystals of Lu(CH₂SiMe₃)₃((CH₃)₂(OCH₂CH₂)₃O) (Lu-G₃). Yield: 78.3 mg, 24% based on Lu.

¹H NMR (600 MHz, C₆D₆) δ 3.23 (t, J = 4.9 Hz, 4H), 3.13 (d, J = 8.5 Hz, 10H), 2.97 (t, J = 5.2 Hz, 4H), 0.42 (s, 27H), −0.82 (s, 6H). ¹³C NMR (151 MHz, C₆D₆) δ 71.53, 69.69, 69.26, 60.51, 38.44, 34.45, 22.73, 14.27, 5.06.

Anal. calcd for LuC₂₀H₅₁Si₃O₄: C, 39.07, H, 8.36. Found: C, 38.69; H, 8.11.

Synthesis of Sc(CH₂SiMe₃)₃(Me₂NCH₂CH₂NMe₂) (Sc-TMEDA). In an Ar-filled glovebox, Sc(CH₂SiMe₃)₃(C₄H₈O)₂ (45.6 mg, 0.101 mmol) was dissolved in 5 mL toluene and cooled to −40 °C in the glovebox freezer. To the cold clear yellow solution TMEDA (11.8 mg, 0.101 mmol, 1 eq.) in 3 mL toluene was added dropwise. The resulting clear yellow mixture was stirred for 1 h at room temperature. Toluene was removed under reduced pressure at room temperature. The remaining yellow oil was extracted with n-hexane (5 × 3 mL). The combined colorless extracts were filtered, concentrated and cooled to −40 °C, yielding colorless block shaped diffraction quality crystals of Sc(CH₂SiMe₃)₃(Me₂NCH₂CH₂NMe₂) (Sc-TMEDA) overnight. Yield: 15 mg, 36% based on Sc(CH₂SiMe₃)₃(THF)₂.

¹H NMR (600 MHz, C₆D₆) δ 1.91 (s, 12H), 1.57 (s, 4H), 0.40 (s, 27H), 0.10 (s, 6H). ¹³C NMR (151 MHz, C₆D₆) δ 56.85, 46.69, 16.93, 4.41.

Anal. calcd for ScC₁₈H₄₉Si₃N₂: C,51.13, H, 11.68; N, 6.63. Found: C, 50.97; H, 11.59; N, 6.57.

Synthesis of Sm(CH₂SiMe₃)₃(Me₂NCH₂CH₂NMe₂) (Sm-TMEDA). In an Ar-filled glovebox, Sm(CH₂SiMe₃)₃(C₄H₈O)₃ (63.6 mg, 0.101 mmol) was dissolved in 5 mL toluene and cooled to −40 °C in the glovebox freezer. To the cold clear yellow solution TMEDA (11.8 mg, 0.101 mmol, 1 eq.) in 1 mL toluene was added dropwise. The resulting clear yellow mixture was stirred for 1 h at room temperature. Toluene was removed under reduced pressure at 0 °C. The remaining yellow oil was extracted with cold hexane. Removal of hexane at reduced temperature (−40 to 0 °C) yielded Sm(CH₂SiMe₃)₃(Me₂NCH₂CH₂NMe₂) (Sm-TMEDA) as a yellow solid in 18% yield, 9.5 mg based on Sm(CH₂SiMe₃)₃(THF)₃.

¹H NMR (600 MHz, C₆D₆) δ 4.50 (s, 6H), 2.10 (s, 12H), 0.21 (s, 27H), −2.68 (s, 4H). ¹³C NMR (151 MHz, C₆D₆) δ 58.45, 46.03.

Synthesis of Lu(CH₂SiMe₃)₃(Me₂NCH₂CH₂NMe₂) (Lu-TMEDA). In an Ar-filled glovebox, Lu(CH₂SiMe₃)₃(C₄H₈O)₂ (58.6 mg, 0.101 mmol) was dissolved in 5 mL toluene and cooled to −40 °C in the glovebox freezer. To the cold clear colorless solution TMEDA (11.7 mg, 0.101 mmol, 1 eq.) in 3 mL toluene was added dropwise. The resulting clear colorless mixture was stirred for 1 h at room temperature. Toluene was removed under reduced pressure at room temperature. The remaining colorless solid was dissolved in hexane/Et₂O (2 mL, 1/1). The resulting solution was stored at −40 °C, yielding diffraction quality crystals of Lu(CH₂SiMe₃)₃(Me₂NCH₂CH₂NMe₂) (Lu-TMEDA) after two days. Yield: 38 mg, 72% based on Lu(CH₂SiMe₃)₃(THF)₂.

¹H NMR (600 MHz, C₆D₆) δ 1.82 (s, 12H), 1.48 (s, 4H), 0.41 (s, 27H), −0.61 (s, 6H). ¹³C NMR (151 MHz, C₆D₆) δ 56.60, 46.23, 45.51, 4.87.

Anal. calcd for LuC₁₈H₄₉Si₃N₂: C, 39.11, H, 8.93; N, 5.07. Found: C, 39.11; H, 8.97; N, 4.99.

Synthesis of Sc(CH₂SiMe₃)₃(C₆H₁₆P₂) (Sc-DMPE). In an Ar-filled glovebox, Sc(CH₂SiMe₃)₃(C₄H₈O)₂ (45.5 mg, 0.101 mmol) was dissolved in 5 mL of toluene and cooled to −40 °C in the glovebox freezer. To the cold clear yellow solution bis(dimethylphosphino)ethane (C₆H₁₆P₂, DMPE) (15 mg, 0.101 mmol, 1 eq.) in 3 mL toluene was added dropwise. The resulting clear yellow mixture was stirred for 1 h at room temperature. Toluene was removed under reduced pressure at room temperature. The remaining yellow oil was extracted with n-hexane (5 × 3 mL). The combined colorless extracts were filtered, concentrated and cooled to −40 °C yielding colorless block shaped diffraction quality crystals of Sc(CH₂SiMe₃)₃(C₆H₁₆P₂) (Sc-DMPE) after five days. Yield: 13 mg, 28% based on Sc(CH₂SiMe₃)₃(THF)₂.

¹H NMR (600 MHz, C₆D₆) δ 0.92 (s, 4H), 0.76 (t, J = 1.6 Hz, 12H), 0.40 (s, 27H), 0.26 (s, 6H). ¹³C NMR (151 MHz, C₆D₆) δ 28.15, 14.04, 14.01, 13.97, 13.94. ³¹P NMR (243 MHz, C₆D₆) δ −36.30.

Anal. calcd for ScC₁₈H₄₉Si₃P₂: C,47.33; H, 10.81. Found: C, 46.95; H, 10.56.

Synthesis of Lu(CH₂SiMe₃)₃(C₆H₁₆P₂)(C₄H₈O) (Lu-DMPE). In an Ar-filled glovebox, Lu(CH₂SiMe₃)₃(C₄H₈O)₂ (58 mg, 0.101 mmol) was dissolved in 5 mL toluene and cooled to −40 °C in the glovebox freezer. To the resulting cold clear yellow solution 1,2-bis(dimethylphosphino)ethane (C₆H₁₆P₂, DMPE) (15 mg, 0.101 mmol, 1 eq.) in 3 mL toluene was added dropwise. The resulting clear colorless mixture was stirred for 1 h at room temperature. Toluene was removed under reduced pressure at room temperature. The remaining colorless solid was dissolved in a 1 : 1 mixture of hexane and Et₂O (2 mL). The resulting solution was stored at −40 °C, yielding diffraction quality crystals of Lu(CH₂SiMe₃)₃(C₆H₁₆P₂)(C₄H₈O) (Lu-DMPE) after three days. Yield: 41 mg, 60% based on Lu(CH₂SiMe₃)₃(THF)₂.

¹H NMR (600 MHz, C₆D₆) δ 3.91 (s, 4H), 1.36–1.28 (m, 4H), 1.12 (t, J = 7.0 Hz, 4H), 0.78 (s, 12H), 0.36 (s, 27H), −0.63 (s, 6H). ¹³C NMR (151 MHz, C₆D₆) δ 25.13, 12.30, 4.77. ³¹P NMR (243 MHz, C₆D₆) δ −38.88.

Anal. calcd for LuC₂₂H₅₇Si₃P₂O: C,40.1; H, 8.72. Found: C, 39.89; H, 8.62.

Data availability

The data supporting this article have been included as part of the ESI.† Crystallographic data are deposited at CCDC, codes 2351142–2351150, 2369795 and can be obtained from https://www.ccdc.cam.ac.uk.

Author contributions

ET, AB, MJ: experiments, analysis, writing & editing. PLA, JK, SP, PF: project conceptualization, funding acquisition, supervision, analysis, writing & editing.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

We thank Dow Chemical for support of this work. Some of the alkyl kinetic studies were also supported by the Catalysis Program of the U.S. Department of Energy (DOE), Office of Science, Office of Basic Energy Sciences, Chemical Sciences, Geosciences, and Biosciences Division at the Lawrence Berkeley National Laboratory under Contract DE-AC02-05CH11231. We thank Dr Cooper Citek and the Catalysis Laboratory in the DOE Catalysis Program for resource and instrumentation used in this work. We acknowledge the National Institutes of Health (NIH) for funding the UC Berkeley College of Chemistry NMR facility under grant no. S10OD024998 and the NIH Shared Instrumentation Grant S10-RR027172 for funding the UC Berkeley College of Chemistry Small Molecule X-ray Crystallography Facility (CheXray).

References

1. M. F. Lappert and R. Pearce, J. Chem. Soc. Chem. Commun., 1973, 126,  10.1039/C39730000126.
2. M. Zimmermann and R. Anwander, Chem. Rev., 2010, 110, 6194–6259.
3. M. Zimmermann, J. Takats, G. Kiel, K. W. Tornroos and R. Anwander, Chem. Commun., 2008, 612–614,  10.1039/b713378b.
4. S. Arndt, T. P. Spaniol and J. Okuda, Angew. Chem., Int. Ed., 2003, 42, 5075–5079.
5. C. Queffelec, F. Boeda, A. Pouilhès, A. Meddour, C. Kouklovsky, J. Hannedouche, J. Collin and E. Schulz, ChemCatChem, 2011, 3, 122–126.
6. S. C. Kosloski-Oh, K. D. Knight and M. E. Fieser, Inorg. Chem., 2023, 63, 9464–9477.
7. J. Zhai, F. You, S. Xu, A. Zhu, X. Kang, Y.-M. So and X. Shi, Inorg. Chem., 2022, 61, 1287–1296.
8. P. L. Arnold, M. W. McMullon, J. Rieb and F. E. Kuhn, Angew. Chem., Int. Ed., 2015, 54, 82–100.
9. H. Liu, S. Saha and M. S. Eisen, Coord. Chem. Rev., 2023, 493, 215284.
10. S.-M. Chen, Y.-Q. Zhang, J. Xiong, B.-W. Wang and S. Gao, Inorg. Chem., 2020, 59, 5835–5844.
11. M. F. Lappert, in Inorganic Compounds with Unusual Properties, American Chemical Society, 1976, ch. 21, vol. 150, pp. 256–265.
12. D. J. H. Emslie, W. E. Piers, M. Parvez and R. McDonald, Organometallics, 2002, 21, 4226–4240.
13. H. Schumann and J. Müller, J. Organomet. Chem., 1979, 169, C1–C4.
14. J. J. Carbó, D. García-López, M. Gómez-Pantoja, J. I. González-Pérez, A. Martín, M. Mena and C. Santamaría, Organometallics, 2017, 36, 3076–3083.
15. K. A. Rufanov, D. M. M. Freckmann, H.-J. Kroth, S. Schutte and H. Schumann, Z. Naturforsch. B Chem. Sci., 2005, 60, 533–537.
16. A. R. Kennedy, J. Klett, R. E. Mulvey and D. S. Wright, Science, 2009, 326, 706–708.
17. A. Mortis, C. Maichle-Mössmer and R. Anwander, Dalton Trans., 2022, 51, 1070–1085.
18. R. Shannon, Acta Crystallogr., Sect. A: Cryst. Phys., Diffr., Theor. Gen. Crystallogr., 1976, 32, 751–767.
19. H. Schumann, D. M. Freckmann and S. Dechert, Z. Anorg. Allg. Chem., 2002, 628, 2422–2426.
20. A. G. B. Getsoian, B. Hu, J. T. Miller and A. S. Hock, Organometallics, 2017, 36, 3677–3685.
21. A. Mortis, F. Kracht, T. Berger, J. Lebon, C. Maichle-Mössmer and R. Anwander, Dalton Trans., 2023, 52, 44–51.
22. R. R. Golwankar, T. D. Curry II, C. J. Paranjothi and J. D. Blakemore, Inorg. Chem., 2023, 62, 9765–9780.
23. A. O. Tolpygin, A. S. Shavyrin, A. V. Cherkasov, G. K. Fukin and A. A. Trifonov, Organometallics, 2012, 31, 5405–5413.
24. W. J. Evans, J. C. Brady and J. W. Ziller, J. Am. Chem. Soc., 2001, 123, 7711–7712.
25. S. Arndt, P. M. Zeimentz, T. P. Spaniol, J. Okuda, M. Honda and K. Tatsumi, Dalton Trans., 2003, 3622–3627.
26. F. P. Gabbaï, P. J. Chirik, D. E. Fogg, K. Meyer, D. J. Mindiola, L. L. Schafer and S.-L. You, Organometallics, 2016, 35, 3255–3256.
27. R. E. H. Kuveke, L. Barwise, Y. van Ingen, K. Vashisth, N. Roberts, S. S. Chitnis, J. L. Dutton, C. D. Martin and R. L. Melen, ACS Cent. Sci., 2022, 8, 855–863.
28. S. D. Robertson, M. Uzelac and R. E. Mulvey, Chem. Rev., 2019, 119, 8332–8405.
29. M. de Giovanetti, S. H. Hopen Eliasson, A. C. Castro, O. Eisenstein and M. Cascella, J. Am. Chem. Soc., 2023, 145, 16305–16309.
30. B. Haag, M. Mosrin, H. Ila, V. Malakhov and P. Knochel, Angew. Chem., Int. Ed., 2011, 50, 9794–9824.
31. D. E. Anderson, A. Tortajada and E. Hevia, Angew. Chem., Int. Ed., 2023, 62, e202218498.

---

Footnote

† Electronic supplementary information (ESI) available: Synthetic and mechanistic details. CCDC 2351142–2351150, and 2369795. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d4sc05983b

---