Bis(phosphinimino)methanide borohydride complexes of the rare-earth elements as initiators for the polymerization of methyl methacrylate: combined experimental and computational investigations

Abstract

Rare earth borohydride complexes are known as efficient initiators for the polymerization of both apolar and polar monomers. Significant contribution of the phosphiniminomethanide ligand on the reactivity of group 3 derivatives was previously established in the polymerization of ε-caprolactone. Investigations of the capability of bis(phosphinimino)methanide rare earth metal bisborohydrides, [{CH(PPh₂NSiMe₃)₂}La(BH₄)₂(THF)] (1) and [{CH(PPh₂NSiMe₃)₂}Ln(BH₄)₂] (Ln = Y (2), Lu (3)) to polymerize methyl methacrylate (MMA) both experimentally and computationally, are reported here. All three metallic compounds allowed the preparation of PMMA at room temperature. However, the overall performances of 1–3 remain quite poor based on experimental observations. DFT investigations on the insertion of the first two MMA molecules revealed that the incoming of the first MMA molecule was the most important step regardless of the nature of the metal center. The nucleophilic attack of MMA leads to the formation of a first adduct B followed by the unprecedented trapping of the liberated BH₃ group by the nitrogen of the phosphiniminomethanide ligand to afford the active enolate species C. The unique significant role played by the phosphiniminomethanide ligand has thus been clearly unveiled and evidenced computationally. This whole first insertion of the MMA process is both kinetically and thermodynamically favorable. Trapping of the BH₃ by the ancillary ligand appeared to make the second MMA insertion more energetically favorable than the first one, especially for lanthanum. The ketoenolate thus formed, Prod., is thermodynamically and kinetically favorable. In the case of yttrium, steric considerations, in addition to energetically comparable first and second MMA insertions, support the experimentally observed difficulty to polymerize MMA. DFT calculations closely corroborate experimental findings.

---

1 Introduction

Poly(methylmethacrylate) (PMMA) is a highly important member of the poly(acrylate)s family. As a transparent thermoplastic, it is often used as a light or shatter-resistant alternative to glass. Other major common applications include paints and optical devices. As a chiral molecule, methyl methacrylate (MMA) may offer, under suitable operating conditions, PMMAs of various tacticities and therefrom, polyacrylates with possibly distinct stereochemistry.¹ The polymerization of polar monomers, among which MMA, mediated by group 3 metal systems has gained significant attention since the pioneering achievements of the group of Yasuda in the polymerization of MMA, as extensively reviewed.^2–5 Both the nature of the metal center as well as that of the surrounding ligand(s), especially in terms of electronic or steric factors and bonding interactions, have been shown to greatly affect the nature of the initiating species thereby influencing the activity, productivity, regioselectivity and stereoselectivity of the catalytic systems. Similarly, the macromolecular chemical features, specially the molar mass, molar mass distribution, and stereoregularity, as well as the PMMA physical properties, in particular its thermal transition temperatures and crystallinity are, to a great extent, governed by the identity of the organolanthanide involved.^6–8

In the early 1990s, MMA has been successfully polymerized, within a short reaction time and with a quantitative conversion, into the highest syndio-enriched (>95% rr) PMMAs ever prepared from a rare earth system. These polymers also exhibited high molar masses (M̄_n ≈ 1.06 g mol⁻¹) and extremely narrow molar mass distributions (M̄_w/M̄_n = 1.02–1.05). These performances were obtained from the bis(pentamethylcyclopentadienyl)lanthanide hydride or methyl complexes [(Cp*)₂Ln(R)] (η⁵-Cp* = C₅Me₅; Ln = Sm, Yb, Y, Lu; R = H, Me).^3,9–13 Other organolanthanide precursors were reported to similarly favor the syndiotactic enrichment of PMMA, however, not to such a large extent.^14–30 On the other hand, isotactic-enriched PMMA prepared from related group 3 metals are much less common.^31–34

Regarding rare earth borohydride derivatives, mono-, bis- or tris-borohydridodiamide-diamine, guanidinate, amino- or alkyl-substituted cyclopentadienyl, and alkoxide complexes have been recently reported to polymerize MMA.^35–40 Molar mass values typically ranged from 7500 up to 100 000 g mol⁻¹ (as high as 615 000 g mol⁻¹) whereas molar mass distributions were quite large (1.2 < M̄_w/M̄_n < 2.9), and even multi-modal in several instances. Reported yields were, although rarely, as high as quantitative, and varied from one catalytic system to another as well as according to the experimental conditions. In general, rare earth borohydrides and their derivatives showed a lower performance than comparable alkyl and hydride systems. These observations are in marked contrast with the valuable activities of rare earth borohydride derivatives observed in the ring-opening polymerization (ROP) of cyclic esters, especially affording the important α,ω-dihydroxy telechelic polyesters.^41–48 Based on these considerations, we were interested in studying the origin of these differences by combining experimental and theoretical methods, in line with our ongoing concerns.⁴⁸

Theoretical work by some of us aimed at establishing polymerization pathways of MMA involving borohydride group 3 derivatives in comparison to the ubiquitous hydrides congeners using [Eu(BH₄)₃], [Cp₂Eu(BH₄)], and [Cp₂EuH], respectively, as models.³⁹ The calculated free-energy profile for the reaction of MMA with [Cp₂EuH] supports the formation of a five-membered ring enolate upon hydride attack on the CH₂ group of MMA. The formation of this enolate ([Cp₂Eu{OC(OMe)=CMe₂}]) was computed to be both thermodynamically (exergonic by 31.1 kcal mol⁻¹) and kinetically favorable (−2.1 kcal mol⁻¹ with respect to the separated reactants). The DFT results further support the mechanism proposed by Yasuda, favoring a MMA polymerization mechanism involving the enolate [Cp*₂Sm{OC(OMe)=CMeCH₂CMe₂C(O)(OMe)}] as suggested from experimental results. A coordination–insertion mechanism which favored coordination of the MMA molecule via the carbonyl oxygen was also confirmed computationally. Regarding [Cp₂Eu(BH₄)], all possible B–H activations including enolate or carboxylate formation and hydroboration reactions were investigated. The thermodynamic results are summarized in Scheme 1.³⁹

Scheme 1 Schematic representation of the thermodynamic results for the reaction of [Cp₂Eu(BH₄)] with MMA.³⁹

Formation of the carboxylate f was found kinetically unlikely whereas neither hydroboration products g or h may initiate the polymerization of MMA because the boranes exhibit an aliphatic carbon chain rather than a classical unsaturated one. Unlike these products of the hydroboration reaction or carboxylate formation, the BH₃-free enolate formation (e, Scheme 1-i) is thermodynamically unfavorable (endergonic by 12.0 kcal mol⁻¹), due to the kinetic low stability of free BH₃. Subsequent trapping of the BH₃ by the oxo group of the enolate would yield [Cp₂Eu(OBH₃)(OMe)C=CMe2] (e-2, Scheme 1). Formation of this species e-2 is thermodynamically favorable (exergonic by 8.1 kcal mol⁻¹) and could provide a pathway for MMA polymerization. Therefore, although formation of the enolate e is disfavored, a small amount of the borane adduct e-2 could in principle initiate the polymerization of MMA. The thermodynamic features are thus consistent with the non-quantitative and poorly controlled polymerization process observed experimentally. Energetically, [Eu(BH₄)₃] behaves in a similar fashion to [Cp₂Eu(BH₄)], but with all the reactions being more favorable.³⁹ Focusing on the enolate formation, the corresponding ensuing borane adduct (equivalent of e-2) is however thermodynamically more favorable (exergonic by 14.4 kcal mol⁻¹).

Experimentally, some of us have been working for some time with the bis(phosphinimino)methanide {CH(PPh₂NSiMe₃)₂}⁻ ligand in rare earth metal chemistry^30,49–58 In this context, the chloride complexes [{CH(PPh₂NSiMe₃)₂}Ln{(Ph₂P)₂N}Cl] (Ln = Y, La, Nd, Yb) were shown to initiate the ROP of CL and the polymerization of MMA.³⁰ Very recently, the bisborohydride analogues, [{CH(PPh₂NSiMe₃)₂}La(BH₄)₂(THF)] (1) and [{CH(PPh₂NSiMe₃)₂}Ln(BH₄)₂] (Ln = Y (2), Lu (3)) were demonstrated as successful initiators for the ROP of CL at 0 °C. The narrowest molar mass distribution values ever obtained from a rare earth metal borohydride initiator were determined.⁴⁸ DFT investigations of this ROP revealed a rather unique mechanism. Indeed, with this latter borohydride complexes, the first B–H activation of BH₄⁻ is achieved in two distinct unprecedented steps,⁴⁸ as opposed to the unique step evidenced with the monoborohydride complexes such as the metallocene [Cp₂Eu(BH₄)].⁵⁹ These differences arise from the simultaneous presence of two BH₄⁻groups along with the beneficial greater electron-donating ability of the phosphiniminomethanide unit.⁴⁸ The positive influence of this {CH(PPh₂NSiMe₃)₂}⁻ ancillary was further suggested by DFT findings that highlighted a reduced charge at the metal center resulting from the charge inherent to the {CH(PPh₂NSiMe₃)₂}⁻ ligand, thereby promoting valuable electrostatic interactions in the intermediates. The favorable signature of the phosphiniminomethanide ligand being thus highlighted, it prompted further studies on phosphiniminomethanide rare earth initiators in polymerization processes.

Based on these observations, we now disclose in the present contribution our results on the polymerization of MMA by using phosphiniminomethanide bisborohydride compounds 1–3 as initiators.

2 Results and discussion

The bis(phosphinimino)methanide rare earth metal bisborohydrides, both solvated [{CH(PPh₂NSiMe₃)₂}La(BH₄)₂(THF)] (1) and unsolvated [{CH(PPh₂NSiMe₃)₂}Ln(BH₄)₂] (Ln = Y (2), Lu (3)), were obtained by two different synthetic approaches, as previously reported.⁴⁸ Compounds 1 and 3 could be best prepared by treatment of [Ln(BH₄)₃(THF)₃]⁶⁰ with K{CH(PPh₂NSiMe₃)₂},⁶¹ whereas compound 2 was obtained by reaction of [{CH(PPh₂NSiMe₃)₂}YCl₂]₂⁴⁹ with NaBH₄ (Scheme 2).⁴⁸

Scheme 2 Phosphiniminomethanide rare earth complexes [{CH(PPh₂NSiMe₃)₂}La(BH₄)₂(THF)] (1) and [{CH(PPh₂NSiMe₃)₂}Ln(BH₄)₂] (Ln = Y (2), Lu (3)) used in the polymerization of MMA.⁴⁸

Compounds 1–3 were evaluated for their ability to polymerize MMA at room temperature. Both apolar (toluene) and polar (THF) solvents were used and the ratio of MMA to rare earth complex was increased from 100 to 800. Representative results are summarized in Table 1.

Table 1 Polymerization of MMA initiated by [{CH(PPh₂NSiMe₃)₂}La(BH₄)₂(THF)] (1) and [{CH(PPh₂NSiMe₃)₂}Ln(BH₄)₂] (Ln = Y (2), Lu (3)) at 23 °C

Entry	Ln	[MMA]0/[Ln]0^a	Solvent	Reaction time^b	Yield^c (%)	M̄ n,SEC ^d/g mol^−1	M̄ w/M̄n^e	Tacticity^f (rr–mr–mm)
a [Ln]0 = 20.0 mmol L^−1. b Reaction times were not necessarily optimized. c Monomer conversion determined by gravimetry. d Determined by SECvs.polystyrene standards. e Molar mass distribution calculated from SEC traces. f The ratio of syndio-, iso- and heterotactic linkages between monomer units determined by ^1H NMR in CDCl3. g Polymer insoluble in THF. h Bimodal SEC peak.
1	2	100	Toluene	1 h 15	<5	—	—	44–31–25
2	2	200	Toluene	1 h 45	<5	—^g	—^g	27–36–37
3	2	800	THF	2 h 10	<5	—^g	—^g	30–45–25
4	2	800	THF	24 h	<5	—^g	—^g	39–44–17
5	2	800	Toluene	24 h	<5	—^g	—^g	30–43–27
6	1	100	Toluene	1 h 15	<5	3200	2.36	31–40–29
7	1	200	Toluene	1 h 45	<5	7000	1.56	27–38–35
8	1	800	THF	2 h 10	9	—^g	—^g	20–45–35
9	1	800	THF	24 h	<5	—^g	—^g	20–42–38
10	1	800	Toluene	24 h	<5	—^g	—^g	22–24–54
11	3	100	Toluene	1 h 15	10	16 000	1.37	61–27–12
12	3	200	Toluene	1 h 45	9	138 600	1.04^h	—
						60 350	1.03	—
13	3	800	THF	2 h 10	8	—^g	—^g	46–34–20
14	3	800	THF	24 h	10	—^g	—^g	51–33–16
15	3	800	Toluene	24 h	11	—^g	—^g	41–38–21

---

The borohydride complexes of both lanthanum, 1, and lutetium, 3, do polymerize MMA to some extent whereas the yttrium derivative, 2, appears less active, whichever the operating conditions (solvent, reaction time). The yields of PMMA recovered remain generally low for all three catalytic systems. In the case of 2, very little polymer could be collected. A similar trend was observed in the polymerization of MMA using the analogous (phosphinimino)methanide chloro complexes combined with the amido ligand, [{CH(PPh₂NSiMe₃)₂}Ln{(Ph₂P)₂N}Cl] (Ln = Y, La).³⁰ This behavior may suggest the presence of impurities in the monomer which would rapidly degrade the borohydride initiators that are known to be quite sensitive to protic impurities, thereby inhibiting the polymerization. Therefore, the absence of contamination in MMA has been verified by the successful polymerization of trimethylene carbonate (TMC) ran in the presence of an equimolar amount of MMA using compound 3.⁶² This, in fact, attempted copolymerization of TMC and MMA only resulted, within the experimental conditions, in the formation of the polycarbonate with the expected features.^47,63

This experiment first showed that MMA was pure enough not to decompose the metallic borohydride complexes within a few minutes.⁶² However, the poor yields obtained in the MMA homopolymerization from 1–3 (Table 1) may suggest that the borohydride complexes 1–3 placed in the presence of MMA (from 100 to 800 equiv.) over more than an hour may indeed decompose to some extent. Note that, no polycarbonate–polyacrylate copolymer was formed within the reaction time allotted. This may not be so surprising, taking into account the failure to copolymerize ε-caprolactone and MMA from the rare earth trisborohydride derivatives [Ln(BH₄)₃(THF)₃].⁶⁴ The PMMAs recovered from the homopolymerization of MMA using 1–3 were not always fully soluble in THF (a behavior previously observed)^30,35,37 precluding systematic SEC characterization and suggesting, in these instances, high molar mass polymers. When possible, SEC analysis of the PMMA in THF revealed most often a unique unimodal peak (unless otherwise stated); however, the molar mass distribution values were quite large. These results suggest a poor initiation efficiency of the initiators.

Rather atactic PMMA was obtained with compounds 1 and 2 whereas 3 afforded rather syndio-enriched PMMA up to rr 61%. Similarly, the previously reported lanthanum analogue [{CH(PPh₂NSiMe₃)₂}La{(Ph₂P)₂N}(NPh₂)] preferentially afforded atactic PMMA.³⁰ However, polymerizations of MMA by borohydrido species used without any co-catalyst were generally rather syndio-specific (up to rr 81.8%).^35–40 Overall, the present results are comparable in terms of yield, activity, molar mass, molar mass distribution and tacticity, to previously reported data on MMA polymerization using neat (no addition of any alkylating agent) rare earth borohydride complexes.^35–40 Such (phosphinimino)methanide complexes were however much more active in the presence of an added alkylating cocatalyst or toward the ROP of cyclic esters.^30,48,62,63

The reaction mechanism was next investigated using DFT methods in order to provide energetic elements that would rationalize the poor efficiency observed experimentally. Particularly, attention was focused on the initiation process, based on the assumption that the rate of the propagation is faster than that of the initiation. The Gibbs free energy profiles have been determined at room temperature for the addition of the first couple of MMA monomer units to both the smaller yttrium, 2 (Fig. 1a), and the larger lanthanum, 1 (Fig. 1b), initiators.

Fig. 1 Calculated free energy profiles for the reaction of MMA with (a) [{CH(PPh₂NSiMe₃)₂}Y(BH₄)₂], 2, and (b) [{CH(PPh₂NSiMe₃)₂}La(BH₄)₂(THF)], 1.

As the profiles (and especially the geometries of the stationary points) are similar for 2 and 1, for the sake of clarity, only the geometries related to 2 are discussed in detail. The reaction begins by the nucleophilic attack of MMA resulting in the MMA coordination to the metal center via the oxygen of the carbonyl, as already reported by some of us for the reaction between MMA and [Cp₂Eu(BH₄)].³⁹ In the resulting intermediate A, the C=O double bond remains formed and the borohydride ligand is in an η³ bonding mode to the metal center in agreement with the X-ray structure.⁴⁸ The coordination is predicted to be endergonic by 4.3 kcal mol⁻¹ for 2 (0.5 kcal mol⁻¹ for 1). The system evolves with the nucleophilic attack of BH₄⁻ onto the CH₂ group of the MMA to reach the transition state (TS) between A and B, TSAB. In doing so, the MMA has rotated by 90° and then lies in the perpendicular plane so as to point the CH₂ toward the incoming hydride. The CH₂ which is thus pyramidalized (sum of the angles around C of 349.5°) is thereby prepared to accept the hydride. So far, the hydride is still coordinated to the boron atom (B–H distance of 1.33 Å and C–H distance of 1.41 Å). The activation barrier is found to be kinetically accessible (18.9 kcal mol⁻¹ for 2 and 16.5 kcal mol⁻¹ for 1). Following the intrinsic reaction coordinate leads to the formation of adduct B in which the BH₄ ligand linked to yttrium is still formed, although the B–H bond distance in B⋯H⋯CH₂ has increased by 0.11 Å (to 1.44 Å) while the C–H bond is fully formed (1.22 Å). The BH₄ ligand is still interacting with the metal center through two hydrogens. A natural bonding (NBO) analysis also indicates the enolate character, reminiscent of the adduct observed in the polymerization of MMA by [Cp₂Eu(BH₄)],³⁹ of this adduct. Indeed, three lone pairs are found on the oxygen concomitantly to a C=C double bond that is relocalized. This adduct B is hardly stabilized (0.1 kcal mol⁻¹ for both 2 and 1) with respect to TSAB almost forming a plateau.

In polymerization processes involving borohydride derivatives, the main problem is not the release of BH₃ from the metal-coordinated BH₄ ligand but the becoming/trapping of the BH₃ moiety. Noteworthy, some of us recently demonstrated that reaction of [Sc(BH₄)₃(THF)₂] and [Lu(BH₄)₃(THF)₃] with an iminomethyl pyrrole-type ligand leads to the formation of a stable borohydride derivative bearing one BH₃ molecule interacting with the metal while remaining trapped by the nitrogen of the iminomethyl pyrrole moiety, as supported by the single crystal X-ray diffraction structure.^65,66 Thus, such a potential trapping of the borane by a nitrogen containing ligand is an important issue which we investigated in the present study. From adduct B, the resulting TS, TSBC, has been located on the Potential Energy Surface (PES). From a geometrical point of view, the TS is classical of a trapping one. Indeed, the BH₃ moiety is equidistant from the methyl and the nitrogen of the bis(phosphinimino)methanide ligand (B–N distance of 2.71 Å and B–C distance of 2.72 Å). In TSBC, the BH₃ is fully formed and it is planar (sum of angles around B is 360°). At the same time, the oxygen coordination to the metal center is reinforced (Y–O distance 2.13 vs. 2.20 Å in B). The energetic barrier is low (3.7 kcal mol⁻¹ with respect to adduct B for 2, and 6.2 kcal mol⁻¹ for 1). This height can be explained by the strong electrostatic interaction that controls the reaction (NBO charge of −1.58 on N and of +0.65 on BH₃). The BH₃ in TSBC has thus adopted an intermediate position, moving from a BH₄ group interacting with the metal in B, to a BH₃ group in TSBC that is still interacting with the center through a bridging hydrogen atom (η¹ coordination mode) and that is getting closer to the phosphiniminomethanide ligand.

Moving further on the reaction coordinate leads to the formation of the enolate complex C in which the BH₃ is then distinctively trapped by the nitrogen of the bis(phosphinimino)methanide ligand. The formation of complex C is predicted to be highly favorable (exergonic by 8.1 kcal mol⁻¹ for 2 and 8.7 kcal mol⁻¹ for 1). In all other studies of the polymerization of polar monomers by rare earth borohydride initiators, the liberated BH₃ moiety typically forms an adduct with a solvent molecule present in the medium (BH₃THF) or with oxygen of the MMA moiety itself.^39,48,59 Such a BH₃N(PPh₂)(SiMe₃) adduct observed here in C, with the nitrogen still bound to the metal center, is the distinctive feature of the first insertion as well as of the whole insertion step of MMA involving borohydride species. Such a trapping of the BH₃ moiety by the ancillary ligand in the coordination sphere of the metal is computationally observed here for the first time. This is further in excellent agreement with the experimental observations mentioned above on the interaction of the BH₃ moiety trapped by an iminomethyl pyrrole-type ancillary.^65,66 This originality is definitively a signature of the quite unique phosphiniminomethanide ligand itself. In contrast to the borane zinc complex [{CH(Ph₂PNSiMe₃)₂}Zn(κ¹-BH₃)ZnMe], the BH₃ molecule is, in the present results, not coordinated to the methine group.⁶⁷ The trapping of the BH₃ by the nitrogen induces rearrangement in the ligand to allow a lone pair of nitrogen to interact with the boron. The Y–C bond is elongated by 0.1 Å (to 2.82 Å) and the P–N is also elongated by 0.06 Å (1.71 vs. 1.65 Å). At this stage, the initiation process is found to be a kinetically and thermodynamically favorable reaction. This is another major difference between the bis(phosphinimino)methanide borohydride complexes and the [Cp₂Eu(BH₄)] or [Eu(BH₄)₃] complexes.³⁹ Indeed, in these latter studies, the formation of the enolate was endergonic in all cases and the only possibility for trapping BH₃ was by the oxygen of the enolate (e-2, Scheme 1), an intermediate which would reduce the reactivity of the enolate.⁴² Thus, the relative low experimental activity of the bis(phosphinimino)methanide complexes most likely does not result from some problem occurring during the initiation process, since the formation of the enolate complex C is favorable. However, trapping of BH₃ by the ligand may influence the subsequent reactivity of the enolate. Therefore, the second insertion of MMA has also been considered.

The second inserted MMA molecule begins by its coordination to the enolate complex C, to form adduct D. The coordination of MMA is again ensured by the oxygen of the carbonyl. In D, this new Y–O distance is longer than the Y–O distance of the enolate (2.34 vs. 2.11 Å). Interestingly, the coordination of the second monomer takes place parallel to the enolate already in position, with the methoxy group of both chains pointing away in the opposite direction. This leads to a syndiotactic arrangement of the two consecutive MMA units.⁶⁸ This orientation is mainly controlled by the steric repulsion between the two methoxy groups. The ancillary BH₄ ligand remains η³ coordinated to the metal, so that the coordination of the second MMA molecule is increasing drastically the steric hindrance around the metal center. The coordination is endergonic by 13.5 kcal mol⁻¹ for 2 (only 8.2 kcal mol⁻¹ for 1, in line with the steric effects since La is bigger than Y⁶⁹). The system is then evolving to Michael transition state TSDP. At the TS, the CH₂ group of the incoming MMA faces the CMe₂ of the enolate already in place. This C–C distance remains long (2.43 Å) and the two groups begin to pyramidalize. CH₂ is slightly pyramidalized inward (sum of the angles around C is 356°) whereas CMe₂ is slightly pyramidalized outward (sum of the angles around C is 352°), suggesting that a coupling will occur. Except these two changes, the geometries of the MMA and enolate are not much affected (e.g. the Y–O distance is reduced to 2.27 Å as compared to 2.34 Å in D). The barrier is found to be accessible (19.5 kcal mol⁻¹ for 2 and 12.2 kcal mol⁻¹ for 1, with respect to C). Following the reaction coordinate, TSDP leads to the formation of the ketoenolate complex Prod. The two Y–O distances are now 2.10 Å (2.27 Å at the TS) and 2.40 Å (2.11 Å at the TS). Moreover, the latter oxygen forms again a double bond with the carbon whereas the other oxygen is now part of the enolate group. The formation of the ketoenolate Prod. from adduct C is exergonic by 1.3 kcal mol⁻¹ for 2 (4.3 kcal mol⁻¹ for 1). This difference is related to the steric hindrance, as already mentioned for adduct D. Thus, the propagation of the polymerization is found to be kinetically and thermodynamically favorable, but the problem of the steric hindrance for complex 2 is significant. Furthermore, in the case of 2, it should also be noticed that the barrier for the second MMA insertion is comparable to the barrier for the first MMA insertion, so that all insertions are basically equivalent thereby indicative of a low activity. These results provide some information as to why the polymerization of MMA from the yttrium complex experimentally hardly gives any PMMA. In the case of 1, the barrier for the second MMA insertion is significantly lower than that for the first MMA insertion (12.2 kcal mol⁻¹vs. 22.5 kcal mol⁻¹, respectively) so that the activity should be higher once a first MMA molecule is added. However, although to a smaller extent than for 2, the steric effect is still important for 1 (the second insertion is less thermodynamically favorable than the first one (4.3 kcal mol⁻¹vs. 8.7 kcal mol⁻¹, respectively)). Again, these observations are in agreement with the poor (although slightly better than for 2) polymerization efficiency of 1 in the polymerization of MMA.

3 Conclusion

In this contribution, we have investigated the capability of the phosphiniminomethanide bisborohydride rare earth complexes 1–3 to polymerize MMA from a comparative experimental and computational approach. All three metallic compounds polymerize MMA at room temperature. The lanthanum, 1, and lutetium, 3, polymerize MMA to a slightly larger extent than the yttrium derivative, 2. However, 1–3 remain poorly active.

Theoretical studies on the polymerization mechanism using DFT calculations essentially focused on the insertion of the first and the second MMA units using the yttrium 2 and the lanthanum 1 derivatives. Regardless of the nature of the metal center, the first insertion involves first the nucleophilic attack of MMA to form adduct B followed by the trapping of the BH₃ entity affording the enolate C. This whole process is both kinetically and thermodynamically favorable. No major energetic difference is observed between the process involving either the lighter or the heavier metal. Formation of the active enolate species C is established as being favorable. However, steric restrictions may impede the polymerization of MMA in the case of 2. The distinctive and original feature of this first MMA insertion lies in the trapping of the BH₃ moiety. Indeed, as observed here for the first time computationally, the BH₃ group in C is being trapped by the phosphiniminomethanide ligand present in the coordination sphere of the metal, and not by the coordinated MMA unit, while still interacting with the metal. This finding is corroborated by the experimental observation of such a BH₃ trapping by the nitrogen of the related (diisopropylphenyl)iminomethyl pyrrole lutetium and scandium bisborohydride complexes, as evidenced by a single crystal X-ray diffraction structure.^65,66 Analysis of the insertion of the second incoming MMA molecule (C to Prod.) revealed an overall energetically more favorable process as compared to the first insertion (reactants to C). The second insertion leading to the ketoenolate species Prod. was thermodynamically and energetically more favorable with the lanthanum complex 1 as compared to the more costly step for the yttrium analogue 2. The first MMA insertion thus rather appears as the limiting step in the polymerization of MMA, in comparison to the second insertion—and most likely to all subsequent insertions—of MMA. Such a difference between the first and the following additions of monomer units is first unveiled in the present studies.

Both thermodynamic and kinetic results obtained from these DFT studies may allow a tentative comparison between the activity of 1, 2 and Cp₂Eu(BH₄).³⁹ Based on all these data, 1 and 2 appear slightly more active than Cp₂Eu(BH₄) in which case the thermodynamics due to the BH₃ trapping has to be overcome; yet the energetic barrier is slightly higher for both 1 and 2 compared to Cp₂Eu(BH₄). A rather similar activity can thus be anticipated for all three complexes, 1, 2, and Cp₂Eu(BH₄).

These results revealed first that complexes 1–3 were not suitable for the polymerization of MMA, similarly to all other rare earth borohydride evaluated to date.^35–40 This is a major difference with the good performances of similar rare earth borohydride derivatives observed in the ROP of other polar monomers such as the cyclic esters.^41–48,63 In addition, the really beneficial influence of the phosphiniminomethanide ligand has been highlighted with its significant (steric, electronic, energetic) contribution to the polymerization process in trapping the liberated BH₃ group. Finally, DFT calculations have once more significantly contributed to the overall understanding and elucidation of key intermediates and major steps in the polymerization process of a polar monomer. Furthermore, these computational investigations have allowed a better understanding of the experimental observations allowing their rationalization.

4 Experimental section

Experimental details: materials

All manipulations were performed under inert atmosphere (argon, <3 ppm of O₂) using standard Schlenk, vacuum line and glove-box techniques. Solvents were thoroughly dried and deoxygenated by standard methods and distilled before use. CDCl₃ was dried over a mixture of 3 and 4 Å molecular sieves. Methyl methacrylate (MMA, 99%, Aldrich) was dried, stored cold over CaH₂ and distilled before use. [{CH(PPh₂NSiMe₃)₂}La(BH₄)₂(THF)] (1) and [{CH(PPh₂NSiMe₃)₂}Ln(BH₄)₂] (Ln = Y (2), Lu (3)) were prepared according to literature procedures.⁴⁸ Deuterated solvents were obtained from Aldrich (99 atom% D).

Instrumentation and measurements

NMR spectra were recorded on a Bruker Avance 400 MHz, Avance II 300 MHz or a Bruker Avance DPX 200 NMR spectrometer. Chemical shifts are referenced to internal solvent resonances and are reported relative to tetramethylsilane (¹H NMR), 15% BF₃Et₂O (¹¹B NMR), and 85% phosphoric acid (³¹P NMR), respectively. IR spectra were obtained on a Shimadzu FTIR-8400 s. Elemental analyses were carried out with an Elementar vario EL or EL III.

Average molar mass (M̄_n) and molar mass distribution (M̄_w/M̄_n) values were determined from chromatogram traces recorded by SEC in THF at 30 °C (flow rate = 1.0 mL min⁻¹) on a Polymer Laboratories PL50 apparatus equipped with a refractive index detector and a PLgel 5 Å MIXED-C column. The recovered polymer samples were dissolved in THF (2 mg mL⁻¹) and filtered; only the soluble fraction was analyzed. All elution curves were calibrated with polystyrene standards.

Typical procedure for MMA polymerisation

MMA (1.7 mL, 1.60 mmol) was added at room temperature onto the initiator 1 (14 mg, 20 μmol) previously placed in the solvent (1.0 mL). The resulting mixture was maintained at the same temperature and stirred over the appropriate reaction time. After drying under vacuum, the crude polymer was dissolved in CH₂Cl₂, purified upon precipitation in cold methanol, filtered and dried under vacuum. The resulting polymer was then analyzed by ¹H NMR and SEC.

Computational details

The calculations were performed in the gas phase. Although the solvent effect on the activity seems to be experimentally important, it corresponds to an energy difference of less than 5 kcal mol⁻¹ (this is also the precision of the computational method). Also, a recent study has shown that including solvent effect by means of single-point CPCM calculations on the gas phase optimized structures was not sufficient.⁷⁰Yttrium and lanthanum were represented with a Stuttgart–Dresden pseudo-potential in combination with their adapted basis set.⁷¹ The basis sets have been augmented by f function (α = 1.0). Silicon and phosphorus were also treated with a Stuttgart–Dresden pseudopotential in combination with their adapted basis set.⁷² The basis sets have been augmented by a set of d polarization function.⁷³Carbon, oxygen, nitrogen, and hydrogen atoms have been described with all electrons 6–31G(d,p) double-ζ quality basis sets.⁷⁴ Calculations were carried out at the DFT level of theory using the hybrid functional B3PW91.^75,76 Geometry optimizations were carried out without any symmetry restrictions; the nature of the extrema (minima and transition states) was verified with analytical frequency calculations. Connectivity of each transition state was determined while following their intrinsic reaction coordinates (IRC). All the computations were performed with the Gaussian 03 suite of programs.⁷⁷ The electronic density was analyzed according to the Natural Population Analysis (NPA) scheme.⁷⁸ The NBO analysis⁷² on lanthanum complexes was carried out using the technique proposed by Clark et al.⁷⁹

Acknowledgements

This work was supported by the Deutsche Forschungsgemeinschaft and the Helmholtz-Kolleg “Energy-Related Catalysis”. We are grateful to the CNRS and UPS for financial support of this work. LM is grateful to the Institut Universitaire de France. CalMip (CNRS, Toulouse, France) and CINES (CNRS, Montpellier, France) are acknowledged for calculation facilities.

Notes and references

1. G. Odian, in Principles of Polymerization, Wiley Inter Science, New York, 4th edn, 2004.
2. Z. Hou and Y. Wakatsuki, Coord. Chem. Rev., 2002, 231, 1–22.
3. H. Yasuda, J. Organomet. Chem., 2002, 647, 128–138.
4. H. Yasuda, Top. Organomet. Chem., 1999, 2, 255–283.
5. H. Yasuda and E. Ihara, Adv. Polym. Sci., 1997, 133, 55–101.
6. H. Yasuda and E. Ihara, Bull. Chem. Soc. Jpn., 1997, 70, 1745–1767.
7. J. Okuda, Dalton Trans., 2003, 2367–2378.
8. Z. Hou and Y. Wakatsuki, J. Organomet. Chem., 2002, 647, 61–70.
9. H. Yasuda, H. Yamamoto, K. Yokota, S. Miyake and A. Nakamura, J. Am. Chem. Soc., 1992, 114, 4908–4910.
10. H. Yasuda, H. Yamamoto, M. Yamashita, K. Yokota, A. Nakamura, S. Miyake, Y. Kai and N. Kanehisa, Macromolecules, 1993, 26, 7134–7143.
11. E. Ihara, M. Morimoto and H. Yasuda, Macromolecules, 1995, 28, 7886–7892.
12. E. Ihara, M. Morimoto and H. Yasuda, Proc. Jpn. Acad., Ser. B, Phys. Biol. Sci., 1995, 71, 126–131.
13. M. Tanabe, T. Sugimura and H. Yasuda, React. Funct. Polym., 2002, 52, 135–141.
14. M. Nodono, T. Tokimitsu, S. Tone, T. Makino and A. Yanagase, Macromol. Chem. Phys., 2000, 201, 2282–2288.
15. E. Kirillov, C. W. Lehmann, A. Razavi and J.-F. Carpentier, Organometallics, 2004, 23, 2768–2777.
16. C. Qian, W. Nie and J. Sun, Organometallics, 2000, 19, 4134–4140.
17. Y. Satoh, N. Ikitake, Y. Nakayama, S. Okuno and H. Yasuda, J. Organomet. Chem., 2003, 667, 42–52.
18. Z. Li-Ying, S. Hong-Ting, Y. Ying-Ming, Z. Yong and S. Qi, J. Organomet. Chem., 2007, 692, 2990–2996.
19. Y. Luo, Y. Yao, Q. Shen, K. Yu and L. Weng, Eur. J. Inorg. Chem., 2003, 318–323.
20. T. J. Woodman, M. Schormann, D. L. Hughes and M. Bochmann, Organometallics, 2004, 23, 2972–2979.
21. S. Wang, X. Tang, A. Vega, J.-Y. Saillard, S. Zhou, G. Yang, W. Yao and Y. Wei, Organometallics, 2007, 26, 1512–1522.
22. L. Mao and Q. Shen, J. Polym. Sci., 1998, 36, 1593–1597.
23. S.-L. Zhou, S.-W. Wang, G.-S. Yang, X.-Y. Liu, E.-H. Sheng, K.-H. Zhang, L. Cheng and Z. X. Huang, Polyhedron, 2003, 22, 1019–1024.
24. K. Zhang, W. Zhang, S. Wang, E. Sheng, G. Yang, M. Xie, S. Zhou, Y. Feng, L. Mao and Z. Huang, Dalton Trans., 2004, 1029–1037.
25. L. Zhou, Y. Yao and Q. Shen, J. Appl. Polym. Sci., 2009, 114, 2403–2409.
26. Y. Nakayama, T. Shibahara, H. Fukumoto, A. Nakamura and K. Mashima, Macromolecules, 1996, 26, 8014–8016.
27. S. Agarwal and A. Greiner, J. Chem. Soc., Perkin Trans. 1, 2002, 2033–2042.
28. Z. Zhu, C. Wang, X. Xiang, C. Pi and X. J. Zhou, Chem. Commun., 2006, 2066–2068.
29. M. D. Bala, J. Huang, H. Zhang, Y. Qian, J. Sun and C. Liang, J. Organomet. Chem., 2002, 647, 105–113.
30. M. T. Gamer, M. Rastätter, P. W. Roesky, A. Steffens and M. Glanz, Chem.–Eur. J., 2005, 11, 3165–3172.
31. M. A. Giardello, Y. Yamamoto, L. Bard and T. J. Marks, J. Am. Chem. Soc., 1995, 117, 3276–3277.
32. E. Ihara, K. Koyama, H. Yasuda, N. Kanehisa and Y. Kai, J. Organomet. Chem., 1999, 574, 40–49.
33. F. Fabri, R. B. Muterle, W. de Oliveira and U. Schuchardt, Polymer, 2006, 47, 4544–4548.
34. H. Yasuda, J. Polym. Sci., Part A: Polym. Chem., 2001, 39, 1955–1959.
35. F. Bonnet, A. C. Hillier, A. Collins, S. R. Dubberley and P. Mountford, Dalton Trans., 2005, 421–423.
36. F. Yuan, Y. Zhu and L. Xiong, J. Organomet. Chem., 2006, 691, 3377–3382.
37. D. Barbier-Baudry, F. Bouyer, A. S. M. Bruno and M. Visseaux, Appl. Organomet. Chem., 2006, 20, 24–31.
38. G. G. Skvortsov, M. V. Yakovenko, G. K. Fukin, A. V. Cherkasov and A. A. Trifonov, Russ. Chem. Bull., 2007, 56, 1742–1748.
39. N. Barros, M. Schappacher, P. Dessuge, L. Maron and S. M. Guillaume, Chem.–Eur. J., 2008, 14, 1881–1890.
40. Z. Jian, W. Zhao, X. Liu, X. Chen, T. Tang and D. Cui, Dalton Trans., 2010, 39, 6871–6876.
41. S. M. Guillaume, M. Schappacher and A. Soum, Macromolecules, 2003, 36, 54–60.
42. I. Palard, A. Soum and S. M. Guillaume, Chem.–Eur. J., 2004, 10, 4054–4062.
43. I. Palard, A. Soum and S. M. Guillaume, Macromolecules, 2005, 38, 6888–6894.
44. I. Palard, M. Schappacher, A. Soum and S. M. Guillaume, Polym. Int., 2006, 557, 1132–1137.
45. S. M. Guillaume, M. Schappacher, N. M. Scott and R. Kempe, J. Polym. Sci., Part A: Polym. Chem., 2007, 45, 3611–3619.
46. M. Schappacher, A. Soum and S. M. Guillaume, Biomacromolecules, 2006, 7, 1373–1379.
47. I. Palard, M. Schappacher, B. Belloncle and S. M. Guillaume, Chem.–Eur. J., 2007, 13, 1511–1521.
48. J. Jenter, P. W. Roesky, N. Ajellal, S. M. Guillaume, N. Susperregui and L. Maron, Chem.–Eur. J., 2010, 16, 4629–4638.
49. M. T. Gamer, S. Dehnen and P. W. Roesky, Organometallics, 2001, 20, 4230–4236.
50. M. T. Gamer, P. W. Roesky, I. Palard, M. Le Hellaye and S. M. Guillaume, Organometallics, 2007, 26, 651–657.
51. M. T. Gamer and P. W. Roesky, J. Organomet. Chem., 2002, 647, 123–127.
52. T. K. Panda, P. Benndorf and P. W. Roesky, Z. Anorg. Allg. Chem., 2005, 631, 81–84.
53. T. K. Panda, M. T. Gamer and P. W. Roesky, Inorg. Chem., 2006, 45, 910–916.
54. T. K. Panda, A. Zulys, M. T. Gamer and P. W. Roesky, Organometallics, 2005, 24, 2197–2202.
55. M. Rastätter, A. Zulys and P. W. Roesky, Chem. Commun., 2006, 874–876.
56. M. Rastätter, A. Zulys and P. W. Roesky, Chem.–Eur. J., 2007, 13, 3606–3616.
57. A. Zulys, T. K. Panda, M. T. Gamer and P. W. Roesky, Chem. Commun., 2004, 2584–2585.
58. J. Jenter, G. Eickerling and P. W. Roesky, J. Organomet. Chem., 2010, 695, 2756–2760.
59. N. Barros, P. Mountford, S. M. Guillaume and L. Maron, Chem.–Eur. J., 2008, 14, 5507–5518.
60. S. M. Cendrowski-Guillaume, M. Nierlich, M. Lance and M. Ephritikhine, Organometallics, 1998, 17, 786–788.
61. M. T. Gamer and P. W. Roesky, Z. Anorg. Allg. Chem., 2001, 627, 877–881.
62. Both MMA (80 equiv.) and TMC (50 equiv.) were added at room temperature to [{CH(PPh₂NSiMe₃)₂}L_n(BH₄)₂] (1 equiv., 0.02 mmol) previously dissolved in THF (1.0 mL). Quenching after 5 min with a toluene solution of acetic acid (ca. 1 mL of a 16.5 mmol L⁻¹ solution in toluene), drying of the crude polymer followed by its dissolution in CH₂Cl₂ and precipitation in cold MeOH resulted in PTMC (M̄_n = 4400 g mol⁻¹; M̄_w/M̄_n = 1.43). Note that these experimental conditions were not optimized for, nor aimed at the synthesis a TMC/MMA copolymer, the objective being here only to verify the purity of MMA.
63. The bis(borohydride) phosphinomethanide rare earth complexes 1–3 successfully initiate the ROP of trimethylene carbonate. The results will be reported in a forthcoming publication.
64. Such (ε-caprolactone)/MMA copolymers could be otherwise sequentially prepared from a combined ROP/atom transfer radical polymerization approach using bromo-functionalized poly(ε-caprolactone) as macroinitiators for MMA polymerization: M. Schappacher, N. Fur and S. M. Guillaume, Macromolecules, 2007, 40, 8887–8896.
65. N. Meyer, J. Jenter, P. W. Roesky, G. Eickerling and W. Scherer, Chem. Commun., 2009, 4693–4695.
66. J. Jenter, N. Meyer, P. W. Roesky, S. K.-H. Thiele, G. Eickerling and W. Scherer, Chem.–Eur. J., 2010, 16, 5472–5480.
67. S. Marks, R. Köppe, T. K. Panda and P. W. Roesky, Chem.–Eur. J., 2010, 16, 7096–7100.
68. In the case of 2, although the calculations on the second inserted monomer unit suggested a syndiotactic preference, the PMMA prepared was found almost atactic. It should be noted that it is difficult to estimate the tacticity from calculations dealing only with the second insertion. Indeed, to suitably assess the syndio vs. atactic preference of the initiator, calculations should be carried out as far as the third and even the fourth MMA insertion. However, our present investigations were rather focused on the activity of the various initiators which remained quite low, prior to considering the polymer tacticity.
69. Ionic radius of yttrium (0.900 Å) and lanthanum (1.032 Å) for a coordination number of six according to R. D. Shannon, Acta Crystallogr., Sect. A: Cryst. Phys., Diffr., Theor. Gen. Crystallogr., 1976, 32, 751–767.
70. N. Susperregui, D. Delcroix, B. Martin-Vaca, D. Bourissou and L. Maron, J. Org. Chem., 2010, 75, 6581–6587.
71. D. Andrae, U. Häussermann, M. Dolg, H. Stoll and H. Preuss, Theor. Chim. Acta, 1990, 77, 123.
72. A. Bergner, M. Dolg, W. Küchle, H. Stoll and H. Preuss, Mol. Phys., 1993, 80, 1431–1441.
73. A. W. Ehlers, M. Böhme, S. Dapprich, A. Gobbi, A. Höllwarth, V. Jonas, K. F. Köhler, R. Stegmann, A. Veldkamp and G. Frenking, Chem. Phys. Lett., 1993, 208, 111–114.
74. P. C. Hariharan and J. A. Pople, Theor. Chim. Acta, 1973, 28, 213–222.
75. A. D. Becke, J. Chem. Phys., 1993, 98, 5648.
76. K. Burke, J. P. Perdew and W. Yang, in Electronic Density Functional Theory: Recent Progress and New Directions, ed. J. F. Dobson, G. Vignale and M. P. Das, Plenum, New York, 1998.
77. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgomery, R. E. Stratman, J. C. Burant, S. Dapprich, J. M. Millam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala, Q. Cui, K. Morokuma, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. Cioslowswi, J. V. Ortiz, A. G. Baboul, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R. Martin, D. J. Fox, T. Keith, M. A. AlLaham, C. Y. Peng, A. Nanayakkara, C. Gonzalez, M. Challacombe, P. M. W. Gill, B. Jonhson, W. Chen, M. W. J. Wong, L. Andres, M. Head-Gordon, E. S. Replogle and J. A. Pople, Gaussian 03, Revision D-02, Gaussian, Inc., Pittsburgh, PA, 2003.
78. A. E. Reed, L. A. Curtiss and F. Weinhold, Chem. Rev., 1988, 88, 899.
79. D. L. Clark, J. C. Gordon, P. J. Hay, R. L. Martin and R. Poli, Organometallics, 2004, 21, 5000.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c1py00133g

---