Chemistry and catalysis advances in organometallic chemistry and catalysis
Download 11.05 Mb. Pdf ko'rish
|
- Bu sahifa navigatsiya:
- 27.3.3 Rare Earth Borohydride/Nonmagnesium Metal Alkyl in Excess Amount
- 27.3.4 Miscellaneous Rare-Earth-Based Catalysts
- Scheme 27.7
- 27.4.1 Statistical Chain Transfer Copolymerization
- TABLE 27.1 Isoprene/Styrene Statistical Copolymerization Results under CCTP (Excess CTA) and Conventional (Ln/Mg 1 : 1) Conditions (Precatalyst
- 27.4.2 Sequenced Copolymerization
- Scheme 27.9
- Figure 27.5
Scheme 27.4 Coordinative chain transfer polymerization (CCTP) of myrcene with neodymium borohydride/dialkyl magnesium catalysts. [Nd](BH
4 ) = Nd(BH 4 ) 3 (THF) 3 , Cp*Nd(BH 4 ) 2 (THF) 2 , MgR 2 = BEM.
[Nd] X MgR 2 [Nd]
PS 1 * MgR 2 [Nd] R + RMgPS 1 RMgPS 2 [Nd]
R RMgPS
2 +
PS 2 MgPS 3 [Nd]
PS 2 * + PS 1 MgPS
3 Scheme 27.5 Mechanism of the transfer polymerization of styrene in the presence of lanthanide borohydride/dialkylmagnesium systems. X = BH
4 group; PS i *
CATALYTIC SYSTEMS AND THEIR APPLICATIONS IN COORDINATIVE CHAIN TRANSFER POLYMERIZATION 351 Ln (BH 4 ) THF (BH 4 ) THF + nBuMgEt n Mg 2 LaCl 3 (THF) 3 Scheme 27.6 Rare earth-catalyzed polystyrene chain growth on magnesium. 85% syndiotactic polystyrene for Cp*Ln(BH 4 )
(THF) 2 , Ln = Nd, La, and atactic polystyrene for LaCl 3 (THF) 3 . Cp*Nd(BH 4 ) 2 (THF) 2 , and Cp*La(BH 4 ) 2 (THF) 2 . The quantitative transfer efficiency observed led us to conclude that there was a CCG on magnesium. Moreover, the reaction remained significantly syndioselective (85%) with the two latter ones, as observed previously when combined with only 1 equiv BEM (Scheme 27.6) [20, 29]. In summary, a CCG process was evidenced for ethylene, isoprene, and styrene, when using Ln/Mg catalytic combinations. In other words, for all these monomers, it was possible to quantitatively produce Mg(polymer) 2 species.
27.3.3 Rare Earth Borohydride/Nonmagnesium Metal Alkyl in Excess Amount While ethylene was easily polymerized in a controlled CCG manner with a lanthanidocene associated to dialkylmagnesium, it was also observed by the group of Boisson that borohydrido ansa-lanthanidocene Me 2 Si(3-Me 3 Si-C
5 H 3 ) 2 Nd(BH 4 )(THF)
2 can be combined with lithium alkylaluminates to yield polyethylene. However, the process is not as controlled as with Mg-cocatalysts regarding molecular weight and polydispersity values that are higher than two [24]. We found that ethylene can be smoothly polymerized by means of catalytic combinations made of a half sandwich of neodymium {Cp*Nd(BH 4 ) 2 (THF)
2 /[TiBA
+ BuLi]} (TiBA = Al(i-Bu) 3 ). From gas chromatography analysis of the resulting oligomers, it was shown that the polymer chain number was related to the alkyl number of the MgR 2 cocatalyst, indicating that a transfer mechanism had to take place. Second, the polydispersities were rather low (circa 1.4), which denotes a high rate of reversible exchange of alkyl chains between neodymium and the cocatalyst [25; unpublished results]. Chain transfer was reported in cis-polymerization of isoprene using Nd(BH 4 ) 3 (THF)
3 /[HNMe
2 Ph][B(C
6 F 5 ) 4 ]/Al(i-Bu) 3 . It was assumed, considering the Mn values of isolated polyisoprene, which strongly decreased in the presence of large excesses of Al cocatalyst, that a nonnegligible part of the Al cocatalyst acts as a transfer agent, with a number of growing chains per metal up to circa 10 in certain conditions. Interestingly, the PDI values remained quite narrow, indicating a rapid exchange between Al and Nd, but to the detriment of the cis-selectivity [30].
CCTP was noticed in a study devoted to polymerization of isoprene with Nd phenate Nd(OAr*) 3 (2,6-di-tert-butyl-OC 6 H 3 ) combined to Al cocatalysts [31]. The transfer efficiency was much lower than previously observed when similar lanthanide phenate compounds were associated to dialkylmagnesium in butadiene polymerization [32]. Keeping in mind the efficiency of aluminum derivatives as cocatalysts and CTA when associated to lanthanide versatate in conjugated diene polymerization, we investigated the behavior of some Nd-carboxylate-based MOF/aluminoxane catalysts versus isoprene polymerization [33]. At elevated temperatures, the value of initiation efficiency (vs neodymium) appeared far higher than 100%, which suggests the probable occurrence of transfer reactions between Nd and Al during the polymerization process. It is worth noting that the cis-selectivity remains close to 80% (Scheme 27.7). [MOF-Nd]-PI + [Al]-R [MOF-Nd]-R + [Al]-PI Scheme 27.7 Transfer reactions between Al CTA and MOF-Nd. 352 COORDINATIVE CHAIN TRANSFER POLYMERIZATION AND COPOLYMERIZATION 27.4 CATALYTIC SYSTEMS AND THEIR APPLICATIONS IN COORDINATIVE CHAIN TRANSFER COPOLYMERIZATION (CCTCOP) Copolymerization presents a one-step, simple, and convenient way of polyolefin functionalization [34]. It implies the use/development of highly versatile catalytic systems able to (co)polymerize a large variety of monomers. In this context, we studied the effect of polymer chain transfer in a coordination copolymerization process. 27.4.1 Statistical Chain Transfer Copolymerization Compared to classical statistical copolymerizations, the occurrence of transfer could have an effect in terms of activity, selectivity, and comonomer contents (modification of reactivity ratios). Transmetalation between the lanthanide and the magnesium was shown to take place efficiently when monosubstituted bis-borohydrides Cp*Ln(BH 4 ) 2 (THF)
n or trisborohydrides Ln(BH 4 )
(THF) 3 (Ln = La, Nd) were combined with magnesium dialkyl in the course of a statistical isoprene/styrene copolymerization: regular decrease of molecular weights (and quite narrow polydispersities) while increasing the CTA amount was noticed. Moreover, and very interestingly, it was found that the transmetalation is accompanied by (i) an increase in the quantity of styrene inserted in the copolymer and (ii) a gradual decrease of the 1,4-trans stereoselectivity of the reaction at the benefit of 3,4-selectivity. Indeed, for the same monomer feed (50/50), the amount of styrene inserted in the copolymer (up to 32 mol%, vs 12 mol%) can be increased by a factor of three under CCTP conditions (using 10 equiv dialkylmagnesium) versus conventional conditions (1 equiv Mg-CTA), Such a result in terms of comonomer content could already be obtained with Ln/Mg (1 : 1) catalysts but a 80/20 styrene/isoprene feed ratio was necessary [27, 35]. In addition, the content of 3,4-isoprene units increases from 2% to 16%, at the expense of the trans-selectivity (Table 27.1). Such results can be interpreted considering the formation of polymetallic Ln/Mg active species (see further), which modify the reactivity ratio of each comonomer, to the benefit of the less active one, styrene. This was also presumably partially attributed to a magnesium-induced co-oligomerization of isoprene and styrene, with a higher reactivity of styrene as compared to neodymium. Finally, we were able to face a challenge: enhance the comonomer content, while maintaining the selectivity, by combining dialkyl magnesium and trialkyl aluminum as cocatalysts/CTA, associated to precatalyst Cp*La(BH
4 ) 2 (THF) 2 . Under such conditions, a 1,4-trans (up to 98%) stereospecific reversible coordinative chain transfer copolymerization (CCTcoP) of isoprene and styrene is observed. The styrene incorporation rate is this time less impressive, around 19%. Although we did not succeed in the homopolymerization of α-olefins with our borohydrido lanthanide-based catalysts [36], the Cp*La(BH 4 ) 2 (THF)
2 /n-butylethylmagnesium catalytic system was nevertheless assessed for the CCTcoP of styrene and 1-hexene. Poly(styrene-co-hexene) statistical copolymers were obtained with up to 46% yield, and 23% 1-hexene content (Scheme 27.8). The occurrence of chain transfer reactions in the presence of excess BEM was established in the course of the statistical copolymerization, through significant molecular weights decrease versus 1 equiv BEM, along with narrowing of the polydispersities. It is noteworthy that these transfer reactions do not modify the activity of the catalyst, by contrast with what was observed for isoprene/styrene copolymerizations. Thanks to this transfer process (and as observed in the course of isoprene/styrene copolymerization), the quantity of 1-hexene in the copolymer is increased substantially, from 8.6% to 23.2%, for 80/20 1-hexene/styrene composition in the feed and in the presence of 10 equiv BEM versus 1 equiv. From a run conducted with BEM alone, which gave, although in low yield—15% conversion in 120 h, copolymer containing TABLE 27.1 Isoprene/Styrene Statistical Copolymerization Results under CCTP (Excess CTA) and Conventional (Ln/Mg 1 : 1) Conditions (Precatalyst = Cp*La(BH 4 ) 2 (THF) 2 , CTA = BEM) S/I (feed) Mg/Ln
Time, h Yield, %
S, % 1,4-trans% 3,4-% 50/50
1 2 48 12 98 2 50/50 10 40 58 32 84 16 80/20
1 a 2 69 b 32 96 4 80/20 10 40 35 49 90 10 Monomer/catalyst = 2000 (S = styrene, I = isoprene), 50 ◦ C, Ln
= La except line 3. a Ln = Nd. b Based on isoprene consumed. CATALYTIC SYSTEMS AND THEIR APPLICATIONS IN COORDINATIVE CHAIN TRANSFER COPOLYMERIZATION (CCTCOP) 353 + nBuMgEt La (BH
4 ) THF (BH 4 ) THF m [Mg]
p +
Scheme 27.8 Half-lanthanocene/BEM-mediated styrene–hexene coordinative chain transfer copolymerization. 37% 1-hexene, it was postulated that a kind of chain shuttling could operate between neodymium and magnesium in the course of the copolymerization [37]. These results extend the range of our concept of a chain transfer-induced control of the composition of statistical copolymers to poly(styrene-co-hexene) copolymers. Chain transfer in the course of a metal-catalyzed statistical copolymerization may thus be viewed as a new and innovative way for the control of the composition of a copolymer. Statistical copolymerization of ethylene and isoprene was achieved by using a borohydrido Cp*(BH 4 )
Nd(THF) 2 ] (Cp* = C 5 Me 5 ) half-lanthanidocene under polymer chain transfer conditions, and with lithium alkyl aluminum as cocatalyst (Al/Nd = 5). Polyisoprene-co-ethylene was received, with ethylene amount incorporated of circa 25 mol%, and the stereospecificity of isoprene enchainments was found to be around 96% trans1,4- [38]. It is noteworthy that, in turn, isoprene/hexene copolymerization did not succeed with the same precatalyst. This was attributed, after theoretical calculations, to a difference of reactivity by comparison with isoprene/ethylene mixtures. It is noteworthy that the absence of chain transfer conditions cannot be advanced to explain this result as we checked that isoprene/hexene copolymerization failed as well with a borohydrido lanthanidocene and excess BEM. 27.4.2 Sequenced Copolymerization The living character of a CCTP process can be advantageously applied for further functionalization [39]. Considering that some of our catalytic systems could be complementary, we targeted to combine them with the aim to prepare novel macromolecular architectures. Practically, as magnesium dialkyls were found adequate to initiate the growth of polymer chains, the idea was to use a bis(polyethylenyl)magnesium [Mg(PE) 2 has been recently used to prepare functionalized polyethylene, similarly to Al(PE) 3 ] compound as cocatalyst in the trans-polymerization of isoprene [16]. For this purpose, as illustrated in Scheme 27.9, Mg(PE) 2 was first prepared by using the already mentioned Cp* 2 Nd(BH 4 )(THF)/BEM-based CCTP of ethylene (or the same with the chloro analog Cp* 2 NdCl
2 Li(OEt
2 ) 2 , first step, Nd-CAT
1 ). Secondly, and without any intermediate polymer isolation, a stoichiometric amount of Nd(BH 4 )
(THF) 3 (Nd- CAT 2 ) was added to the in situ formed Mg(PE) 2 and this mixture was immediately submitted to isoprene polymerization. As expected, the result of this dual process was a sequenced (PE)-b-(trans-PI), as established by careful analyses for the determination of the junction motives by gas chromatography, MALDI-TOF, and 13 C NMR analysis of oligomers. Differential scanning analysis of the sequenced copolymer formed confirmed the loss of crystallinity of the polyethylene block, in line with the soft-matter character observed (Fig. 27.5). Nd-CAT 2
Linear PE trans-PI Mg n n + Mg-CTA
Nd-CAT 1
Preparation of PE-TPI block copolymer by a controlled dual-catalytic process.
354 COORDINATIVE CHAIN TRANSFER POLYMERIZATION AND COPOLYMERIZATION Figure 27.5 Picture of the PE-TPI block copolymer, showing its soft gummy character. 27.5 DISCUSSION 27.5.1 Precatalyst/Cocatalyst Adequacy Reversible and fast chain transfer is a prerequisite in order to control a polymerization process. In this regard, relative amount and synergy between each catalytic partner, that is, precatalyst and cocatalyst/CTA, are of primary importance. This was the case with ethylene, which has been by far the most studied monomer in this frame, as shown by Gibson [9] with Fe/Zn catalysts, after the pioneering studies of Mortreux [8] with Ln/Mg and more recently Kempe with Y/Al systems [40]. Regarding the rare earths, and particularly in the “early” series (the biggest ones, from La to Sm), magnesium dialkyl seems to be the best partner, having at the same time a high transfer efficiency allowing to involve all magnesium in the process, along with very fast ability to transfer a polymer chain, finally resulting in narrow polydispersities, and hence the highest degree in mastering molecular weights. This was established for a variety of monomers: ethylene, isoprene, myrcene, styrene, showing in particular the high versatility of borohydrido lanthanide/alkyl magnesium catalytic combinations. Aluminum CTAs are less prone to transfer reactions than magnesium CTAs when combined to rare earth borohydrides (this is not the case for yttrium alkyls as demonstrated by Kempe), with lower efficiencies and broader polymolecularities. However, aluminum alkyls are mandatory for selective polymerizations that require cationic-like active species, that is, those producing syndiospecific polystyrene and 1,4-cis stereoregular polyisoprene. A judicious combination of several CTAs/cocatalysts may eventually be a solution to fully control a process, like in the
dialkyl to prepare poly-trans-isoprene under CCTP conditions, but the trans-rate is limited to circa 90% [41]. On the other hand, the molecular structure of the precatalyst may eventually require an adjustment in the cocatalyst/CTA nature: thus, whereas Cp* 2 Nd(BH
4 )(THF) is fully adapted to dialkyl magnesium for polyethylene CCG, lithium aluminum alkyl seems better suited to half-lanthanidocene Cp*Nd(BH 4 ) 2 (THF)
2 (see earlier in the text). Regarding the effect of the molecular structure, it was claimed in the case of polyethylene, that CCG should be more favorable for more sterically hindered catalysts [9]. On the basis of steric considerations, we tentatively drew an analogy between lanthanidocene- catalyzed polyethylene chain growth and half-lanthanidocene-catalyzed polystyrene chain growth on magnesium, considering an interaction with the aromatic ring and keeping in mind the secondary insertion of styrene that we observed (Scheme 27.10) [29]. Ln
n [Ln]
R' Mg-X
R' Mg-X
Scheme 27.10 Analogy between lanthanidocene-mediated polyethylene catalyzed chain growth (CCG) and half-lanthanidocene-mediated polystyrene CCG.
DISCUSSION 355 27.5.2 Active Species in CCTP One intriguing feature in CCTP processes is the nature of the active species, as numerous metals—precatalyst, cocatalyst, CTA—are involved. Proposals have been made by several authors to clarify the nature of the active species in the frame of a CCTP, that is, in excess of the second metal [9, 40, 42]. The in situ formation of heterobimetallic complexes, [M]( μ-R) 2
and [M]( μ-R)
2 [Al] (M precatalyst), depending on the cocatalyst, was advanced, whereas in the presence of a third metal, an equilibrium could take place between bridged bimetallics such as [Fe]( μ-R)
2 [Zn] and [Fe]( μ-R) 2
propagating species is supposed to be a monometallic [M]-R alkyl derivative. However, it has just been proposed that the propagating active species could also be a singly bridged one, on the basis of calculations [43]. From our results in isoprene polymerization with Ln/Mg catalysts (1,4-trans to 3,4- switch) and in copolymerization (increase of comonomer content), it seems obvious that the active species should be different in dual-component catalytic systems, depending on whether the cocatalyst/CTA is in stoichiometric quantity or in excess. Actually, it is accepted that 3,4-polymerization is connected to steric hindrance around the metal catalyst, which will impose a single coordination of the monomer [44]. The 1,4-trans selectivity is supposed to result either from single - η 2 or s-trans- η 4 coordination that affords the trans-product, the most thermodynamically stable, whereas the cis-product is the kinetic one [45]. Alternatively such selectivity may be connected to limited possibilities of coordination due to the presence of an additional compound in the coordination sphere of the catalyst metal [46]. An s-trans- η 4
active [ {Me
2 Si(C
5 H 4 )(C 13 H 8 ) }NdR] species during ethylene–butadiene copolymerization has been proposed to account for the alternate copolymer isolated [47]. Regarding the role of the cocatalyst, magnesium cocatalyst has been known for decades to favor trans-polymerization [45]. In our opinion, the very high level ( >98%) of trans-selectivity encountered with Ln(BH
4 )/Mg catalysts was connected to the existence of heterobimetallic structures (borohydrido bridged Ln/Mg bimetallics were isolated [21]), which association was strengthened by the presence of the borohydride bridge [18a]. As represented in Scheme 27.11, monometallic species A and bimetallics B1, B2 would be prone to propagate the polymerization in a 1,4-trans manner, whereas B3 would be a dormant form. C1, C2, and C3 trimetallic species result from the reaction with an additional BEM molecule; among them, C2 and C3 may preferentially lead to 3,4-polyisoprene, while other possible dormant forms are not represented. The more BEM is present in the mixture, the more 3,4-polymer is formed by displacement of equilibria toward polymetallic species. Note that the growth of several polymer chains per catalyst metal is not considered here. The same reasoning can be considered to account for higher comonomer incorporation in statistical copolymerizations: the higher the steric hindrance due to polymetallic associations, the more the difference of reactivity ratio will decrease in favor of a more bulky comonomer, as the latter is able to compete more efficiently with an η 2
than with an η 4 -coordinated one [35]. Another possibility to explain both the 3,4- tendency in selectivity (isoprene homopolymerization) and the increase of comonomer incorporation (in copolymerizations) is to consider the propensity of MgR 2 to propagate a polymer chain by itself. Indeed, we were able to reproduce Yasuda’s experiments, showing that magnesium-isoprene “adducts” (actually Mg- isoprenyl compounds) do polymerize isoprene over long reaction times [48], under our experimental conditions, to afford low yields of polyisoprene with 86% 3,4-selectivity [28]. Therefore, we believe that, under isoprene CCTP conditions, the transferred macromolecular chain can grow mainly in a 1,4-trans fashion on the lanthanide atom, and it may also be able to grow in a 3,4-fashion on the magnesium atom, leading to the observed modification of the regioselectivity and microstructure. It can thus be considered that at least a part of the 3,4-regioselectivity can be due to the growth of the polyisoprenyl chain on the magnesium atom. A consequence of this process is that isoprenyl magnesium moieties may no longer be considered as dormant species, and the overall polymerization process may result from a highly efficient transmetalation of a growing macromolecular chain between two catalysts exhibiting different catalytic behavior (Scheme 27.12). The fact that BEM too shows a propensity to incorporate rather high amounts in styrene (46%) under isoprene–styrene CCTcoP conditions also supports this hypothesis of a participation of Mg to the polymer growth. When more than two different metals are associated in a catalytic combination, the nature of the active species will depend on the role played by each metal. In other words, Gibson postulated equilibrium between Fe/Al (alkylation and propagation) and Fe/Zn (transfer) bimetallic species, aluminum and zinc having their own task to accomplish. It is also the case in the chain-shuttling processes, where two metal catalysts interact each on their own side with the same metal-CTA. In the particular situation of our three components Ln/Mg/Al trans-stereospecific isoprene polymerization catalyst, trimetallic active species may be involved in the process. From our experience with all these catalytic systems, it is clear that the selectivity is rather connected to the presence of the magnesium, while the transfer of the growing polymer chains would be a role devoted to aluminum CTA. One can postulate that the “strong” [Ln]( μ-BH
4 )[Mg] association is not perturbed by the polymer chain transfer between Ln and Al. The driving force of such interactions between metallic species present in the mixture could
|
ma'muriyatiga murojaat qiling