Polyisoprene (PIP) with well-defined microstructure represents an essential class of functional material in elastomers and plastics [1-3]. The huge demand of them thereby promotes an impressive development of rare-earth metal catalysts [4-6], which achieved the selective synthesis of cis-1, 4-PIP, [7-26] trans-1, 4-PIP [27-33] and 3, 4-PIP [34-39]. It is noteworthy that some catalytic systems were reported to enable a switch of the selectivity by a slight modification of the metal center, ancillary ligand, or cocatalysts [40-54]. For example, Anwander found that the lanthanum fluorenyl half-sandwich catalyst efficiently produced trans-1, 4-PIP (80%), while the lutetium complex gave access to higher cis-1, 4-contents (78.2%) [40]. Arnold discovered that half-sandwich scandium borohydrides complexes with different NHC ligands could swtich the stereoselectivity in the isoprene polymerization (94% trans-1.4 and 80% cis-1.4) [41]. Hou and Zhang reported that the yttrium amidinate complex, by addition of AlMe₃, were transformed to heterotrinuclear Y/Al complex, which could dramatically switch the regio- and stereoselectivity of isoprene polymerization (from 3, 4-isospecific to 1, 4-cis selective) [46, 47]. In these cases, the precise control on the regio- and stereoselectivity definitely displays the state of the art in catalyst design. A tailor-made catalyst with new structure, good stability, high catalytic activity and selectivity, however, is still highly desired in both academic and industrial fields.

Dialkyl rare-earth complexes bearing a monoanionic ancillary ligand have recently drawn much attention, serving as efficient catalysts for the polymerization and functionalization reactions of alkenes [55-60]. Most of these ligands feature delocalized carbanions or multidentate nitrogen anions structures (Scheme 1A). According to the strong oxophilicity of rare-earth metals, oxyanions are supposed to have great potential for acting as suited ancillary ligands [61-75]. Based on this assumption, Evans synthesized dialkyl rare-earth complexes adopting steric bulky 2, 6-di‑tert-butylphenol as an ancillary ligand [61, 62]. Subsequently, functionalized phenols [63-67], silanols [68, 69], alkoxy NHCs [70, 71] were also introduced into the ancillary ligand family (Scheme 1B). Anwander reported the synthesis of various alkylated carboxylate rare-earth complexes that catalyzed cis-1, 4-polymerization of isoprene [72]. These oxyanion ligands moved down a new path for the synthesis of dialkyl rare-earth complexes, whereas most of these complexes were not stable enough, presumably because of the easy ligand redistribution and complex aggregation reactions. Nevertheless, further studies about dialkyl rare-earth complexes bearing oxyanion ligands remain sluggish due to a lack of ligand diversity. Although some lanthanide carboxylates [76-79] and phosphates [80-85] were used as catalyst procursors in polymerization process, the well-defined alkyl complexes are yet undeveloped and the role of ligands is still obscure. Thus, we are quite curious about that if phosphates could be applied as anion ligands for dialkyl rare-earth complexes; a positive answer to this question could provide further insight about oxyanion ancillary ligands and also an attractive alternative catalyst. Herein, we reported the synthesis of well-defined phosphate yttrium dialkyl complexes and their catalytic application in stereo-controllable 1, 4-polymerization of isoprene (Scheme 1C).

The oxophilicity of rare-earth metals would definitely provide a strong interaction between the metal ion and the phosphate anion. This strong bonding, however, together with the small steric hindrance of the oxyanions, would probably result in the bridging interaction and ligand redistribution. To avoid the possible side reactions, we proposed to start with tris(aminobenzyl) yttrium complex, whose alkyl anion has a "built-in" chelating amino group. Thinking of the widely used N-heterocyclic carbene ligands, we synthesized two N-heterocyclic phosphorodiamidic acids (NHPAs) bearing similar skeleton structures and tunable steric parameters [86, 87]. The acid-base reaction between NHPAs and Y(CH₂C₆H₄NMe₂-o)₃ took place smoothly at 60 ℃ in toluene and afforded the N-heterocyclic phosphate yttrium bis(aminobenzyl) complexes PYR₂ after 3 h (Scheme 2). Both complexes were fully characterized by ¹H, ¹³C and ³¹P NMR spectroscopy, and their structures in solid state were determined by single-crystal X-ray diffraction (XRD). These characterizations reveal the similar structures of the two phosphate yttrium complexes, as shown in Fig. 1. The phosphate anion is bonded to the yttrium center in κ² fashion through two oxygen atoms. The yttrium atom is well located on the OPO rigid frame to form a planar quadrilateral. In a simplified view, the phosphate ligand creates a well-defined steric pocket, along with the κ² coordination from the aminobenzyl ligands, conducing to enhance the stability of PYR₂. Through comparison with the two phosphate yttrium complexes, it is showed that the Y-O bond lengths in P²YR₂ (2.306(2) Å) are longer than those in P¹YR₂ (2.282(2) Å). We speculated that the larger steric hindrance from the isopropyl groups increases the distance between phosphate and yttrium atom.

As part of our continuing investigations into isoprene polymerization, we then used the PYR₂ as catalysts for further exploration. The neutral P¹YR₂ alone proved to be inactive for the isoprene polymerization (Table 1, entry 1). However, once activited with 1 equiv. of borate reagent [Ph₃C][B(C₆F₅)₄], P¹YR₂ at room temperature demonstrated high catalytic activity: 1000 equiv. of isoprene were converted quantitatively into PIP within 20 min (entry 2). Remarkably, the polymer product features a high cis-1, 4 content (93.7%). GPC curve indicated that the PIP obtained is unimodal with a M_n of 2.6 × 10⁵ and a narrow molecular weight distribution (PDI = 1.4). The catalytic polymerization under lower temperature took place smoothly and gave PIP with even higher cis-1, 4 content (96.5%) as well as higher molecular weight (M_n = 3.6 × 10⁵) (entry 3). We next carried out the catalytic polymerization with various [Isoprene]/[P¹YR₂] ratios varying from 600 to 5000 (entries 4–9). All the reactions produced PIP with high efficiency (yields > 98%) and excellent selectivity (cis-1, 4 selectivity: 92.6%−96.5%). Furthermore, the molecular weight of the resultant PIP increases linearly with the amount of the isoprene monomers (see Supporting information for details), while the molecular weight distribution still keeps narrow (1.4–1.6). The phosphate yttrium complex here displayed a very high catalytic activity and thus provided an approach for the synthesis of high-molecular-weight cis-1, 4-PIP (M_n up to 1.06 × 10⁶). Considering the important roles of alkylaluminum compounds in polymerization, we further carried out the catalytic polymerization reaction in the presence of trialkylaluminum reagents (entries 10 and 11). Al(iBu)₃, a well-known chain transfer reagent, led to the decrease of the molecular-weight (M_n = 1.9 × 10⁵), as expected. Unlike some known findings [46, 48] that addition of AlMe₃ could increase the cis-1, 4 content of PIP, the addition of 5 equiv. of AlMe₃ herein greatly increased the trans-1, 4 content (44.1%).

The catalytic system P²YR₂/[Ph₃C][B(C₆F₅)]₄ exhibited lower cis-1, 4 selectivity in the polymerization of isoprene, but higher initiation efficiency (42.1%), compared to the catalytic system P¹YR₂/[Ph₃C][B(C₆F₅)]₄ (25.9%), producing the PIP product with 83.3% cis-1, 4, 5.3% trans-1, 4 and 11.4% 3, 4 contents (Table 2, entry 1). This result suggested that the steric effect of the two ortho-substituted N-aryl rings might be a contributing factor of the cis-1, 4 selectivity. The bulky o-isopropyl in N-aryl rings led to the decrease of cis-1, 4 selectivity [11, 16]. The catalytic system containing 5 equiv. of AlMe₃ surprisingly switched the stereoselectivity from cis-1, 4 to trans-1, 4-polymerization, thus affording the polymers with contents of 89.4% trans-1, 4 units (entry 6). The trans-1, 4-polymerization took place much more slowly (90 min). The significantly increased molecular weight (M_n = 3.6 × 10⁵) revealed an obvious decreased efficiency of the catalyst (18.7%), which hinted a dramatic structure change of the catalytic species. We speculated that the heteronuclear Y/Al complex formed in-situ might be the possible cataytic species in the trans-1, 4-polymerization [9, 28, 46]. To find out more details, we carefully examined the polymerization with different [AlMe₃]/[P²YR₂] ratios. 1 or 2 equiv. of AlMe₃ did not obviously change the regioselectivity (entries 2 and 3). Using 3 equiv. of AlMe₃ as additive resulted in a noticeable increase of the trans-1, 4 selectivity from 5.3% to 75.0% (entry 4). When 4 equiv. of AlMe₃ or more was added, the PIP was obtained with higher trans-1, 4 content (85.1%−92%) (entries 5 and 7). It is suggested that 4 equiv. of AlMe₃ was necessary for the formation of the heterotrinuclear Y/Al complex [46]. It is noteworthy that 10 equiv. of AlMe₃ as additive gave the PIP with lower molecular weight (M_n = 1.3 × 10⁵) but similar narrow molecular weight distribution (PDI = 1.5), suggesting that the excess amount of AlMe₃ may act as a chain transfer reagent to interrupt the growing polymer chain and tune the molecular weight. For comparison, the PIP obtained with the system P²YR₂/[Ph₃C][B(C₆F₅)]₄/AlEt₃ or Al(iBu)₃ failed to achieve high trans-1, 4 contents (entries 8 and 9). This observation showed that the trans-1, 4 selectivity might be influenced by the steric effect of the aluminum alkyls (AlMe₃ > AlEt₃ > Al(iBu)₃) [21]. Control experiments were conducted that the isoprene polymerization could not be initiated without either borate reagent or P²YR₂, indicating that the cationic Y/Al species may account for the trans-1, 4-polymerization of isoprene (entries 10 and 11). To gain further insight into the possible heterotrinuclear Y/Al species, we carried out the reaction between phosphate yttrium dialkyl complexes and 4 equiv. of AlMe₃. A crystal of a phosphate bis(tetramethylaluminate) yttrium complex [P¹Y(AlMe₄)₂]₂ suit for X-ray diffraction analysis was surprisingly obtained (Fig. 2). Compared to the crystal structure of P¹YR₂, this dimer complex [P¹Y(AlMe₄)₂]₂ features a quite different coordination environment around the yttrium metal center. The huge change of the catalyst structure might result in the different coordination and insertion mode of isoprene, leading to trans-1, 4-polymerization of isoprene (Table 1, entry 11). The synthesis and isolation of heteronuclear Y/Al complex [P²Y(AlMe₄)₂]₂ and catinoic species of [P¹Y(AlMe₄)₂]₂ were tried but failed, possibly because of their thermal instability.

In conclusion, N-heterocyclic phosphate ancillary ligand was designed and proved to be suitable for the synthesis and isolation of yttrium dialkyl complexes, which could be conveniently prepared via the acid-base reactions between tris(aminobenzyl) yttrium complex and N-heterocyclic phosphoric acid. Activated by [Ph₃C][B(C₆F₅)]₄, the yttrium complexes PYR₂ exhibit high catalytic activities for cis-1, 4-selective (up to 96.5%) polymerization of isoprene. The stereoselectivity was dramatically switched from cis-1, 4 to trans-1, 4-polymerization (up to 92.0%) when using AlMe₃ as an additive. Considering various easily available phosphoric acids including the chiral ones, we could expect a nice synthetic value and diversified catalytic applications of phosphate ligated metal complexes in near future. Along this line, the synthesis of N-heterocyclic phosphate complexes with other metals and also with chiral phosphoric acids are currently ongoing in our laboratory.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

We gratefully acknowledge the support from the Fudan University. We thank Prof. Li Pan (Tianjin Univ.), Prof. Jian-Hua Cheng (CIAC, China) for their helpful suggestions. This project was supported by National Natural Science Foundation of China (Nos. 21890722 and 21950410519), National Natural Science Foundation of Tianjin (No. 19JCYBJC20100).

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2021.11.064