Ring-Opening Polymerization of Cyclic Ester Catalyzed by Metal Complexes Containing Phosphorus and Nitrogen Bonds

Xing Wang; Xiaopan Xue; Youshu Jiang; Wenjuan Zhang; Yanping Ma; Wen-Hua Sun

doi:10.7536/PC240307

Progress in Chemistry ›› 2024, Vol. 36 ›› Issue (10) : 1425-1442 DOI: 10.7536/PC240307

research-article

Scan to read & Share

Ring-Opening Polymerization of Cyclic Ester Catalyzed by Metal Complexes Containing Phosphorus and Nitrogen Bonds

Author information +

History +

PDF

Abstract

Aliphatic polyesters are important biodegradable polymers and bulk materials, and their monomers of cyclic esters can come from biomass. The catalysts are the key for converting cyclic esters into polyesters, in which the new homogenous catalysts not only initiate the ring-opening polymerization (ROP) of cyclic esters with high activities but also finely tune the molecular weights and structures of the obtained polyesters and thus improve the polymer properties. Currently metal complexes ligated by Schiff-bases have attracted much attention in the ROP of cyclic esters, in contrast,metal complexes containing the phosphorus-nitrogen bond are very limited but display unique catalytic properties and show the potential for industry. The nitrogen and phosphorus possesssimilar electronic structure but with different electronegativity, which helps to adjust the coordination ability of ligand to metal and thus improve thecatalytic performance in the ROP of cyclic esters. The phosphorus-nitrogen bond can be a single bond or a double bond, and the corresponding metal complexes are commonly formed with nitrogen coordination than phosphine coordination, in which by regulating the steric and electronic of phosphine groups the environment around the adjacent coordinatingnitrogen could be adjusted with resulting in good activity but also in controllability to tailor the microstructure of the resultant polyesters. This review collects the recent progress of such multidentate metal complex catalysts containing phosphine nitrogen bond for ring-opening polymerization of cyclic esters, and summarizes the relationship between ligand structure, catalytic performance, and the microstructure of the resulting polyester, which favor the exploring the new efficient complex catalysts and guiding the industry to select the practical catalysts to promote the scientific development as well as industrialization of related technologies.

Contents

1 Introduction

2 Construction methodology of phosphorus nitrogen double bonds (P=N) and single bonds (P—N) in organic compounds

3 Metal complexes containing P-N single bonds for ring-opening polymerization of lactones

3.1 Alkali(alkaline) metal complexes containing P—N single bonds

3.2 Rare earth metal complexes containing P—N single bonds

4 Metal complexes containing P=N double bonds for ring-opening polymerization of lactones

4.1 N^N bidentate metal complexes containing P=N double bonds for ROP

4.2 N^N^X Tridentate metal complexes containing P=N double bonds for ROP

4.3 Ion pair metal complexes with P=N bond for ROP

4.4 N^N^O^O tetradentate metal complexes containing P=N double bonds for ROP

5 Conclusion and outlook

Graphical abstract

Keywords

phosphorus nitrogen bond / organometallics / homogeneous catalysis / ring-opening polymerization (ROP) / biodegradable polyester

Cite this article

Download citation ▾

Xing Wang, Xiaopan Xue, Youshu Jiang, Wenjuan Zhang, Yanping Ma, Wen-Hua Sun. Ring-Opening Polymerization of Cyclic Ester Catalyzed by Metal Complexes Containing Phosphorus and Nitrogen Bonds. Progress in Chemistry, 2024, 36(10): 1425-1442 DOI:10.7536/PC240307

登录浏览全文

4963

注册一个新账户忘记密码

1 Introduction

The contradiction between the drastically reduced primary forests and the continuously increasing population, under the global information interconnection background, the universal pursuit of living and transportation convenience has led to more energy consumption, urgently requiring the synthesis of polymer materials that can degrade after use to reduce survival pollution. Apart from natural polymer materials, the important types of biodegradable materials are mainly aliphatic polyesters, including synthetic aliphatic polyesters, biopolymers such as polyhydroxyalkanoates (PHA), etc.; among them, polylactic acid and polycaprolactone, as typical aliphatic polyesters, have monomers that can be derived from biomass conversion, with vast resources, and are considered to have the greatest development potential and market value^[1]. The preparation methods for aliphatic polyesters generally include two: direct condensation of bifunctional compounds and ring-opening polymerization of cyclic esters, where the condensation reaction of bifunctional compounds involves a chemical equilibrium process, and the polymers prepared by this method have relatively low molecular weight and a wide molecular weight distribution, which will affect the quality and performance of the material. The synthesis of poly(butylene succinate) (PBS) is a typical example of using this method^[2,3]. In the case of the ring-opening polymerization method, the cyclic ester monomers can be produced through biological fermentation, and the ring-opening polymerization process is easy to control, enabling the control of the molecular weight of the resulting polyester, thus becoming the focus of current research on preparing aliphatic polyesters. Typically, the metal elements in the metal catalysts used for ring-opening polymerization of cyclic esters include alkali (earth) metals^[4], aluminum^[5], indium^[6], rare earth metals^[7], etc. These metal compounds, as initiators, have high catalytic efficiency and good selectivity in terms of monomer and polymer structure^[8,9]. At present, the monomers suitable for ring-opening polymerization, which are extensively studied in the literature (as shown in Figure 1), mainly include lactones and lactides. The metal compound-catalyzed ring-opening polymerization of these cyclic esters mostly follows a "coordination-insertion" mechanism: first, the carbonyl oxygen atom of the cyclic ester monomer coordinates with the metal center of the catalyst, further activating the monomer; then, the metal-alkoxide (M-OR) in the complex attacks the carbon atom on the carbonyl to form a four-membered ring intermediate; the monomer undergoes a ring-opening process, forming a new metal-alkoxide bond, entering the next cycle of "coordination and insertion" for another cyclic ester^[10⇓~12]. Therefore, the stereostructure of the ligands in the metal complexes and the coordination ability of the coordinating atoms significantly influence the catalytic activity, polymerization controllability, and stereoselectivity of the ring-opening polymerization.

Literature reports that in metal complexes used for the ring-opening polymerization of cyclic esters, their ligands mostly contain carbon-nitrogen single bonds C—N, carbon-nitrogen double bonds C=N, and oxyanions O^- (phenoxide or alkoxy groups); among them, the tetradentate salen-Al compounds based on salicylaldimine structures have received extensive attention^[13,14], showing good catalytic activity for both lactide and ε-caprolactone^{[15⇓⇓⇓⇓⇓⇓⇓⇓~24]}. Particularly noteworthy is that chiral salen-Al exhibits good isotactic selectivity (with P_m up to 0.9) in the ring-opening polymerization of racemic lactide, resulting in stereoblock copoly(lactic acid) with a melting point much higher than that of isotactic poly(L-lactic acid) (PLLA). In the ring-opening polymerization of racemic lactide, due to the two different chiral monomers, the chemical structure of the obtained polymer has stereoselectivity, with the degree of stereo regularity represented by the probability of forming racemic or meso chains (the probability of forming new racemic (syndiotactic) or meso (isotactic) dyads) P_r and P_m^[25]. For the ring-opening polymerization of racemic lactide, P_r =1.00 (P_m =0.00) and P_m =1.00 (P_r =0.00) represent perfectly atactic and isotactic polymers, respectively, and generally, the higher the P_m value, the better the isotactic selectivity. Some literature also uses P_i to denote isotactic selectivity (i: isotactic), which is the same as P_m; similarly, syndiotactic selectivity P_s = P_r. Salen-Al compounds show low catalytic efficiency in the ring-opening polymerization of racemic lactide, typically requiring several days to achieve complete conversion of 100 times the amount of monomer^[18]. In recent years, Auffrant et al.^[26] found through DFT calculations that introducing phosphorus into the salen model (containing a phosphinimine structure, referred to as Phosphasalen) after coordination with metals can efficiently catalyze the ring-opening polymerization of racemic lactide while maintaining good stereoselectivity, driving research on using P=N ligand-containing metal complexes for the preparation and catalysis of cyclic ester ring-opening polymerization^[27]. Studies have shown that, unlike the C=N bond in salen ligands, the P=N bond in Phosphasalen ligands, due to the difference in electronegativity between phosphorus and nitrogen atoms, exhibits high ionic character. Additionally, its nitrogen atom, having two lone pairs of electrons, gives the P=N bond strong σ and π electron-donating capabilities, thereby reducing the Lewis acidity of the coordinated metal center, making the metal-alkoxy bond looser and more active, facilitating the insertion of cyclic ester monomers, and thus enhancing the activity of ring-opening polymerization^{[28⇓⇓⇓~32]}; moreover, in terms of coordination geometry, the phosphorus atom in phosphinimine (P=N), with two substituents, presents a pyramidal configuration, which is different from the planar triangular structure of sp²-hybridized C=N, and they have significantly different bond lengths (P=N ~1.58 Å vs C=N ~1.28 Å). The highly polarized P=N double bond in phosphinimine mainly coordinates with transition metals through approximately sp²-hybridized nitrogen atoms to form stable complexes, exhibiting unique catalytic properties^[33,34].

Organic compounds containing P=N double bonds have long been reported, but the study of metal complexes containing P=N double bonds and their catalytic ring-opening polymerization of cyclic esters is relatively recent. Research on metal compounds containing P—N single bonds and their catalytic properties is even less. However, metal compounds containing phosphorus-nitrogen bonds exhibit unique coordination structures, which indeed enhance their catalytic activity in the ring-opening polymerization of cyclic esters and show good stereoselectivity. Based on this, this paper reviews the construction of ligands and metal complexes based on phosphorus-nitrogen bonds, as well as their progress in the catalytic ring-opening polymerization of cyclic esters.

2 Methods for Constructing P=N and P-N Bonds in Organic Compounds

P—N or P=N bonds are widely present in various bioactive molecules, ligands, and catalysts. Phosphorus and nitrogen are in the same group and have similar electron configurations. Therefore, the properties of organophosphorus compounds are somewhat similar to those of the corresponding organic nitrogen compounds. However, phosphorus can also use 3d orbitals for bonding in addition to 3s and 3p orbitals, leading to many special high-valence organophosphorus compounds. Most coordination chemists and homogeneous catalysis chemists believe that low-coordination-number phosphorus compounds will have chemical properties more similar to carbon compounds, hence they have been trying to synthesize novel organophosphorus compounds based on sp and sp²hybridization forms of phosphorus atoms^[27]. However, due to the significant difference in spatial distribution between the 3p and 3s orbitals of phosphorus, their overlap is very weak^[31], making it difficult for phosphorus atoms to adopt sp and sp² hybridization. According to the Pauling electronegativity scale, compared to carbon (2.5) and nitrogen (3.04), phosphorus (2.1) is more electropositive. Moreover, because the electronegativity difference between P and N is greater than that between C and N, unlike the imine (C=N) bond in Schiff bases, the phosphinimine bond (P=N) exhibits a highly polarized characteristic.

Currently, there are mainly two methods for constructing phosphorus-nitrogen double bonds: one is the Staudinger reaction^{[35⇓⇓~38]} (as shown in Figure 2-1), where a nucleophilic azide group (R-N₃) attacks a trivalent phosphorus center (PR₃), followed by the release of N₂ to form P=N. However, this process requires the use of compounds containing azide groups, which poses an explosive risk; the other method is the Kirsanov reaction^{[39⇓⇓~42]} (as shown in Figure 2-2), which primarily involves primary amines as nucleophiles attacking the phosphorus center in pentavalent phosphorus chloride compounds to form P=N.

Due to the explosive hazards associated with the Staudinger reaction for preparing P=N double bond compounds, most research groups mainly refer to the Kirsanov reaction to prepare organic compounds containing P=N double bonds. For example, Le Floch et al^{[43⇓⇓~46]} proposed a synthesis route for the unilateral bromination of diphosphine ligands (as shown in Figure 3, path A). One method involves the reaction of an active intermediate, a bromide salt, with a primary diamine; another method involves the reaction of the active intermediate with one equivalent of a primary amine and diazabicyclooctane, followed by a simple deprotonation reaction of the mixed phosphorus-amine phosphonate to prepare the corresponding iminophosphorane. This method can be used to prepare mono/di-P=N-R compounds containing different R groups (such as alkyl, aryl, or benzyl). Among these, A1, A2 are prepared from triphenylphosphine, which is first brominated to form a P—N bromide compound, then reacted with 3-methylbutylamine or cyclohexanediamine to produce a lithium compound containing P=N, and finally reacted with PPh₂Cl to prepare a neutral ligand. If further acidified with hydrochloric acid, a chloride salt is formed^[45]. The product of unilateral bromination, when reacted with ethylenediamine, produces A4, a compound containing two P=N double bonds^[44]. In addition, using the same method, o-phenylenediamine is reacted to prepare an amino-containing P, N, N type ligand A3, which can form tridentate coordination with metals^[46].

Réau et al.^[47] prepared 1,2-diphosphinoamine chiral ligands (as shown in Figure 3, path B) through the Kirsanov reaction using commercial diamines and dibromophosphine derivatives. When the substituents attached to the phosphorus atom are phenyl or amino, the resulting compounds containing P=N bonds exhibit good stability.

Currently, there have been some reports on ligands containing phosphorus-nitrogen double bond structures^[48], with more common structures including monophosphoramidates (R₂P=NR')^[49], diphosphoramidates [(Ph₂P)₂N]^[50⇓~52], phosphinimides (R₃P=N)^[53,54], methylene-bridged phosphinimides [(RNPR'₂)₂CH]^[55⇓~57], and diamidophosphates [R₂P(NR')]^[58].

Compared to the relatively uniform methods for constructing P=N double bonds, there are multiple pathways for the formation of P—N single bonds in organic compounds, and compounds containing P—N single bonds also have a wide range of applications^[59⇓~61]. The following will introduce the progress in the ring-opening polymerization of cyclic esters catalyzed by metal complexes containing P—N and P=N bonds.

3 Catalytic Ring-Opening Polymerization of Cyclic Esters by Metal Complexes Containing P-N Single Bonds

Metal complexes containing P—N single bonds for ring-opening polymerization of cyclic esters have been reported in a limited number of studies, mainly including two categories: alkali (earth) metal complexes and rare earth metal complexes.

3.1 Alkali (Earth) Metal Complexes Containing P–N Single Bonds

Panda et al^[62]reported the sodium and potassium complexes 1a-bof 2, 6-diisopropylanilinoborane dimer with a P—N single bond structure (as shown in Figure 4). The molecular structure of 1aindicates that the phosphorus atom on the ligand coordinates with the sodium ion through the nitrogen atom in aniline (η¹) and borane (η¹). In the potassium complex 1b,the potassium ion prefers to coordinate with the π electrons of the aniline aromatic ring in an η⁶mode and with the BH₃group in the ligand in an η³mode. Lithium complex 2was prepared by reacting bis(borane-diphenylphosphinamine) with bis(trimethylsilyl)amido lithium; the lithium atom coordinates with the nitrogen atom in the amine and the hydrogen atom in the BH₃group of the ligand in an η²mode, forming a distorted tetrahedral configuration. Its catalytic properties have not yet been reported^[63].They also synthesized a sodium complex 3of N-(2,6-dimethylphenyl)diphenylphosphinamide^[64];its structure shows that the sodium atoms in this dimer are coordinated through oxygen atoms, and the sodium atoms are arranged in a tetrahedral manner. Subsequently, they also reported a series of alkali metal complexes of diphenylphosphine (sulfur or selenium) amides, in which the phosphorus atoms did not participate in coordination in the structures of the above-mentioned alkali metal compounds 4~6containing a P—N single bond^[65].

In 2015, Panda et al.^[66]synthesized and characterized chiral alkaline earth metal strontium and barium complexes 7and 8(containing P—N single bond, as shown in Figure 5) and studied their performance in catalyzing the ring-opening polymerization of caprolactone. Among them, strontium complexes 7aand 7bare enantiomers. When [ɛ-CL]₀/[I]₀=500 (in toluene, 25 ℃), 7acan efficiently catalyze the ring-opening polymerization of caprolactone within 15 minutes, achieving a monomer conversion rate of 75% with a narrow molecular weight distribution PDI=1.09; while 7bshows higher activity, reaching a monomer conversion rate of 90% within 10 minutes. The barium compounds 8a, 8bwith the same ligand exhibit even higher polymerization activity. Under the same conditions, 8aachieves a monomer conversion rate of 80% within 15 minutes, whereas 8breaches a monomer conversion rate of 98% within 10 minutes, which may be attributed to the larger ionic radius of barium compared to strontium. They subsequently reported a series of 2-methylamino-diphenylphosphine selenide pyridine-based alkaline earth metal complexes 9~12containing P-N single bonds^[65]. The structural analysis of the larger ionic radius alkaline earth metal complexes 9~11(Ae=Ca, Sr, Ba) shows that they are all N^N^Se tridentate bidentate chelating structures, while the smaller ionic radius magnesium compound 12with the same ligand is an N^N bidentate chelating structure, where the selenium atom in the ligand does not participate in coordination. These alkaline earth metal complexes 9~12can all catalyze the homopolymerization and copolymerization of racemic lactide (rac-LA) and caprolactone (ɛ-CL). Among them, the magnesium complex 12has lower polymerization activity, with a monomer conversion rate of 47% after 24 hours when [rac-LA]₀/[Mg] =100, while complexes 9~11show high catalytic activity and isotactic selectivity for the ring-opening polymerization of rac-LA at room temperature. For example, when [M]₀/[I]₀=100, calcium complex 9and strontium complex 10catalyze 99% monomer conversion within 30 minutes, whereas barium complex 11achieves the same conversion rate in just 5 minutes. The resulting polymers have molecular weights ranging from 14.6 to 15.9 kg/mol, with a narrow molecular weight distribution, PDI=1.24~1.35. The calcium complex 9with a smaller ionic radius produces PLA with the highest isotactic selectivity (P _m=0.89), significantly higher than the isotactic selectivity of polymers prepared under the same conditions by 10, 11(P _m=0.79, 0.70). Moreover, complexes 9~11can also highly actively catalyze the ring-opening polymerization of caprolactone ( ɛ-CL). Within 5 minutes, 11can completely convert 1000 equivalents of ɛ-CL, yielding a polymer with a molecular weight of 110 kg/mol and a narrow molecular weight distribution, PDI=1.27. Additionally, metal complexes 10, 11can efficiently catalyze the random copolymerization of rac-LA and ɛ-CL at room temperature. The random copolymerization of lactide and caprolactone is quite challenging due to the significant difference in reactivity between these two types of monomers, often resulting in block copolymers. The results show that when caprolactone/lactide copolymerize at a molar ratio of 1:1, the average sequence lengths of capryloyl and lactic acidoyl in the random copolymer catalyzed by 11are L _CL=1.9 and L _LA=2.7, respectively, and the resulting polymer has a unimodal molecular weight distribution (PDI =1.47 or 1.38), confirming it as a random copolymer, with the barium complex showing higher catalytic activity than the strontium complex. In contrast, the calcium complex exhibits low catalytic activity, producing only homopolymers.

In 2017, Panda et al.^[67]further synthesized and characterized a series of alkaline earth metal complexes 13~15(as shown in Figure 5). Using toluene as the solvent and at room temperature, when [rac-LA]₀/[I]₀=100, calcium and strontium complexes 13and 14catalyzed ring-opening polymerization, achieving a monomer conversion rate of 99% within 1 hour; whereas barium complex 15, under the same polymerization conditions, required only 10 minutes to achieve 99% monomer conversion, with the molecular weight of the resulting polymers close to the theoretical value, PDI < 1.6. The study showed that the initiation activity of these metal complexes increases with the increase in the ionic radius of the metal, Ba > Sr > Ca. When the coordinating atom in the complex changes from S to Se, the catalytic activity of the metal complex significantly improves. Calcium and strontium complexes 13, 14catalyze the ring-opening polymerization of racemic lactide, producing polymers with highly isotactic structures and good stereoselectivity, but the polymer produced by barium complex 15is atactic. Among them, the polymer obtained from calcium complex 13ahas a highly isotactic structure (P _i=0.87~0.78), the P _iof the polymer from barium complex 14is 0.75~0.63, while the P _iof the polymer from 15ais < 0.50, indicating that the isotacticity decreases as the size of the metal center increases, due to the reduction in steric hindrance leading to lower stereoselectivity. When the sulfur atoms in the ligands are replaced by selenium atoms, the stereoselectivity rapidly decreases (P _m=0.82, 13bvs P _m=0.71, 14b). In addition, calcium and strontium complexes 13, 14show very high catalytic activity for caprolactone in toluene ([ ɛ-CL]₀/[I]₀=100, T=25 ℃, achieving a monomer conversion rate of up to 99% within 5 minutes). Barium complex 15shows even higher catalytic activity, achieving a 99% monomer conversion rate of caprolactone within 1 minute. When other conditions remain unchanged and the monomer ratio is further increased to [ ɛ-CL]₀/ [I]₀=1000, calcium and barium complexes 13, 15can completely convert the caprolactone monomer within 30 seconds, demonstrating extremely high catalytic activity.

The bis(diphenylphosphine) amine (Ph₂PNHPPh₂) ligand, due to its isoelectronic relationship with bis(diphenylphosphine) methane (Ph₂PCH₂PPh₂) and the capability of the ligand to form a bite angle, can form binuclear or multinuclear metal complexes^{[68⇓⇓⇓⇓~73]}. Mallik and Panda et al. ^[64] reported the synthesis of lithium and corresponding alkaline earth metal complexes of bis(borane-diphenylphosphine) amide using silyl amine method or metal salt substitution method 16~18 (as shown in Figure 5). Among them, complex 18 adopts an η³-coordination mode centered on barium ions. When used for the ring-opening polymerization of ε-caprolactone, with [ɛ-CL]₀ /[I]₀ = 100:1, the monomer conversion rate reaches 98% within 4 minutes, the molecular weight of the polymer M_n = 9.8 kg/mol, PDI=1.86.

In the structures of the above-mentioned alkaline earth metal complexes containing P—N single bonds, the phosphorus atoms are not coordinated to the metal.

3.2 Rare Earth Metal Complexes Containing P—N Single Bonds

Compared to alkali (earth) metal complexes, rare earth complexes containing P-N structures have been studied relatively early. The Eklelman group^[50] reported the synthesis of neodymium organometallic compounds containing P—N bonds in 1989. The Roesky group^[74] synthesized lanthanide metal complexes with phosphorus-nitrogen single bonds 19~22 in 2002 and investigated their catalytic ring-opening polymerization performance (as shown in Figure 6). X-ray diffraction indicated that these complexes all adopted a hexacoordinate trigonal prismatic geometry. At room temperature, complexes 19~22 efficiently catalyzed the ring-opening polymerization of caprolactone, achieving monomer conversion rates of 95% to 99% within 1 min when [M]₀ /[I]₀ = 150. Among them, the molecular weights of the polymers obtained from 19, 20, and 22 were slightly higher than the theoretical values, with M_n ranging from 20.7 to 26.2 kg/mol, and a broad molecular weight distribution (PDI=1.58~1.83); whereas, the molecular weight of the polymer obtained from the neodymium complex 21 was very close to the theoretical value, M_n = 17.2 kg/mol, and had a narrow molecular weight distribution (PDI=1.12). To explore the initiation mechanism of these rare earth metal complexes, they crystallized complexes 19~22 in oxygen-containing solvents, and the results showed that the Ln-P bond did not break. Therefore, it is speculated that during the ring-opening polymerization process, the carbonyl group on the caprolactone coordinates with the metal, forming a heptacoordinate structure around the metal center, further generating a metal alkoxide to initiate the ring-opening polymerization of caprolactone, which is a typical coordination-insertion polymerization mechanism.

In 2007, Williams et al.^[75] prepared and characterized yttrium metal complexes 23~26 containing phosphorus-nitrogen single bond structures (as shown in Figure 7). X-ray diffraction indicated that the central metal of complex 23 had a planar tetrahedral coordination structure, while 24~26 were dimeric with oxygen-bridged structures. These yttrium complexes all showed high polymerization activity in the ring-opening polymerization of lactide. When dichloromethane was used as the polymerization solvent, complex 24 could achieve a monomer conversion rate of 99% within 70 s ([M]_o/[I]₀=200, T=25 ℃), much higher than when tetrahydrofuran was used as the solvent. The reason for this is that tetrahydrofuran tends to coordinate with the metal, reducing the reactivity of the complex. Complex 23 exhibited good controllability during the polymerization process, yielding polymers with number-average molecular weights very close to theoretical values, and there was a linear relationship between the molecular weight and the monomer conversion rate. When catalyzing the ring-opening polymerization of rac-LA, the homonuclear decoupled NMR of the resulting polymers showed a certain degree of syndiotactic selectivity (P_r =0.61~0.64), and this value was independent of the solvent used in the polymerization. In 2011, they^[76] also investigated the effect of linkers between diamines on catalytic performance. A series of yttrium complexes 27~31 containing different phosphorus substituents (alkyl, alkoxy) and various linkers in bis(phosphoramidate) ligands were synthesized and characterized (as shown in Figure 7). These complexes all showed high apparent rate constants (k _obs=(1.7~6.8)×10^-2 s^-1) for the ring-opening polymerization of lactide, producing polymers with syndiotactic selectivity (P _r=0.61~0.85). In CH₂Cl₂ solvent, at 298 K, [LA]₀/[I]₀ =400, complex 31 could catalyze the ring-opening polymerization of rac-LA, achieving a monomer conversion rate of 99% within 1 min, with k _obs=6.8×10^-2 s^-1, and the resulting polymer had a molecular weight of 37.2 kg/mol and a narrow molecular weight distribution (PDI=1.19). Compared to yttrium amido compounds, yttrium alkoxide complexes (X =OⁱPr) exhibited higher initiation activity and better controllability. Yttrium aryloxide complexes showed slower polymerization rates and poorer stereo-selectivity, and large steric hindrance substituents on phosphorus also led to lower polymerization rates for phosphoramidate metal complexes. This may be due to the lower nucleophilicity of the amide group compared to the aryloxide, as well as the steric effects from the amide and aryl oxide groups. Yttrium amido complexes with tert-butyl phosphorus substituents showed good stereoselectivity towards racemic lactide, generating stereocomplexed polylactides.

Yttrium metal complexes with similar structures 32~33 ^[77]show very high catalytic activity for racemic lactide. When [M]₀/[I]₀=200, T=25 ℃, 100% monomer conversion can be achieved within 10 minutes. Kinetic studies show that ln([LA]₀/[LA]_t) has a linear relationship with the polymerization time t. In dichloromethane, the polymerization rate constant is very large, k _obs=17.8×10^-3s^-1, much higher than the polymerization rate constant in tetrahydrofuran (THF), k _obs=8.36 ×10^-3 s^-1. This is consistent with previous conclusions^[25]. The mononuclear yttrium complex 33yields syndiotactic polymers P _r=0.85 in THF, while the binuclear yttrium complex 32mainly produces atactic polylactide. This stereoselectivity is attributed to the chain-end control mechanism. This process is mainly controlled by the steric hindrance of the ligand, that is, to minimize the steric hindrance during insertion, the two methyl groups should have opposite chirality; therefore, the chain end with RRconfiguration is more favorable for inserting the SSenantiomer, ultimately leading to a polymer with an RRSSRRSSconfiguration. Therefore, in the catalytic process of the mononuclear yttrium complex 33, THF coordination is very important for selective polymerization because after coordination, the environment around the metal center becomes more crowded, limiting the direction of attack of the lactide monomer, thereby controlling the bonding or insertion rate of the next lactide through the stereostructure of the inserted lactide. On the other hand, the binuclear complex 32, due to the lack of a single active center with spatial steric hindrance, does not restrict the coordination and attack of lactides with different chiralities, and thus cannot control the stereostructure through steric effects.

In 2012, Williams et al.^[78] used a high molecular weight polymer with hydroxyl end groups as a macromolecular initiator to study the ring-opening polymerization of LA catalyzed by yttrium metal complexes containing phosphorus-nitrogen single bonds 34 (as shown in Figure 7), successfully preparing ABA triblock copolymers, with monomer conversion rates of lactide as high as over 95%. GPC analysis showed that the molecular weight of this triblock copolymer was 51 kg/mol (PDI=1.30), and it exhibited a unimodal distribution. DSC data indicated that the obtained triblock copolymer had two T _g values, with the lower T _g being attributed to the movement of the diblock copolymer segments serving as external alcohol in the block copolymer, which also demonstrated the block nature of the polymer.

To further investigate the impact of electronic and steric effects on the catalytic behavior of complexes, Williams et al.^[79] prepared a series of yttrium metal complexes containing P—N single bond structures 35~42 (as shown in Figure 7). These complexes also exhibited good catalytic polymerization activity for rac-LA, with the rate of ring-opening polymerization decreasing in the order of isopropyl, phenyl, and ethoxy substituents on the phosphorus site. When the substituent was phenyl or ethoxy, it reduced the polymerization reaction rate, thus proving that the rate-determining step in the ring-opening polymerization process was the insertion of lactide monomers between the yttrium metal and the alkoxide. The phosphorus substituents in the complexes affected the chemical bonding force between the yttrium metal and the alkoxide, thereby influencing the initiation rate of the complexes. The bridging group of the diamine in the complex structure had a significant influence on stereoselectivity. In most of the polymerizations, there was a linear relationship between the molecular weight of the polymer and the monomer conversion, indicating good controllability of the polymerization process. However, during the ring-opening polymerization of rac-LA initiated by complexes 35~40, no stereoselectivity was observed, resulting in atactic polylactides; while 41, 42 showed some syndiotactic stereoselectivity in the ring-opening polymerization of rac-LA (P_r values were 0.68 and 0.79, respectively).

4 Catalytic Ring-Opening Polymerization of Lactones by Metal Complexes Containing P=N Structures

Compared with metal complexes containing P—N single bond structures, there are a wide variety of metal complexes that contain phosphinimine P=N double bond structures, and the ligand structures are also diverse. Based on this, this section will introduce the application progress of their catalytic ring-opening polymerization according to the types of coordinating atoms in the complexes.

4.1 Catalysis of Lactide Polymerization by Metal Complexes with N^N Bidentate Ligands Containing P=N Double Bonds

In 2006, Cui Dongmei et al.^[80]reported anilinophosphinimine rare earth metal complexes 43~46 (as shown in Figure 8), where the molecular structure of complex 43was characterized by X-ray diffraction. The crystal structure indicated that one C-H on the phenyl ring of the anilinophosphinimine ligand was activated, thus forming a dianionic ligand. This ligand coordinates with the yttrium atom in a rare tridentate C^N^N mode, forming a distorted trigonal bipyramidal coordination configuration. The lutetium complex 44has a similar structure to the yttrium complex 43, also exhibiting a tridentate C^N^N coordination configuration. When complex 43is treated with 2,6-diisopropylaniline in toluene at room temperature, it yields N^N-coordinated bis(imino) complex 45; the hydrolysis product of 43yields dinuclear yttrium complex 46. Yttrium complex 43can catalyze the ring-opening polymerization of rac-LA with high activity, under conditions of [LA]₀/[Y] = 200 at room temperature, achieving a monomer conversion rate of 95.5% within 8 minutes, with a polymer molecular weight of 26 kg/mol and PDI=1.83. Lutetium complex 44with a similar structure yields almost no polymer under the same conditions, suggesting that the smaller the ionic radius of the rare earth metal center, the lower the catalytic activity.

Roesky et al.^[81]reported a series of rare earth metal complexes 47~50 (containing phosphorus-nitrogen double bonds and phosphorus-nitrogen single bonds, as shown in Figure 8). These complexes can catalyze the ring-opening polymerization of caprolactone and the addition polymerization of methyl methacrylate at room temperature. When [ε-CL]₀/[I]₀=240~440, without the need to add other cocatalysts, these complexes can highly actively catalyze the ring-opening polymerization of caprolactone, achieving a maximum conversion rate of up to 100% within 5 minutes, and the catalytic activity depends on the ionic radius of the central metal. However, complexes 47and 50(M = Y, Yb) quickly become gel-like after the monomer is added, thus limiting further increases in monomer conversion rates, so the molecular weight distribution of the resulting polycaprolactone is relatively broad, with PDI=1.8~4.4. At low monomer conversion rates, the molecular weight of the obtained polymers is three times the theoretical value, while at high conversion rates, the molecular weight of the polymers is relatively low. They believe that this is likely due to one complex molecule being able to continuously initiate the chain growth of multiple molecular chains.

In 2011, Lee et al.^[82] prepared a series of amino phosphinimine alkylzinc complexes 51a~c and 52a~c (as shown in Figure 9), but the alkoxo zinc derivatives were not obtained. The alkylzinc complexes 51a~c and 52a~c could efficiently catalyze the ring-opening polymerization of ε-CL. In toluene as solvent, when [ε-CL]₀ / [I]₀ = 125, and the reaction temperature was 80 ℃, the monomer conversion rate reached up to 100% within 1.5 h, with a polymer molecular weight of 60 kg/mol and PDI=1.31. In contrast, the catalytic efficiency of zinc acetate complex 53 for the ring-opening polymerization of ε-caprolactone was lower; in toluene, when [ε-CL]₀ / [I]₀ = 100, and the reaction temperature was 80 ℃, after 1.5 h, the monomer conversion rate was only 48%.

4.2 Metal Complexes with N^N^X (X = N, O, C) Tridentate Ligands Containing a P=N Double Bond

In 2007, Wang Zhongxia et al.^[83]first reported the synthesis of magnesium and zinc complexes containing phosphinimine-quinoline structures 54~58(as shown in Figure 10)and their application in the ring-opening polymerization of ε-CL. The X-ray diffraction results of the magnesium complex 54indicated that it had an N^N^C tridentate coordination structure. When toluene or tetrahydrofuran was used as the solvent, the magnesium complex 54could catalyze the ring-opening polymerization of ε-CL, regardless of whether benzyl alcohol was used as a co-initiator. In the presence of benzyl alcohol, the polymerization reaction proceeded very quickly; at 25 ℃, with a monomer-to-metal ratio of 200:1, the polymerization system rapidly turned into a gel, entrapping the catalyst. ¹H NMR analysis showed a monomer conversion rate of 67%, yielding a polymer with a molecular weight M _n=213.4 kg/mol and PDI=1.3. To reduce the reaction rate, the reaction temperature was lowered to 0 ℃, and the polymerization was carried out in 20 mL of toluene or tetrahydrofuran. The results indicated that the polymerization proceeded faster, producing polymers with higher molecular weights. The complex 54reacted with an equivalent amount of benzyl alcohol in toluene to form product 59, indicating that the benzyl alcohol underwent addition to the ligand rather than elimination with Mg-Et to form an alkoxymagnesium compound. This is speculated to be due to the high ring strain in the complex 54. Without benzyl alcohol, the complex 54could also catalyze the ring-opening polymerization of ε-CL. At [ε-CL]₀/[I]₀=200 and a reaction temperature of 0 ℃, the monomer could be completely converted within 3 minutes in toluene, yielding a polymer with a molecular weight of 255 kg/mol and a narrow molecular weight distribution (PDI=1.16). Zinc complexes 55~58with similar structures exhibited lower catalytic activity for the ring-opening polymerization of ε-CL in the presence of benzyl alcohol compared to the complex 54/benzyl alcohol system. At 60 ℃, 90% monomer conversion required 60~156 minutes, and the resulting polymers had molecular weights M _n=26.2~39.3 kg/mol and a narrow molecular weight distribution (PDI =1.01~1.09). ¹H NMR confirmed that the zinc complex 58also reacted with benzyl alcohol to form 60, where benzyl alcohol only coordinated with the zinc metal without undergoing elimination to release ethane and form an alkoxizinc compound, which is often considered the active center for ring-opening polymerization. Additionally, the kinetic studies of zinc complexes 55~58in toluene in the presence of benzyl alcohol showed that ln([M]₀/[M]_t) had a good linear relationship with time t, indicating that the polymerization rate was first-order with respect to the monomer concentration. However, the complexes 56~58all exhibited an induction period before catalyzing the ring-opening polymerization, while the zinc complex 55did not require an induction period. This may be primarily due to the electronic effect of the phenyl group in Zn-Ph on the central metal. The kinetic curves of complexes 57and 58were almost parallel, suggesting that the substituents on the phosphorus atom mainly affected the induction period of the polymerization reaction but did not influence the rate constant. In the 58/BnOH catalytic system, M _nshowed a linear relationship with the monomer conversion, and the resulting polymers had a narrow molecular weight distribution (PDI=1.02~1.05), indicating that the polymerization exhibited characteristics of "living polymerization."

They also prepared and characterized a series of zinc, aluminum, and tin complexes with quinoline derivatives containing phosphino-imine structures as ligands 61~71 (as shown in Figure 10) ^[84~85]. Among them, the complexes 61~67 are all N^N^N tridentate coordination structures. In the presence of benzyl alcohol, these complexes 61~67 catalyze the ring-opening polymerization of ɛ-CL to produce polymers with controllable molecular weight and molecular weight distribution. The aluminum complexes exhibit higher polymerization activity than the zinc complexes, with the obtained polymers having molecular weights very close to the theoretical values and a narrow distribution (PDI=1.08~1.38), indicating that the active sites of the catalysts are quite stable during the polymerization process, making the process controllable. Meanwhile, regardless of the presence or absence of benzyl alcohol, the aluminum complexes cannot catalyze the polymerization of LA. Similarly, N^N^O coordinated complexes 68~71 can also efficiently catalyze the ring-opening polymerization of caprolactone, where the polymerization activities of complexes 70, 71 are comparable, suggesting that the R group has a minor impact on the catalytic activity of the alkylaluminum complexes 70, 71. At 95 ℃, under conditions where [CL]₀:[Al]₀:[BnOH]₀ =200∶1∶1, they achieve 100% monomer conversion within 40 minutes, yielding polymers with M_n =33 kg/mol and a molecular weight distribution PDI=1.24. Tin complexes with the same ligands show lower catalytic activity than the aluminum complexes, achieving monomer conversion rates of 26%~48% after 20 hours at 85 ℃. This is possibly due to the Sn-N(SiMe₃)₂ group having a higher steric hindrance compared to Al-Me, which hinders the coordination process between caprolactone and the metal center, thus reducing reactivity. The reason for the lower activity of aluminum complex 68 compared to 69 lies in the higher steric hindrance when the R group is a tert-butyl, thereby reducing the catalytic activity of the complex. Polymerization kinetic studies have shown that during the ring-opening polymerization of caprolactone catalyzed by complexes 61~71, the reaction exhibits a clear first-order reaction towards the concentration of caprolactone monomers, indicating that the active centers are relatively stable during the reaction. There is a linear relationship between the molecular weight of the polymers and the conversion rate of the caprolactone monomers, with a narrow molecular weight distribution (PDI =1.16~1.30), indicating good controllability of this type of catalyst in the ring-opening polymerization of caprolactone.

In 2010, Williams et al.^[86] synthesized and characterized a series of neodymium metal complexes with N^N^C tridentate coordination containing phosphinimine ligands 72~75 (as shown in Figure 11). Among these, neodymium complexes 72~73 could efficiently catalyze the ring-opening polymerization of rac-LA. When [M]₀ / [I]₀ = 200, the solution viscosity of LA increased significantly within seconds at room temperature, with monomer conversion rates reaching over 95%. Additionally, the molecular weight of the polymers showed a linear relationship with the conversion rate and had a very narrow distribution, indicating that the polymerization process was controlled. Characterization of the polymers by MALDI-TOF indicated that the polymer ends were N(SiMe₃)₂ structures. However, the molecular weight of the obtained polymers was ten times the theoretical calculated value, suggesting that only 10% of the initiators were activated and effective. The bidentate coordinated neodymium complex 74 did not show initiating activity for lactide ring-opening polymerization, possibly due to the lack of an initiating group in its structure. Nevertheless, when an equivalent amount of 2-propanol was added, complex 74 exhibited polymerization activity with better controllable polymerization characteristics. Therefore, an alkoxyneodymium complex 75 was prepared and applied to the ring-opening polymerization of rac-LA. The results showed that under the same polymerization conditions, complex 75 demonstrated higher initiating activity than 73, achieving a monomer conversion rate of up to 90% within 90 seconds and complete monomer conversion within 6 minutes when [M]₀ / [I]₀ = 200. These findings indicate that metal alkoxide M-OR as an initiating group often exhibits higher reactivity than M-N(SiMe₃)₂ due to lower steric hindrance and higher nucleophilicity.

Recently, Mehrkhodavandi^[87] and others reported on tert-butyl indium complexes with N^N^O tridentate coordination based on a salicylaldimine structure 76, 77, as well as similar compounds containing phosphinimine structures 78 (as shown in Figure 12,). They explored the effect of ligand variations on the copolymerization of rac-LA and ε-CL (with different monomer addition sequences and simultaneous addition). The results showed that all three complexes exhibited moderate activity for the ring-opening polymerization of rac-LA, achieving 100% monomer conversion within 24 hours in dichloromethane solvent, yielding poly(lactic acid) with a narrow molecular weight distribution (PDI = 1.05-1.10). Under the same conditions, the order of activity for the ring-opening polymerization of rac-LA was 76 > 78 > 77, indicating that in this catalytic system, indium complexes with imine ligands are more stable and thus more active than those with phosphinimine ligands. Experimental results also found that complexes 76 and 78 can effectively catalyze the block copolymerization of rac-LA and ε-CL, but exhibit different catalytic behaviors when the addition sequence of rac-LA and ε-CL is varied. When ε-CL and rac-LA were added sequentially, the indium complex 76 containing the salicylaldimine ligand could catalyze the formation of PCL-b-PLA; whereas, under the same conditions, the complex 78 with the phosphinimine ligand only produced a mixture of PCL homopolymer and PLA homopolymer. When rac-LA was added first followed by ε-CL, the polymerization behavior was entirely different. After 100% conversion of rac-LA, complex 76 could not continue to polymerize ε-CL, resulting in only the homopolymer PLA; while the complex 78 with the phosphinimine structure, after complete conversion of rac-LA, could continue to catalyze the ring-opening polymerization of ε-CL, forming PLA-b-PCL block copolymers, albeit over a longer period. When rac-LA and ε-CL were added simultaneously, only the indium complex 76 containing the salicylaldimine ligand could form PLA-b-PCL copolymers. The conversion of ε-CL was only observed after the complete conversion of rac-LA. This difference in polymerization may be explained by two reasons: (1) the difference in the relative stability of the intermediates formed by complexes 76 and 78 (the imine ligand forms a more stable intermediate in such complexes); (2) the difference in the stability of the lactide-chelated intermediates during the polymerization of lactide. Although there is no direct evidence for the formation of lactide chelates, this view is supported by comparing the polymerization reactions with sequential and simultaneous monomer addition. Experiments show that complex 76 has much higher reactivity for the ring-opening polymerization of ε-CL than LA, but when added simultaneously, rac-LA always polymerizes first. Furthermore, if rac-LA is polymerized first, complex 76 cannot further polymerize ε-CL, indicating that it is difficult to insert ε-CL after rac-LA in this system. In contrast, the complex 78 with the phosphinimine ligand can catalyze the polymerization of ε-CL after the complete polymerization of rac-LA, which may be due to the greater instability of the lactide chelate in this system. However, it is still unclear whether the steric effect (presence of PPh₂ groups) or the electronic effect of the P=N double bond is dominant. Later, Williams et al.^[88] used computational methods to compare the catalytic performance differences between indium complexes with P=N and C=N in the same structure, finding that indium complexes with imine ligands are more stable and active, while P=N complexes have an advantage in regulating the activity of copolymerization.

4.3 Ion-Pair Type Metal Complexes Containing P=N Structure

4.3.1 Ion-Pair Metal Complexes with N^O Coordination Containing a P=N Structure

Hayes et al^[89,90]reported a class of ionic complexes containing P=N structures, with phosphino imine substituents linked to a dibenzofuran (dbf) backbone79~84(as shown in Figure 13).

The ligand coordinates with the metal zinc in a bidentate manner through the nitrogen in the phosphinimine and the oxygen in the furan, forming a six-membered ring coordination structure^[89]. ³¹P NMR indicates that the phosphinimine is highly sensitive to the chemical environment, showing significant shifts when coordinated with Lewis acids, so the ³¹P NMR of complexes 79 and 80 indicates that the phosphinimine is tightly bound to the zinc center. The ¹⁹F NMR spectrum of 79 is similar to that of the non-coordinating anion [B(C₆F₅)₄]^-, and even at high temperatures for extended periods, no signals indicative of fluorine or C₆F₅transfer to the metal center were detected. Crystal data of complex 80 show that the strongly coordinating triflate anion binds to the zinc center, with significant disorder present in the structure; moreover, due to enhanced steric repulsion between the ethyl group and the bulky 2,6-diisopropylphenyl (Dipp) groups, the Zn—O bond is not as strong as the Zn—N bond, which may play a critical role in chemical reactions. When complexes 79 and 80 were applied to the ring-opening polymerization of L-LA, it was found that under initial lactide concentrations of 1 M and a catalyst concentration of 1% (molar fraction), complex 79 achieved 90% conversion after 6 h at 100 ℃, whereas 80 required 9 h to reach 85% monomer conversion under the same conditions. Its lower catalytic activity may be due to competitive coordination by the triflate anion, but this is the first report of a cationic zinc complex serving as a catalyst for the ring-opening polymerization of lactide.

They then synthesized a series of methylphenyl-substituted phosphinimine ligands with the same backbone and the corresponding ionic ethylzinc and lactate zinc complexes 81~84^[90], all with non-coordinating B(C₆F₅)₄^- as the anion. When complex 81 and 83 were used to catalyze the ring-opening polymerization of lactide, even under very stringent conditions, the ionic ethylzinc complex 81 in the form of ion pairs showed almost no activity towards the ring-opening polymerization of lactide, possibly due to the poor initiating ability of the metal alkyl (Zn-Et). In contrast, the methyl lactate zinc complex 83 exhibited relatively high activity towards lactide, achieving a conversion rate of 90% within 9 hours at a monomer-to-metal ratio of 100:1, [M]=0.5 mol/L. At M/I ratios of 200-400, the molecular weight of the obtained polymers ranged from 17 to 36 kg/mol, with a broad molecular weight distribution, PDI=1.81-1.98, indicating a significant amount of transesterification side reactions.

4.3.2 Ion-Pair Metal Complexes with N^N^O Coordination Containing P=N Structure

In 2010, the Hayes group^[91,92]prepared bis(phosphinimine)benzofuran cationic zinc complexes 85~87 (as shown in Figure 14),the structures of which indicated that the ligands adopted an N^N^O tridentate coordination model. The alkyl zinc ionic complexes 85 and 86 could not initiate LA polymerization; however, complex 87 could catalyze the ring-opening polymerization of lactide with high activity at room temperature. In CD₂Cl₂ solvent at 25 ℃, with a polymerization time of 50 min, a conversion rate of 90% for 200 equivalents of rac-LA was achieved, yielding polymers with a certain degree of syndiotacticity P _r = 0.63, and an apparent rate constant of 8.65×10^-4 s^-1. The measurement of reaction rates confirmed that the polymerization reaction was first-order with respect to monomer concentration. The polymerization results indicated that the methyl lactate zinc complex 87 was the first cationic metal complex to initiate “coordination-insertion” polymerization of lactide at room temperature, although the polymerization mechanism proved it was not cationic. When complex 87 catalyzed the ring-opening polymerization under [LA]₀/[Zn]₀= 100 and 200 conditions, the molecular weight of the obtained polymers was close to the calculated value; however, at higher monomer-to-metal ratios, there was a significant decrease in the molecular weight of the resulting polymers. For example, when [LA]₀/[Zn]₀ = 1000, the number-average molecular weight was slightly over 50 kg/mol, only about 35% of the theoretical value, possibly due to the role of monomer impurities as chain transfer agents during the polymerization at high monomer concentrations.

They also reported a cationic magnesium complex with a similar structure 88, 89^[93], which can catalyze the ring-opening polymerization of ε-CL with high activity. Although there have been many reports on neutral magnesium complexes catalyzing the polymerization of caprolactone, cationic magnesium complexes catalyzing the ring-opening polymerization of ε-CL are rarely reported. When comparing the ³¹P NMR spectra of magnesium complexes 88 and 89 with their corresponding ligands, it was found that the phosphorus shift of the ligand changed from δ=-17.6 to 23.0 ppm (for 88) and from δ=-17.6 to 23.2 ppm (for 89). This result is consistent with previously reported monophosphinimine zinc complexes and other phosphinimine magnesium complexes. Magnesium complexes 88 and 89 catalyze the ring-opening polymerization of ε-CL, achieving a monomer conversion rate as high as 90% within 4 minutes at a complex concentration of about 4 mM (equivalent to 0.77 mol% of ε-CL monomer). Due to the extremely weak coordinating ability of the BR₄^-anion, it is speculated that it will not significantly affect the polymerization process. Moreover, complex 89 still maintains high catalytic activity even when the concentration is reduced to 0.37 mM (0.19 mol% relative to ε-CL monomer). Furthermore, magnesium complex 89 remains catalytically active at -40 ℃; however, below 0 ℃, the increased viscosity of the reaction system makes it difficult for the monomer to be completely converted.

In 2012, Hayes et al. reported a series of novel bisphosphinimine ligands and their corresponding cationic alkylzinc complexes 90, 91^[94], which can initiate the ring-opening polymerization of rac-LA to obtain syndiotactic polymers (P_r = 0.51~0.63) with molecular weights ranging from 60.5 to 209 kg/mol and a relatively broad molecular weight distribution (PDI = 1.69~2.62). However, the conversion rate is high (when the monomer-to-metal ratio is 200~1000, after 2~4 h, the monomer conversion rate is above 92%). The details of the polymerization mechanism for these complexes remain unclear, and kinetic data indicate that rac-90 and meso-90 form active centers with the monomer in solution (as shown in Figure 15).

4.3.3 Ion-Pair Metal Complexes with N^N^N Coordination Containing a P=N Structure

In 2021, the Hayes group ^[95]also reported a series of neutral gallium and indium complexes LMCl ₂ 92, 93(as shown in Figure 16).Reacting AlMe₃with the same ligand yielded the related N^N^N tridentate organoaluminum complex LAlMe ₂, which upon treatment with the Lewis acid B(C₆F₅)₃produced the ionic aluminum methyl complex [LAlMe]⁺ [MeB(C₆F₅)₃]^- (94). Cui et al. ^[96]synthesized and characterized a series of tridentate nitrogen zinc complexes 95~103containing phosphinimine ligands (Figure 16). When these complexes were applied to the ring-opening polymerization of racemic lactide, it was found that the catalytic activity of complexes 95, 96, 98was low, with polymerization conditions being [M]₀/[I]₀=200, 70 ℃, THF as solvent, and monomer conversion rates of 84%~86% after 10 hours, yielding atactic polymers P _m=0.40~0.45, while the zinc chloride complex 97showed almost no activity. In contrast, complexes 99~102exhibited high polymerization activity at lower temperatures (30 ℃), with monomer conversion rates of 73%~91% over 10~36 hours, producing isotactic poly(lactic acid) (P _m=0.68~0.74) with stereoselectivity. Due to transesterification reactions, the resulting polymers had a broad molecular weight distribution (PDI =1.27~2.01), but the measured molecular weights of the polymers were very close to the theoretical calculated values. Among them, 99, 101showed higher activity than 100, 102, indicating that the influence of groups connected to the zinc metal is greater than that of the ligands. The benzyloxy zinc complex 103demonstrated even higher catalytic activity under the same conditions, suggesting that the Zn-O bond is more easily activated in the reaction than the Zn-N bond. Through ¹H NMR and MALDI-TOF characterization of the polymers, it was found that the mechanism by which complexes 95~103initiate the ring-opening polymerization of lactide differs, where complexes 95, 96and 99~102follow an anionic or cationic mechanism, while 103catalyzes the polymerization process via a "coordination-insertion" mechanism.

4.4 N^N^O^O Tetracoordinate Metal Complexes Containing P=N Double Bonds

The preparation of Salen-type ligands based on salicylaldehyde imine structure is relatively simple, and it is easy to change the electronic environment around the metal center by varying the substituents. Therefore, there are many reports on the use of metal complexes containing Salen-type ligands for ring-opening polymerization of cyclic esters, especially aluminum complexes, which show excellent stereoselectivity in the ring-opening polymerization of rac-LA, but their catalytic efficiency is relatively low. Phosphasalen (also known as Palen) type ligands have a very similar structure to salen-type ligands (as shown in Figure 17)^[28]. When the phosphorus imine structure (P=N) replaces the imine structure (C=N), there is a significant difference in their stereochemistry and electron-donating properties compared to the imine structure. The presence of two substituents on the tetrahedral structure of the phosphorus atom gives them greater electron-donating ability, higher flexibility, and larger steric hindrance. Moreover, the electron-donating nature of the ligand reduces the Lewis acidity of the central metal, making the metal-alkoxide bond more reactive, thus accelerating the ring-opening polymerization process of lactones.

The synthesis of phosphasalen ligands is mainly achieved through the Kirsanov reaction between dibenzylamines and diphenylphosphinophenols. Due to the limited variability of the basic phosphinophenol structure, the catalytic performance of the catalysts is typically regulated by changes in the substituents at the ortho- and para-positions of the phenol and modifications in the linking groups of the diamines (as shown in Figure 18).

Phosphasalen complexes have many sites that can be modified: 1) the diamine bridge and the central part of the diamine bridge X; 2) the phosphorus substituent R; 3) the ortho- and para-substituents R¹ of the phenol group; 4) the central coordinating metal M; 5) the alkyl group R² in the metal alkoxy. Among them, the modification of the central part X of the diamine bridge and the diamine bridge is relatively simple, as there are a series of diamines that can be applied to the Kirsanov reaction, and the influence of different X groups largely causes changes in the structure/conformation of the complex. The ortho- and para-substituents (R¹) of the phosphinophenol have a significant impact on the electronic and spatial structure of the catalytic metal center, especially the ortho-substituents of the phosphinophenol, due to their proximity to the catalytically active center, have a greater steric hindrance effect than the para-substituents, thus further affecting the catalytic activity and stereoselectivity. There are many reports about Phosphasalen-type metal complexes, and the following will introduce their performance in catalyzing the ring-opening polymerization of lactones from the perspective of structural variability.

4.4.1 Phosphasalen Metal Complexes with Different Amine Bridge Linkages

Ferrocene has received extensive attention due to its unique redox properties, and diamines with ferrocene as a linker can also be used for the construction of Phosphasalen ligands. In 2011, Diaconescu et al.^[97,98] synthesized four-coordinate Phosphasalen-type yttrium, indium, and cerium complexes 104~108 (as shown in Figure 19) with ferrocene-structured amine bridges. To investigate the influence of the central metal valence on the ring-opening polymerization of L-LA, (3,5-bis(trifluoromethylphenyl))ferrocene boronic acid (FeBAr^F) and cobaltocene (Cp₂Co) were used as oxidizing and reducing agents, respectively, to achieve the conversion between complexes 104~108 and their oxidized state complexes (104-BAr^F~108-BAr^F). Mössbauer spectroscopy and XANES results for complexes 104~107 (104-BAr^F~107-BAr^F) showed that during the conversion between 104~107 and 104-BAr^F~107-BAr^F, the valence of iron in the ferrocene group changed. For the cerium complex 108, during the transition from 108 to its oxidized product 108-BAr^F, the valence of the central metal cerium changed from +3 to +4, while the iron in the ferrocene remained at +2. At room temperature, the yttrium complex 104 catalyzed the ring-opening polymerization of 100 equivalents of L-LA, achieving a monomer conversion rate of 74% after 3 hours, with a polymer molecular weight M_n = 7.5 kg/mol and PDI = 1.07. However, the corresponding oxidized state complex 104-BAr^F had no catalytic activity. When THF was used as the solvent, at 70 ℃, complex 104 catalyzed the ring-opening polymerization of 100 equivalents of rac-LA, reaching a monomer conversion rate of 95% after 8 hours, but the resulting polymer had no stereoselectivity. To further study the effect of metal valence, under the same polymerization conditions, complex 104 first catalyzed the ring-opening polymerization of L-LA for 1 hour, achieving a monomer conversion rate of 24%. Then, FeBAr^F was added, temporarily stopping the polymerization reaction. After half an hour, Cp₂Co was added, and it was observed that the L-LA ring-opening polymerization resumed, indicating that the ferrocene(II)-bridged yttrium complex 104 has higher catalytic activity for lactide ring-opening polymerization. The molecular weight of the polymer obtained by complex 104 was very close to the theoretical value, and the molecular weight distribution was narrow (PDI = 1.1), suggesting that the lactide polymerization process catalyzed by complex 104 conforms to the characteristics of controlled polymerization. Additionally, ¹H NMR data of the polymer indicated that the polymer end contained a tert-butanol group, suggesting that the lactide monomer was inserted into the Y-O^tBu bond. However, the indium complex 107-BAr^F exhibited higher activity than the yttrium complex 104-BAr^F, contrary to the trend observed for ferrocene(II)-containing yttrium complexes, indicating that the activity of ferrocene-based metal complexes with redox properties depends on the type of central metal. Furthermore, when [LA]₀/[I]₀ = 100 and T = 0 ℃, complex 108 catalyzed the ring-opening polymerization of rac-LA with moderate syndiotactic selectivity (P_r = 0.84).

In 2012, Auffrant and Williams et al.^[27] synthesized Phosphasalen-type yttrium complexes 109 with high stereoselectivity and high activity for the ring-opening polymerization of rac-LA (as shown in Figure 20). When [M]₀ /[I]₀ =1000, complex 109 (R¹ =R² =R=^tBu), using THF as the solvent, could achieve a monomer conversion rate of 97% after 0.75 min of reaction, with the molecular weight of the polymer M_n =223 kg/mol and PDI =1.34.

Using aminoethylenediamine as the raw material, they also prepared a series of five-coordinated Phosphasalen-type yttrium complexes 110~113 (Figure 20). Complex 110 [Y(Psalen^R)(OR’)] catalyzes the polymerization of rac-LA at room temperature, with [M]₀ /[I]₀ =500, achieving 100% monomer conversion in just a few seconds, yielding syndiotactic polymers (P_r> 0.87). Complex 111 [Y(PsalenN^tBu)(OR’)] also exhibits high activity, with [M]₀ /[I]₀=500, reaching 100% monomer conversion after 1 h of reaction. Polymerization kinetics studies show that the polymerization rate of complex 111 [Y(PsalenN^tBu)(OR’)] is linearly related to the monomer concentration, yielding isotactic polymers (P_m=0.84). They attribute the different stereochemical structures of the polymers obtained from the same metal center but different ligands to steric effects. The yttrium complex 110 [Y(PsalenN^t-Bu)(OR’)] has higher steric hindrance, which limits the spatial hindrance of the active center, thus favoring the formation of polymers with better stereoselectivity^[26].

In 2013, Williams et al.^[30]designed and synthesized a series of complexes 114~119with different diimine bridges (Figure 20). These complexes showed high activity and good syndiotactic selectivity in the ring-opening polymerization of rac-LA. When [LA]₀/[Y]₀=1000, at 298K in THF solvent, 118catalyzed a monomer conversion rate of 86% within 20 seconds, yielding a polymer with a molecular weight of up to 165 kg/mol, PDI=1.5, and exhibiting good syndiotactic selectivity (P _r=0.87). Under the same conditions, a similar structured complex 109(R¹=R²=R’= ^tBu) achieved a monomer conversion rate of 97% within 45 seconds, producing a high molecular weight polymer (M _n =223 kg/mol, PDI =1.34), also with good syndiotactic selectivity (P _r=0.86). The polymerization kinetics of the catalysts indicated that the yttrium complex 118had a faster polymerization rate, with k _obs=7.9×10^-2 s^-1, which was higher than the rate constants of other catalysts 114~117. The polymerization activities of complexes with different amine bridges varied, where the complexes 114, 117with a two-carbon (C₂) amine bridge had significantly higher polymerization rate constants (9.4×10^-3 s^-1, 6.7×10^-3 s^-1) compared to the complexes 115, 116with a three-carbon (C₃) amine bridge (2.0×10^-3 s^-1, 3.7×10^-3 s^-1). X-ray diffraction analysis revealed that when connected by C ₂, the metal center adopted a five-coordinated square pyramidal structure, whereas with C ₃connection, it exhibited a six-coordinated octahedral structure. These different coordination configurations may be the reason for their activity differences. Furthermore, the polymers prepared by these yttrium complexes initiating the ring-opening polymerization of rac-LA were all syndiotactic, with an order of 118 > 117>114>115. The stereoselectivity in this system is related to the bridging structure, suggesting a chain-end control mechanism (stereoregularity controlled by the configuration of the terminal monomer). The choice of solvent significantly affects the polymerization performance; dichloromethane as a solvent leads to easy decomposition of the catalyst, while toluene requires high temperatures for polymerization. Additionally, the syndiotactic selectivity of the polymers obtained using THF as a solvent was generally higher than that of the polymers obtained using toluene as a solvent. This is related to the substituents on the ligands and the THF solvent, possibly due to the acceleration of the coordination/insertion reaction of a certain l-lactide enantiomer after THF coordinates with the metal center.

Williams et al.^[99]used yttrium complex 119to catalyze the ring-opening polymerization of macrolactones (C15-C23).The results showed that this type of yttrium catalyst can highly actively catalyze the ring-opening polymerization of pentadecalactone (PDL), nonadecalactone (NDL), and tricosalactone (TCL) (as shown in Figure 21),where the TOF (Turnover frequency) value for NDL can reach 200~400 h^-1(100°C, [NDL] = 0.3~0.75 mol/L, toluene solution), while the TOF value for TCL (100°C, [TCL] = 0.3 mol/L, toluene) is 34~175 h^-1.The reaction rate is fast, with high monomer conversion rates (NDL: 86%, TCL: 97%), much higher than the activity when using Al-Salen to catalyze their ring-opening polymerization, and the polymerization controllability is good, with the molecular weight of the obtained polymers increasing linearly with the monomer conversion, resulting in polymer molecular weights in the range of 10~60 kg/mol, but with a polydispersity index around 2.0. The reason is that due to the lack of strain or very low strain in the macrolactone monomers, their ring-opening polymerization is an entropy-driven process, so during the polymerization, the chain growth rate and the side reaction rate of transesterification are very close, ultimately leading to a polydispersity index of about 2 for the products of macrolactone ring-opening polymerization.

Spassky et al.^[14] previously reported that Al-Salen has excellent stereoselectivity for the ring-opening polymerization of rac-LA, but its major drawback is low reactivity and the need for a high metal loading. It was later found that using indium, a heavier homolog of aluminum, not only accelerates the reaction rate but also maintains good controllability. Therefore, in the Phosphasalen system, Williams^[31] also developed two indium complexes [In(PsalenR)(OR’)] 120, 121 (Figure 22), which exhibit high activity, high controllability, low dosage, and high isotactic selectivity (P_m=0.92) for rac-LA without the need for additional chiral additives. In improving the catalytic performance of the catalyst, the steric hindrance and electronic effects of the substituents on the ligand backbone have certain influences on the catalytic properties of the complex. Both catalysts induce highly stereoregular polymerization of rac-LA, with complex 120 [In(PsalenPh)(OEt)] achieving a monomer conversion of 96% within 985 minutes at a monomer-to-metal ratio of 500:1 and a lower temperature, showing good isotactic selectivity (P_m = 0.92, 258K). The selectivity of [In(PsalenMe)(O^tBu)] (121) is lower (P_m=0.75), mainly due to the larger steric protection of the bulky isopropyl substituent in the ortho position around the metal center in [In(PsalenPh)(OEt)] (120). The ortho-phenyl substituent is also important in modulating high isotactic selectivity, as these complexes can achieve high stereoselectivity without the need for additional chiral additives or chiral ligands. In 2018, Williams et al.^[88] synthesized indium complexes 122~126 [In(Psalen^tBu)(OX)] with similar structures, which can catalyze the copolymerization of carbon dioxide and epoxides.

Phosphasalen ligands act as both σ donors and π donors, thus the increase in electron cloud density enhances the instability of the metal-alkoxide bond, accelerating the monomer insertion reaction. Compared to Salen-type metal complexes, in addition to improving the polymerization rate, phosphinimine ligands, which have two phosphorus atoms, can greatly influence the steric effects of the active center and its coordination with lactones, thereby imparting stereoselectivity to the monomers. Later, they^[100] synthesized five new non-chiral, meso-metal indium complexes 127~131 (as shown in Figure 22). These complexes exhibit high activity (TOF = 100 h^-1, [rac-LA]₀: [I]₀ = 500:1, THF, 25 ℃) and high isotactic selectivity (P_m > 0.90, 25 ℃) in the ring-opening polymerization of rac-LA. At the same time, these catalysts also show good activity and control over the ring-opening polymerization of ε-CL, β-butyrolactone, ε-decanolactone, and δ-valerolactone (γ-methyl-δ-valerolactone), demonstrating potential for copolymerization.

4.4.2 Phosphasalen-Type Complexes with Different Metal Centers

In 2014, Long and Williams et al.^[101] conducted a systematic study on the effect of metal radius on the selectivity of rac-LA ring-opening polymerization. They synthesized a series of Phosphasalen ligands and the corresponding rare earth complexes 132 (as shown in Figure 23), and found that the lutetium complex [Lu(PsalenN^t^-Bu)(OEt)] exhibited fast polymerization rate, highly controllable polymerization, and high isotactic selectivity (P_m = 0.81~0.84 at 298 K; P_m = 0.89 at 257 K). In contrast, [La(PsalenN^t^-Bu)(O^tBu)] showed moderate syndiotactic selectivity (P_r = 0.74 at 298 K). These results indicate that both the polymerization rate and stereoregularity of the complexes are significantly influenced by the metal radius. In the same ligand structure, the reaction rates of different metal complexes are inversely proportional to the metal radius (La > Y > Lu). Therefore, in the same ligand structure, the central metal is the key factor determining the stereo-selectivity.

They further synthesized and characterized Phosphasalen-type scandium and yttrium metal complexes133 ^[102], and studied their catalytic performance in the ring-opening polymerization of racemic lactide. The results showed that at a monomer-to-metal ratio of 500:1, with THF as the solvent and at 298 K, the scandium complex had no catalytic activity towards lactide, with no monomer conversion after 24 h, even when the polymerization temperature was increased to 70 ℃, there was no change after 24 h. However, under high monomer concentration (8 mol/L) and a monomer-to-metal ratio of 100, after 48 h at 298 K, the monomer conversion rate reached 98%. The theoretical and experimental values of the molecular weight of the obtained polymers were very close, indicating that the scandium complex exhibited good control over the ring-opening polymerization of ε-CL. They further investigated the tracking reaction of [Sc(PsalenN^R)(OR’)] with stoichiometric amounts of lactide, finding that due to the stability of the generated monomeric adduct intermediate, more lactide monomers could not continue to undergo coordination-insertion, thus the scandium complex was inactive for the ring-opening polymerization of lactide. However, the scandium complex could effectively initiate the ring-opening polymerization of ɛ-CL with good controllability, possibly because the intermediate formed by the opened ε-CL and the metal active center was unstable, which facilitated the coordination-insertion of caprolactone molecules. Compared to the scandium complex, the yttrium complex [Y(PsalenN^R)(OR’)] was a highly active catalyst for the ring-opening polymerization of lactide, achieving complete monomer conversion within 0.5 h at a monomer-to-metal ratio of 500, with THF as the solvent and at 298 K. When a methoxy group was introduced at the para position of phenol, its strong electron-donating ability enhanced the electron-donating capacity of the ligand to the metal, thereby reducing the Lewis acidity of the metal and increasing the reactivity of the metal-alkoxide bond, doubling the rate of lactide ring-opening polymerization. However, the scandium complex with the same ligand containing a methoxy group remained inactive. These findings further indicate that for Phosphasalen-type complexes, simple modifications to the ligand backbone have a certain impact on the rate and control of lactide ring-opening polymerization, but the matching between the central metal and the ligand has a greater influence on the ring-opening polymerization.

Using similar ligands, Le Gendre et al^[103] synthesized and characterized a series of Phosphasalen titanium and zirconium metal complexes 134~139, and for the first time reported the stereoselective ring-opening polymerization of lactide with cationic group IV metal complexes (as shown in Figure 24). X-ray diffraction and NMR results indicate that the Phosphasalen ligand in this type of metal complex wraps around the metal ion in a trans or cis-β-configuration, thus proving that compared to the Salen structure, Phosphasalen has more variable configurations and can form out-of-plane coordination with metals. Moreover, due to the strong electron-donating ability of the Phosphasalen ligand, it is conducive to stabilizing electron-deficient titanium and zirconium metal complexes, thereby successfully isolating Phosphasalen zirconium complexes 135~137 in ionic pair form. The catalytic activity of neutral titanium complex 139 and zirconium complexes 134, 138 for the ring-opening polymerization of lactide is poor. In contrast, ionic pair-type zirconium complexes 135~137 exhibit good catalytic activity, achieving a monomer conversion rate of 92% when the molar ratio of monomer to metal is 100:1, yielding isotactic polylactide with controllable molecular weight and narrow molecular weight distribution.

From the above results, it can be seen that in ligands with the same structure, the metal center has a significant impact on the polymerization activity and selectivity.

4.4.3 Metal Complexes of Phosphasalen Derivatives Based on Asymmetric Structures

The above-mentioned ligands in Phosphasalen metal complexes are all symmetric structures. In 2021, Shi et al.^[104] reported the synthesis of a series of novel asymmetric bridged Phosphasalalen ligands and rare earth metal complexes 140~143 (as shown in Figure 25). The structures indicate that 140~142 are dinuclear metal complexes containing two different oxygen-bridging structures (with oxygen atoms from alkoxy and hydroxyl groups as bridges). However, when trimethylsilyloxy is involved, it forms a dinuclear metal complex 143 with two silyloxy bridges. These metal complexes can catalyze the selective ring-opening polymerization of rac-LA and exhibit high catalytic activity; at room temperature, with a [rac-LA]₀/[I]₀ ratio of 1000, the monomer conversion rate reaches 97% within 36 minutes, and they show good syndiotactic selectivity (P_r up to 0.77). Although most metal chlorides in the literature are not active for the ring-opening polymerization of lactones, the yttrium and lutetium complexes 140~142 can effectively promote the ring-opening polymerization of rac-BBL to form syndiotactic polymers (P_r up to 0.73). In the ring-opening polymerization of rac-BBL, the complexes 140~142 show high polymerization activity, possibly due to the chelating Phosphasalalen ligands having better electron-donating ability than their Salen counterparts. Analysis of the end groups of the polymers and polymerization kinetics studies suggest that the dinuclear structure of the alkoxy-hydroxy-bridged metal complexes remains intact during the polymerization process.

5 Conclusions and Future Prospects

Phosphorus and nitrogen are in the same group, having similar electronic layer structures. Due to the greater difference in electronegativity between P and N elements compared to that between C and N elements, metal complexes containing phosphorus-nitrogen bonds exhibit unique properties in coordination structure and catalytic lactide polymerization. Currently, metal complexes based on phosphorus-nitrogen single bonds mainly focus on alkali (earth) metal and rare earth metal complexes, where coordination primarily occurs through the nitrogen atom, with only a small number of ligands' phosphorus atoms participating in coordination, which can improve the activity of catalytic lactide ring-opening polymerization compared to pure nitrogen-containing ligands. Unlike the planar triangular structure of C=N, the phosphinimine (P=N) shows a pyramidal configuration for the phosphorus atom, not only differing in bond length (P=N ~1.58 Å vs C=N ~1.28 Å), but also featuring a highly polarized P=N double bond, which mainly forms complexes with transition metals through the nitrogen atom. Adjusting the substituents at the phosphorus position can control the steric hindrance and electronic effects of the adjacent coordinating nitrogen, effectively regulating catalytic activity and product regularity. Particularly in the multidentate coordinating Phosphasalen structure, replacing the C=N double bond with a P=N double bond enhances catalytic activity and improves the stereoselectivity of the products. Existing results indicate that metal complexes formed by ligands containing asymmetric P=N double bonds can effectively achieve stereoselective polymerization of racemic lactide. The preparation of ligands constructed from P=N bonds will be further developed along with their applications, promoting innovation and diversification in synthetic routes. To achieve rapid development in the biodegradable resin industry, it is not only necessary to reduce the cost of monomer production but also to pay more attention to the performance of the resins formed, making the development of new and efficient catalysts a critical step.

References

Publishing order | Descend order by publishing year | Descend order by cited within

[1]	Kerr R A. Science, 2011, 331(6024): 1510.

[2]	Mecking S. Angew. Chem. Int. Ed., 2004, 43(9): 1078.

[3]	Inkinen S, Hakkarainen M, Albertsson A C, Södergård A. Biomacromolecules, 2011, 12(3): 523.

[4]	Huang Y, Tsai Y H, Hung W C, Lin C S, Wang W, Huang J H, Dutta S, Lin C C. Inorg. Chem., 2010, 49(20): 9416.

[5]	Pilone A, Press K, Goldberg I, Kol M, Mazzeo M, Lamberti M. J. Am. Chem. Soc., 2014, 136(8): 2940.

[6]	Aluthge D C, Yan E X, Ahn J M, Mehrkhodavandi P. Inorg. Chem., 2014, 53(13): 6828.

[7]	Lyubov D M, Tolpygin A O, Trifonov A A. Coord. Chem. Rev., 2019, 392: 83.

[8]	Wu J C, Yu T L, Chen C T, Lin C C. Coord. Chem. Rev., 2006, 250(5-6): 602.

[9]	Radano C P, Baker G L, Smith M R. J. Am. Chem. Soc., 2000, 122(7): 1552.

[10]	Klitzke J S, Roisnel T, Kirillov E, de L Casagrande O, Carpentier J F. Organometallics, 2014, 33(1): 309.

[11]	Marshall E L, Gibson V C, Rzepa H S. J. Am. Chem. Soc., 2005, 127(16): 6048.

[12]	Ryner M, Stridsberg K, Albertsson A C, von Schenck H, Svensson M. Macromolecules, 2001, 34(12): 3877.

[13]	Ovitt T M, Coates G W. J. Am. Chem. Soc., 2002, 124(7): 1316.

[14]	Spassky N, Wisniewski M, Pluta C, Le Borgne A. Macromol. Chem. Phys., 1996, 197(9): 2627.

[15]	Ruan J M, Xiao A G, Wu H W, Yang H L. Des. Monomers Polym., 2014, 17(4): 345.

[16]	Shaw S, White J D. Chem. Rev., 2019, 119(16): 9381.

[17]	Alaaeddine A, Thomas C M, Roisnel T, Carpentier J F. Organometallics, 2009, 28(5): 1469.

[18]	Zhong Z Y, Dijkstra P J, Feijen J. J. Am. Chem. Soc., 2003, 125(37): 11291.

[19]	Thongkham S, Monot J, Martin-Vaca B, Bourissou D. Macromolecules, 2019, 52(21): 8103.

[20]	Stewart J A, McKeown P, Driscoll O J, Mahon M F, Ward B D, Jones M D. Macromolecules, 2019, 52(15): 5977.

[21]	Thomas C M. Chem. Soc. Rev., 2010, 39(1): 165.

[22]	Wei Y, Wang S W, Zhou S L. Dalton Trans., 2016, 45(11): 4471.

[23]	Dijkstra P J, Du H Z, Feijen J. Polym. Chem., 2011, 2(3): 520.

[24]	Consiglio G, Oliveri I P, Failla S, Di Bella S. Molecules, 2019, 24(13): 2514.

[25]	Singha Roy S, Sarkar S, Chakraborty D. Chem. Rec., 2021, 21(8): 1968.

[26]	Cao T P A, Labouille S, Auffrant A, Jean Y, Le Goff X F, Le Floch P. Dalton Trans., 2011, 40(39): 10029.

[27]	Cao T P A, Buchard A, Le Goff X F, Auffrant A, Williams C K. Inorg. Chem., 2012, 51(4): 2157.

[28]	Bakewell C, Cao T P A, Long N, Le Goff X F, Auffrant A, Williams C K. J. Am. Chem. Soc., 2012, 134(51): 20577.

[29]	Marín I M, Auffrant A. Eur. J. Inorg. Chem., 2018, 2018(15): 1634.

[30]	Bakewell C, Cao T P A, Le Goff X F, Long N J, Auffrant A, Williams C K. Organometallics, 2013, 32(5): 1475.

[31]	Myers D, White A J P, Forsyth C M, Bown M, Williams C K. Angew. Chem. Int. Ed., 2017, 56(19): 5277.

[32]	Reddi Y, Cramer C J. ACS Catal., 2021, 11(24): 15244.

[33]	Steiner A, Zacchini S, Richards P I. Coord. Chem. Rev., 2002, 227(2): 193.

[34]	Witt M, Roesky H W. Chem. Rev., 1994, 94(5): 1163.

[35]	Saxon E, Armstrong J I, Bertozzi C R. Org. Lett., 2000, 2(14): 2141.

[36]	Gololobov Y G, Zhmurova I N, Kasukhin L F. Tetrahedron, 1981, 37(3): 437.

[37]	Staudinger H, Meyer J. Helv. Chim. Acta, 1919, 2(1): 635.

[38]	Zhang G Q, Liu T G, Song J Y, Quan Y Y, Jin L, Si M Y, Liao Q. J. Am. Chem. Soc., 2022, 144(6): 2444.

[39]	Peganova T A, Kalsin A M. Synthesis, 2020, 52(3): 433.

[40]	Tolmachev A A, Chaikovskaya A A, Smaliy R V, Kudrya T N, Yurchenko A A, Pinchuk A M. Heteroatom Chem., 1999, 10(4): 343.

[41]	Horner L, Oediger H. Justus Liebigs Ann. Chem., 1959, 627(1): 142.

[42]	Zimmer H, Singh G. J. Org. Chem., 1963, 28(2): 483.

[43]	Boubekeur L, Ricard L, Mézailles N, Le Floch P. Organometallics, 2005, 24(6): 1065.

[44]	Buchard A, Komly B, Auffrant A, Le Goff X F, Le Floch P. Organometallics, 2008, 27(17): 4380.

[45]	Boubekeur L, Ricard L, Mézailles N, Demange M, Auffrant A, Le Floch P. Organometallics, 2006, 25(12): 3091.

[46]	Boubekeur L, Ulmer S, Ricard L, Mézailles N, Le Floch P. Organometallics, 2006, 25(2): 315.

[47]	Sauthier M, Forniés-Cámer J, Toupet L, Réau R. Organometallics, 2000, 19(4): 553.

[48]	García-Álvarez J, García-Garrido S E, Cadierno V. J. Organomet. Chem., 2014, 751: 792.

[49]	Kasani A, McDonald R, Cavell R G. Organometallics, 1999, 18(19): 3775.

[50]	Recknagel A, Witt M, Edelmann F T. J. Organomet. Chem., 1989, 371(3):C40.

[51]	Tannoux T, Mazaud L, Cheisson T, Casaretto N, Auffrant A. Dalton Trans., 2023, 52(34): 12010.

[52]	Knorr M, Strohmann C. Organometallics, 1999, 18(2): 248.

[53]	Dehnicke K, Krieger M, Massa W. Coord. Chem. Rev., 1999, 182(1): 19.

[54]	Dehnicke K, Weller F. Coord. Chem. Rev., 1997, 158: 103.

[55]	Panda T K, Roesky P W. Chem. Soc. Rev., 2009, 38(9): 2782.

[56]	Kamalesh Babu R P, Cavell R G, McDonald R. Chem. Commun., 2000(6): 481.

[57]	Aparna K, Kamalesh Babu R P, McDonald R, Cavell R G. Angew. Chem., 2001, 113(23): 4535.

[58]	Reißmann U, Poremba P, Noltemeyer M, Schmidt H G, Edelmann F T. Inorg. Chim. Acta, 2000, 303(2): 156.

[59]	Fenske D, Maczek B, Maczek K. Z. Anorg. Allg. Chem., 1997, 623(7): 1113.

[60]	Kühl O, Koch T, Somoza F B, Junk P C, Hey-Hawkins E, Plat D, Eisen M S. J. Organomet. Chem., 2000, 604(1): 116.

[61]	Wetzel T G, Dehnen S, Roesky P W. Angew. Chem. Int. Ed., 1999, 38(8): 1086.

[62]	Naktode K, Bhattacharjee J, Chakrabarti A, Panda T K. J. Chem. Sci., 2015, 127(2): 265.

[63]	Bhattacharjee J, Das S, Reddy T D N, Nayek H P, Mallik B S, Panda T K. Z. Anorg. Allg. Chem., 2016, 642(2): 118.

[64]	Naktode K, Kottalanka R K, Jana S K, Panda T K. Z. Anorg. Allg. Chem., 2013, 639(6): 999.

[65]	Harinath A, Bhattacharjee J, Sarkar A, Nayek H P, Panda T K. Inorg. Chem., 2018, 57(5): 2503.

[66]	Kottalanka R K, Harinath A, Panda T K. RSC Adv., 2015, 5(47): 37755.

[67]	Bhattacharjee J, Harinath A, Nayek H P, Sarkar A, Panda T K. Chem., 2017, 23(39): 9319.

[68]	Bhattacharyya P, Woollins J D. Polyhedron, 1995, 14(23-24): 3367.

[69]	Puddephatt R J. Chem. Soc. Rev., 1983, 12(2): 99.

[70]	Knorr M, Hallauer E, Huch V, Veith M, Braunstein P. Organometallics, 1996, 15(18): 3868.

[71]	Knorr M, Braunstein P, Tiripicchio A, Ugozzoli F. J. Organomet. Chem., 1996, 526(1): 105.

[72]	Mague J T. J. Clust. Sci., 1995, 6(2): 217.

[73]	Ogawa T, Kajita Y, Wasada-Tsutsui Y, Wasada H, Masuda H. Inorg. Chem., 2013, 52(1): 182.

[74]	Roesky P W, Gamer M T, Puchner M, Greiner A. Chem. Eur. J., 2002, 8(22): 5265.

[75]	Platel R H, Hodgson L M, White A J P, Williams C K. Organometallics, 2007, 26(20): 4955.

[76]	Platel R H, White A J P, Williams C K. Inorg. Chem., 2011, 50(16): 7718.

[77]	Platel R H, White A J P, Williams C K. Chem. Commun., 2009(27): 4115.

[78]	Kember M R, Copley J, Buchard A, Williams C K. Polym. Chem., 2012, 3(5): 1196.

[79]	Hodgson L M, Platel R H, White A J P, Williams C K. Macromolecules, 2008, 41(22): 8603.

[80]	Liu B, Cui D M, Ma J, Chen X S, Jing X B. Chem.Eur.J., 2007, 13(3): 834.

[81]	Gamer M T, Rastätter M, Roesky P W, Steffens A, Glanz M. Chem., 2005, 11(10): 3165.

[82]	Liang L C, Tsai T L, Li C W, Hsu Y L, Lee T Y. Eur. J. Inorg. Chem., 2011, 2011(19): 2948.

[83]	Wang Z X, Qi C Y. Organometallics, 2007, 26(9): 2243.

[84]	Ma W A, Wang Z X. Organometallics, 2011, 30(16): 4364.

[85]	Ma W A, Wang Z X. Dalton Trans., 2011, 40: 1778.

[86]	Buchard A, Platel R H, Auffrant A, Le Goff X F, Le Floch P, Williams C K. Organometallics, 2010, 29(13): 2892.

[87]	Diaz C, Fu J, Soobrattee S, Cao L R, Nyamayaro K, Goonesinghe C, Patrick B O, Mehrkhodavandi P. Inorg. Chem., 2022, 61(8): 3763.

[88]	Thevenon A, Cyriac A, Myers D, White A J P, Durr C B, Williams C K. J. Am. Chem. Soc., 2018, 140(22): 6893.

[89]	Wheaton C A, Ireland B J, Hayes P G. Organometallics, 2009, 28(5): 1282.

[90]	Sun H S, Ritch J S, Hayes P G. Inorg. Chem., 2011, 50(17): 8063.

[91]	Wheaton C A, Hayes P G. Chem. Commun., 2010, 46(44): 8404.

[92]	Wheaton C A, Hayes P G. Dalton Trans., 2010, 39(16): 3861.

[93]	Ireland B J, Wheaton C A, Hayes P G. Organometallics, 2010, 29(5): 1079.

[94]	Sun H S, Ritch J S, Hayes P G. Dalton Trans., 2012, 41(13): 3701.

[95]	Chisholm D T, Hayes P G. New J. Chem., 2021, 45(33): 15043.

[96]	Mou Z H, Liu B, Wang M Y, Xie H Y, Li P, Li L, Li S H, Cui D M. Chem. Commun., 2014, 50(77): 11411.

[97]	Broderick E M, Guo N, Vogel C S, Xu C L, Sutter J, Miller J T, Meyer K, Mehrkhodavandi P, Diaconescu P L. J. Am. Chem. Soc., 2011, 133(24): 9278.

[98]	Broderick E M, Guo N, Wu T P, Vogel C S, Xu C L, Sutter J, Miller J T, Meyer K, Cantat T, Diaconescu P L. Chem. Commun., 2011, 47(35): 9897.

[99]	Myers D, Witt T, Cyriac A, Bown M, Mecking S, Williams C K. Polym. Chem., 2017, 8(37): 5780.

[100]

Yuntawattana N, McGuire T M, Durr C B, Buchard A, Williams C K. Catal. Sci. Technol., 2020, 10(21): 7226.

[101]

Bakewell C, White A J P, Long N J, Williams C K. Angew. Chem. Int. Ed., 2014, 53(35): 9226.

[102]

Bakewell C, White A J P, Long N J, Williams C K. Inorg. Chem., 2015, 54(5): 2204.

[103]

Normand A T, Malacea-Kabbara R, Lapenta R, Dajnak A, Richard P, Cattey H, Bolley A, Grassi A, Milione S, Auffrant A, Dagorne S, Le Gendre P. Dalton Trans., 2020, 49(21): 6989.