Pd(II)-catalyzed ethylene/methyl acrylate copolymerization: toward catalyst recovery and recycling

Abstract

The paradigm shift from a linear to a circular economy is a very important goal, which involves several fields of chemistry, including the recycling of plastic materials and of critical raw materials, including platinum-group metal catalysts. The introduction of polar functional groups into the polyolefin skeleton, to yield functionalized polyolefins (FPOs), might open new and easier pathways for plastic depolymerization processes compared to simple polyolefins. However, the controlled copolymerization of ethylene with polar vinyl monomers via homogeneous catalysis to produce FPOs is a highly challenging reaction; at present, Pd–α-diimine complexes are the most promising catalysts for this reaction. In this contribution, we have revisited the catalytic behaviors of three known palladium complexes with benchmark α-diimines (N–N), with the general formula [Pd(N–N)(Me)(NCMe)][PF₆], by carrying out, for the first time, the copolymerization of ethylene with methyl acrylate in trifluoroethanol/dichloromethane mixtures of different compositions. We found that the solvent composition had an unprecedented effect on catalyst productivity, the content of inserted polar monomers and the mode of incorporation. Detailed NMR investigations in CD₂Cl₂/TFE-d₃ mixtures of the reaction of one of the Pd-complexes with methyl acrylate allowed us to correlate the observed catalytic behavior with the organometallic intermediates present in solution. Moreover, at specific solvent compositions, the spontaneous formation of two phases occurred at the end of the catalytic runs. We exploited this phenomenon to successfully perform, for the first time in this field, the recovery and recycling of the catalyst.

---

1. Introduction

The major class of macromolecules that currently accounts for the highest percentage of the world's thermoplastic material production consists of polyolefins such as polyethylene and polypropylene.¹ This is due to the excellent properties that make them suitable for a wide range of applications, from food packaging to the production of medical devices, automotive parts, and insulation materials. However, the chemical inertness of polyolefins, one of the main desirable properties of thermoplastic products, becomes a major drawback when considering their end-of-life fate.² Functionalized polyolefins (FPOs) are a class of promising macromolecules that can offer a valid and more sustainable alternative to simple polyolefins, both in terms of material properties^3,4 and depolymerization processes.^5–8 The most powerful and environmentally friendly approach for obtaining FPOs is through the controlled homogeneously catalysed copolymerization of ethylene with polar vinyl monomers. Since the 1990s, following Brookhart's breakthrough in the field,⁹ a wide range of palladium complexes capable of catalysing the copolymerization of ethylene (E) with methyl acrylate (MA), considered a model reaction, has been reported.^10,11 Most investigated complexes contain either α-diimines (L1, Chart 1)¹² or phosphino–sulfonate ligands (L2, Chart 1),¹³ but other molecules such as phosphino–phosphonates (L3, Chart 1),^14,15 phosphine–phosphonic amides (L4, Chart 1),^16,17 phosphine–benzenamines (L5, Chart 1),¹⁸ N-heterocyclic carbenes (NHCs) (L6, Chart 1),¹⁹ α-diamines (L7, Chart 1),²⁰ and pyridyl–pyridylidene amides (L8, Chart 1) have been investigated as ancillary ligands.²¹

Chart 1 Examples of reported ancillary ligands.

Even though Pd(II) complexes with α-diimine ligands are considered as the most promising catalysts for E/MA copolymerization, they have two main drawbacks for potential industrial applications: low catalyst productivity, combined with the high cost of palladium metal; and the formation of branched macromolecules with the polar monomers almost exclusively located at the ends of the branches (T(MA) in Scheme 1), due to the chain-walking processes that take place after the insertion of both MA and ethylene.²² Indeed, a desirable FPO should consist of linear macromolecules with the polar monomer inserted into the main chain (M(MA) in Scheme 1). Therefore, control of the macromolecule microstructure is highly relevant, and it has been preferentially addressed by varying the steric hindrance of the α-diimine ligands. The best results were obtained with α-diimines bearing highly sterically hindered substituents such as the double-decker α-diimine L9, L10 – which bears benzhydryl groups – and L11 with a benzobarrelene-derived backbone (Chart 1). For instance, a dinuclear Pd catalyst based on L9 (Chart 1) led to E/MA copolymers with an M(MA) : T(MA) ratio of 85 : 15,²³ whereas 98% M(MA) was achieved by using highly encumbered L11 (Chart 1); nevertheless, in the latter case, the product was a hyperbranched copolymer.²⁴ Moreover, Pd-catalysts with α-diimines bearing either benzhydryl groups (L10, Chart 1)^25–31 or benzobarrelene-derived backbones (L11, Chart 1)³² have also been extensively studied to afford catalysts with increased thermal stability compared to those with less hindered α-diimines.

Scheme 1 The microstructure of ethylene/methyl acrylate copolymers: M(MA) = methyl acrylate into the main chain (M = main); and T(MA) = methyl acrylate at the ends of the branches (T = terminal).

The introduction of cyclopentyl or cyclohexyl substituents on the aryl rings of α-diimines with either rigid dibenzobarrelene or more-flexible cyclohexyl backbones was applied for ethylene and propene copolymerization with MA (L12–L14, Chart 1). Despite the increase in steric hindrance around palladium, these catalysts led to hyperbranched macromolecules with the polar monomers located at the ends of the long branches.^33–35

As an alternative approach to ligand modification, we recently discovered that the solvent used for E/MA copolymerization plays a fundamental role in controlling the mode of methyl acrylate incorporation.³⁶ We found that when E/MA copolymerization is carried out in dichloromethane using [Pd(1)(Me)(NCMe)][PF₆] (Pd1b, Scheme 2) as a precatalyst, the polar monomers are inserted almost exclusively at the ends of the branches, in agreement with the literature;⁹ however, when 2,2,2-trifluoroethanol (TFE) is the solvent, acrylate is inserted both in the main chain and at the ends of the branches with M(MA) : T(MA) = 40 : 60. Thorough room-temperature NMR studies of the reaction between Pd1b and both comonomers—ethylene and MA—revealed different catalysts resting states in the two solvents. In CD₂Cl₂, the expected six-membered palladacycle MC6 is observed. In contrast, in TFE-d₃, the catalyst resting state is the open-chain intermediate OC, in which both acetonitrile and the organic moiety formed by the migratory insertion of MA into the Pd–Me bond are coordinated to palladium (Scheme 3).³⁶ When the growth of the polymeric chain occurs on OC, the acrylate is trapped as part of the main chain, whereas MC6 is responsible for the enchainment of the polar monomers at the ends of the branches.²²

Scheme 2 Investigated complexes and their synthesis.

Based on these results, we re-investigated the catalytic behaviors during E/MA copolymerization of three Pd-complexes, namely [Pd(1)(Me)(NCMe)][PF₆] (Pd1b), [Pd(2)(Me)(NCMe)][PF₆] (Pd2b), and [Pd(3)(Me)(NCMe)][PF₆] (Pd3b), that bear the benchmark α-diimines 1–3 (Scheme 2). The reactions were performed in trifluoroethanol/dichloromethane mixtures of different compositions. The use of mixtures of solvents for this reaction is unprecedented in the case of Pd catalysts. Only very recently has one example of an E/MA copolymerization reaction carried out in toluene/diethyl ether mixtures, using a neutral Ni complex (Ni1, Chart 2) as precatalyst, been reported.⁸ The precatalyst is activated in situ by the simultaneous addition of MBArF and M′BArF in 1 : 1 and 2 : 1 ratios with respect to the nickel catalyst (M and M′ = alkali metal cations; BArF = tetrakis-((3,5-trifluoromethyl)phenyl)borate). The nature of the obtained copolymer, in terms of the amount of incorporated MA and the molecular weight distribution, was found to depend on M and M′ and on the solvent composition.

Scheme 3 The proposed mechanism for ethylene/methyl acrylate copolymerization in trifluoroethanol (GP = growing polymer chain).

Our current investigation reveals that the composition of the reaction medium remarkably affects, in an unpredictable way, catalyst productivity and copolymer features, such as molecular weight, the amount of inserted MA, and the mode of enchainment. Moreover, we found that for specific solvent compositions, the spontaneous formation of two phases occurs at the end of the catalytic runs, affording partial catalyst recovery and recycling, two processes that are very difficult to implement in coordination–insertion polymerizations. To the best of our knowledge, the only reported example of a catalyst recycling protocol concerns a titanocene dichloride compound (Ti1, Chart 2) used for ethylene homopolymerization. An appropriate functionalization of one cyclopentadienyl ligand allowed the separation of the complex from the polymer by simple filtration at the end of the catalytic run and its reuse in a second run of polymerization, thus demonstrating the feasibility of catalyst recycling.³⁷

Chart 2 The nickel and titanium complexes mentioned herein.

2. Materials and methods

2.1 Experimental

Anhydrous dichloromethane was obtained by freshly distilling it over CaH₂ under an argon or nitrogen atmosphere. Deuterated solvents (Cambridge Isotope Laboratories, Inc. (CIL) and Merck) were stored as recommended by the manufacturers. Ethylene (purity ≥ 99.9%), supplied by SIAD, was used as received. Methyl acrylate, trifluoroethanol and all other reagents and solvents were purchased from Merck and used without further purification for synthetic, spectroscopic and catalytic purposes. [Pd(cod)Cl(Me)] was synthesized according to a literature procedure.³⁸ [Pd(OAc)₂] was purchased from Johnson Matthey. NMR spectra of ligands, complexes and catalytic products were recorded on a Varian 500 spectrometer at 500 MHz (¹H) and 125.68 MHz (¹³C) or on a Varian 400 spectrometer at 400 MHz (¹H) and 100 MHz (¹³C). In situ NMR experiments to investigate the reactions between Pd1b and methyl acrylate in different solvent mixtures were performed on a Bruker Avance NEO 600/54 Onebay spectrometer at 600 MHz (¹H) and 150 MHz (¹³C). The resonances are reported in ppm (δ) and referenced to the solvent residual peak versus Si(CH₃)₄: CDCl₃ at δ 7.26 (¹H) and δ 77.0 (¹³C); CD₂Cl₂ at δ 5.32 (¹H) and δ 54.0 (¹³C). NMR experiments were performed employing the automatic software parameters. In the case of NOESY experiments, a mixing time of 600 ms was used. The average molecular weights (M_n and M_w) and polydispersity values (M_w/M_n) of copolymer samples were measured with a GPC-IR apparatus (PolymerChar). All the measurements were carried out at 150 °C in 1,2,4-trichlorobenzene (TCB), the solution concentration was 1.5 mg mL⁻¹ (at 150 °C), and 0.3 g L⁻¹ 2,6-di-tert-butyl-p-cresol was added to prevent degradation. A set of four PLgel Olexis mixed-bed columns (Polymer Laboratories) and an IR5 infrared detector (PolymerChar) were used. The dimensions of the columns were 300 × 7.5 mm, and their particle size was 13 μm. The mobile phase flow rate was kept at 1.0 mL min⁻¹. For GPC calculations, a universal calibration curve was obtained using 12 polystyrene (PS) standard samples supplied by PolymerChar (peak molecular weights ranging from 266 to 1 220 000 g mol⁻¹).

2.1.1 Synthesis of ligands 1–3 and palladium complexes Pd1a–Pd3a and Pd1b–Pd3b. Ligands 1–3 were synthesized by a slight modification of literature procedures.^39,40

The synthesis of all complexes was performed according to the literature,^40–42 using standard Schlenk techniques, under an argon atmosphere, at room temperature.

General procedure to obtain the neutral complexes [Pd(Me)Cl(N–N)] Pd1a–Pd3a. To a stirred solution of [Pd(cod)Cl(Me)] (1.37 mmol, 363.2 mg) in anhydrous dichloromethane (1 mL), under an Ar atmosphere, a solution of α-diimine (1.1 equiv., 1.51 mmol) in dichloromethane (2 mL) was added. After 4 h at room temperature, if the product did not spontaneously precipitate, the reaction mixture was concentrated under reduced pressure to ca. 1 mL, and then n-hexane was added to favour precipitation. The solid was filtered, washed with n-hexane and dried under vacuum.

Pd1a (yellow solid, yield = 80%) ¹H NMR (500 MHz, CD₂Cl₂, 298 K): δ = 7.15 (m, 6H, CH^m and CH^p), 2.25 (s, 6H, Me^Ar), 2.22 (s, 6H, Me^Ar), 2.00 (s, 3H, Me^DAB), 1.96 (s, 3H, Me^DAB), 0.26 (s, 3H, Pd-Me).

Pd2a (orange solid, yield = 87%) ¹H NMR (500 MHz, CD₂Cl₂, 298 K): δ = 7.28 (m, 6H, CH^m and CH^p), 3.06 (sept, 2H, CH^iPr), 3.00 (sept, 2H, CH^iPr), 2.04 (s, 3H, Me^DAB), 2.02 (s, 3H, Me^DAB), 1.40 (d, 6H, Me^iPr), 1.35 (d, 6H, Me^iPr), 1.18 (m, 12H, Me^iPr), 0.37 (s, 3H, Pd–Me).

Pd3a (yellow solid, yield = 73%) ¹H NMR (500 MHz, CD₂Cl₂, 298 K): δ = 6.98 (d, 4H, CH^Ar), 2.33 (s, 6H, Me^p), 2.17 (s, 12H, Me^o), 1.97 (d, 6H, Me^DAB), 0.25 (s, 3H, Pd–Me). ¹³C NMR (125 MHz, CD₂Cl₂, 298 K, derived from the ¹H,¹³C HSQC NMR spectrum): δ = 129.10 (CH^Ar), 128.65 (CH^Ar), 20.74 (Me^p), 19.42 (Me^DAB), 18.60 (Me^DAB), 17.80 (Me^o), 0.44 (Pd–Me).

General procedure to obtain the cationic complexes [Pd(Me)(NCMe)(N–N)][PF₆] Pd1b–Pd3b. To a stirred solution of a neutral derivative, Pd1a–Pd3a (0.530 mmol, 238.1 mg for Pd1a; 0.490 mmol, 273.4 mg for Pd2a; 0.210 mmol, 100 mg for Pd3a), in anhydrous dichloromethane (6 mL for Pd1a; 3 mL for Pd2a; 5 mL for Pd3a), a solution of AgPF₆ (1.2 equiv.) in anhydrous acetonitrile was added. The reaction was stirred for 2 h in the dark, then filtered over Celite. The solution was then concentrated under reduced pressure. The addition of n-hexane resulted in the precipitation of the complex, which was recovered by filtration and dried under vacuum.

Pd1b (yellow solid, yield = 92%) ¹H NMR (500 MHz, CD₂Cl₂, 298 K): δ = 7.22 (m, 6H, CH^m and CH^p), 2.28 (s, 6H, Me^Ar), 2.21 (s, 3H, Me^DAB), 2.20 (s, 6H, Me^Ar), 2.19 (s, 3H, Me^DAB), 1.82 (s, 3H, Pd-NCMe), 0.34 (s, 3H, Pd–Me).

Pd2b (yellow solid, yield = 87%) ¹H NMR (500 MHz, CD₂Cl₂, 298 K): δ = 7.36 (m, 6H, CH^m and CH^p), 2.93 (sept, 2H, CH^iPr), 2.89 (sept, 2H, CH^iPr), 2.27 (s, 3H, Me^DAB), 2.27 (s, 3H, Me^DAB), 1.82 (s, 3H, Pd–NCMe), 1.39 (d, 6H, Me^iPr), 1.34 (d, 6H, Me^iPr), 1.27 (d, 12H, Me^iPr), 1.23 (d, 12H, Me^iPr), 0.48 (s, 3H, Pd-Me). ¹³C NMR (125 MHz, CD₂Cl₂, 298 K, derived from the ¹H,¹³C HSQC NMR spectrum): δ = 128.62 (CH^m), 124.52 (CH^Ar), 29.22 (CH^iPr) 23.79 (Me^iPr), 23.69 (Me^iPr), 23.47 (Me^iPr), 23.03 (Me^iPr), 21.65 (Me^DAB), 20.20 (Me^DAB), 6.60 (Pd–Me), 2.32 (Pd–NCMe).

Pd3b (yellow solid, yield = 93%) ¹H NMR (500 MHz, CD₂Cl₂, 298 K): δ = 7.02 (d, 4H, CH^Ar), 2.33 (d, 6H, Me^p), 2.23 (s, 6H, Me^o), 2.19 (s, 3H, Me^DAB), 2.17 (s, 3H, Me^DAB), 2.15 (s, 6H, Me^o), 1.84 (s, 3H, Pd–NCMe), 0.34 (s, 3H, Pd–Me). ¹³C NMR (125 MHz, CD₂Cl₂, 298 K, derived from the ¹H,¹³C HSQC NMR spectrum): δ = 129.60 (CH^Ar), 129.43 (CH^Ar), 20.87 (Me^p), 21.11 (Me^DAB), 19.10 (Me^DAB), 17.92 (Me^o), 17.77 (Me^o), 4.42 (Pd-Me), 2.40 (Pd–NCMe).

2.1.2 Ethylene/methyl acrylate copolymerization reaction – standard procedure. The catalytic experiments were performed in a Büchi “tinyclave” reactor equipped with an interchangeable 50 mL glass vessel. The vessel was loaded with the Pd(II) complex (21 μmol), the solvent (TFE, distilled DCM, or a mixture of the two solvents; total volume: 21 mL) and methyl acrylate (2.26 mL). The reactor was placed in a preheated oil bath at T = 308 K and connected to the ethylene tank. Ethylene was bubbled for 10 minutes. Then, the reactor was sealed and pressurized with ethylene to the desired pressure. The reaction mixture was stirred at constant temperature and pressure. After the correct amount of time, the reactor was cooled to room temperature and vented. The reaction mixture was poured into a 50 mL round-bottom flask, and the volatiles were evaporated under reduced pressure. The flask containing the copolymer was then placed in a desiccator connected to a vacuum pump until constant weight. The copolymer, obtained as a viscous oil, was then analysed by NMR spectroscopy and GPC.

A few selected experiments were performed in duplicate, showing excellent reproducibility.

2.1.3 General procedure for recycling experiments.
2.1.3a A first protocol for catalyst recovery and recycling was defined for the catalytic tests performed at χ_TFE = 0.17 for Pd1b and Pd3b and at χ_TFE = 0.09 for Pd2b. The catalytic test in run 1 was performed in the appropriate solvent mixture following the procedure described above. At the end of the run, after venting the reactor, the reaction mixture was poured into a plastic centrifuge tube. The mixture was centrifuged at 5000 revolutions per minute (rpm) for 5 minutes, and the two phases were separately collected using a Pasteur pipette. The upper phase containing the copolymer was transferred to a round-bottom flask, and the residual volatile fraction was removed under reduced pressure. The copolymer was dried under vacuum to constant weight.

The lower phase containing the catalyst was stored overnight at 253 K. The following day, it was transferred into the glass vessel of the “tinyclave” reactor, and both methyl acrylate and a mixture of trifluoroethanol and dichloromethane with the same composition as used in run 1 were added. The amount of MA to be added was calculated on the basis of the amount of MA incorporated in the copolymer obtained in the experiment performed under the same conditions but with no catalyst recovery. The amount of the solvent mixture added corresponded to that required to restore a total volume equal to that of run 1. From this point on, the procedure was the same as for run 1.

At the end of run 2, the reaction mixture was treated following the protocol reported above, and a third catalytic run was performed.

At the end of run 3, the phase separation was too small to allow for efficient separation; thus, the standard work-up procedure was applied to the whole reaction mixture.

This protocol was tested twice, obtaining data in very good agreement.

2.1.3b A second protocol for catalyst recovery and recycling was defined for the catalytic tests performed in neat TFE with Pd1b. The catalytic test in run 1 was performed following the standard procedure reported above. At the end of the run, after venting the reactor, 30 mL of dichloromethane was added to the reaction mixture that was left standing for 30 min. Then, this was transferred to a centrifuge tube and centrifuged at 5000 rounds per minute for 5 minutes. The two phases formed were separately collected with a Pasteur pipette. The upper phase was treated as reported in the procedure in 2.1.3a. The lower phase was concentrated under reduced pressure at room temperature to a total volume of approximately 15 mL, to remove dichloromethane, and stored overnight at 253 K. The following day, it was transferred to the glass vessel of the “tinyclave” reactor, and both methyl acrylate and trifluoroethanol were added in agreement with the protocol in procedure 2.1.3a. From this point on, the procedure was the same as for run 1.

At the end of run 2, the reaction mixture was treated following the same protocol as reported above, and a third catalytic run was performed.

At the end of run 3, the standard workup was applied to the whole reaction mixture.

2.1.4 General procedure for the in situ NMR investigation of the reactivity between Pd1b and methyl acrylate in different solvent mixtures. A 10 mM solution of Pd1b in a CD₂Cl₂/TFE-d₃ mixture with the composition χ_TFE = 0.17 was prepared, and a first ¹H NMR spectrum was recorded. After the addition of 2 equivalents of methyl acrylate, the reaction was monitored by NMR spectroscopy at T = 298 K until the disappearance of the signals of the Pd complex. The same study was performed on the CD₂Cl₂/TFE-d₃ mixture with the composition χ_TFE = 0.79.

2D NMR spectra were recorded after all Pd1b had reacted, that is at t = 30 min for χ_TFE = 0.17 and at t = 5 min for χ_TFE = 0.79.

The proton assignments for the detected species, reported in the SI, were based on 1D and 2D NMR spectra (Fig. S31–S42).

3. Results and discussion

3.1 Synthesis and characterization of palladium complexes

The α-diimine ligands 1–3 (ref. 39 and 40) and the corresponding neutral Pd-complexes [Pd(Me)Cl(N–N)] (Pd1a–Pd3a; N–N = 1–3; Scheme 2) were obtained according to published procedures,^40–42 and their NMR characterization data in solution were consistent with literature data (Fig. S1–S8).^40–42

The corresponding cationic derivatives [Pd(Me)(NCMe)(N–N)][PF₆] (Pd1b–Pd3b, Scheme 2) were obtained from Pd1a–Pd3a through a dehalogenation reaction with AgPF₆ in the presence of acetonitrile (Scheme 2). The complexes were isolated as yellow solids in excellent yields. The solution NMR characterization data of Pd1b were in agreement with the literature (Fig. S9).⁴¹ Pd2b and Pd3b were characterized by 1D and 2D NMR spectroscopy in CD₂Cl₂ solution, and their spectral features were totally consistent with the expected structures (Fig. S10–S15). The choice of acetonitrile was dictated by our recent results showing that this ligand has a relevant role in determining the microstructures of the produced macromolecules.²¹

3.2 Ethylene/methyl acrylate copolymerization experiments

The cationic complexes Pd1b–Pd3b were tested as catalysts for the target ethylene/methyl acrylate copolymerization using dichloromethane/trifluoroethanol mixtures of different compositions as the reaction medium, ranging from neat DCM to neat TFE with a constant total volume (21 mL). Solvent mixtures with 10 different compositions were tested for both Pd1b and Pd2b, and 5 examples were tested for Pd3b. All the other reaction parameters were kept constant in each catalytic run, including the amount of catalyst (21 μmol), ethylene pressure (5 bar), [MA]/[Pd] ratio (1188), temperature (308 K) and reaction time (6 h) (Fig. 1, Tables S1–S3). At the end of each catalytic run, the obtained copolymer was characterized by NMR spectroscopy in solution and via gel permeation chromatography.†

Fig. 1 Ethylene/methyl acrylate copolymerization: the effect of the reaction medium composition on: a) catalyst productivity in kg CP per mol Pd (kilograms of copolymer per mol of palladium) and b) the amount of inserted MA in mol% calculated from ¹H NMR spectra of isolated products. Reaction conditions: n_Pd = 2.1 × 10⁻⁵ mol, V_sol = 21 mL, V_MA = 2.26 mL, [MA]/[Pd] = 1188, T = 308 K, P_E = 5 bar, t = 6 h.

The catalytic behaviors of Pd1b in neat DCM and in neat TFE have already been reported by us,³⁶ whereas its catalytic performances in the solvent mixtures together with the study of Pd2b and Pd3b are presented here for the first time. In addition, the complexes [Pd(N–N)(Me)(OEt₂)][BArF] (N–N = 1, 2), differing from Pd1b and Pd2b in terms of the labile monodentate ligand (diethyl ether vs. acetonitrile) and the counterion (BArF vs. PF₆), were investigated in neat dichloromethane only, and – even if the reaction conditions are not the same – they allow for useful comparison.⁹

In all the tested DCM/TFE mixtures, the three Pd complexes showed catalytic activity for the synthesis of E/MA copolymers. In neat solvents, each complex had similar catalytic performance in terms of productivity (with the exception of Pd2b, Fig. 1a) and the amount of inserted MA (Fig. 1b). In each case, the productivity showed a sawtooth trend with respect to the molar fraction of trifluoroethanol (χ_TFE) in the solvent mixture, and two maxima were reached at two different compositions (Fig. 1a). For catalysts Pd1b and Pd3b in all the solvent mixtures, the productivity was higher than in the neat solvents, and two maxima of similar intensity were achieved at χ_TFE = 0.17 and 0.47 (Fig. 1a, Tables S1 and S2). The behavior of Pd2b was similar but with two remarkable differences: 1) the first maximum of productivity was found at χ_TFE = 0.09; and 2) for (ca.) 0.15 < χ_TFE < 0.3, the productivity was lower than in the neat solvents. At χ_TFE = 0.47, Pd2b showed the highest productivity, corresponding to 262.3 kg CP per mol Pd (Fig. 1a, Table S3), almost twice as high as those of Pd1b and Pd3b and of that found for the catalyst obtained from [Pd(2)(Me)(OEt₂)][BArF], even though slightly different reaction conditions were used (solvent: neat DCM; 1 × 10⁻⁴ mol of catalyst; 6 bar of ethylene; total volume of solvent and MA: 100 mL; 18.5 h; and [MA]/[Pd] = 5800).⁹

For all three catalysts, the formation of inactive palladium metal was observed at the end of the copolymerization experiments (6 h) carried out in neat DCM and in the solvent mixtures characterized by χ_TFE ≤ 0.17, whereas for higher concentrations of TFE, no catalyst decomposition was observed. Even though the stabilizing effect of TFE is known from the literature,^41,43,44 our findings indicate that even relatively small amounts of this solvent are sufficient for observing this effect.

The trend relating to the content of inserted methyl acrylate inversely mirrored that of productivity (Fig. 1b), that is – consistent with the literature data for copolymerizations carried out in neat solvents^22,36,45 – the highest amounts of inserted MA were found in the macromolecules obtained in the lowest yields. For Pd1b and Pd3b, in all solvent mixtures, the amount of MA was lower than in neat solvents. The copolymers synthesized with Pd2b have the lowest amount of inserted MA for all solvent compositions, due to the steric hindrance of the iso-propyl substituents in 2 (Fig. 1b; Tables S1–S3).

The concentration of TFE in the reaction mixture also affected the molecular weight of the isolated macromolecules. All three catalysts share the same behavior: macromolecules with the shortest chains were produced in neat dichloromethane, whereas copolymers with the highest molecular weights were obtained in neat trifluoroethanol (Fig. 2, Tables S1–S3). The increase in molecular weight with increasing χ_TFE was not linear but followed a two-step trend, increasing for 0 < χ_TFE < 0.17 and then again for χ_TFE > 0.47. For all the macromolecules, monomodal GPC curves were obtained with polydispersity values typical of single-site catalysts (Fig. S17–S19, Tables S1–S3).

Fig. 2 Ethylene/methyl acrylate copolymerization: the effect of the reaction medium composition on M_n. Reaction conditions: see Fig. 1.

A remarkable dependence on the nature of the α-diimine ligand was also observed: moving from neat DCM to neat TFE, the molecular weight of the copolymer increased ca. 2-fold for Pd3b, ca. 5-fold for Pd1b, and as much as 35-fold for Pd2b, reaching a value of 101 kDa. The positive effect of iso-propyl substituents on the molecular weights of copolymers had already been observed with similar Pd complexes – though with less spectacular results – both in neat DCM^9,34 and in neat TFE.⁴⁵ In the current case, moving from Pd1b to Pd2b, M_n remained almost unaffected when the copolymerizations were carried out in dichloromethane (5.0 and 2.9 kDa, respectively), whereas it rose from 23.4 to 101.2 kDa for the copolymers obtained in trifluoroethanol. This enhancement indicates that a combination of iso-propyl substituents on the ligand with trifluoroethanol as a reaction medium results in a remarkable positive effect in terms of slowing down the chain transfer reaction.

In addition to the molecular weight, the concentration of TFE in the reaction medium affected the T(MA)/M(MA) ratio (Fig. 3, Tables S1–S3), which was determined from the ¹³C NMR spectra of the isolated copolymers recorded in CDCl₃ solution. On the basis of literature data,^36,45 the signals of the carbonyl and the methoxy moieties of the ester group and of the methine carbon atom of the polymer chain are serve as diagnostic signals for distinguishing between M(MA) and T(MA). For all three catalysts, the macromolecules obtained in solvent mixtures having a low concentration of TFE (χ_TFE ≤ 0.35) had a low amount of methyl acrylate in the main chain, M(MA), which was almost unaffected by the reaction medium composition. The highest M(MA) value of 22% was achieved in neat dichloromethane with Pd3b, almost twice that afforded by Pd1b. On the basis of the pK_a values of the precursor anilines (2,4,6-trimethyl-aniline and 2,6-dimethyl-aniline for Pd3b and Pd1b, respectively),⁴⁶ ligand 3 should have higher Lewis basicity than 1, thus resulting in higher electron density at the palladium center, which, in turn, becomes less oxophilic, disfavoring the formation of the metallacycle MC6 and leading to a higher amount of polar monomers being inserted into the main chain (M(MA)).

Fig. 3 Ethylene/methyl acrylate copolymerization: the effect of the reaction medium composition on the amount of M(MA). Reaction conditions: see Fig. 1.

Conversely, in the copolymers produced in solvent mixtures with χ_TFE > 0.47, the value of M(MA) progressively increased up to the highest value of 49%, achieved with Pd3b in neat TFE (Fig. 3, Tables S1–S3, Fig. S24–S26).

These data indicate that the effect of the fluorinated alcohol³⁶ in slowing down the chain walking process following methyl acrylate insertion becomes evident when TFE is the major component of the solvent mixture.

Finally, the solvent composition has almost no effect on the branching degree, which is in the range of 89–104 branches for 1000 C atoms, depending mainly on the ancillary ligand 1–3 (Tables S1–S3). This result indicates that the chain walking process that follows the insertion of ethylene is unaffected by the solvent composition.

Our findings point out the unprecedented effect of the reaction medium composition on the catalyst performance, such as productivity and control of the macromolecule features. A threshold value was found at χ_TFE = 0.47, where the highest productivity is obtained. For χ_TFE > 0.47, the copolymer molecular weight, the amount of incorporated MA, and the proportion of MA incorporated as M(MA) all increase.

Even though several variables, e.g. ethylene and polymer solubility, should be considered to rationalize the catalytic results, an initial analysis can be made. Previous mechanistic investigations pointed out that in neat dichloromethane, the 6-membered metallacycle MC6 was the catalyst resting state (Scheme 3),²² whereas in trifluoroethanol, in addition to MC6, the open chain intermediate OC was demonstrated to be a novel catalyst resting state.³⁶ Moreover, MC6 is responsible for the enchainment of the polar monomers at the ends of the branches, while for OC, the growth of the polymer chain takes place with methyl acrylate located in the main chain.

Starting from these considerations, the catalytic data obtained in the solvent mixtures might be the result of a combination of the activities of both resting states, MC6 and OC, with the OC/MC6 ratio depending on the concentration of TFE in the reaction mixture. The observed increase in the amount of M(MA) with the TFE concentration for χ_TFE > 0.47 suggests a corresponding increase in the concentration of the open-chain intermediate compared to that of the metallacycle. This hypothesis is supported by the observation that the highest amount of M(MA) is found in the copolymers obtained with Pd3b, as a result of the combination of the electron donor ability of ligand 3 and the presence of trifluoroethanol.

The enhancement in the molecular weight can also be explained through an increase in the OC/MC6 ratio. Indeed, for OC, the growth of the polymer chain requires the simple substitution of acetonitrile by ethylene, whereas for MC6, ethylene has to cleave the metallacycle. Thus, it is reasonable to expect that the coordination-migratory insertion of ethylene is faster with OC than with MC6. In addition, trifluoroethanol, being more coordinating than dichloromethane, should disfavor β-hydrogen elimination.⁴⁵ The latter effect, combined with an increase in the OC/MC6 ratio, is expected to increase the ratio between the propagation and termination rates, leading to macromolecules with higher molecular weight.

Finally, the decrease in productivity and the increase in the content of inserted MA for χ_TFE > 0.47 might be explained by analyzing the TON values for the two co-monomers. Moving from χ_TFE = 0.47 to neat TFE for all three catalysts, the TON of ethylene remarkably decreases (by 66% for Pd1b and 45% for both Pd2b and Pd3b, Fig. S16a), whereas that of MA is much less affected (by 40% for Pd1b, 13% for Pd2b, and 28% for Pd3b; Fig. S16b), thus suggesting that the trends affecting the productivity and content of MA might be due to the lower solubility of ethylene in the liquid phase enriched with the polar solvent.

To shed light on the effect of the solvent composition on catalyst performance, in situ NMR investigations on the reaction between Pd1b and methyl acrylate in two CD₂Cl₂/TFE-d₃ mixtures were performed at room temperature and compared with our previous studies carried out with the same precatalyst in the two neat solvents.³⁶ In the first experiment, 2 equiv. of methyl acrylate were added to a 10 mM solution of Pd1b in a CD₂Cl₂/TFE-d₃ mixture with the composition χ_TFE = 0.17, corresponding to the first maximum of productivity. ¹H NMR spectral variations with time indicate the rapid reaction of Pd1b, whose signals basically disappear within 30 min (Fig. S31). Furthermore, in the spectrum recorded 1 min after the addition of MA, three singlets are evident for NCMe, which are assigned to free acetonitrile (2.01 ppm), to Pd–NCMe of residual Pd1b (1.82 ppm) and to Pd–NCMe of the new species OC1 (1.76 ppm, see below). 2D NMR experiments recorded after reaction completion indicate that the main species present in solution are the expected metallacycles MC6 and MC5 (in trace amounts) and the open chain intermediate OC1, in the ratio MC6 : MC5 : OC1 = 65 : 5 : 30 (Table S4, Fig. S32–S36). OC1 is the result of the migratory insertion reaction, with secondary regiochemistry, of methyl acrylate into the Pd–Me bond, and it has acetonitrile bound to the metal center (Scheme 4).

Scheme 4 The proposed mechanism for the reaction of Pd1b with methyl acrylate in CD₂Cl₂/TFE-d₃ mixtures. MC4 was not detected. c.w. = chain walking.

A similar NMR investigation was performed in a CD₂Cl₂/TFE-d₃ mixture with the composition χ_TFE = 0.79, which involves reverse concentrations of the two solvents with respect to χ_TFE = 0.17. In this case, Pd1b fully reacts within 5 min (Fig. S37), affording a solution that contains MC6, MC5 (in traces) and OC1, in the ratio MC6 : MC5 : OC1 = 53 : 4 : 43 (Table S4, Fig. S38–S42).

In both experiments, MC4 is not detected, and MC6 and MC5 are in equilibrium at a slow rate on the NMR time scale, as pointed out by the exchange peaks present in the corresponding ¹H,¹H NOESY spectra (Fig. S36 and S42).

Therefore, our findings show that upon moving from χ_TFE = 0.17 to χ_TFE = 0.79, the OC1/MC6 ratio increases from 0.46 to 0.81, thus supporting our hypothesis regarding the mechanism of copolymerization in the different solvent mixtures. In addition, the comparison with the same studies carried out in the two neat solvents points out that the reaction of Pd1b with methyl acrylate becomes progressively faster upon increasing the TFE concentration when moving from neat DCM to neat TFE (Table S4).³⁶ Moreover, in neat DCM, no open-chain intermediate was detected, in agreement with the almost exclusive incorporation of terminal MA in the obtained macromolecules.

3.3 Catalyst recycling experiments

For each Pd catalyst and for all the investigated solvent mixtures, the initial reaction mixture was a clear yellow solution. Then, as the reaction progressed, a suspension of the copolymer formed in all cases, except in neat dichloromethane and in the solvent mixture characterized by χ_TFE < 0.09 (Fig. 4a).

Fig. 4 The appearance of the reaction mixture at the end of catalytic runs carried out in different solvent compositions with Pd1b. Reaction conditions: see Fig. 1.

At the end of the catalytic run, a cloudy milk-white suspension of the copolymer was obtained for mixtures with χ_TFE > 0.47 (Fig. 4c). However, for the reactions carried out in solvent mixtures within the two maxima of productivity, that is, 0.17 ≤ χ_TFE ≤ 0.47 for Pd1b and Pd3b and 0.09 ≤ χ_TFE ≤ 0.47 for Pd2b, the unprecedented phenomenon of the spontaneous formation of two phases was observed (Fig. 4b). The two phases appeared as an off-white cloudy layer on top and a yellow layer at the bottom.

The two phases were separated by centrifugation, and their compositions were analysed by NMR spectroscopy in CDCl₃ solution at room temperature. The ¹H NMR spectrum of the upper phase showed signals mainly due to the E/MA copolymer; whereas in the spectrum of the lower phase, in addition to signals from residual copolymer, resonances attributed to the coordinated ligand of the catalyst were clearly evident in the aromatic region (Fig. 5).

Fig. 5 ¹H NMR spectra (CDCl₃, 298 K) of the two phases separated by centrifugation of the reaction mixture at the end of the catalytic run with Pd1b at χ_TFE = 0.17: a) upper phase and b) lower phase. The two spectra have normalized intensities, but in each of them, the aliphatic and aromatic regions are not to scale.

On the basis of this NMR analysis, we decided to develop a protocol for catalyst recovery and recycling. Thus, as detailed in the Experimental section, at the end of catalytic run 1, the two layers were separated by centrifugation. The lower phase, presumably containing almost all of the catalyst in its resting state, was stored overnight at 253 K. The following day, no decomposition to palladium metal was evident. Thus, this solution was used to carry out catalytic run 2 following the protocol reported in the Experimental section. In particular, methyl acrylate and solvent mixture were added to create reaction conditions as similar as possible to those of run 1, although not identical. At the end of the second run, the same protocol was applied, thus enabling the recovery and recycling of the catalyst for a second time. However, at the end of catalytic run 3, the formation of the two phases was insufficiently clear to allow for efficient separation, and the entire mixture was therefore processed by following the standard workup procedure to isolate the copolymer.

This protocol was tested with Pd1b–Pd3b in catalytic runs carried out in one of the solvent mixtures with peak productivity, that is, χ_TFE = 0.17 for Pd1b and Pd3b and χ_TFE = 0.09 for Pd2b (Table 1). The three catalysts shared some common features: i) all of them were still active after the separation procedure and could be reused for up to two additional copolymerization experiments carried out on two consecutive days; ii) the amount of polymer isolated at the end of run 1 was lower than that obtained in the relevant reference experiment when the two phases were not separated, confirming that some copolymer remained dissolved in the catalyst-containing phase; iii) the yield of the isolated copolymer decreased from run 1 to 3, suggesting that some of the catalyst was deactivated and/or remained dissolved in the copolymer-containing phase after separation; iv) for each catalyst, the overall yield of the recycling experiments was higher than that obtained in the reference experiment (single run at the same solvent composition, with no catalyst recycling); the increase, ranging from 20% to 57%, depended on the nature of the catalyst; and v) NMR and GPC characterization of the macromolecules obtained from each recycle run showed that they have very similar microstructures (amount of inserted MA, M(MA) : T(MA) ratio (Fig. S27–S29), branching degree, molecular weight and polydispersity), thus indicating that the catalyst was the same species in each run. In addition, the aromatic region of the ¹H NMR spectrum of the lower phase isolated at the end of run 1 with Pd1b indicates that ligand 1 is coordinated to palladium (its resonances are different from those of the free ligand and of 2,6-dimethyl aniline, Fig. S43).

Table 1 Ethylene/methyl acrylate copolymerization: catalyst recycling^a

Entry	Run	Yield [g]	mol% MA^b	M(MA) : T(MA)^c	Mn^d [kDa] (Mw/Mn)	Bd^e
Precatalyst: Pd1b–Pd3b.a Reaction conditions: nPd = 2.1 × 10^−5 mol, Vsol = 21 mL, VMA = 2.26 mL, [MA]/[Pd] = 1188, T = 308 K, PE = 5 bar, t = 6 h.b Calculated by ^1H NMR spectroscopy analysis of the isolated product.c Calculated by ^13C NMR spectroscopy analysis of the isolated product.d The molecular weight (Mn and Mw) and the molecular weight distribution (Mw/Mn) were measured by GPC.e Bd = branching degree: branches per 1000 carbon atoms, calculated by ^1H NMR spectroscopy analysis of the isolated product.f Value from entry 3, Table S1.g Value from entry 2, Table S2.h Value from entry 2, Table S3.
Pd1b (χTFE = 0.17)
1	1	2.58	2.2	17 : 83	18.8 (1.44)	103
2	2	1.08	1.9	19 : 81	17.6 (1.42)	104
3	3	1.01	2.8	14 : 86	12.3 (1.65)	102
	Total	4.67
Ref. 1^f		2.97	2.4	15 : 85	18.2 (1.72)	101
Pd3b (χTFE = 0.17)
4	1	1.82	2.5	19 : 81	22.1 (1.74)	99
5	2	0.83	2.2	19 : 81	14.1 (2.53)	98
6	3	0.49	2.4	17 : 83	15.0 (2.10)	98
	Total	3.14
Ref. 3^g		2.61	2.6	21 : 79	19.5 (1.80)	100
Pd2b (χTFE = 0.09)
7	1	2.63	0.9	9 : 91	19.2 (1.87)	93
8	2	1.72	0.7	8 : 92	16.6 (1.58)	94
9	3	1.91	0.8	5 : 95	14.4 (1.82)	98
	Total	6.23
Ref. 2^h		4.51	0.9	12 : 88	18.4 (1.67)	95

---

The recycling experiments were performed using a solvent mixture composition leading to macromolecules with acrylate inserted preferentially at the ends of the branches. In order to fully exploit the effect of TFE on the mode of MA incorporation, we decided to perform a second series of catalyst recycling experiments with Pd1b by carrying out the copolymerizations in neat TFE as the reaction medium. During catalysis, the polymer precipitated, and at the end of the catalytic run 1, a cloudy suspension was obtained, with no spontaneous formation of two phases. Thus, we defined a new protocol for catalyst recovery (see the Experimental section), which involved the addition of dichloromethane to extract the formed copolymer, while the catalyst should remain confined in trifluoroethanol. Following the procedure detailed in the Experimental section, it was possible to recycle the catalyst twice, for a total of three catalytic runs (Table 2).

Table 2 Ethylene/methyl acrylate copolymerization in neat TFE: catalyst recycling^a

Entry	Run	Yield [g]	mol% MA^b	M(MA) : T(MA)^c	Mn^d [kDa] (Mw/Mn)	Bd^e
Precatalyst: Pd1b.a Reaction conditions: nPd = 2.1 × 10^−5 mol, solvent: TFE, VTFE = 21 mL, VMA = 2.26 mL, [MA]/[Pd] = 1188, T = 308 K, PE = 5 bar, t = 6 h.b Calculated by ^1H NMR spectroscopy analysis of the isolated product.c Calculated by ^13C NMR spectroscopy analysis of the isolated product.d The molecular weight (Mn and Mw) and the molecular weight distribution (Mw/Mn) were measured by GPC.e Bd = branching degree: branches per 1000 carbon atoms, calculated by ^1H NMR spectroscopy analysis of the isolated product.f Value from entry 10, Table S1.
1	1	0.65	3.1	36 : 64	31.0 (1.96)	99
2	2	1.18	2.4	33 : 67	21.4 (2.08)	101
3	3	0.98	2.7	30 : 70	9.9 (2.29)	103
	Total	2.81
Ref. 1^f		1.02	4.0	40 : 60	23.4 (1.19)	93

---

As observed in the recycling experiments carried out at χ_TFE = 0.17, even in this case, the amount of polymer isolated at the end of run 1 was lower than that obtained in the reference experiment, thus indicating that some copolymer remained in the lower phase after the addition of DCM. However, unlike in the previous set of recycling experiments, in this case, the yield of isolated polymer increased from run 1 to 3, with a simultaneous decrease in the M(MA) : T(MA) ratio (Fig. S30) and in the molecular weight of the isolated macromolecules. On the basis of the results obtained upon performing catalysis in DCM/TFE mixtures, these trends suggest that some residual dichloromethane was present in the reaction mixture. Nevertheless, the overall yield from the recycling experiments was 2.7 times as high as that from the reference experiment with no catalyst recycling.

4. Conclusions

In this contribution we investigated in detail the catalytic behaviors of three known Pd(II) complexes with the general formula [Pd(N–N)(Me)(NCMe)][PF₆], namely Pd1b, Pd2b and Pd3b, with three benchmark α-diimines (N–N), during ethylene/methyl acrylate copolymerization by performing, for the first time, the catalytic reaction in dichloromethane/trifluoroethanol mixtures of different compositions. We discovered that the concentrations of the two solvents remarkably affect the catalyst performances both in terms of productivity and control of the features of the produced E/MA copolymers. In particular, we found that:

i. catalyst productivity, copolymer molecular weight, methyl acrylate content and the type of incorporation do not linearly vary with an increase in the concentration of trifluoroethanol in the solvent mixture;

ii. in the used DCM/TFE mixtures, catalyst productivity was, in almost all cases, higher compared to that found in the two neat solvents;

iii. a molar fraction of trifluoroethanol of 0.47 represents a threshold value, where the highest productivity of 262.3 kg CP per mol Pd was achieved using Pd2b;

iv. for χ_TFE > 0.47, the copolymer molecular weight, the amount of incorporated MA, and its incorporation into the main chain (M(MA)) increase. The highest molecular weight, corresponding to 101 kDa, was achieved in neat TFE for the macromolecules synthesized using Pd2b. The highest amount of methyl acrylate inserted as M(MA), 49%, was found in the macromolecules produced in neat TFE with the catalyst Pd3b and;

v. at the end of the catalytic runs carried out in solvent mixtures with a composition of 0.17 ≤ χ_TFE ≤ 0.47 for Pd1b and Pd3b and 0.09 ≤ χ_TFE ≤ 0.47 for Pd2b, the unprecedented phenomenon of the spontaneous formation of two phases occurred.

We exploited the formation of the two phases to perform, for the first time for E/MA copolymerization, the recovery and recycling of the Pd catalyst. To recover the catalyst, we defined two different protocols, depending on the used solvent composition. In both protocols, no decomposition to inactive palladium metal was evident after storing the catalyst-containing solution for 14 h at 253 K, thus highlighting the high stability of the catalyst resting state in the applied solvent mixtures and allowing the catalyst to be reused for at least three consecutive catalytic runs. The overall yield of the isolated copolymer was higher than that obtained in a single run with no catalyst recovery, with the increase depending on the catalyst. No variations in the microstructure of the copolymers isolated at the end of the recycling runs were observed, thus indicating that the catalyst was the same species in each run.

As a general conclusion, the found effect for these solvents might be also considered for already reported catalysts bearing other α-diimines, and copolymers with different microstructures compared to those obtained in dichloromethane could be produced. Finally, it cannot be ruled out that the protocol for catalyst recovery and recycling can also be applied to other catalysts in the literature.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI).

Supplementary information: NMR characterization of ligands, complexes, and polymers; catalytic data; GPC curves of polymers; in situ NMR studies. See DOI: https://doi.org/10.1039/d6cy00332j.

Acknowledgements

This work was supported by Università degli Studi di Trieste (FRA2022). The PhD Fellowship of K. A. H. was supported by PON 2014–2020 CCI 2014IT16M2OP005. We acknowledge Progetto Competitivo CMPT231981 for the post-doc fellowship to K. A. H. Dr. Isabella Camurati and Dr. Gabriele Tani from LyondellBasell are gratefully acknowledged for the GPC measurements of the synthesized copolymers.

References

1. Plastics – the fast Facts 2024, https://plasticseurope.org/knowledge-hub/plastics-the-fast-facts-2024/.
2. F. Vidal, E. R. van der Marel, R. W. F. Kerr, C. McElroy, N. Schroeder, C. Mitchell, G. Rosetto, T. T. D. Chen, R. M. Bailey, C. Hepburn, C. Redgwell and C. K. Williams, Nature, 2024, 626, 45–57.
3. W. Qu, Z. Bi, C. Zou and C. Chen, Adv. Sci., 2024, 11, 2307568.
4. Z. Balzade, F. Sharif and S. R. Ghaffarian Anbaran, Macromolecules, 2022, 55, 6938–6972.
5. K. Li, L. Cui, Y. Zhang and Z. Jian, Macromolecules, 2023, 56, 915–922.
6. X. Zhang, Y. Zhao, M. Chen, M. Ji, Y. Sha, K. Nozaki and S. Tang, J. Am. Chem. Soc., 2024, 146, 24024–24032.
7. B. Lu, K. Takahashi, J. Zhou, S. Nakagawa, Y. Yamamoto, T. Katashima, N. Yoshie and K. Nozaki, J. Am. Chem. Soc., 2024, 146, 19599–19608.
8. T. Ganguly, L. C. Ruiz De Castilla, R. Adhikary and L. H. Do, Polym. Chem., 2025, 16, 4731–4744.
9. L. K. Johnson, S. Mecking and M. Brookhart, J. Am. Chem. Soc., 1996, 118, 267–268.
10. J. T. Medina, Q. H. Tran, G. G. Ramachandru, M. Brookhart and O. Daugulis, Acc. Chem. Res., 2025, 58, 2770–2780.
11. H. Zheng, Z. Qiu, D. Li, L. Pei and H. Gao, J. Polym. Sci., 2023, 61, 2987–3021.
12. F. Wang and C. Chen, Polym. Chem., 2019, 10, 2354–2369.
13. A. Nakamura, T. M. J. Anselment, J. Claverie, B. Goodall, R. F. Jordan, S. Mecking, B. Rieger, A. Sen, P. Van Leeuwen and K. Nozaki, Acc. Chem. Res., 2013, 46, 1438–1449.
14. Y. Mitsushige, B. P. Carrow, S. Ito and K. Nozaki, Chem. Sci., 2016, 7, 737–744.
15. N. D. Contrella, J. R. Sampson and R. F. Jordan, Organometallics, 2014, 33, 3546–3555.
16. X. Sui, S. Dai and C. Chen, ACS Catal., 2015, 5, 5932–5937.
17. W. Zhang, P. M. Waddell, M. A. Tiedemann, C. E. Padilla, J. Mei, L. Chen and B. P. Carrow, J. Am. Chem. Soc., 2018, 140, 8841–8850.
18. L. Cao, Z. Cai and M. Li, Organometallics, 2022, 41, 3538–3545.
19. R. Nakano and K. Nozaki, J. Am. Chem. Soc., 2015, 137, 10934–10937.
20. L. Pei, H. Gao, C. Wang, C. Zhang, B. Ning, H. Zheng and H. Gao, Chem. – Eur. J., 2025, e02580,  DOI:10.1002/chem.202502580.
21. C. Alberoni, E. Reusser, G. Balducci, E. Alessio, M. Albrecht and B. Milani, Dalton Trans., 2025, 54, 6876–6886.
22. S. Mecking, L. K. Johnson, L. Wang and M. Brookhart, J. Am. Chem. Soc., 1998, 120, 888–899.
23. S. Takano, D. Takeuchi, K. Osakada, N. Akamatsu and A. Shishido, Angew. Chem., Int. Ed., 2014, 53, 9246–9250.
24. Y. Zhang, C. Wang, S. Mecking and Z. Jian, Angew. Chem., Int. Ed., 2020, 59, 14296–14302.
25. S. Dai, X. Sui and C. Chen, Angew. Chem., Int. Ed., 2015, 54, 9948–9953.
26. S. Dai, S. Zhou, W. Zhang and C. Chen, Macromolecules, 2016, 49, 8855–8862.
27. Q. Muhammad, C. Tan and C. Chen, Sci. Bull., 2020, 65, 300–307.
28. Y.-S. Liu and E. Harth, Angew. Chem., Int. Ed., 2021, 60, 24107–24115.
29. Y. Liu, G. Yang, C. Li, C. Tan and M. Chen, Polym. Chem., 2024, 15, 40–45.
30. X. Wu, J. Jiang, M. Zou, H. Wang and S. Dai, Polymer, 2024, 312, 127617.
31. Y. Wang and S. Dai, Catal. Sci. Technol., 2025, 15, 2822–2828.
32. H. Zhou, C. Feng, H. Zheng, G. Tu, X. Xiao and H. Gao, Catalysts, 2025, 15, 127.
33. J. Dai and S. Dai, J. Polym. Sci., 2025, 63, 1527–1535.
34. H. Zhang and S. Dai, New J. Chem., 2025, 49, 18430–18435.
35. H. Sun, M. Ma, J. Dai and S. Dai, New J. Chem., 2025, 49, 15776–15783.
36. C. Alberoni, M. C. D'Alterio, G. Balducci, B. Immirzi, M. Polentarutti, C. Pellecchia and B. Milani, ACS Catal., 2022, 12, 3430–3443.
37. P. Jutzi and T. Redeker, Organometallics, 1997, 16, 1343–1344.
38. J. Durand, E. Zangrando, M. Stener, G. Fronzoni, C. Carfagna, B. Binotti, P. C. J. Kamer, C. Muller, M. Caporali, P. W. N. M. van Leeuwen, D. Vogt and B. Milani, Chem. – Eur. J., 2006, 12, 7639–7651.
39. H. A. Zhong, J. A. Labinger and J. E. Bercaw, J. Am. Chem. Soc., 2002, 124, 1378–1399.
40. L. K. Johnson, C. M. Killian and M. Brookhart, J. Am. Chem. Soc., 1995, 117, 6414–6415.
41. V. Rosar, T. Montini, G. Balducci, E. Zangrando, P. Fornasiero and B. Milani, ChemCatChem, 2017, 9, 3402–3411.
42. L. Guo, J. Li, W. Zhao, P. Wei, Y. Ju, X. Cui, L. Yuan, M. Ji and Z. Liu, Inorg. Chem., 2024, 63, 17809–17827.
43. B. Milani, A. Anzilutti, L. Vicentini, A. Sessanta o Santi, E. Zangrando, S. Geremia and G. Mestroni, Organometallics, 1997, 16, 5064–5075.
44. A. Scarel, J. Durand, D. Franchi, E. Zangrando, G. Mestroni, B. Milani, S. Gladiali, C. Carfagna, B. Binotti, S. Bronco and T. Gragnoli, J. Organomet. Chem., 2005, 690, 2106–2120.
45. A. Dall'Anese, V. Rosar, L. Cusin, T. Montini, G. Balducci, I. D'Auria, C. Pellecchia, P. Fornasiero, F. Felluga and B. Milani, Organometallics, 2019, 38, 3498–3511.
46. A. N. Lange, in Langes' Handbook of Chemistry, McGraw-Hill, New York, 15th edn, 1999, vol. 8.

---

Footnote

† The catalytic tests at χ_TFE = 0.17 and 0.35 were performed in duplicate, obtaining excellent reproducibility. In addition, very similar trends were found for the three investigated catalysts, thus indicating the reliability of the data.

---