Dispersity-controlled ring-opening polymerization of epoxide

Abstract

Dispersity or molar mass distribution (Ð_M) profoundly influences polymer properties, but its control remains a challenge, especially for ring-opening polymerization (ROP). We report here a simple approach to tailor Ð_M of aliphatic polyethers, e.g., poly(ethylene oxide) and poly(propylene oxide), by introducing an editable chain-transfer reaction in organocatalyzed ROP of epoxides. Trifluoroacetate, acting as the chain-transfer agent (CTA), brings about non-uniform growth of polyether chains while ensuring complete initiation efficiency and readily removable CTA residue. Triple control over Ð_M (1.05–2.00), molar mass (3.6–17.7 kg mol⁻¹), and end group is achieved simply by regulating the feed ratio, conforming well with the theoretical modeling. Experimental and calculation results reveal the orthogonality between chain growth and transesterification which is key to the three-modular ROP for independent Ð_M control.

Introduction

Dispersity or molar mass distribution (Ð_M), used to quantify heterogeneity of chain lengths, has profound effects on polymers properties^1,2. For more than half a century, the rapid development of living/controlled (chain-growth) polymerization techniques have been continuously empowering chemists to obtain low-Ð_M polymer samples, which are usually favored by polymer physicists because the structure-property relationship can be easily clarified without being affected by dispersity^3,4. On the other hand, high-Ð_M polymers, usually obtained by step-growth polymerization and non-controlled chain-growth polymerization, have exhibited advantages over the low-Ð_M ones in e.g., microphase separation, self-assembly, interfacial properties, melt rheology and processing. Therefore, Ð_M control constitutes an indispensable section of macromolecular engineering^{5,6,7,8,9,10,11}.

In principle, a high-Ð_M polymer can be rationally designed and obtained through blending of low-Ð_M ones with the same composition but different molar mass. Yet, application of this strategy has been limited by the long and costly multistep preparation as well as the difficulty in achieving continuous and unimodal distribution^12,13. Recently, chemical strategies for the synthesis of Ð_M-controlled polymers have been reported. Fors et al. tailored Ð_M in living anionic and controlled radical polymerization of vinyl monomers by intermittent feeding of initiators^14,15. Boyer et al. used continuous flow reactor to control Ð_M by adjusting the initiator-monomer mixing process^16,17,18,19. Chiu et al. used photochromic initiators or mediators to control Ð_M through reversible photo-deactivation^20,21,22. Anastasia et al. tailored Ð_M by mixing two chain-transfer agents (CTAs) with distinct activities in reversible addition-fragmentation chain-transfer (RAFT) polymerization²³, which was further developed into a strategy with in-situ switched CTA activity^24,25,26. Matyjaszewski et al. demonstrated that Ð_M could be regulated by the structure and concentration of catalyst, ligand or initiator in atom transfer radical polymerization (ATRP)^27,28,29. Goto et al. exploited temperature-selective reversible complexation-mediated polymerization to control Ð_M by using two monomers with distinct activities⁸. Besides, some theoretical studies dealt with the principles of Ð_M control^30,31. So far, all these studies have focus on vinyl-addition polymerizations¹², while Ð_M-controlled ring-opening polymerization (ROP) has not been reported.

Aliphatic polyethers such as poly(ethylene oxide) (PEO) and poly(propylene oxide) (PPO) are among the most important heterochain polymers in industry^32,33,34. The easy production from bulk chemicals (epoxides), the flexible and oxygen-rich polymer chains with tunable polarity and hydrophilicity have enabled PEO/PPO to be extensively utilized as the main constituents of water reducers, polyurethanes, surfactants, drug modifiers/carriers, solid-state electrolytes, etc^32,33,35. Low-Ð_M polyethers are usually synthesized through living/controlled anionic ROP^{36,37,38,39,40,41,42}. In the past two decades, they were widely achieved via organocatalytic ROP^43,44. For example, Lewis pair type organocatalyst comprising a strong phosphazene base (e.g., 1-tert-butyl-2,2,4,4,4-pentakis(dimethylamino)−2λ⁵,4λ⁵-catenadi(phosphazene); ^tBuP₂ in Fig. 1a) and excess triethylborane (Et₃B) fulfilled highly efficient and chemoselective ROP of epoxides under mild conditions^45,46,47. Polyethers with controlled molar mass, well-defined (end group) structure, and low Ð_M (<1.1) were obtained even though the basic catalytic component was added far less than the protonic initiator, which has been attributed to a potent Lewis acid/base cooperative catalytic effect⁴⁴.

Fig. 1: Mechanism of Ð_M-controlled ECT-ROP.

a Proposed pathway of ECT-ROP of epoxides comprising proton transfer (purple) and editable chain transfer (blue), with the angular velocities of the colored gears indicating the relative reaction rates. b Illustration of the mechanism for modular ECT-ROP, R_ti is the chain-transfer rate at initiation stage, with the proportions of pie-charts indicating the weights of the three modules effecting Ð_M control.

In this study, we have synthesized polyethers via organocatalyzed ROP of epoxides with trifluoroacetate (TFA) as a CTA, where the chain-transfer rate (R_t) is regulated to be moderately lower than the chain-propagation rate (R_p) to ensure the non-uniform chain growth. Efforts are made to realize ternary-control over structure, molar mass, and Ð_M in ROP.

Results

Selection of CTA for ROP of epoxides

We have investigated into the relation between Ð_M and R_t or R_p (Supplementary Figs. 3 and 4). It shows that only when R_p and R_t perfectly match, i.e., R_t is moderately lower than R_p (ca. 10R_t < R_p < 1000R_t), dual control over molar mass and Ð_M can be realized. Since the interchange between activated chain end (ACE, alkoxide) and dormant chain end (DCE, hydroxyl) is a quick proton transfer (Fig. 1a), it is a great challenge to harness R_t for Ð_M control.

Transesterification is a common hydroxyl interchange reaction in anionic ROP^48,49. It was inhibited (R_t → 0) during ROP of epoxides catalyzed by ^tBuP₂-Et₃B for aliphatic esters in our previous work⁵⁰. Considering that TFA lies between aliphatic ester and activated epoxide in surface electrostatic potential (ESP) (Supplementary Figs. 5 and 6), it should afford an additional chain transfer with matching and editable R_t (blue gear in Fig. 1a). In the present work, TFA was used as the CTA to enable an editable-chain-transfer ROP (ECT-ROP) with three mutually orthogonal reactions, i.e., two chain transfers and one propagation (Fig. 1a; and Supplementary Fig. 7), with three independently regulatable modules for tuning Ð_M (Fig. 1b).

Polymerization

ROP of ethylene oxide (EO) is performed in THF catalyzed by ^tBuP₂-Et₃B Lewis pair with two-component initiation-transfer system comprising methanol (MeOH) and methyl trifluoroacetate (MTFA) yielding PEO (EO₁₆₀¹-EO₁₆₀⁴; the subscript and superscript indicate targeted DP and experiment number in a series, respectively) with the degree of polymerization (DP) of 160 (Supplementary Table 1). EO is fully consumed in 1 h as verified by ¹H NMR. The dual control over molar mass with M_n,th = [EO]₀/([MeOH]₀ + [MTFA]₀)M_EO and Ð_M from 1.05 to 1.74 is confirmed by SEC traces and ¹H NMR spectra (Fig. 2a). More PEOs are synthesized with targeted DPs of 80 (EO₈₀¹-EO₈₀⁴), 240 (EO₂₄₀¹-EO₂₄₀⁴), and 400 (EO₄₀₀¹-EO₄₀₀³) by regulating the feed of EO (Supplementary Figs. 8–10). In the range of DP from 80 to 400, measured molar mass is close to the theoretical value M_n,th (Fig. 2d), and Ð_M increases with the feed ratio of MTFA and MeOH, agreeing well with those fit from modified Müller equation³⁰.

$${{\mbox{-\!-}}{\hskip-4pt{D}}}_{{{\rm{M}}}}=\frac{1+\beta }{\beta+\alpha }+{{\rm{C}}}$$

(1)

where α is [OH]₀/([TFA]₀ + [OH]₀) (OH=hydroxyl group), β is (R_t/[TFA])/(R_p/[epoxide]), and our modeled results (Fig. 2d–h). The theoretical Ð_M at α = 1 should be close to 1, so C = Ð_M,_α=1-1 is used to rule out the instrument error in experimental data. High Ð_M of 1.95 and 2.05 can be achieved when the ratio of [MeOH]₀/[MTFA]₀ is 0.15/0.85 (EO₈₀⁵ and EO₄₀₀⁴). However, the molar mass obtained by SEC (M_n,SEC) is higher than the value determined by NMR (M_n,NMR), which is ascribed to the partial participation of MTFA at such a low ratio of [MeOH]₀/[MTFA]₀. This agrees with the modeling results (white zones in Supplementary Figs. 39 and 40, in which initiation efficiency is lower than 90%), and also supported by the observation of residual MTFA in the crude product of EO₈₀⁵ (Supplementary Fig. 11).

Fig. 2: Polyethers with dual control over molar mass and Ð_M.

SEC traces of dual-controlled PEOs with DP of 160 (a EO₁₆₀^1–4), dual-controlled PPOs with DP of 80 (b PO₈₀^1–4), and dual-controlled PAGEs with DP of 60 (c AGE₆₀^1–4). d [MTFA]₀/[MeOH]₀ dependence of Ð_M (top) and controlled number-average molar mass from SEC traces (bottom). e–h The fitting of experimental Ð_M with the modeled ones and those from Eq. (1), the fitting parameters for modified Müller equation are on the left.

Monomer is then extended to propylene oxide (PO) and allyl glycidyl ether (AGE). The ROP is carried out in bulk and the reaction time is prolonged to adapt to the decreased monomer activity. PO conversion reaches above 95% in 48 h (Supplementary Table 2) with controlled molar mass. As [MTFA]₀ increases, Ð_M increases from 1.05 to 1.69 for DP = 80 (PO₈₀¹-PO₈₀⁴), from 1.06 to 1.63 for DP = 160 (PO₁₆₀¹-PO₁₆₀⁴), and from 1.07 to 1.69 for DP = 240 (PO₂₄₀¹-PO₂₄₀⁴) (Fig. 2b and Supplementary Figs. 12–14), This trend agrees with modeled results and Eq. (1) (Supplementary Figs. 15–17). Due to the bulky substituent, conversion of AGE only reaches 85% at the targeted DP of 60 in 48 h (Supplementary Table 3). The functional (pendant allyl) groups, molar mass, and Ð_M (1.09–1.59) are all controlled (Fig. 2c and Supplementary Figs. 19 and 22).

¹H NMR spectra of the products demonstrate terminal residues of trifluoroacetate ester (CF₃COOCH_1/2−). It can readily convert into hydroxyl group after polymerization along with the purification process because of its high reactivity (Supplementary Figs. 20–22). Mass spectrometry of crude products provides further evidence for controlled Ð_M, defined repeating unit and end group compositions of EO₁₆₀¹-EO₁₆₀⁴ and PO₈₀¹-PO₈₀⁴ (Supplementary Figs. 23 and 24). Clearly, MTFA can serve as a valid CTA for tuning Ð_M of polyethers by [MeOH]₀/[MTFA]₀. Only small amounts of residual TFA chain ends are discernable in some of the mass spectra because the TFA groups are readily hydrolyzed during sample preparation and measurement (e.g., the product of PO₈₀³). Complete ester removal (liberation of hydroxy chain ends) can be achieved via alcoholysis (Supplementary Fig. 25). We also provide here the modeled mass spectra, which share good similarity with the experimentally acquired ones (Supplementary Figs. 26–33).

Experiments are also performed at different feed ratios of the two catalytic components. Ð_M increases with [Et₃B]₀ while [^tBuP₂]₀ remains constant (Fig. 3a, b). Specifically, Ð_M is 1.88, 1.69, and 1.60 when [^tBuP₂]₀/[Et₃B]₀ is 0.05/0.15 (PO₈₀⁶), 0.05/0.1 (PO₈₀⁴), and 0.05/0.075 (PO₈₀⁷), respectively. However, the molar mass of PPO of PO₈₀⁶ is uncontrolled, and a significant low-molar-mass tailing is observed in SEC. We have then modeled the SEC traces based on the parameters including R_p/R_t, DP, and [ACE]/[DCE] (Supplementary Table 4). All modeled traces do not agree with the experimental results due to the tailing (Supplementary Figs. 34–36), indicating some undetected factors in the polymerization.

Fig. 3: Effects of the catalyst ratio and CTA structure on Ð_M.

a, b SEC traces of PPOs regulated by [^tBuP₂]₀/[Et₃B]₀. c–e EO₁₆₀⁴, AGE₆₀⁴, and PO₈₀⁴ vs. modeled traces from our program. f Effect of CTA structure on Ð_M, PO₈₀⁴ (MeOH-MTFA) vs. PO₈₀^B4 (BnOH-BTFA). g, h SEC traces of PPOs regulated by [BnOH]₀/[BTFA]₀ with symmetric distribution (PO₈₀^B1-PO₈₀^B6).

An equilibrium reaction between TFA and alcohol is conducted to compare the activities of CTAs in different forms. ¹H NMR spectrum of the product indicates that the ether-based TFA is more active than the original MTFA (Supplementary Fig. 37), which is also corroborated by comparative DFT calculations of transesterification reactions (Supplementary Fig. 50). Namely, the chain-transfer rate at initiation stage (R_ti) is lower than that at propagation stage (R_t). After revising our model with this assumption (parameterized by R_t/R_ti), a good fitting between experimental and modeled trace is achieved (Fig. 3c–e). Clearly, R_t/R_ti determined by the structure of CTA can be utilized to regulate the extent of the low-molar-mass tailing (Supplementary Fig. 38 and Supplementary Table 4). Note that too high R_t/R_ti would reduce the initiation efficiency, leading to uncontrolled molar mass (Supplementary Figs. 39 and 40).

To rule out slow transfer at initiation stage, benzyl trifluoroacetate (BTFA) is used as the CTA (Supplementary Table 5). As expected, the low-molar-mass tailing is removed and the obtained PPOs exhibit symmetric distribution (Fig. 3f and Supplementary Fig. 41). Ð_M is also regulated by [OH]₀/[TFA]₀, following the Eq. (1) and our model (Fig. 3g, h, Supplementary Fig. 42). At [OH]₀/[TFA]₀ = 0.1/0.9, controlled molar mass and Ð_M of 1.80 are obtained by using benzyl alcohol (BnOH) and BTFA (PO₈₀^B5). Increasing the feed of Et₃B ([^tBuP₂]₀/[Et₃B]₀ = 0.05/0.2; PO₈₀^B6) still results in controlled molar mass and Ð_M of 2.00 (Fig. 3g, h, Supplementary Fig. 43b). In contrast, MeOH-MTFA leads to uncontrolled molar mass with multimodal distribution (PO₈₀⁵) (Supplementary Fig. 44).

DFT calculations

DFT calculation is performed on the simplified structures in EO polymerization. Figure 4 shows the energy profile (the calculational details can be found in Supplementary Information). ACE can be generated via the deprotonation of the alcohol by ^tBuP₂ and coordination with Et₃B (exoergic process of ΔG = −5.9 kcal mol⁻¹; Supplementary Fig. 45b). The strong coordination of Et₃B exceedingly decreases the nucleophilicity of the alkoxide species, resulting in inhibited transesterification on regular aliphatic ester (AE) by a high energy barrier of 44.4 kcal mol⁻¹ (IN1 → TS2-H → IN3-H). Thanks to the effective activation of EO by free Et₃B (endoergic process of ΔG = 8.1 kcal mol⁻¹; Supplementary Fig. 45c), the epoxy ring can still be opened by overcoming a reasonable energy barrier of 24.4 kcal mol⁻¹ (IN1 → TS1 → IN2; R_p=k_p[epoxide][ACE][Et₃B]_f, [Et₃B]_f = [Et₃B]₀-[ACE]; also see Supplementary Fig. 46 for more details).

Fig. 4: Energy profile of the elementary reactions in ECT-ROP of EO.

Each energy and structure are acquired by DFT calculation including propagation (red lines) and chain transfer via transesterification on TFA (blue lines) with the optimized 3D structures of the main transition states. The inhibited transesterification on regular carboxylic ester (gray lines) is shown for comparison, and a matching zone for R_p and R_t is shown as a light purple area.

Fortunately, the strong electron-withdrawing effect of the CF₃ group substantially enhances the electrophilicity of carbonyl group of TFA, reducing the energy barrier of transesterification from 44.4 kcal mol⁻¹ to 28.4 kcal mol⁻¹ (IN1 → TS2-F → IN3-F). Chain transfer to TFA is moderately slower than the propagation (24.4 kcal mol⁻¹), which is consistent with the ESP rank of the reactants (Supplementary Figs. 5 and 6). The gap between the two energy barriers (4.0 kcal mol⁻¹) is located at the matching zone estimated from our model (<5.0 kcal mol⁻¹). It strongly supports that TFA is a valid CTA for Ð_M control. In the second step of chain transfer (IN3-F → TS2’-F → IN1’), the assisting Et₃B can be a new Et₃B from the solution (first-order dependence on [Et₃B]_f) or the one released in the first step of TS2-F → IN3-F (Pathway 1 and 2 in Supplementary Fig. 47, respectively). Since the latter pathway (intramolecular Et₃B migration) does not depend on the concentration of Et₃B, the dependence of transfer reaction on [Et₃B]_f is thus a linear combination of zero-order and first-order based on the equation R_t = (k_t,0-order+k_t,1-order[Et₃B]_f)[ACE][TFA]. Clearly, Et₃B impacts chain propagation and transfer rather differently, similar to the Janus catalytic effect previously observed in a copolymerization⁵¹. In other words, the catalyst ratio can be utilized to regulate R_p/R_t, affording another module for controlling Ð_M (Supplementary Fig. 49).

Based on this assumption, β in Eq. (1) can be estimated as,

$$\begin{array}{c}\begin{array}{c}\beta=\frac{{k}_{{{\rm{t}}},0-{{\rm{order}}}}+{k}_{{{\rm{t}}},1-{{\rm{order}}}}{\left[{{{\rm{Et}}}}_{3}{{\rm{B}}}\right]}_{{{\rm{f}}}}}{{k}_{{{\rm{p}}}}{\left[{{{\rm{Et}}}}_{3}{{\rm{B}}}\right]}_{{{\rm{f}}}}}\end{array}\end{array}$$

(2)

According to our experimental protocol, [Et₃B]_f = [epoxide]₀/DP×(0.1-0.05), then

$$\begin{array}{c}\begin{array}{c}\beta=\frac{{k}_{{{\rm{t}}},1-{{\rm{order}}}}}{{k}_{{{\rm{p}}}}}+\frac{{20k}_{{{\rm{t}}},0-{{\rm{order}}}}\times {{\rm{DP}}}}{{k}_{{{\rm{p}}}}{\left[{{\rm{epoxide}}}\right]}_{0}}\end{array}\end{array}$$

(3)

Therefore, β should increase with targeted DP, explaining for the increase of experimentally calculated β for EO₁₆₀, EO₂₄₀, EO₄₀₀ groups (0.88, 1.02, and 1.30; Supplementary Fig. 48) as well as the significant decrease in Ð_M (1.34) and increase in β (2.34) for EO₈₀₀¹ (Supplementary Table 1). Elucidation for the derivation of β value is detailed in Supplementary Information.

Several possible chain transfers in ECT-ROP are examined. The 31.4 kcal mol⁻¹ barrier of transfer at initiation stage (Supplementary Fig. 50a) is higher than the 28.4 kcal mol⁻¹ barrier at propagation stage, substantiating the slow initiation of MTFA-MeOH. This agrees with our experimental results and establishes a module for controlling Ð_M by CTA structure.

Kinetics studies

PO₂₄₀²-PO₂₄₀⁴ in Supplementary Table 2 are monitored as a function of reaction time (Fig. 5 and Supplementary Figs. 51–58). When [MeOH]₀/[MTFA]₀ is 0.75/0.25 (PO₂₄₀²), molar mass increases linearly with PO conversion (Fig. 5a). Ð_M increases when conversion is <50% but slightly decreases when conversion is >50%, finally arriving at 1.13 (Fig. 5b). PO₂₄₀³ shows kinetics similar to PO₂₄₀², with the correlation of molar mass and conversion slightly deviating from linearity (Fig. 5a). When [MeOH]₀/[MTFA]₀ is 0.25/0.75 (PO₂₄₀⁴), molar mass increases rapidly at early stages but plateaus when conversion is >50%, and Ð_M continuously increases until conversion reaches 75% (Fig. 5a, b), indicating that low [MeOH]₀/[MTFA]₀ significantly decreases the uniformity of propagation and prolongs the initiation stage. MTFA is fully consumed at a PO conversion of 96% (Fig. 5c, d at 24 h), confirming 100% initiation efficiency and molar mass control.

Fig. 5: Effects of [MeOH]₀/[MTFA]₀ on kinetics of ECT-ROP.

a, b Dependence of M_n,SEC or Ð_M on the conversion of PO. c–f Time-dependence of the consumption of PO or MTFA. PO₂₄₀² (red points), PO₂₄₀³ (purple points) and PO₂₄₀⁴ (blue points) in Supplementary Table 2.

Figure 5e, f presents the evolution of the consumption of PO and MTFA. In principle, R_p=k_p[PO][ACE][Et₃B]_f, and k_p,obs=k_p[ACE][Et₃B]_f. Considering that [ACE] is roughly equal to [^tBuP₂]₀, [Et₃B]_f can be estimated as [Et₃B]₀-[^tBuP₂]₀, k_p,obs is thus constant and independent of [MeOH]₀/[MTFA]₀, which is confirmed by Fig. 5e. However, the kinetic plots for MTFA in Fig. 5f are not linear. Actually, there are two kinds of ACEs at the initiation stage. The consumption of MTFA is only driven by the ACE during chain propagation (ACE_p) in that the interchange between MTFA and un-initiated ACE (methoxide) do not consume MTFA (Supplementary Fig. 50), entitling the MTFA consumption rate to be R_MTFA=k_ti[ACE_p][MTFA]. Apparently, [ACE_p] is positively correlated with the conversion of MTFA, the slopes of the plots in Fig. 5f thus increase at the early stages. When MTFA is consumed up, only ACE_p exists, which is reasonably estimated as [^tBuP₂]₀ and the same for the three experiments. Therefore, the three plots exhibit similar slopes at late stages irrespective of the feed ratio, and the chain-transfer rate can be generally expressed as R_t=k_t[ACE][TFA].

Kinetics studies are conducted in PO₈₀⁴ and PO₈₀⁶ to insight into the effect of catalyst ratio on propagation and transfer (Supplementary Figs. 57–59). Notably, the fact that k_p,obs of PO₈₀⁶ is 1.84 times as large as that of PO₈₀⁴ (Supplementary Fig. 59e) indicates the first-order dependence of propagation on [Et₃B]_f as k_p,obs=k_p[Et₃B]_f[ACE]. The increase of k_ti,obs (observed rate constant of chain transfer at initiation stage) with [Et₃B]_f deviates from first-order dependence (Supplementary Fig. 59f), which supports the above-proposed kinetics equation depicting a linear combination of first-order and zero-order dependence on [Et₃B]_f, and explains the difference in Ð_M of PO₈₀⁴ and PO₈₀⁶ conducted with varied amount of Et₃B. Accordingly, regulating catalyst ratio is another module for Ð_M control.

Discussion

Chain-transfer reaction can be used to synthesize Ð_M-controlled aliphatic polyethers. The key lies in transesterification on the elaborately selected ester, TFA, as CTA in the ROP of epoxides catalyzed by acid-excess organic Lewis pairs. The rates of transfer and propagation reactions are well-matched and simultaneously editable by simply varying the feed ratio. The TFA groups can be readily removed from the products. Ternary control over Ð_M, molar mass, and end group can hence be achieved in a broad range. The program and equations can be designed to model ECT-ROP and control Ð_M by optimizing the feed ratio. This study provides a facile approach to Ð_M-tailored polymers by fulfilling multidimensionally controlled ROP.

Methods

See Supplementary Information for more details of experimental and modeled methods.

ECT-ROP of EO

A typical procedure is given as follows (EO₁₆₀⁴ in Supplementary Table 1). A dried reaction flask was transferred into a glovebox, where THF (4.0 mL, 49.4 mmol), MeOH (5.1 μL, 0.126 mmol), MTFA (38.1 μL, 0.379 mmol), Et₃B (1.0 M, 50.5 μL, 0.051 mmol), and ^tBuP₂ (2.0 M, 12.5 μL, 0.025 mmol) were successively loaded. The flask was connected to the vacuum line, and purified EO (4.0 mL, 80.0 mmol) was cryo-condensed in the flask at −70 °C ([EO]₀ = 10.0 M). Then the reaction mixture was warmed up to 0 °C, then kept stirring at 0 °C until a large amount of white solid (crystallized PEO) appeared (in ca. 1 h). ¹H NMR (400 MHz, CDCl₃): δ/ppm = 4.51–4.48 (CF₃COOCH₂CH₂O−), 3.82–3.45 (−OCH₂CH₂O−), 3.74–3.71 (−OCH₂CH₂OH), 3.39–3.37 (−OCH₃). M_n,NMR = 7.3 kg mol⁻¹. MeOH (1.0 mL, 24.7 mmol) was added into the flask to release the hydroxyl end groups (60 °C, 4 h), followed by precipitation in cold diethyl ether and vacuum drying overnight at 40 °C. M_n,SEC (THF, PEO standards)=7.2 kg mol⁻¹, Đ_M = 1.74. ¹H NMR (400 MHz, in CDCl₃): δ/ppm = 3.82–3.45 (−OCH₂CH₂O−), 3.74–3.71 (−OCH₂CH₂OH), 3.39–3.37 (−OCH₃). M_n,NMR = 7.3 kg mol⁻¹.

ECT-ROP of PO and AGE

A typical procedure is given as follows (PO₈₀³ in Supplementary Table 2). A dried reaction flask was transferred into a glovebox, where PO (6.0 mL, 85.7 mmol), MeOH (21.7 μL, 0.536 mmol), CF₃COOMe (53.9 μL, 0.536 mmol), Et₃B (1.0 M, 107 μL, 0.107 mmol), and ^tBuP₂ (2.0 M, 26.8 μL, 0.054 mmol) were loaded. The mixture was stirred at room temperature for 4 h. Aliquot (ca. 50 μL) was withdrawn for ¹H NMR and SEC analysis. Conversion of PO was >99%. M_n,SEC (THF, PS standards) = 7.4 kg mol⁻¹, Đ_M = 1.36. ¹H NMR (400 MHz, in CDCl₃): δ/ppm = 5.28–5.19 (CF₃COOCH(CH₃)CH₂O−), 3.97–3.89 (−OCH₂CH(CH₃)OH), 3.64–3.46 (−OCH₂CH(CH₃)O−), 3.46–3.28 (−OCH₂CH(CH₃)O−), 1.35–1.28 (−OCH₂CH(CH₃)OH), 1.18–1.09 (−OCH₂CH(CH₃)O−).