A simple, efficient and selective catalyst for closed-loop recycling of PEF in situ towards a circular materials economy approach

Abstract

Developing plastics from biomass and performing chemical recycling are two essential strategies in circular materials economy. Herein, we present an innovative technique for the closed-loop, in situ chemical recycling of bio-derived poly(ethylene 2,5-furandicarboxylate) (PEF), utilizing the exceptional capabilities of monodisperse nano γ-Ga₂O₃ with tunable oxygen vacancy density. This framework enables seamless cycling of bio-based plastics from polymerization to de-polymerization and re-polymerization, promoting a sustainable polymer economy. The introduction of oxygen vacancy defects in the structure of gallium oxide, a low toxicity and transparent metal oxide, is considered to be an effective strategy for improving catalytic activity. The polymerization process was controlled by using novel oxygen vacancy-defective Ga₂O₃, which catalyzed the reaction between bio-based 2,5-furandicarboxylic acid and ethylene glycol to produce high molecular weight PEF. (M_n = 41 kg mol⁻¹). This PEF can then undergo efficient in situ glycolysis, achieving complete de-polymerization under moderate conditions without the need for external catalysts. The glycolysis derivatives of PEF can be directly re-polymerized to polyester rPEF, achieving a significant molecular weight (M_n = 43 kg mol⁻¹) and a remarkable yield (93%). Notably, γ-Ga₂O₃ with nano oxygen vacancy defects exhibits the ability to selectively de-polymerize PEF within composite material systems containing commercial PET. This research highlights the significant utility of a green catalyst in in situ closed-loop recycling processes.

---

Introduction

Increasing waste volumes and greenhouse gas emissions have demonstrated that the current linear plastics economy, characterized by extraction, usage, and disposal, compromises global sustainability and circularity goals.¹ Enhancing the circularity of plastics is thus essential for transitioning to a future centered on sustainable materials. This necessitates developing effective recycling and reuse strategies. Among these, chemical recycling, which transforms waste polymers into valuable chemicals through processes like solvent decomposition or chemical breakdown, has garnered significant attention.² To date, chemical recycling strategies have been widely applied to polyester plastics.^3–9 These strategies not only provide an effective means to mitigate environmental pollution from plastic waste but also offer a new way to recover valuable chemicals,^10–14 thus creating a recycling process aligned with the vision of circular plastics economy.^15–18

Polyesters, characterized by their polar main-chain ester bonds, already play a crucial role in today's polymer plastics.^19,20 They are poised to become even more significant in a future sustainability-focused circular plastics economy²¹ due to their ability to undergo chemical or biological cleavage of the main-chain ester bond on demand, offering enhanced recycling options for discarded plastics.^22,23 In addition to chemical recycling, developing polymer materials derived from biomass resources represents another effective strategy for achieving sustainable resource utilization as part of a sustainable economy. Among the promising bio-based polymers, poly(ethylene 2,5-furandicarboxylate) (PEF) stands out due to its low permeability,²⁴ high-performance properties²⁵ and compatibility with existing processing and recycling infrastructures.²⁶ Thus, PEF is expected to be a more sustainable alternative to traditional fossil-based aromatic polyesters such as PET.^27,28

The preparation and recycling of PEF are crucial for its future applications. A range of catalysts have been reported to support PEF production, including tetrabutyl titanate,^29,30 titanium(IV) isopropoxide,³¹ tin(II) and alkyl tin(IV) salts,^26,32 antimony oxide,³³ zinc acetate and aluminum acetylacetonate.³⁴ Although effective, these catalysts have notable drawbacks. Titanium-based catalysts exhibit poor stability and produce darker-colored products, while organotin compounds are highly toxic, limiting a material's further application in biomedicine and applications involving food contact.^35,36 Among various catalysts, metal oxide catalysts are extensively utilized in transesterification reactions due to their high catalytic activity, stability, and strong resistance to hydrolysis. Research has demonstrated that the acidic sites on metal oxide catalysts, such as metal cations, can readily interact with hydroxyl oxygen and ester bond oxygen to form multi-ring structures. These multi-ring structures exhibit strong coordination capabilities, which can redistribute the charges on their groups, rendering the ester carbonyl carbon positively charged. This positive charge accelerates nucleophilic reactions between groups, thereby enhancing the reaction rate.³⁷ Oxygen vacancy defects, a common type of point defect, are widely present in metal oxides. They influence the local geometric and electronic structure of the material, generating acidic sites that serve as numerous catalytic active sites. By modulating the density of oxygen vacancy defects on the surface of metal oxides, it is possible to quantitatively control the number of surface acidic sites, thereby regulating the activity of polymerization reactions. Thus, the abundant acidic sites and oxygen vacancy defects on the surface of metal oxides offer a promising means for achieving efficient synthesis and de-polymerization of PEF.

Research on chemical recycling strategies for PEF is increasing annually, with processes such as methanolysis,^6,38 glycolysis^39,40 and hydrolysis^41,42 being reported. Among these, glycolysis is favored due to its mild reaction conditions and the ease of separating the target product bis(hydroxyethyl)-2,5-furandicarboxylate (BHEFDC), which can be used for re-polymerization. Gabirondo et al.³⁹ reported the glycolysis of commercial PEF (M_n = 12.4 kg mol⁻¹) using an organocatalyst system (DBU : benzoic acid = 1 : 1) at 180 °C with 5 wt% catalyst, achieving 92% conversion to BHEFDC and some BHEFDC oligomers in 5 h. The purified BHEFDC was re-polymerized via melt polycondensation and solid-state polymerization, producing rPEF with M_n = 11.2 kg mol⁻¹. Agostinho et al.⁴⁰ used a eutectic solvent (DES, urea and Zn(OAc)₂ = 4 : 1) to recycle low molecular weight PEF (catalyzed by titanium butoxide, viscosity 0.268 dl g⁻¹). At 180 °C with 2 wt% DES, 85% glycolysis conversion to BHEFDC was achieved in 1 hour. Without purification, rPEF (viscosity 0.409 dl g⁻¹) was obtained via direct re-polymerization using vacuum and heat, with DES still present. However, current de-polymerization studies of polyesters typically require substantial additional amounts of specific catalysts, which hinder catalyst recycling within the closed-loop system, complicate the overall process, and lead to increased costs. This issue is likely due to the disparity in activity and selectivity between the catalysts used for polyester synthesis and those used for de-polymerization.

In this work, we present a novel in situ closed-loop recycling strategy for bio-based PEF, leveraging the potential of monodisperse γ-Ga₂O₃ with tunable oxygen vacancy defect density as an innovative and highly efficient green catalyst. The chemistry described here is inspired by the catalytic versatility of Ga₂O₃, which includes its low toxicity,^43–48 transparent properties,⁴⁹ and low Ga–O bond dissociation energy, facilitating the formation of controllable oxygen-vacancy defects and potential catalytic activity.⁵⁰ This catalytic system enables controlled polymerization and efficient de-polymerization of PEF in a continuous cycle, whilst promoting its re-polymerization into rPEF. Experimental results and DFT calculations confirm the feasible reaction pathways for both PEF synthesis and glycolysis on the catalyst surface. Additionally, under this strategy, PEF can also undergo selective glycolysis within mixed polyester systems, leading to a truly practical sustainable circular economy approach.

Results and discussion

Synthesis and characterization of nano γ-Ga₂O₃ with oxygen vacancy defects

We synthesized sub-stable Ga₂O₃via thermal injection and calcination using anhydrous gallium chloride, ethylene glycol, and deionized water (Fig. 1A). The experimental details are described in the ESI.† This method is simpler, more cost-effective, and more suitable for large-scale industrial production compared to hydrothermal methods. X-ray powder diffraction (XRD) was performed on Ga₂O₃ catalyst samples synthesized via 3 h, 6 h, and 12 h reflux reactions to study their crystal structures. As depicted in Fig. 1B, all Ga₂O₃ catalysts exhibited similar XRD patterns, with characteristic diffraction peaks at 36.2° and 64.1°, corresponding to the (311) and (440) crystals of cubic spinel-type γ-Ga₂O₃ with defects (JCPDS#20-0426).⁵¹ This similarity suggests that the duration of the reflux reaction does not significantly affect the crystal structure. The broad, well-defined peaks indicate that the synthesized γ-Ga₂O₃ exhibits low crystallinity and consists of small-sized particles or very tiny microcrystals. The solution exhibited a Tyndall effect under UV light at the reaction endpoint (Fig. S1 in the ESI†), indicating the presence of small γ-Ga₂O₃ particles. To study nanoparticle morphology, transmission electron microscopy (TEM) was employed for direct observation. TEM images (Fig. 1C–E) reveal that γ-Ga₂O₃-n, where n represents the reflux reaction time in hours, is monodisperse and granular with irregular edges, with particle sizes ranging from a few nanometers to several dozen nanometers.

Fig. 1 (A) Preparation process of Ga₂O₃ with oxygen vacancy defects; (B) XRD patterns of γ-Ga₂O₃-n and commercial Ga₂O₃ catalysts (n represents the reflux reaction time in hours); and TEM images of (C) γ-Ga₂O₃-3, (D) γ-Ga₂O₃-6, and (E) γ-Ga₂O₃-12.

Oxygen vacancies in the γ-Ga₂O₃ samples were detected during synthesis via electron paramagnetic resonance (EPR) analysis at low temperatures. The strong EPR signals in γ-Ga₂O₃-n samples at g = 2.003 are characteristic of oxygen vacancies with trapped electrons (Fig. 2C).⁵² A stronger characteristic signal of single-electron captured oxygen vacancies (O_V) was noted in γ-Ga₂O₃-6 compared to γ-Ga₂O₃-3 and γ-Ga₂O₃-12. However, γ-Ga₂O₃-3 and γ-Ga₂O₃-12 exhibited similar EPR intensity levels.

Fig. 2 (A) O 1s spectra, (B) Ga 2p spectra, (C) EPR spectra of γ-Ga₂O₃-n and β-Ga₂O₃ catalysts and (D) NH₃-TPD profiles of γ-Ga₂O₃-n catalysts (n represents the reflux reaction time in hours).

We analyzed catalyst samples for oxygen, defects, and metal states using X-ray photoelectron spectroscopy (XPS). Distinct Ga 2p and O 1s signals were identified in the catalyst samples, with no impurity peaks detected. The XPS spectra of O 1s (Fig. 2A) in the γ-Ga₂O₃-n samples can be split into three peaks: lattice oxygen (O_L, 530.6 eV), vacant oxygen (O_V, 532.2 eV), and hydroxyl oxygen (O_OH−, 531.5 eV).⁵³ The peak area ratio of O_V to O_L in γ-Ga₂O₃-6 is 0.59, which is higher than those in γ-Ga₂O₃-3 (0.36) and γ-Ga₂O₃-12 (0.41) (Table S1 in the ESI†). This indicates a higher concentration of oxygen vacancy defects in γ-Ga₂O₃-6, which aligns well with the results from the EPR spectra. In addition, the ratios of O_OH− to O_V in γ-Ga₂O₃ samples—specifically, γ-Ga₂O₃-3 (0.75), γ-Ga₂O₃-6 (0.38), and γ-Ga₂O₃-12 (0.60) —exhibit a distinct trend. This variation indicates a decrease in the O_OH− content and a substantial increase in O_V, suggesting that thermal treatment induces a dehydroxylation reaction, leading to the formation of oxygen-vacancy defects within the gallium oxide structure. However, no information on the O_V was observed in either the XPS or EPR spectrum of the commercial Ga₂O₃ sample, suggesting that the commercial Ga₂O₃ has no capacity for bearing surface O_V (Fig. 2A and C). This leads to its poor catalytic performances. Furthermore, the O_V peak of the γ-Ga₂O₃-n material with oxygen vacancy defects exhibits a considerable blue-shift in its binding energy compared to β-Ga₂O₃ with oxygen vacancy-free defects (Fig. 2A). This shift can be attributed to the significant influence of oxygen vacancies on the electron density and chemical potential of the O_L sites.⁵⁴ The interaction between the O_V and the surrounding gallium and oxygen atoms alters the electronic properties of the material, resulting in the observed blue-shift in the O_L peak's binding energy. This alteration in binding energy can significantly impact the catalytic activity, as it is directly related to the ease of electron transfer in chemical reactions. Additionally, the observed blue-shift indicates that the γ-Ga₂O₃-n material with oxygen vacancy defects may exhibit enhanced electron-attracting properties and acid strength compared to its counterparts, making it a promising candidate as a catalyst for polyester synthesis.

The Ga 2p_3/2 and Ga 2p_1/2 XPS peaks of β-Ga₂O₃ are centered at binding energies of 1117.6 eV and 1144.4 eV, respectively (Fig. 2B), which is characteristic of Ga³⁺–O bonds in Ga₂O₃.⁵⁵ In contrast, the Ga 2p XPS spectra of γ-Ga₂O_3−x with non-stoichiometric defects exhibit different features, where the Ga 2p_3/2 and Ga 2p_1/2 peaks shift to higher binding energies at 1118.0 eV and 1144.8 eV, respectively, representing a shift of 0.4 eV (Table S2 in the ESI†). Upon conducting a thorough analysis, it is reasonable to deduce that the positive shift in the Ga 2p energy level is attributable to the significant interplay between Ga³⁺ and oxygen vacancies. This interaction pattern is analogous to the lattice contraction induced by oxygen vacancies in titanium dioxide (TiO₂). Specifically, in TiO₂ crystals, the presence of oxygen vacancies induces a contraction in the lattice structure, facilitated by the strong interaction between oxygen vacancies and Ti⁴⁺ ions.^56–59 This lattice adjustment subsequently influences the catalytic properties of the material, including photocatalytic efficiency and electron mobility. Analogously, in γ-Ga₂O₃, the removal of an O atom causes the nearest Ga atoms to relax away from the vacancy, strengthening their bonding with the surrounding lattice. This outward relaxation decreases the overlap of the Ga dangling bonds and shortens the Ga–O bond length. The Ga³⁺–O bond binding energy increases to maintain the crystal structure and stability, potentially impacting the material's catalytic performance. Therefore, a thorough investigation into the mechanism behind the Ga 2p level shift is crucial for enhancing the performance and application potential of Ga₂O₃ catalysts.

The surface acidity of the γ-Ga₂O₃ catalyst was measured using temperature-programmed desorption of NH₃ (NH₃-TPD). All three γ-Ga₂O₃-n catalysts exhibited two distinct NH₃ desorption peaks. The first peak, observed in the low temperature range (298 K to 473 K), corresponded to weak acid centers. The second peak, detected in the medium temperature range (473 K to 673 K), indicated medium–strong acidic centers (Fig. 2D). This suggests that γ-Ga₂O₃-n catalysts possess both weak and medium–strong acidic sites on their surfaces. The surface acidity of the γ-Ga₂O₃ catalysts can be effectively controlled and adjusted by varying the duration of the reflux reaction (Table S3 in the ESI†), and the total acidic site abundance followed a specific order: Ga₂O₃-6 > Ga₂O₃-12 > Ga₂O₃-3. The highest concentration of surface acidic sites in Ga₂O₃-6 is attributed to its robust capacity for single-electron captured oxygen vacancies, as predicted in our previous analyses.

A thermal gravimetric analyzer (TGA) was used to assess the thermal stability of the catalysts by measuring the percentage of mass loss at high temperatures (Fig. S2 in the ESI†). The TGA curves of the three catalysts show minimal mass loss between 30 °C and 600 °C, indicating that the nano γ-Ga₂O₃ catalysts with defects exhibit high thermal stability below 600 °C with negligible heat loss. Notably, the TGA curves display a sharp decline after 600 °C, corresponding to the phase transition from γ-Ga₂O₃ to β-Ga₂O₃.⁶⁰ These results demonstrate that the synthesized γ-Ga₂O₃ catalysts maintain stability under the thermal conditions required for polyester synthesis.

In situ catalytic PEF closed-loop chemical recovery performance

The γ-Ga₂O₃-n catalysts were employed for the polymerization of renewable 2,5-furandicarboxylic acid (FDCA) and ethylene glycol (EG) through a direct esterification method that involves a two-step process: esterification followed by polycondensation. The esterification was performed under a nitrogen atmosphere at temperatures of 190 °C for 1 h, 200 °C for 1 h, and 210 °C for 1 h, respectively. Subsequently, the polycondensation was conducted at 240 °C for 3 hours under high vacuum to yield PEF. The results of γ-Ga₂O₃-6 with oxygen-vacancy defects and the conventional polyester catalyst tetrabutyl titanate (TBT) in synthesizing PEF are shown in Fig. 3A and B and Table 1. The ¹H NMR spectra of PEFs revealed chemical shifts at 4.69 ppm (c) and 4.17 ppm (d), attributed to the methylene proton in the diethylene glycol furandicarboxylate (DEGF) unit of PEF (Fig. 3A). This unit, a by-product of etherification side reactions involving EG or hydroxyethyl ester end-groups, has its molar fraction (φ_DEGF) calculated accordingly (eqn (1)). Consequently, the catalysts γ-Ga₂O₃-6 and TBT can be found to exhibit similar selectivity (φ_DEGF, 4.90% vs. 4.40%) in catalyzing the PEF synthesis reaction. The FTIR spectrum shown in Fig. 3C confirmed the expected PEF structure, with characteristic peaks evident for the furan's =CH stretching at 3116 cm⁻¹, the ester carbonyl C=O stretching vibration at 1720 cm⁻¹, and the C–O stretching of the ester carbonyl group at 1265 cm⁻¹. Differential scanning calorimetry (DSC) and TGA results (Table 1 and Fig. S4 in the ESI†) indicated that the thermal properties of PEF synthesized using γ-Ga₂O₃-6 as the catalyst, such as the glass transition temperature (T_g), 5% degradation temperature (T_d-5%), and maximum degradation temperature (T_d-max), are comparable to those of PEF synthesized using TBT as the catalyst. It is noteworthy that PEF catalyzed by γ-Ga₂O₃-6 exhibited lower yellowness and higher transmittance (Fig. 3D), suggesting that γ-Ga₂O₃-6 not only achieves slightly better catalytic activity compared to TBT but also significantly ameliorates the yellow-brown coloration issue of PEF products, resulting in a product with enhanced optical properties.

φ_DEGF = A_d/(A_d + A_b) × 100% (1)

where A_b represents the areas of the b peaks in the NMR spectra of PEF and A_d represents the areas of the d peaks in the NMR spectra of PEF.

Fig. 3 (A) ¹H NMR of PEFs catalyzed by TBT, nano γ-Ga₂O₃-n and β-Ga₂O₃ with oxygen-vacancy defects (the inset is a structural formula of PEF); (B) ¹H NMR local magnification (4.7–4.0 ppm) of PEFs; (C) FTIR spectrum of PEF catalyzed by nano γ-Ga₂O₃ with oxygen-vacancy defects; and (D) appearance of the synthesized PEFs catalyzed by γ-Ga₂O₃ and TBT, respectively.

Table 1 Comparison of catalytic performances of nano γ-Ga₂O₃-n with oxygen-vacancy defects, TBT and β-Ga₂O₃ in PEF synthesis^a

Samples	Catalysts	Concentration of catalyst (ppm)	M n (kg mol^−1)	Đ	φ DEGF (%)	T g ( °C)	T d-5% ( °C)	T d-max ( °C)	Yield (%)
a Reaction conditions: transesterification stage, under a nitrogen atmosphere at temperatures of 190 °C for 1 h, 200 °C for 1 h, and 210 °C for 1 h, respectively; polycondensation stage, 240 °C for 3 h under high vacuum; molar ratio of EG to FDCA was 1.7 : 1. b Đ: polydispersity index of PEFs determined by gel permeation chromatography. c Diethylene glycol furandicarboxylate (DEGF) unit content in PEF determined by ^1H NMR and calculated by using eqn (1). d TBT: traditional polyester catalyst tetrabutyl titanate. e β-Ga2O3: commercial gallium oxide with oxygen vacancy-free defects.
PEF-1	TBT^d	1600	38	1.77	4.40	86.3	375.5	417.3	87.5
PEF-2	γ-Ga2O3-3	880	35	1.60	5.88	84.6	370.6	425.6	85.3
PEF-3	γ-Ga2O3-6	880	41	1.79	4.90	86.0	374.1	424.4	88.1
PEF-4	γ-Ga2O3-12	880	39	1.81	5.37	85.8	370.2	424.1	85.1
PEF-5	β-Ga2O3^e	880	24	1.46	6.00	80.0	373.5	418.0	83.0

---

To investigate the relationship between the density of oxygen vacancy defects in γ-Ga₂O₃ and its catalytic activity, we compared the performance of γ-Ga₂O₃-n catalysts with commercial β-Ga₂O₃, which lacks oxygen-vacancy defects (Fig. 1B, 2A and C). All γ-Ga₂O₃-n catalysts demonstrated activity (Fig. 3A, B and Table 1). The number average molecular weights (M_n) of PEF catalyzed by γ-Ga₂O₃-3, γ-Ga₂O₃-6, and γ-Ga₂O₃-12 were 35 kg mol⁻¹, 41 kg mol⁻¹, and 39 kg mol⁻¹, respectively. This trend aligns with the initial increase and subsequent decrease in the density of surface oxygen vacancy defects and the concentration of surface acidic sites on γ-Ga₂O₃, demonstrating a positive correlation between these two characteristics of the catalyst and its catalytic activity, as previously predicted. Notably, the M_n of PEF catalyzed by β-Ga₂O₃ was significantly lower than that achieved by γ-Ga₂O₃-n catalysts, attributable to the lower surface acidity and absence of oxygen vacancies in β-Ga₂O₃ (Fig. 2A and Table S3 in the ESI†). This result suggests that oxygen-vacancy defects can indeed increase the number of acidic sites on nano-oxides, thereby enhancing catalytic activity. Among the γ-Ga₂O₃-n catalysts, γ-Ga₂O₃-6, which is monodisperse and has the highest surface acidity, exhibits the best catalytic activity.

In the catalytic glycolysis study, we conducted an in situ glycolysis of PEF, synthesized using the γ-Ga₂O₃-6 catalyst. This approach obviates the requirement for additional catalysts during the glycolysis of the polyester material in ethylene glycol solution, enhancing process efficiency. We discovered that in situ glycolysis of PEF can be efficiently facilitated by using trace amounts of the nano γ-Ga₂O₃ catalyst with residual oxygen-vacancy defects in PEF at 180 °C for 2 h under atmospheric pressure. A comparison of the reaction solution before and after glycolysis reveals that the PEF powder/glycol mixture became clear and transparent after 2 h of reaction, indicating the near-complete de-polymerization of PEF powder (Fig. S6 in the ESI†). To determine the degradation efficiency, we analyzed the original solution after the glycolysis reaction using high-performance liquid chromatography (HPLC). Fig. S10† confirms that trace amounts of γ-Ga₂O₃ (829 ppm, see Table S5†) facilitated the complete glycolysis of PEF, resulting in a yield of 75.7% bis(hydroxyethyl)-2,5-furandicarboxylate (BHEFDC) and 24.3% BHEFDC dimer.

The de-polymerization product was further processed by cooling and filtering twice, resulting in the separation of two distinct solids: solid I and solid II. In particular, upon completion of the glycolysis reaction, the mixture was rapidly cooled to quench the reaction. The resulting precipitate was stored at −10 °C for a period of time, then filtered and washed with deionized water to obtain solid I, and solid I can be completely dissolved in DMSO (Fig. S7 in the ESI†). The filtrate was then concentrated using a rotary evaporator and stored at −10 °C, and the subsequent precipitate was filtered and washed to obtain solid II. In addition, the absence of ¹H NMR signals corresponding to PEF, coupled with the appearance of new singlet peaks at δ 4.60 ppm and 4.14 ppm attributed to the methylene protons of the terminal group derived from ethylene glycol (Fig. 4A), indicates complete de-polymerization of PEF. The signal in the ¹H NMR spectrum at δ 4.74–4.80 ppm corresponds to the chemical shifts of four protons associated with the two methylene groups located at the center of the 2,5-furandicarboxylate dimer, indicating that solid I is composed of oligomers. The average degree of solid I was determined to be 2, based on the peak area in the ¹H NMR spectrum (Fig. 4A). In contrast, the ¹H NMR peak distribution of solid II resembles that of solid I. However, based on NMR calculations, the average degree of polymerization for solid II is found to range between 1 and 2. The glycolysis products of PEF were further evaluated in the ESI, with HPLC-MS results presented in Fig. S8 and S9, and summarized in Table S4 in the ESI.† While this work primarily focuses on developing a new strategy for in situ closed-loop recovery of PEF, the recovery of BHEFDC following glycolysis was also achieved through simple refrigeration and filtration procedures. These findings further underscore the exceptional efficiency of γ-Ga₂O₃ with defects in the in situ glycolysis of PEF.

Fig. 4 (A) ¹H NMR spectra of PEF glycolytic products of solid I and solid II and (B) ¹H NMR spectra of the rPEF and PEF.

Given the excellent performance of γ-Ga₂O₃ with defects in catalyzing both PEF polymerization and glycolysis, we utilized it for the in situ re-polymerization recycling of PEF to its original polymer form (rPEF). We conducted the glycolysis reaction of ground PEF (2 g) with EG (8 g) in a three-necked round-bottomed flask. Without isolating the main intermediate glycolysis product, bis(hydroxyethyl)-2,5-furandicarboxylate (BHEFDC), and removing the γ-Ga₂O₃ catalyst, we utilized the unpurified reaction mixture and the residual of γ-Ga₂O₃ to catalyze the polymerization of BHEFDC to rPEF. The re-polymerized product was identified as PEF, exhibiting a promising M_n of 43 kg mol⁻¹ (Fig. 4B and Table S6 in the ESI†). Additionally, the thermal properties of rPEF show no degradation and are nearly identical to those of the original PEF polymer (Fig. S11 and Table S4 in the ESI†). The efficacy of this innovative in situ closed-loop recycling method is demonstrated by its seamless transitions from polymerization to de-polymerization and subsequent re-polymerization.

Given that PET is a major plastic present in waste streams and subject to recycling, the ability to efficiently separate and recycle PEF from PET is crucial to prevent cross-contamination of either recycling stream. Equal amounts of commercial PET and PEF synthesized using γ-Ga₂O₃-6 were mixed and subjected to glycolysis under the conditions of 180 °C and atmospheric pressure, for 2 h. Notably, PEF can be depolymerized completely to its monomer, while PET is retained as large white particles. The PET granules were collected, rinsed with deionized water and anhydrous ethanol, and then dried under vacuum at 60 °C for 12 h (Scheme 1). Post-reaction gravimetric measurements indicated no significant mass change in PET, confirming its resistance to de-polymerization by the γ-Ga₂O₃-6 catalyst under the specified glycolysis conditions. In contrast, PEF was completely depolymerized, as evidenced by the total disappearance of PEF particles from the reaction mixture. The distinct reactivity difference between PEF and PET under identical glycolysis conditions highlights the selective catalytic efficiency of γ-Ga₂O₃ with defects for the PEF glycolysis reaction. The unique catalytic properties of oxygen vacancy-deficient γ-Ga₂O₃ offer a viable solution to enhance the sustainability of complex polymer system recycling processes.

Scheme 1 Glycolysis process of PEF catalyzed by nano γ-Ga₂O₃ with defects and a commercial PET polyester mixture.

Density functional theory (DFT) calculation on the glycolysis reaction of PEF

To investigate the glycolysis pathway of PEF in the presence of the γ-Ga₂O₃ catalyst with oxygen-vacancy defects, we constructed a model of nano-Ga₂O₃ surfaces with oxygen-vacancy defects using DFT calculations (Fig. S12 in the ESI†). Ethane-1,2-diyl bis(furan-2-carboxylate) (EDBFC) was used as model molecules to simplify the calculation process (Fig. S13 in the ESI†). As depicted in Fig. 5, EDBFC initially approaches the catalyst surface to form *C₁₂H₁₀O₆ (steps A1 to B1), with an adsorption energy of −42.49 kJ mol⁻¹. Subsequently, *C₂H₆O₂ forms on the catalyst surface through the interaction with ethylene glycol (steps B1 to C1), with an adsorption energy of −52.12 kJ mol⁻¹. During steps C1 to D1, a nucleophilic substitution reaction occurs on the surface of the nano-catalyst with defects, reinforced by the coordination between the strongly acidic Ga³⁺ ions and the carbonyl oxygen of the ester bond in the EDBFC. This facilitates the nucleophilic attack of EG on the carbonyl carbon, forming a tetrahedral transition state (D1) from TS 1. The departure of Ga facilitates the protonation of the carbonyl group in TS 2, leading to the formation of 2-hydroxyethyl carboxyfuran (steps D1 to E1). The energy barrier for the formation of TS 1 is 146.56 kJ mol⁻¹, while that for TS 2 is 110.53 kJ mol⁻¹, indicating that the formation of TS 1 is the rate-determining step of the reaction. Ga³⁺ ions located at the oxygen vacancy defects on the nano-sized Ga₂O₃ surfaces serve as the active sites for catalyzing the glycolysis reaction of PEF. The formation of these active sites is further enhanced by electrostatic interactions between the Ga³⁺ ions and the oxygen-deficient sites. This conclusion is derived from a detailed analysis of the electronic structure and bonding configurations altered by the presence of these ions on the nanostructured material surfaces. The intentionally created oxygen vacancy defects on the Ga₂O₃ surfaces provide a unique environment where gallium ions are more readily available to interact with reactant molecules, creating a favorable microenvironment for the adsorption and activation of PEF molecules, which enhances the affinity of Ga³⁺ to PEF glycolytic substrates, facilitating the chemical transformations required for the breakdown of PEF into simpler compounds. The same mechanism explains the pathway of Ga₂O₃ in catalyzing the glycolysis of PET (see Fig. S14 and S15 in the ESI†) and its high activity in the catalytic polymerization of PEF (see Fig. S16 and S17 in the ESI†). Finally, based on DFT calculations, it was found that the adsorption energy of H₂O on Ga₂O₃ in the reaction system is 0.302 eV, indicating that the likelihood of adsorption is low and confirming that the interaction between H₂O and Ga is sufficiently weak, implying that water will not poison the reaction medium. Additionally, the adsorption energies of Ga₂O₃ with oxygen in furan, hydroxyl oxygen in EG, and carbonyl oxygen are −0.148 eV, −0.305 eV, and −0.583 eV, respectively (see Table S7 in the ESI†), indicating that carbonyl oxygen has the strongest affinity. These results further support our proposed mechanism, suggesting that the C=O–Ga interaction is the most favorable.

Fig. 5 (A) Modelled activation mechanism for the glycolysis of PEF, using ethane-1,2-diyl bis(furan-2-carboxylate) to mimic PEF, ethylene glycol as a nucleophile, and nano-Ga₂O₃ with oxygen-vacancy defects as the catalyst, yielding 2-hydroxyethyl carboxyfuran as products; (B) energy curve and reaction structure formula of the ethane-1,2-diyl bis(furan-2-carboxylate) glycolysis reaction pathway on the nano-Ga₂O₃ surface with oxygen-vacancy defects.

Conclusion

In this study, we have demonstrated that Ga₂O₃ monodisperse nanoparticles with oxygen vacancy defects γ- can facilitate an innovative in situ closed-loop chemical recycling process for bio-based PEF. This process achieves a seamless transition from catalytic polymerization to de-polymerization and re-polymerization. Modulating the oxygen vacancy density of γ-Ga₂O₃ enables highly active, controllable polymerization reactions, resulting in satisfactory PEF appearance. We designed an in situ-assisted glycolysis reaction using γ-Ga₂O₃ nanoparticles, achieving 100% conversion of PEF into monomers that meet the purity standards required for polymer synthesis. This process eliminates the need for additional specific catalysts or energy-intensive purification steps. Furthermore, the removal of γ-Ga₂O₃ is unnecessary during the transition from PEF to rPEF. The integration of this strategy with the non-toxic γ-Ga₂O₃ green catalyst aligns seamlessly with circular plastics economy goals and exhibits potential applicability to other polyesters. Notably, these nanoparticles can selectively catalyze the glycolysis of PEF in mixed plastics streams, paving new pathways for catalyst design and sustainable polymer recycling. This advancement significantly promotes circular economy within the polymer industry.

Author contributions

Shaowei Wu: conceptualization, data curation, formal analysis, investigation, methodology, validation, visualization, writing – original draft, and writing – review and editing. Lu Li: funding acquisition, resources, supervision, and writing – review and editing. Lei Song: investigation, visualization, and writing – review and editing. Guannan Zhou: investigation and writing – review and editing. Lixin Liu: validation, and writing – review and editing. Hailan Kang: visualization and writing – review and editing. Rui Wang: conceptualization, funding acquisition, project administration, resources, supervision, and writing – review and editing. Guangyuan Zhou: funding acquisition, project administration, and resources.

Data availability

All data generated or analyzed during this study are included in this published article and its ESI.†

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This study was supported by the National Key Research and Development Programme of China (2022YFC2105800), the Science and Technology Programme of Liaoning Province (No. 2022JH24/10200024) and the Medical-Industrial Joint Innovation Fund (No. DMU-1 & DICP UN202210).

References

1. 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.
2. C. Shi, E. C. Quinn, W. T. Diment and E. Y. X. Chen, Chem. Rev., 2024, 124, 4393–4478.
3. Q. Dong, A. D. Lele, X. Zhao, S. Li, S. Cheng, Y. Wang, M. Cui, M. Guo, A. H. Brozena, Y. Lin, T. Li, L. Xu, A. Qi, I. G. Kevrekidis, J. Mei, X. Pan, D. Liu, Y. Ju and L. Hu, Nature, 2023, 616, 488–494.
4. S. Zhang, Q. Hu, Y. Zhang, H. Guo, Y. Wu, M. Sun, X. Zhu, J. Zhang, S. Gong, P. Liu and Z. Niu, Nat. Sci., 2023, 6, 965–973.
5. F. Wu, Y. Wang, Y. Zhao, M. Tang, W. Zeng, Y. Wang, X. Chang, J. Xiang, B. Han and Z. Liu, Sci. Adv., 2023, 9, eade7971.
6. X. Qu, G. Zhou, R. Wang, B. Yuan, M. Jiang and J. Tang, Green Chem., 2021, 23, 1871–1882.
7. S. Westhues, J. Idel and J. Klankermayer, Sci. Adv., 2018, 4, eaat9669.
8. L. Monsigny, J.-C. Berthet and T. Cantat, ACS Sustainable Chem. Eng., 2018, 6, 10481–10488.
9. Q. Wang, X. Yao, Y. Geng, Q. Zhou, X. Lu and S. Zhang, Green Chem., 2015, 17, 2473–2479.
10. L. T. J. Korley, T. H. Epps, B. A. Helms and A. J. Ryan, Science, 2021, 373, 66–69.
11. A. Kulkarni, G. Quintens and L. M. Pitet, Macromolecules, 2023, 56, 1747–1758.
12. M. D. de Dios Caputto, R. Navarro, J. L. Valentín and Á. Marcos-Fernández, J. Polym. Sci., 2022, 60, 3269–3283.
13. P. D. Goring and R. D. Priestley, JACS Au, 2023, 3, 2609–2611.
14. W. Zeng, Y. Zhao, F. Zhang, R. Li, M. Tang, X. Chang, Y. Wang, F. Wu, B. Han and Z. Liu, Nat. Commun., 2024, 15, 160.
15. M. Hong and E. Y. X. Chen, Green Chem., 2017, 19, 3692–3706.
16. G. W. Coates and Y. D. Y. L. Getzler, Nat. Rev. Mater., 2020, 5, 501–516.
17. D. E. Fagnani, J. L. Tami, G. Copley, M. N. Clemons, Y. D. Y. L. Getzler and A. J. McNeil, ACS Macro Lett., 2021, 10, 41–53.
18. F. M. Lamberti, L. A. Román-Ramírez and J. Wood, J. Polym. Environ., 2020, 28, 2551–2571.
19. M. Rabnawaz, I. Wyman, R. Auras and S. Cheng, Green Chem., 2017, 19, 4737–4753.
20. R. Nisticò, Polym. Test., 2020, 90, 106707.
21. C. Jehanno, J. W. Alty, M. Roosen, S. De Meester, A. P. Dove, E. Y. X. Chen, F. A. Leibfarth and H. Sardon, Nature, 2022, 603, 803–814.
22. L. Cederholm, P. Olsén, M. Hakkarainen and K. Odelius, Polym. Chem., 2023, 14, 3270–3276.
23. V. Tournier, S. Duquesne, F. Guillamot, H. Cramail, D. Taton, A. Marty and I. André, Chem. Rev., 2023, 123, 5612–5701.
24. M. R. B. Mermet-Guyennet, J. Gianfelice de Castro, H. S. Varol, M. Habibi, B. Hosseinkhani, N. Martzel, R. Sprik, M. M. Denn, A. Zaccone, S. H. Parekh and D. Bonn, Polymer, 2015, 73, 170–173.
25. S. K. Burgess, J. E. Leisen, B. E. Kraftschik, C. R. Mubarak, R. M. Kriegel and W. J. Koros, Macromolecules, 2014, 47, 1383–1391.
26. K. Loos, R. Zhang, I. Pereira, B. Agostinho, H. Hu, D. Maniar, N. Sbirrazzuoli, A. J. D. Silvestre, N. Guigo and A. F. Sousa, Front. Chem., 2020, 8, 585.
27. J.-G. Rosenboom, D. K. Hohl, P. Fleckenstein, G. Storti and M. Morbidelli, Nat. Commun., 2018, 9, 2701.
28. A. F. Sousa, R. Patrício, Z. Terzopoulou, D. N. Bikiaris, T. Stern, J. Wenger, K. Loos, N. Lotti, V. Siracusa, A. Szymczyk, S. Paszkiewicz, K. S. Triantafyllidis, A. Zamboulis, M. S. Nikolic, P. Spasojevic, S. Thiyagarajan, D. S. van Es and N. Guigo, Green Chem., 2021, 23, 8795–8820.
29. G. Z. Papageorgiou, V. Tsanaktsis and D. N. Bikiaris, Phys. Chem. Chem. Phys., 2014, 16, 7946–7958.
30. A. Pellis, K. Haernvall, C. M. Pichler, G. Ghazaryan, R. Breinbauer and G. M. Guebitz, J. Biotechnol., 2016, 235, 47–53.
31. S. Thiyagarajan, W. Vogelzang, R. J. I. Knoop, A. E. Frissen, J. van Haveren and D. S. van Es, Green Chem., 2014, 16, 1957–1966.
32. Z. Terzopoulou, E. Karakatsianopoulou, N. Kasmi, M. Majdoub, G. Z. Papageorgiou and D. N. Bikiaris, J. Anal. Appl. Pyrolysis, 2017, 126, 357–370.
33. L. Papadopoulos, A. Zamboulis, N. Kasmi, M. Wahbi, C. Nannou, D. A. Lambropoulou, M. Kostoglou, G. Z. Papageorgiou and D. N. Bikiaris, Green Chem., 2021, 23, 2507–2524.
34. M. B. Banella, J. Bonucci, M. Vannini, P. Marchese, C. Lorenzetti and A. Celli, Ind. Eng. Chem. Res., 2019, 58, 8955–8962.
35. G. J. Gruter, L. Sipos and M. A. Dam, Comb. Chem. High Throughput Screening, 2012, 15, 180–188.
36. Z. Terzopoulou, E. Karakatsianopoulou, N. Kasmi, V. Tsanaktsis, N. Nikolaidis, M. Kostoglou, G. Z. Papageorgiou, D. A. Lambropoulou and D. N. Bikiaris, Polym. Chem., 2017, 8, 6895–6908.
37. I. Shigemoto, T. Kawakami, H. Taiko and M. Okumura, Polymer, 2011, 52, 3443–3450.
38. C. Alberti, K. Matthiesen, M. Wehrmeister, S. Bycinskij and S. Enthaler, Chem. Sel., 2021, 6, 7972–7975.
39. E. Gabirondo, B. Melendez-Rodriguez, C. Arnal, J. M. Lagaron, A. Martínez de Ilarduya, H. Sardon and S. Torres-Giner, Polym. Chem., 2021, 12, 1571–1580.
40. B. Agostinho, A. J. D. Silvestre and A. F. Sousa, Green Chem., 2022, 24, 3115–3119.
41. F. Kawai, Y. Furushima, N. Mochizuki, N. Muraki, M. Yamashita, A. Iida, R. Mamoto, T. Tosha, R. Iizuka and S. Kitajima, AMB Express, 2022, 12, 134.
42. H. P. Austin, M. D. Allen, B. S. Donohoe, N. A. Rorrer, F. L. Kearns, R. L. Silveira, B. C. Pollard, G. Dominick, R. Duman, K. El Omari, V. Mykhaylyk, A. Wagner, W. E. Michener, A. Amore, M. S. Skaf, M. F. Crowley, A. W. Thorne, C. W. Johnson, H. L. Woodcock, J. E. McGeehan and G. T. Beckham, Proc. Natl. Acad. Sci. U. S. A., 2018, 115, E4350–E4357.
43. S. A. Younis, E. Amdeha and R. A. El-Salamony, J. Environ. Chem. Eng., 2021, 9, 104619.
44. A. M. Deliormanlı, J. Mater. Sci.: Mater. Med., 2015, 26, 67.
45. C. H. Nguyen, J. A. Field and R. Sierra-Alvarez, J. Environ. Sci. Health, Part A: Toxic/Hazard. Subst. Environ. Eng., 2020, 55, 168–178.
46. A. A. Ismail, I. Abdelfattah, M. Faisal and A. Helal, J. Hazard. Mater., 2018, 342, 519–526.
47. C. H. Nguyen, C. Zeng, S. Boitano, J. A. Field and R. Sierra-Alvarez, Int. J. Toxicol., 2020, 39, 218–231.
48. R. Raphael, W. F. Aswathy and E. I. Anila, MRS Commun., 2024 DOI:10.1557/s43579-024-00647-z.
49. X. Li, X. Wang, J. Luo, J. Jiang, K. Ding, L. Ye, Y. Xiong, D. Pang, H. Li, P. Yu, C. Kong, L. Ye, H. Zhang and W. Li, Mater. Today Commun., 2023, 35, 106118.
50. W. Xu, T. Peng, Y. Li, F. Xu, Y. Zhang, C. Zhao, M. Fang, S. Han, D. Zhu, P. Cao, W. Liu and Y. Lu, Nanomaterials, 2022, 12, 1125.
51. D. Wawrzynczyk, M. Nyk and M. Samoc, Chem. Phys., 2015, 456, 73–78.
52. Y. Xu, X. Wang, D. Yang, Z. Tang, M. Cao, H. Hu, L. Wu, L. Liu, J. McLeod, H. Lin, Y. Li, Y. Lifshitz, T.-K. Sham and Q. Zhang, ACS Catal., 2021, 11, 10159–10169.
53. Y. Zhang, S. Jiao, J. Zhang, S. Liu, D. Wang, S. Gao and J. Wang, CrystEngComm, 2022, 24, 1789–1794.
54. G. Mu, Y. Zeng, Y. Zheng, Y. Cao, F. Liu, S. Liang, Y. Zhan and L. Jiang, Green Energy Environ., 2023, 8, 831–841.
55. L. Kong, J. Ma, C. Luan, W. Mi and Y. Lv, Thin Solid Films, 2012, 520, 4270–4274.
56. A. Janotti, J. B. Varley, P. Rinke, N. Umezawa, G. Kresse and C. G. Van de Walle, Phys. Rev. B: Condens. Matter Mater. Phys., 2010, 81, 085212.
57. B. J. Morgan and G. W. Watson, J. Phys. Chem. C, 2010, 114, 2321–2328.
58. X. Pan, M.-Q. Yang, X. Fu, N. Zhang and Y.-J. Xu, Nanoscale, 2013, 5, 3601–3614.
59. A. Naldoni, M. Allieta, S. Santangelo, M. Marelli, F. Fabbri, S. Cappelli, C. L. Bianchi, R. Psaro and V. Dal Santo, J. Am. Chem. Soc., 2012, 134, 7600–7603.
60. J.-L. Chiang, B. K. Yadlapalli, M.-I. Chen and D.-S. Wuu, Nanomaterials, 2022, 12, 3601.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4gc03803g

---