Catalyst of a Mytilus galloprovincialis shell substrate: optimization of alcoholysis conditions for the catalytic depolymerization of waste polyester films

Abstract

Catalysts fabricated from waste Mytilus galloprovincialis shells feature high resource utilization, environmental friendliness and low cost. In this work, a powdered Mytilus galloprovincialis shell substrate was used as a catalyst for the glycolysis of a de-aluminized PET film to synthesize bis(2-hydroxyethyl) terephthalate (BHET), and favorable experimental results were obtained. Taking the BHET yield as the response value, a Box–Behnken response surface design was carried out on the basis of single-factor experiments to explore the influencing factors of the BHET yield as well as their interaction effects, and a quadratic regression model was established. Analysis of variance (ANOVA) verified that the established model was highly significant. The optimized glycolysis conditions for PET film depolymerization with ethylene glycol were determined as follows: a catalyst loading of 0.49%, reaction temperature of 198 °C, reaction duration of 1.5 h, and ethylene glycol dosage of 15.56 g; under these conditions, the actual BHET yield reached 83.95%. The Mytilus galloprovincialis shell-based catalytic system delivered a high BHET yield for PET film glycolysis. This strategy realizes the high-value concurrent utilization of waste biomass and waste PET resources, providing a green, low-cost chemical recycling pathway for PET with promising industrial application potential.

In this study, CaO/ZnO composite catalyst (HH-800) was prepared by calcining waste Mytilus galloprovincialis shells. It catalyzes PET depolymerization into high-value BHET monomers, enabling dual recycling of marine food waste and plastic waste.

1. Introduction

An aluminum-coated polyester (PET) film is a vital material widely applied in the agriculture and decoration industries. It exhibits metallic luster and integrates the advantages of plastics and metals, reflecting 70–80% of the incident light and delivering outstanding light reflectivity. At present, pesticides are extensively applied to control pests and diseases affecting fruit trees and fruits, although pesticide application poses severe threats to human health and the environment. Fruit bagging technology has been widely adopted to mitigate these adverse impacts. As an auxiliary measure for fruit bagging, fruit growers lay aluminum-coated PET films on the orchard ground during fruit ripening, which accelerates fruit pigmentation and raises the sugar content, thus improving the economic benefits of fruits. Such agricultural practices greatly boost the consumption of aluminum-coated PET films. However, PET polymer possesses stable chemical structures and is hardly biodegradable. Discarded PET films accumulate in the natural environment and aggravate white pollution. Nevertheless, both the aluminum layer and PET substrate in waste aluminized PET films possess high recycling value. Therefore, developing efficient, eco-friendly and low-cost technologies for recycling waste aluminum-coated PET films is critically required to mitigate agricultural plastic pollution and realize resource recovery of solid waste.

Existing PET recycling technologies fall into two categories: physical recycling and chemical recycling. Physical recycling struggles to produce pure polyester pellets, and even pelletization often fails to meet industrial standards. In contrast, chemical recycling offers flexible application pathways and enables the production of high-value-added chemical products. The main chemical recycling strategies include glycolysis,^1,2 hydrolysis^3–5 and methanolysis.^6–8 Glycolysis is regarded as the most promising recycling route,⁹ as it operates under relatively mild conditions by avoiding extremely high temperatures and pressures, cutting energy consumption and production costs.¹⁰ However, glycolysis proceeds at an extremely low reaction rate without catalysts.^11,12 Under non-catalytic conditions, the reaction rate and efficiency of PET glycolysis remain far below industrial requirements. Hence, high-efficiency catalysts are essential to achieve the effective depolymerization of PET using ethylene glycol. Catalytic glycolysis enables waste PET to be selectively depolymerized into high-purity monomers and intermediates.¹³ These high-purity products can be further utilized for fabricating diverse high-end chemical materials, showing great potential for resource valorization.

Bis(hydroxyethyl) terephthalate (BHET) serves as a critical monomer intermediate with considerable economic value in the PET glycolysis recycling system. Commercial-grade virgin PET is synthesized via the polymerization of petroleum-derived terephthalic acid and ethylene glycol, which renders its production heavily reliant on fossil resources. In contrast, BHET production through ethylene-glycol-mediated depolymerization of waste PET can markedly cut petrochemical feedstock consumption and lower fossil-fuel dependence. Unlike downgraded regenerated PET products from physical recycling, purified BHET monomers can be repolymerized to yield polyester materials whose properties are comparable to those of virgin PET. This overcomes the inherent quality limitations of physical recycling and enables genuine closed-loop chemical recycling of PET. Furthermore, BHET monomers exhibit broad application prospects. Beyond direct reproduction of PET resin, they can be applied to synthesize diverse high-value-added polymers including polyurethane, unsaturated polyester and alkyd resins, which find extensive use in packaging, textile fibers, coatings and other industrial sectors. Against the global policy backdrop of advancing plastic circular-economy strategies, the PET chemical-recycling industry keeps expanding, accompanied by a growing market demand for high-purity BHET. Therefore, the selective conversion of waste PET into BHET holds great industrial-scale implementation potential and economic significance.

Literature reports indicate that catalysts for PET glycolysis mainly consist of metal salts, metal oxides and ionic liquids. Early polyester degradation catalysts were metal salts, including zinc acetate, zinc chloride,¹⁴ sodium carbonate,¹⁵ sodium bicarbonate,¹⁶ and triethylamine.¹⁷ Zinc acetate was the first metal acetate catalyst applied in this field. Later, various metal oxides and modified metal oxides were introduced into PET depolymerization research, such as ZnO,¹⁸ γ-Fe₂O₃,¹⁹ Fe₃O₄,²⁰ mixed metal oxide spinels containing Mn, Co, Zn and other metallic elements,²¹ and SO₄²⁻/ZnO-TiO₂ composite oxides.²² Bifunctional composite catalysts have also been developed, such as CoFe₂O₄@ZIF-8/ZIF-67 (ref. 23) and Mg-Al-O@Fe₃O₄ microspherical catalysts.²⁴ Most high-performance metal oxide and metal salt catalysts contain transition metals, which cause environmental pollution and equipment corrosion during their use. In addition, ionic liquid catalysts developed in recent years suffer from complicated synthesis processes and high costs; at elevated temperatures, they decompose into nitrogen oxides and sulfur oxides, which endanger human health and ecosystems.²⁵ Hence, it is critical to develop low-cost, readily available and eco-friendly catalysts.

In this study, waste Mytilus galloprovincialis shells were recycled to prepare biomass catalysts. After a series of pretreatment procedures, waste mussel shells were ground into a fine bio-based powder featuring low cost and non-toxicity. This catalyst was applied to catalyze the glycolysis of de-aluminized waste PET films (transparent PET films obtained by chemical aluminum stripping of metallized PET films) from agricultural and decorative wastes to synthesize bis(2-hydroxyethyl) terephthalate (BHET). Single-factor experiments combined with response surface methodology were adopted to investigate the single-factor effects and interaction effects on the glycolysis of de-aluminized waste PET film to determine the optimal alcoholysis parameters. This research not only realizes the high-value utilization of waste biomass resources but also achieves excellent recovery performance of waste PET film.

2. Experimental section

2.1. Experimental materials and reagents

Raw materials included waste reflective PET films stripped off aluminum (cut into 10 × 10 mm fragments) and discarded Mytilus galloprovincialis shells from food waste. The chemical reagents used in the experiments were ethylene glycol (analytical grade) and sodium hydroxide (analytical grade).

2.2. Analytical instruments and equipment

The instruments used included a Fourier-transform infrared (FT-IR) spectrometer (Tensor II), X-ray diffractometer (XRD, D8 Venture), ¹H nuclear magnetic resonance (¹H-NMR) spectrometer (AVANCE, Bruker AG, Germany), scanning electron microscope (SEM, Tescan-MIRALMS) coupled with energy-dispersive X-ray spectroscopy (EDS) (Thermo Fisher Scientific), and thermogravimetric analyzer (TG, HCT-1, Beijing Hengjiu Scientific Instruments Co., Ltd).

3. Preparation and characterization of mussel shell-based catalysts

3.1. Preparation of the Mytilus edulis shell-based catalyst

First, waste Mytilus galloprovincialis shells were rinsed with clean water to remove surface impurities and residual attachments. Next, the shells were repeatedly washed with 5% sodium hydroxide aqueous solution and distilled water to eliminate inorganic salts and soluble impurities that could interfere with the experimental results. Subsequently, the fully cleaned mussel shells were placed in an oven and dried at 100 °C for 4 h. The dried mussel shell residues were transferred to a clean, dry crucible and then heated in a muffle furnace under an air atmosphere at a set temperature of 800 °C for 8 h of calcination. Calcination removed the organic matter the inside mussel shells and triggered crystalline phase transformation of the inorganic components. After calcination, the solids were ground and sieved to obtain the Mytilus galloprovincialis shell-based catalyst HH-800 (Fig. 1).

Fig. 1. Preparation process of the Mytilus galloprovincialis shell-based catalyst.

3.2. Characterization results of the Mytilus galloprovincialis shell-based catalyst

3.2.1. Thermogravimetric analysis results of Mytilus galloprovincialis shells

The thermal stability of Mytilus galloprovincialis shells was tested under a nitrogen atmosphere with a heating rate of 10 °C min⁻¹ from 25 °C to 850 °C, and the test curves are displayed in Fig. 2. The mass of the mussel shells gradually declined as the temperature increased. The mass loss reached 4.84% at 660 °C, which was attributed to the decomposition of organic components in the shells. When the temperature further increased from 660 °C to 800 °C, an additional mass loss of 43.78% occurred, mainly originating from the oxidation of residual carbon species. The total mass loss of the raw mussel shells was 48.62%, indicating that the inorganic component available for catalyst preparation accounted for 51.38% of the raw shell material.

Fig. 2. Thermogravimetric analysis of Mytilus edulis shells.

3.2.2. FT-IR characterization results of the catalyst

FT-IR spectroscopy was adopted to compare the as-prepared HH-800 catalyst and the pure calcium oxide reference sample. As shown in the FT-IR spectra (Fig. 3), the HH-800 catalyst exhibited characteristic absorption peaks consistent with calcium oxide, proving that calcium oxide serves as the main active component of the HH-800 catalyst.

Fig. 3. FT-IR spectra of the HH-800 catalyst and calcium oxide samples.

3.2.3. SEM and EDS characterization results of the catalyst

SEM coupled with EDS was employed to analyze the microscopic morphology and elemental distribution of the HH-800 catalyst, and the results are presented in Fig. 4. The catalyst is mainly composed of C, O and Ca, among which Ca reaches 39.96 wt% and acts as the core characteristic element. The raw Mytilus galloprovincialis shells are dominated by calcium carbonate with a mass fraction of approximately 95%. Upon high-temperature calcination, calcium carbonate transforms into CaO, which exhibits catalytic activity toward PET glycolysis and serves as the principal active component of this biomass-derived catalyst.

Fig. 4. SEM and EDS characterization results of HH-800.

Furthermore, trace amounts of Zn and K were detected in the sample, with the mass fraction of Zn being merely 0.14 wt%. These trace elements originate intrinsically from natural Mytilus edulis shells rather than from external artificial addition.

3.2.4. XRD characterization of the catalyst

The XRD pattern of the HH-800 catalyst is shown in Fig. 5, which shows sharp and distinct diffraction peaks and demonstrates the favorable crystallinity of the catalyst. By comparing the obtained peaks with the standard PDF cards (CaO: PDF#74-1226; ZnO: PDF#80-0074), we confirmed that the HH-800 biomass catalyst contained crystalline phases of both CaO and trace ZnO.

Fig. 5. XRD characterization results for HH-800.

4. Ethylene glycol depolymerization of PET films catalyzed by HH-800 based on the Mytilus galloprovincialis shell matrix

4.1. Experimental methods

3 g of the dried de-aluminized PET film was weighed and added to a three-necked flask. A certain amount of ethylene glycol and the HH-800 catalyst were sequentially loaded into the flask, followed by continuous stirring and a constant-temperature depolymerization reaction. After the reaction was complete, the mixture was naturally cooled to 100 °C and poured into a beaker prefilled with deionized water. The mixture was first held at 90 °C for vacuum filtration to separate the undepolymerized PET residues. The filtrate was then cooled to 65 °C and filtered under vacuum to remove oligomer byproducts. Afterwards, the filtrate was further cooled to 45 °C for a second vacuum filtration to eliminate residual oligomers. The final filtrate was cooled to 5 °C and crystallized for 12 h. The precipitated crystals were collected by filtration and dried in a vacuum oven at 78 °C under 0.05 MPa vacuum for 10 h. The dried BHET crystals were weighed to calculate the BHET yield. The complete experimental flow is illustrated in Fig. 6.

Fig. 6. Ethylene glycol depolymerization process of the PET film after dealumination.

The calculation formula of BHET yield (Y) is shown below:where m_BHET = mass of the recovered BHET, g; m_PET0 = initial mass of the PET film, g; M_BHET = molar mass of BHET, 254.24 g mol⁻¹; M_PET = molar mass of the PET repeating unit, 192.16 g mol⁻¹.

4.2. Characterization results of BHET

4.2.1. FT-IR characterization results of BHET

The FT-IR spectra of the glycolysis product BHET and the raw PET film are compared in Fig. 7. For BHET, the absorption peak at 3276.28 cm⁻¹ corresponds to the stretching vibration of terminal hydroxyl (–OH) groups, which is absent in the PET spectrum. The peaks at 2943.80 cm⁻¹ and 2873.24 cm⁻¹ are assigned to aliphatic –CH₂– stretching vibrations. The strong absorption at 1708.06 cm⁻¹ belongs to the stretching vibration of ester C O bonds, while the peak at 1389.85 cm⁻¹ originates from the C–H vibration of para-substituted benzene rings.

Fig. 7. FT-IR spectra of BHET and PET.

4.2.2. NMR characterization results of BHET

¹H-NMR spectra of the glycolysis product were recorded with CDCl₃ as the deuterated solvent (400 MHz), as shown in Fig. 8. ¹H-NMR (400 MHz, CDCl₃) δ: 8.12–8.04 (s, 4H), 4.50–4.49 (m, 4H), 3.91–3.64 (m, 4H), and 2.58 (s, 2H). The characteristic peak distribution fully verifies that the depolymerized product is pure BHET.

Fig. 8. ¹H-NMR spectra of BHET.

4.2.3. DSC characterization results of BHET

The DSC characterization results of the depolymerization product BHET are shown in Fig. 9. It can be seen from the DSC curve of BHET that the sample only exhibits a single sharp exothermic crystallization peak (T_c = 71.17 °C) and a single melting endothermic peak (T_m = 113.18 °C). There is no low-temperature secondary melting peak, shoulder peak or multiple crystallization signals of oligomers. The regular peak profiles and stable baseline demonstrate that the synthesized BHET product possesses a uniform chemical composition and high purity.

Fig. 9. DSC thermogram of BHET.

4.2.4. HPLC characterization of BHET

High-performance liquid chromatography (HPLC) was utilized to analyze the depolymerized BHET product. The detection was carried out using a C₁₈ column at a wavelength of 254 nm. A methanol–water mixture (70 : 30, v/v) was adopted as the mobile phase, with a flow rate of 0.5 mL min⁻¹ and an injection volume of 20 µL. The corresponding HPLC results are illustrated in Fig. 10. A prominent characteristic peak was observed at a retention time of 6.39 min, occupying 97.67% of the total peak area. These results confirm that the product purified via gradient low-temperature filtration is predominantly high-purity BHET monomers with negligible oligomer byproducts. This further verifies that the HH-800 catalyst enables the efficient and highly selective directional depolymerization of PET.

Fig. 10. HPLC spectra of BHET.

4.3. Single-factor experiment on the glycolysis of the PET film

Single-factor experimental results indicate that four key parameters affect PET glycolysis efficiency: catalyst dosage, ethylene glycol dosage, reaction temperature and reaction time (Fig. 11). The preliminary optimal single-factor levels are summarized as follows: a reaction temperature of 195 °C, reaction time of 2.0 h, HH-800 catalyst dosage of 0.5 wt%, and ethylene glycol dosage of 16 g.

Fig. 11. Experimental results of the single-factor experiment on ethylene glycol depolymerization of the PET film. (a) The influence of reaction temperature. (b) The influence of reaction time. (c) The influence of catalyst dosage. (d) The influence of EG dosage.

4.4. Response surface methodology experimental design and determination of optimal conditions

4.4.1. Response surface methodology experimental design

Based on the single-factor results, the BHET yield was selected as the response value. Design-Expert 13.0 software was used to establish a Box–Behnken central composite design. The four independent variables and their corresponding coding levels are listed in Table 1.

Table 1. Factor levels and encoding values for the response surface.

Independent variable	Unit	Levels and encoding values
		−1	0	1
A-HH-800 dosage	%	0.25	0.50	0.75
B-Time	h	1.5	2.0	2.5
C-Temperature	°C	190	195	200
D-Ethylene glycol usage	g	14	16	18

Based on the factor levels and code values for factors affecting the BHET yield in Table 1, the experimental design and results are shown in Table 2.

Table 2. Response surface design and experimental results.

No.	Factor				BHET yield/%
	A	B	C	D
1	1	0	0	1	79.84
2	0	0	0	0	81.98
3	0	−1	0	−1	81.32
4	0	1	0	−1	72.07
5	0	0	0	0	83.26
6	0	−1	−1	0	78.87
7	0	0	−1	−1	69.83
8	1	0	1	0	75.50
9	−1	0	1	0	73.10
10	0	0	1	−1	73.78
11	0	0	1	1	81.09
12	0	0	0	0	83.19
13	0	1	1	0	69.59
14	−1	0	0	1	77.91
15	0	1	−1	0	83.62
16	−1	0	0	−1	74.78
17	0	0	−1	1	83.72
18	1	−1	0	0	77.09
19	0	−1	0	1	82.60
20	0	−1	1	0	82.35
21	0	1	0	1	80.36
22	1	0	−1	0	75.17
23	1	1	0	0	77.67
24	−1	−1	0	0	77.95
25	0	0	0	0	83.59
26	1	0	0	−1	72.85
27	−1	1	0	0	76.47
28	0	0	0	0	82.19
29	−1	0	−1	0	74.72

4.4.2. Quadratic equation of the response surface model and analysis of variance

A quadratic regression analysis was performed on the experimental data in Table 2, and the regression equation for the BHET yield (Y) is given below:Y = 82.84 + 0.2646A − 1.70B − 0.8775C + 3.41D + 0.5149AB + 0.4871AC + 0.9653AD − 4.38BC + 1.75BD − 1.65CD − 4.47A² − 1.11B² − 3.43C² − 2.33D²

Analysis of variance (ANOVA) was conducted to evaluate the statistical significance of the regression model and each variable term, and the results are listed in Table 3.

Table 3. Results of the analysis of variance^a.

Item	Sum of squares	Degrees of freedom	Mean square	F-Value	P-Value
Model	472.55	14	33.75	7.83	0.0002
A	0.8404	1	0.8404	0.1950	0.6655
B	34.63	1	34.63	8.04	0.0132
C	9.24	1	9.24	2.14	0.1652
D	139.26	1	139.26	32.31	<0.0001
AB	1.06	1	1.06	0.2461	0.6275
AC	0.9490	1	0.9490	0.2202	0.6461
AD	3.73	1	3.73	0.8649	0.3681
BC	76.64	1	76.64	17.78	0.0009
BD	12.28	1	12.28	2.85	0.1136
CD	10.84	1	10.84	2.51	0.1351
A 2	129.53	1	129.53	30.05	<0.0001
B 2	7.92	1	7.92	1.84	0.1966
C 2	76.34	1	76.34	17.71	0.0009
D 2	35.24	1	35.24	8.18	0.0126
Residual	60.34	14	4.31
Lack of fit	58.31	10	5.83
Pure error	2.03	4	0.5063
Cor total	532.88	28

^a P < 0.05 is significant, P < 0.01 is highly significant, R² = 0.8868, adj R² = 0.7735, Std. dev. = 2.08, C.V.% = 2.66, and A_deq precision = 8.1392.

The regression model possessed a P-value of 0.0002, R² of 0.8868 and adjusted R² of 0.7735, demonstrating outstanding fitting performance between the model prediction and experimental data. The lack-of-fit term was negligible, indicating that the quadratic regression equation could reliably describe the experimental results instead of raw discrete data points. The low coefficient of variation (C.V.% = 2.66) reflected the high precision and repeatability of the experimental system, while the standard deviation of 2.08 showed a good correlation between the measured and predicted values. A smaller P-value and a larger F-value correspond to higher statistical significance, which collectively confirmed that the regression model was highly significant. The single-factor term D and quadratic terms A² and C² exerted extremely significant effects on BHET yield (P < 0.01). According to the ANOVA results, the descending order of factor influence strength on catalytic glycolysis efficiency was ethylene glycol dosage (D) > reaction time (B) > reaction temperature (C) > catalyst dosage (A).

4.4.3. Response surface analysis of BHET yield

The three-dimensional response surface diagrams for pairwise factor interactions are displayed in Fig. 12. Fig. 12(a) describes the interactive effect between the catalyst dosage (A) and reaction time (B), where the BHET yield peaks at a catalyst loading of roughly 0.5% and a reaction time of 1.5 h. Fig. 12(b) reflects the combined influence of the catalyst dosage (A) and reaction temperature (C); at a fixed temperature, the BHET yield first rises and then declines with further catalyst addition. Fig. 12(c) reveals the interaction between the catalyst dosage (A) and ethylene glycol dosage (D). With constant catalyst loading, additional ethylene glycol effectively boosts the BHET yield, yet the optimal solvent dosage remains below 16 g. Fig. 12(d) illustrates how the reaction time (B) interacts with the reaction temperature (C), showing that a higher temperature greatly elevates the monomer yield under short-duration reactions. Fig. 12(e) compares the variables of reaction time (B) and ethylene glycol dosage (D), confirming that ethylene glycol dosage imposes a far stronger influence on BHET yield than reaction duration. Fig. 12(f) presents the interaction between the reaction temperature (C) and ethylene glycol dosage (D), and the maximum BHET yield is obtained at approximately 198 °C and an ethylene glycol dosage of 16 g.

Fig. 12. 3D Plot of the factors affecting BHET yield. (a) The interaction effect between catalyst dosage and time. (b) The interaction effect between catalyst dosage and temperature. (c) The interaction effect between catalyst dosage and EG dosage. (d) The interaction effect between time and temperature. (e) The interaction effect between time and EG dosage. (f) The interaction effect between temperature and EG dosage.

4.4.4. Adequacy test of the Box–Behnken model

Numerical optimization based on the regression model offered the optimal reaction parameters: an HH-800 catalyst loading of 0.49%, reaction temperature of 198 °C, reaction time of 1.5 h, and ethylene glycol dosage of 15.56 g, with a theoretically predicted BHET yield of 84.02%. Three parallel validation trials were carried out under these optimized conditions, and the average measured BHET yield reached 83.95%, showing good consistency with the model prediction. Model adequacy tests were further conducted to validate the predictive capability of the regression model, and the test results are presented in Fig. 13. Fig. 13(a) displays the relationship between the normal probability and residuals used to test the normality assumption. The experimental values are close to a straight line, indicating that the response surface exhibits a varianceless normal distribution. Fig. 13(b) is a scatter plot of the experimental BHET yield against the model-predicted values; all data points cluster around the diagonal fitting line, indicating that the model accurately captures the correlation between the response value and independent variables and delivers a stable predictive performance. Overall, the established regression model is suitable for the optimization of PET glycolysis parameters.

Fig. 13. (a) Normal probability plot of the RSM model and (b) comparison of experimental and model results in BBD.

The quadratic regression model established using the Box–Behnken response surface methodology in this study presents excellent generality and scalability. This multi-factor optimization framework is not only feasible for the mussel shell-derived calcium-based catalytic system developed herein but also extendable to other catalytic systems and diverse waste polyester feedstocks, thereby providing broad technical references. In terms of catalytic adaptability, the four selected variables—reaction temperature, reaction time, catalyst dosage, and ethylene glycol dosage—represent universal key process parameters for PET glycolysis, which are independent of specific catalyst types. For commonly reported PET depolymerization catalysts, including zinc acetate, metal composite oxides, ionic liquids, and emerging biomass solid bases, only minor adjustments to the parameter ranges are required according to differences in catalyst activity, while the model structure remains unchanged. This approach realizes efficient multi-parameter optimization and substantially reduces experimental costs.

In terms of feedstock applicability, beyond the de-aluminized agricultural PET film adopted in this work, the established optimization model is applicable to various waste polyester resources, such as bottle-grade PET pellets, waste polyester textiles, and composite packaging PET films. Despite discrepancies in the polymerization degree, impurity content, and surface morphology among different polyester substrates, the quadratic regression model can accurately fit product yield trends through simple feedstock pretreatment and parameter range modifications, effectively guiding the optimization of the glycolysis process for various polyester materials. Overall, the proposed response surface optimization model possesses strong universality and promising engineering application potential.

4.5. Mechanism of catalytic disaggregation

Based on the experimental characterization and theoretical analysis, the catalytic glycolysis mechanism of PET mediated by the HH-800 catalyst is proposed in Fig. 14. First, Ca²⁺ active sites on the catalyst surface form coordination bonds with the carbonyl oxygen atoms of the PET macromolecules. This coordination weakens the C O double bonds, lowers the activation energy of ester cleavage and activates the PET polymer chains. Such a coordination-activation effect of metal ions on ester carbonyl groups has been documented in previous studies concerning calcium- and zinc-based solid-base catalysts for polyester glycolysis. The unoccupied outer-shell orbitals of Ca²⁺ accept the lone-pair electrons from carbonyl oxygen to form coordinate bonds, which induces an electron-cloud migration toward oxygen. Consequently, the electron density of the carbonyl carbon decreases, rendering the carbonyl carbon more susceptible to nucleophilic attack by ethylene glycol.^15,26 Ethylene glycol molecules act as nucleophiles and attack the activated carbonyl carbon atoms, triggering transesterification reactions that gradually break long PET polymer chains and generate BHET monomers.

Fig. 14. Mechanism of the HH-800-catalyzed ethylene glycol depolymerization reaction.

In practical reaction systems, PET depolymerization cannot solely produce BHET monomers. Since transesterification proceeds stepwise, intermediate oligomers and dimers continuously react with ethylene glycol monomers. Therefore, besides the target BHET monomer, a certain amount of oligomeric byproducts is generated simultaneously. These oligomers can undergo further glycolysis to convert into BHET monomers with a prolonged reaction time, ultimately realizing complete depolymerization and high-efficiency monomer recovery of waste PET.

4.6. Comprehensive performance of the Mytilus galloprovincialis shell-based catalyst

Mytilus galloprovincialis represents one of the major mariculture species in China, with a domestic annual output of approximately 570 000 metric tons and a global annual production ranging from 2.0 to 2.2 million metric tons. Shells account for 40–50% of the total body weight of Mytilus edulis, and substantial quantities of waste shell are produced during mariculture and processing. At present, most waste shells are subjected to low-value applications, including soil amendment and feed additives. Such routes deliver extremely low resource-utilization efficiency and low product added value. Meanwhile, the random disposal of waste shells may trigger coastal ecological pollution, which highlights the urgent demand for developing high-value resource-recovery technologies for shell wastes.

In this work, waste Mytilus galloprovincialis shells were adopted as raw materials to fabricate a green catalyst for PET glycolysis, realizing high-value valorization of aquatic solid waste. Thermogravimetric analysis reveals that 51.38% of the shell mass can be converted into catalytic active powder dominated by CaO via high-temperature calcination. The catalyst can be synthesized by straightforward procedures, including washing, calcination and grinding. This strategy efficiently converts low-value aquatic solid waste into functional materials for PET chemical recycling, which complies well with circular-economy principles.

Compared with conventional synthetic metal-based catalysts, the as-prepared shell-derived catalyst exhibits prominent overall merits. Using waste shells as feedstock eliminates raw material purchasing costs and brings considerable economic benefits. Its main active constituent is non-toxic CaO, together with naturally occurring trace Zn and K. The reaction residues are free from heavy-metal contamination risks and permit repeated utilization, showing favorable environmental compatibility. Moreover, catalyst preparation avoids complicated synthesis or modification steps, making it promising for large-scale manufacturing. Under the optimized conditions (198 °C, 1.5 h, 0.49% catalyst loading, and 15.56 g ethylene glycol), the BHET yield from waste PET-film glycolysis reached 83.95%, and the catalytic activity was comparable to that of commercial calcium-based catalysts.

Cycling stability serves as a critical metric for evaluating the engineering application potential of catalysts. Reusability tests were carried out to explore the reusability of the HH-800 catalyst, and the corresponding results are presented in Fig. 15. Only a slight decline in the BHET yield was observed after three repeated cycles, which is mainly caused by the loss of fine catalyst particles during solid–liquid separation. The BHET yield remained above 75% even after three successive cycles, demonstrating the satisfactory reusability and reaction stability of HH-800.

Fig. 15. Effect of catalyst recycling.

Nevertheless, this natural-biomass-derived catalyst possesses some intrinsic drawbacks. On the one hand, natural compositional variations in Mytilus galloprovincialis shells originating from different growth environments lead to fluctuations in the active-component content after calcination, which may cause batch-to-batch performance inconsistency. Therefore, standardized protocols for raw-material pretreatment and calcination are required for practical large-scale deployment. On the other hand, calcined shell powder suffers from low mechanical strength; it tends to fragment and pulverize under high-temperature stirring, increasing the difficulty of post-reaction solid–liquid separation. Future targeted optimizations can be implemented to tackle these deficiencies: standardize shell feedstock quality via raw-material screening and acid-washing pretreatment to mitigate batch-dependent performance deviation; introduce inert carriers to improve mechanical strength, anti-passivation capacity and anti-leaching performance of the catalyst; and perform multi-cycle recycling-regeneration experiments to systematically clarify the catalyst deactivation mechanism, so as to offer theoretical foundations for the engineering implementation of shell-based catalysts in the green recycling of waste polyesters.

5. Conclusions

This study explored the green glycolysis process of waste PET catalyzed by a biomass-based HH-800 catalyst derived from Mytilus galloprovincialis shells. The proposed strategy realizes the efficient conversion of difficult-to-recycle de-aluminized agricultural PET films into high-value BHET monomers. Single-factor experiments were performed to clarify the influence of catalyst dosage, reaction temperature, reaction time, and ethylene glycol (EG) dosage on BHET yield, whereby the reasonable variation range of each factor was determined. On this basis, Box–Behnken response surface methodology was adopted for the multi-factor collaborative optimization, and a quadratic regression model correlating the BHET yield with the four process parameters was successfully established. ANOVA demonstrated that the interactive significance of the factors affecting the BHET yield followed the order EG dosage > reaction time > reaction temperature > HH-800 catalyst dosage. The optimal process conditions were determined as 0.49% catalyst dosage, 198 °C, 1.5 h, and 15.56 g EG. Under such conditions, the average experimental BHET yield reached 83.95%, which was in good agreement with the model's theoretical prediction.

The developed Mytilus galloprovincialis shell-based catalytic system exhibits a substantial advantage in resource utilization, environmental sustainability, and economic performance. Thermogravimetric results confirmed that the catalytically active inorganic components account for 51.38% of the raw Mytilus galloprovincialis shell mass. Utilizing waste Mytilus galloprovincialis shell byproducts to fabricate functional catalysts achieves the harmless disposal and high-value reutilization of low-value aquatic solid waste, providing a green alternative to conventional chemically synthesized catalysts. In addition, the glycolysis process presents nearly 100% atomic economy, enabling the selective conversion of PET into BHET, while the unreacted EG solvent can be recovered and reused, further improving the environmental friendliness of the overall process. Compared with zinc acetate and transition metal oxide catalysts, HH-800 is dominated by non-toxic CaO with only trace amounts of endogenous Zn and K elements. The spent catalyst can be recycled or directly applied as a soil amendment, imposing negligible environmental risks.

In terms of energy efficiency and economic benefits, the optimal reaction temperature of this system is 198 °C, which is lower than the typical temperature range (210–240 °C) of conventional metal-catalyzed PET glycolysis systems, reducing the unit thermal energy consumption by approximately 12%. Moreover, waste Mytilus galloprovincialis shells require almost no procurement cost, cutting the raw material cost of the catalyst by over 95% compared with commercial zinc acetate catalysts. In conclusion, the waste Mytilus galloprovincialis shell-based catalytic system provides an efficient, low-cost, and feasible technical route for the green chemical recycling of waste agricultural PET films.

Author contributions

Guoliang Shen and Tiejun Xu conceived, planned, and supervised the experiments. Xingzhu Han and Linlin Zhao were responsible for the main part of the experiment, including the preparation of the catalyst and its application in PET film degradation. Ruiyang Wen, Sijin Jiang and Haichen Wang were responsible for the characterization and analysis of the catalyst. Xingzhu Han and Linlin Zhao wrote the first manuscript, and all authors reviewed the manuscript.

Conflicts of interest

The authors declare that they have no conflicts of interest.

Data availability

All data supporting the findings of this study are available within the paper.