Monomer recycling of virgin polycarbonate (PC), end-of-life PC and PC-ABS blends by Ni-catalyzed reductive depolymerization

Abstract

The reductive depolymerization of polycarbonate (PC) using affordable, heterogeneous and non-noble Ni catalysts was achieved, efficiently breaking down PC back to its monomer, bisphenol-A (BPA), in high yields. This breakthrough was accomplished under mild reductive conditions and in short reaction times. The method also showed remarkable success with end-of-life PC and PC-ABS blends.

---

Green foundation

1. Existing methods for the reductive depolymerization of polycarbonate (PC) rely on non-recyclable homogeneous catalysts and the use of toxic additives. The heterogeneously catalyzed reductive depolymerization of PC to its monomer would help divert PC waste from landfills and decrease the dependency on fossil fuels, lowering the carbon footprint of PC manufacturing.

2. By proposing a heterogeneous system based on a non-noble abundant metal, the recovery of the catalysts is ensured, reducing waste generation and resource depletion. While homogeneous systems offer a slightly higher selectivity towards the desired product, they involve complex separation processes, adding cost and environmental impact.

3. Recycling processes still face challenges in terms of energy consumption and solvent usage. Here, the use of THF, despite its environmental impact, is justified by its beneficial solubilization properties, chemical stability, and low boiling point. Greener solvent alternatives are currently being studied.

Plastics are ubiquitous in our everyday life due to their remarkable physical properties, relatively low cost, versatility, and durability. However, the linear economic model of plastics generates vast amounts of waste.¹ A way to circumvent this issue is the reprocessing or recycling of plastic waste. Among traditional methods, mechanical recycling is the most common, where plastics are generally turned into less valuable products using heat and mechanical shear.² An interesting alternative is the direct transformation of plastic waste into their raw materials in a chemical recycling to monomer (CRM) process, aiming at an ideal circular polymer economy.³

Poly-(bisphenol A carbonate) (PC) is one of the most commercially used polymers with global production reaching an estimated 5 million tons annually by 2022.⁴ The special properties of PC plastics make them a popular choice for durable engineering applications. However, scattered PC waste streams and lack of defined recycling strategies lead to the accumulation of PC wastes in the environment. Prolonged exposure to the weathering conditions causes the degradation of PC and potential leakage of bisphenol A (BPA) into the ecosystem. Therefore, CRM of PC can be highly beneficial in terms of resource recovery and prevention of environmental pollution.⁵ Chemical depolymerization of PC primarily includes solvolysis (hydrolysis, methanolysis and aminolysis) and reductive methods.^4,6–8 Among them, reductive depolymerization has emerged as an interesting CRM alternative for PC waste.^9,10 Despite their high cost and poor recyclability, most routes for reductive depolymerization of PC rely on the use of boranes and silanes as reducing agents^11–14 and homogeneous catalysts, like Ir-, Ru- and Fe-pincer complexes.^15–19 As an alternative, the introduction of a heterogeneous system would be valuable to simplify catalyst recovery. In this sense, some reports have studied the use of heterogeneous bimetallic systems for the hydrodeoxygenation of PC to jet-fuel grade cycloalkanes.^20–24 It is worth noting that most of the aforementioned studies investigate the depolymerization of pure PC. In reality, PC wastes contain various additives and other polymers as the blend component. These additives can lead to catalyst poisoning and impede the depolymerization reaction. To address the problems associated with current catalytic processes, we report here the use of alumina supported Ni catalysts for the reductive depolymerization of PC and also end-of-life PC and PC-ABS blends. In this work, we aim to maximize the BPA selectivity while maintaining a fast depolymerization rate. Catalysts developed in this work offer significant cost benefits and reusability. To the best of our knowledge, this is the first report on heterogeneously catalyzed CRM reductive depolymerization of PC (Table 1).⁶

Table 1 Chemical recycling to monomer reductive depolymerization of PC

P H2 (bar)	Reaction time (h)	Catalyst	Additive	Yield (%)	Ref.
45	24	Fe pincer catalyst	—	99	14
200	16	[Ru(triphos)tmm]	HNTf2	99	16
45	24	Milstein catalyst	KOtBu	91	17
40	5	Ni/Al2O3	—	85	This work

---

We initially studied commercially available catalysts supported on alumina and silica-alumina. These catalysts contain transition metals such as Pd, Ru, and Ni, which are known for their effectiveness in reductive deconstruction reactions (Table 2, entries 1–3).⁹ As observed, under these initial, mild reaction conditions, the catalysts only achieved PC conversions lower than 40%. However, considering the relatively low cost of Ni compared to the other studied (noble) metals, we were encouraged to continue our investigation on Ni-supported catalysts. An increase in metal loading in the reaction mixture resulted in full PC conversion after a reaction time of 2.5 h at 200 °C (Table 2, entry 4). Nonetheless, this increase in conversion was accompanied by the formation of significant amounts of BPA decomposition products, resulting in a selectivity towards BPA of 36%. An effort to improve BPA selectivity by decreasing the reaction temperature resulted in incomplete conversion of PC (Table S1, ESI†). Prompted by these initial findings, the preparation of metal-supported catalysts with different Ni loadings was planned. Aiming to improve the selectivity towards BPA at high PC conversions, catalysts with 5, 10, 15 and 20 wt% NiO over Al₂O₃ were prepared, which after a reduction treatment afforded the catalysts 5, 10, 15, and 20 wt% Ni/Al₂O₃ (see the ESI† for experimental details and characterization). The impact of these treatments on the catalytic activity was further investigated, and the results are presented in Table 2, entries 5–7. The reduction of NiO to Ni⁰ positively affected the activity of the catalyst. The most favorable catalytic outcome was obtained when the sample was reduced at 500 °C, indicating a correlation between the Ni particle size and the overall catalyst performance (vide infra), as suggested by transmission electron microscopy (TEM) analysis (Fig. S6 and S7, ESI†).

Table 2 Initial reaction and catalyst screening for the reductive depolymerization of PC^a

Entry	Catalyst^b	Loading^c (%)	X (%)	S BPA  (%)
a Reaction conditions: 4.6 g of PC in 46 g of THF at 200 °C for 2.5 h under H2 at 20 bar. b Number represents the loading of the metal in the catalyst in wt%. c Metal loading in the reaction in wt%. d Conversion determined by the amount of PC obtained after methanol precipitation (ESI†). e BPA selectivity determined by GPC (ESI†). f Sample reduced at 800 °C. g Sample reduced at 500 °C. h Conversion after 5 h.
1	5Ru/Al2O3	0.25	38	13
2	5Pd/Al2O3	0.25	18	<5
3	66Ni/SiO2–Al2O3	0.25	10	<5
4	66Ni/SiO2–Al2O3	3.33	100	36
5	5NiO/Al2O3	0.25	16	<5
6	5Ni/Al2O3^f	0.25	23 (47)^h	<5
7	5Ni/Al2O3^g	0.25	28 (58)^h	<5

---

The effect of Ni loading on the support in the catalytic activity was then studied and the results are presented in Table 3 and the ESI.† Under mild reaction conditions, a volcano-shaped trend was obtained (Fig. S8, ESI†), with the 10Ni/Al₂O₃ and 15Ni/Al₂O₃ samples outperforming their counterparts. This result suggests an optimal size and dispersion of the Ni particles on the support for the reductive depolymerization of PC. When optimized conditions were applied (Fig. S9 and Table S2, ESI†), both catalysts achieved 100% PC conversion with 41% and 38% selectivity towards BPA for 10Ni/Al₂O₃ and 15Ni/Al₂O₃, respectively.

Table 3 Reductive depolymerization of PC under optimized conditions^a

Entry	Catalyst^b	S BPA  (%)	S BPA+  (%)	S BPA−  (%)
a Reaction conditions: 4.6 g of PC in 46 g of THF at 250 °C for 5 h under H2 at 20 bar (full conversion was reached with 1 wt% metal loading in the reaction). b Number represents the loading of the metal in the catalyst in wt%. c BPA selectivity determined by GC (ESI†). d 85% BPA selectivity under H2 at 40 bar.
1	10Ni/Al2O3	72 (85)^d	2	26
2	15Ni/Al2O3	48	40	12
3	5Ru/Al2O3	45 (48)^d	53	2
4	5Pd/Al2O3	22 (28)^d	76	2

---

Conversely, oligomeric, trimeric and dimeric products (henceforth referred to as BPA⁺) accounted for 49% and 50% of the product mixture, while the main BPA decomposition product (4-cumylphenol, BPA⁻) accounted for 10% and 12% for 10Ni/Al₂O₃ and 15Ni/Al₂O₃, respectively. Remarkably, increasing the metal catalyst loading in the reaction from 0.5 wt% to 1 wt% led to an increase in the selectivity towards BPA to 72% for 10Ni/Al₂O₃ and 48% for 15Ni/Al₂O₃. In contrast, the selectivity to BPA⁺ and BPA⁻ remained at 2% and 26% for 10Ni/Al₂O₃ and 40% and 12% for 15Ni/Al₂O₃, as observed in Table 3, entries 1 and 2. The differences in the catalytic activity of the samples can be attributed to the varying sizes of the Ni-supported particles of each catalyst. TEM analysis (Fig. 1) revealed that the average diameter of the Ni particles on the samples 5Ni/Al₂O₃, 10Ni/Al₂O₃, 15Ni/Al₂O₃ and 20Ni/Al₂O₃ is 3.30 nm, 4.73 nm, 7.32 nm and 7.62 nm, respectively, showing a volcano-like trend of activity (Fig. S10, ESI†). Intuitively, hydrogenolysis, a structure-sensitive reaction, typically benefits from larger metal clusters for bond breaking.^25,26 Reports suggest, however, that a medium Ni particle size is often optimal for achieving the best hydrogenolysis yields. Smaller particles may be better dispersed on the support surface but can be challenging to reduce, leading to lower catalytic activity in some cases.^27,28 Conversely, larger particles may contain relatively inactive facets in higher proportion.^27,29 Our results indicate that medium Ni particle size (4.73 nm) is beneficial for the reductive depolymerization of PC to BPA.

Fig. 1 TEM images of selected catalysts in (a) light mode and (b) dark mode.

Since the sample 10Ni/Al₂O₃ was able to catalyze the reductive depolymerization of PC without significant formation of oligomers, dimers and trimers (S_BPA+ < 3%), resulting in the highest BPA yield (72%) at 1 wt% Ni loading in the reaction mixture, a time profile under these conditions was subsequently gathered (Fig. 2). This kinetic profile is consistent with our hypothesized reaction network (Scheme 1). It demonstrates that the first fragmentation products, BPA⁺, are formed from the start of the reaction and, as time progresses, are then converted into BPA. Moreover, as observed in Fig. 2, the amount of BPA decomposition products in the mixture (BPA⁻, mostly 4-cumylphenol) increases with the reaction time, reaching a maximum of 33% after 24 h. This hypothesis was further verified by additional experiments using BPA as the reactant, where after 5 h of reaction under the optimized conditions, BPA was found to degrade to lower molecular weight compounds.

Fig. 2 Time course of the reductive depolymerization of polycarbonate (PC) using the 10Ni/Al₂O₃ catalyst. Reaction conditions: 4.6 g of PC in 46 g of THF at 250 °C under H₂ at 20 bar (metal loading in the reaction: 1 wt%).

Scheme 1 Reaction model for the reductive depolymerization of polycarbonate (PC); early stages of the reaction show the formation of dimers and trimers, which are later converted to BPA and other decomposition products.

Considering our data and previous studies found in the literature,^22,24,30 it is hypothesized that a plausible mechanism for the reaction could entail the dissociative adsorption of H₂ on the Ni surface as a first step, followed by the nucleophilic attack of the hydride to break the C–O bond present in the PC chain, initially forming oligomers, trimers and dimers, and subsequently releasing BPA as the major reductive depolymerization product. Formation of 4-cumylphenol could occur by the H-induced cleavage of the C_Ar–OH bond, which takes place as the reaction progresses (Fig. S11, ESI†).

The reaction conditions were further optimized for the 10Ni/Al₂O₃ catalyst. First, the effect of the reaction temperature was investigated (Fig. S12, ESI†). At 200 °C, the PC conversion and BPA selectivity decreased to 84% and 24%, respectively, with an increase in dimer, trimer and oligomer formation (S_BPA+ = 74%). At 225 °C, full PC conversion was achieved, but S_BPA decreased to 27% and S_BPA+ increased to 69%. These results indicate that, although lower temperatures can actively break down the PC chains into lower molecular weight products, T = 250 °C is necessary to effectively convert the dimers, trimers and oligomers (BPA⁺ products) into BPA at the studied reaction time (5 h). Next, the Ni loading in the reaction mixture was increased from 1 wt% to 1.5 wt% (Fig. S13, ESI†). This led to BPA formation in smaller amounts (S_BPA = 40%) even at full PC conversion and an increase in BPA⁻ product yield (S_BPA = 49%). However, for 15Ni/Al₂O₃, better results were obtained with 1.5 wt% Ni loading, showing an increase in S_BPA from 48% to 59%. Lastly, the effect of H₂ pressure was investigated (Fig. S14, ESI†). A minimum H₂ pressure of 10 bar appears to be necessary for an adequate hydrogenolysis of PC chains to oligomers, trimers and dimers (full conversion), albeit with a BPA selectivity of 35%. At 20 bar, the BPA yield drastically increased to 72%, a value that can be further increased when 40 bar of H₂ are used during the reaction (85%, TON = 21.1).

A series of solvents were also screened as promising green alternatives to THF (Table S3, ESI†). The results show that THF outperforms all the tested solvents in terms of BPA recovery, which can be related to its known solubilization properties.³¹ Nonetheless, other solvents did exhibit potential for the process, such as acetone, with which a respectable 73% BPA yield was achieved. The performance of the 10Ni/Al₂O₃ catalysts was also tested for the depolymerization of end-of-life PC waste (EOL-PC). To this end, EOL-PC was collected from roof panels used in agricultural applications, shredded, and washed with water followed by vacuum-drying at 45 °C for 16 h before being subjected to reductive depolymerization. Under optimized conditions (Fig. 3a), 10Ni/Al₂O₃ achieves 100% PC conversion and a selectivity towards BPA of 65%. Remarkably, upscaling of the experiments using 100 g of PC waste provided the recovery of BPA in equally exceptional yields (68%, Fig. 4). Analogously, EOL PC-ABS waste (TV waste) was collected, shredded, water-washed and dried under vacuum at 45 °C for 16 h. Once dried, the blend was subjected to a decolorization process using THF and activated carbon in order to remove the anthraquinone-like colorants present in the TV waste.³² Complete depolymerization of the color-free polymer and a selectivity to BPA of 74% were achieved after 5 h of reaction (Fig. 3b). These results highlight the versatility of the proposed protocol for the efficient recovery of BPA from PC waste, irrespective of the composition of the streams.

Fig. 3 BPA recovery from (a) EOL-PC and (b) EOL PC-ABS (TV waste) with a decolorization step by reductive depolymerization. Reaction conditions: 4.6 g of polymer in 46 g of THF at 250 °C under H₂ at 20 bar (metal loading in the reaction: 1 wt%).

Fig. 4 Upscaling of BPA recovery from EOL-PC by reductive depolymerization over 10Ni/Al₂O₃. Reaction conditions: 100 g of polymer in 1 kg of THF at 250 °C under H₂ at 20 bar (metal loading in the reaction: 1 wt%).

The commercial Ru and Pd-based catalysts were then tested for the reductive depolymerization of PC under optimized conditions. The results reveal that these noble metals can also catalyze the reductive depolymerization of PC for BPA recovery, exhibiting full PC conversion after 5 h of reaction (Table 3, entries 3 and 4). However, a substantial decrease in BPA selectivity is observed, which considering their higher cost, highlighting Ni, a cheap and Earth-abundant transition metal, as the most attractive option.

Finally, reusability studies showed that the heterogeneous 10Ni/Al₂O₃ catalyst can be used for at least 8 consecutive reaction runs (Fig. S15, ESI†) with only a minor loss of BPA yield (from 72% to 62%, with a cumulative TON of 131.36). This is consistent with the characterization of the spent catalyst, which revealed a limited Ni leaching, a homogeneous dispersion of the Ni particles on the support (before and after reaction), a negligible increase in the average Ni particle size, and the preservation of the thermal properties of the catalyst, namely S_BET and S_ext (Fig. S16–S18, ESI†).

All in all, the protocol reported here represents a green advance when compared to existing recycling processes and proposed CRM reductive depolymerization routes for PC. First, it is important to consider that current rates of PC recycling are staggeringly low,^31,33 with most of the PC waste ending up in landfills or incinerated. The heterogeneously catalyzed reductive depolymerization of PC to BPA would help divert PC waste from landfills since it is selective towards BPA, offers catalyst recycling and is operated with commonly used solvents. Further, the recycling of PC back to BPA helps in decreasing the dependency on fossil fuels, lowering the carbon footprint of PC manufacturing, and enabling a closed-loop system. It is estimated that 3 million tons of PC are produced annually, consuming around 24 million barrels of crude oil and 526 trillion BTUs of energy.³² However, it is essential to point out that even though CRM processes generally embody a promising step towards sustainability, they still face challenges in terms of energy consumption and solvent usage. In our case, the use of THF, despite its environmental impact, is justified by its beneficial solubilization properties, allowing concentrated solutions and the use of lower amounts of solvent. Moreover, it is compatible with continuous processing (THF is commonly used in polymer science and technology),³¹ is chemically stable under the studied reaction conditions and has a low boiling point, lowering the energy impact of solvent recovery via distillation. Available literature on reductive depolymerization routes for PC focuses on the use of homogeneous catalysts, mostly based on rare and expensive elements at the risk of depletion. By proposing a heterogeneous system based on a non-noble abundant metal, the recovery of the catalysts is ensured, which in turn reduces waste generation and resource depletion. Furthermore, while homogeneously catalyzed systems can still offer a slightly higher selectivity towards the desired product, they generally involve more complex separation processes, adding cost and environmental impact due to the energy consumption and solvent use. For example, in the case of the alkali catalyzed methanolysis of PC, the main drawbacks are often related to catalyst recovery (which is seldom achieved), resulting in the generation of highly alkaline products and waste streams that require additional treatments, involving neutralization steps with concentrated solutions of corrosive and hazardous HCl.^32,34 The need for such neutralization is eliminated in this proposed heterogeneous system, where catalyst recovery and reuse (up to 8 reaction runs) can be easily realized serving as a green advance in the field of CRM of PC technologies, especially aiming at a bigger-scale application of the protocol.

Conclusions

In conclusion, Ni-based Al₂O₃ heterogeneous catalysts have been identified as highly effective materials for the reductive depolymerization of polycarbonate (PC) to bisphenol-A (BPA) in a CRM process. These catalysts also offer the flexibility to fine-tune their activity toward BPA production by adjusting the metal loading and varying key reaction parameters such as temperature and hydrogen pressure. Among the tested catalysts, 10Ni/Al₂O₃ showed the highest activity, achieving a combined yield of 98% for BPA and 4-cumylphenol. Promising results were also obtained with real end-of-life PC samples and PC-ABS blends, following a decolorization step. To the best of our knowledge, this is the first report of heterogeneously catalyzed reductive depolymerization (CRM hydrogenolysis) of PC. Given the lower cost and wide availability of Ni compared to noble metals, this approach represents an attractive alternative for the valorization of PC and PC-ABS waste streams. Future work will focus on further optimizing reaction conditions to enhance BPA recovery rates by studying other catalyst supports and greener solvents in order to improve the circularity of the PC life cycle.

Author contributions

Carlos Marquez: writing – review & editing, writing – original draft, visualization, methodology, investigation, formal analysis, data curation, and conceptualization. Annelore Aerts: writing – review & editing, investigation, and data curation. Dambarudhar Parida: writing – review & editing, investigation, and conceptualization. Illian Glassee: investigation and data curation. Harisekhar Mitta: investigation and data curation. Lingfeng Li: investigation and data curation. Kevin M. Van Geem: writing – review & editing and funding acquisition. Karolien Vanbroekhoven: writing – review & editing and funding acquisition. Elias Feghali: writing – review & editing, funding acquisition, and conceptualization. Kathy Elst: writing – review & editing, funding acquisition, and conceptualization.

Data availability

The data supporting this article have been included as part of the ESI.†

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors would like to acknowledge the financial support from the Flemish Government and Flanders Innovation & Entrepreneurship (VLAIO) through the Moonshot project (CYCLOPS HBC.2021.0584). The authors acknowledge Peter De Rechter and Jan Jordens (VITO) for their support in GC and GPC analyses, respectively, and Pieter Weltens for helping with the preparation of the samples.

References

1. C. Marquez, C. Martin, N. Linares and D. De Vos, Mater. Horiz., 2023, 10, 1625.
2. A. Ravve, Principles of Polymer Chemistry, Springer, New York, 3rd edn, 2012, pp. 567–693.
3. G. W. Coates and Y. D. Y. L. Getzler, Nat. Rev. Mater., 2020, 5, 501.
4. R. A. Clark and M. P. Shaver, Chem. Rev., 2024, 124(5), 2617.
5. T. Do, E. R. Baral and J. G. Kim, Polymer, 2018, 143, 106.
6. D. Parida, A. Aerts, K. Vanbroekhoven, M. Van Dael, H. Mitta, L. Li, W. Eevers, K. M. Van Geem, E. Feghali and K. Elst, Prog. Polym. Sci., 2024, 149, 101783.
7. Y. Liu and X.-B. Lu, J. Polym. Sci., 2022, 60(24), 3256.
8. H. Mitta, L. Li, M. Havaei, D. Parida, E. Feghali, K. Elst, A. Aerts, K. Vanbroekhoven and K. M. Van Geem, Green Chem., 2025, 27, 10–40.
9. A. C. Fernandez, Green Chem., 2021, 23, 7330.
10. C. Wang and O. El-Sepelgy, Curr. Opin. Green Sustainable, 2021, 32, 100547.
11. M. Kobylarski, J.-C. Berthet and T. Cantat, Chem. Commun., 2022, 58, 2830.
12. E. Feghali, L. Tauk, P. Ortiz, K. Vanbroekhoven and W. Eevers, Polym. Degrad. Stab., 2020, 179, 109241.
13. E. Feghali and T. Cantat, ChemSusChem, 2015, 8, 980.
14. D. L. Lourenço, D. F. Oliveira and A. C. Fernandes, Adv. Sustainable Syst., 2024, 2300381.
15. L. Monsigny, J.-C. Berthet and T. Cantat, ACS Sustainable Chem. Eng., 2018, 6, 10481.
16. C. Alberti, E. Fedorenko and S. Enthaler, ChemistryOpen, 2020, 9, 818.
17. E. M. Krall, T. W. Klein, R. J. Andersen, A. J. Nett, R. W. Glasgow, D. S. Reader, B. C. Dauphinais, S. P. Mc Ilrath, A. A. Fischer, M. J. Carney, D. J. Hudson and N. J. Robertson, Chem. Commun., 2014, 50, 4863.
18. S. Westhues, J. Idel and J. Klankermayer, Sci. Adv., 2018, 4, eaat9669.
19. C. Alberti, J. Kessler, S. Eckelt, M. Hofmann, T.-O. Kindler, N. Santangelo, E. Fedorenko and S. Enthaler, ChemistrySelect, 2020, 5(14), 4231.
20. L. Wang, F. Han, G. Li, M. Zheng, A. Wang, X. Wang, T. Zhang, Y. Cong and N. Li, Green Chem., 2021, 23, 912.
21. L. Wang, G. Li, Y. Cong, A. Wang, X. Wang, T. Zhang and N. Li, Green Chem., 2021, 23, 3693.
22. A. K. Manal, G. V. Shanbhag and R. Srivastava, Appl. Catal., B, 2023, 338, 123021.
23. J.-H. Luo and J. Deng, ACS Sustainable Chem. Eng., 2023, 11, 17120.
24. J. Wei, M. Zhu, B. Liu, N. Wang, J. Liu, K. Tomishige, S. Liu and G. Liu, Angew. Chem., Int. Ed., 2023, 62, e202310505.
25. H. Wang and J. Lu, Chin. J. Chem., 2020, 38, 1422.
26. M. Che and C. O. Bennett, Adv. Catal., 1989, 36, 55.
27. M. F. Guilleux, J. A. Dalmon and G. A. Martin, J. Catal., 1980, 62, 235.
28. D. Nazimek, Appl. Catal., 1984, 12, 227.
29. A. Sárkány and P. Tétényi, React. Kinet. Catal. Lett., 1979, 12, 297.
30. C. M. Osmundsen, K. Egeblad and E. Taarning, in New Futur. Dev. Catal, ed. S. L. Suib, Elsevier, Amsterdam, 2013, pp. 73–89.
31. T. Spychal, D. Lath and D. Berek, Polymer, 1979, 20, 437–442.
32. D. Parida, A. Aerts, L. V. Perez, C. Marquez, S. Vloemans, K. Vanbroekhoven, E. Feghali and K. Elst, Chem. Eng. J., 2024, 497, 154390.
33. G. S. Weeden Jr., N. H. Soepriatna and N.-H. L. Wang, Environ. Sci. Technol., 2015, 49, 2425–2433.
34. J. J. Rubio Arias, E. Barnard and W. Thielemans, ChemSusChem, 2022, 15, e202200625.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4gc06438k

---