Antioxidant-induced transformations of a metal-acid hydrocracking catalyst in the deconstruction of polyethylene waste†

Zachary R. Hinton‡ ^a, Pavel A. Kots‡ ^a, Mya Soukaseum ^ab, Brandon C. Vance ^ab, Dionisios G. Vlachos ^ab, Thomas H. Epps III *^abc and LaShanda T. J. Korley *^abc
aCenter for Plastics Innovation, University of Delaware, Newark, DE 19716, USA. E-mail: thepps@udel.edu; lkorley@udel.edu
^bDepartment of Chemical and Biomolecular Engineering, University of Delaware, Newark, DE 19716, USA
^cDepartment of Materials Science and Engineering, University of Delaware, Newark, DE 19716, USA

First published on 10th August 2022

---

---

Introduction

The increased global consumption of plastics is a direct result of the numerous benefits of polymers (e.g., barrier properties, high strength-to-weight ratio).^1,2 This accelerating demand for plastics is accompanied by manufacturer and consumer needs for an amalgam of features (e.g., visual appeal, thermal stability, versatile processability) that drive rising plastics diversity. Polymer types, chemical additives, inorganic co-materials (e.g., metal, paper), form factors, advanced processing and thermal histories, etc. contribute to highly varied waste streams that challenge the creation of economic value and reduction of waste accumulation.^1,3,4 For example, within sorted, post-consumer high-density polyethylene (HDPE), numerous colorants, processing aids, macromolecular architectures, fillers, and crystallinities can be present and can introduce significant heterogeneity among recycling inputs (to either mechanical or chemical processes) that limits product value and process scalability.^4–6 Valorization routes that consider both the macromolecular and the compositional diversity inherent to all plastics are likely to become widespread strategies that address real, discarded plastics.^3,4,7

Plastics waste contains a vast array of small-molecule additives,^8,9 as many commercial polymers require additives to facilitate processability and improve performance.^6,10,11 For example, HDPE can often contain processing aids that enable molding or extrusion, and certain applications (e.g., outdoor furniture, electronics packaging) necessitate the inclusion of additives, such as ultraviolet (UV) stabilizers or antistatic agents.^9,10 Similar formulation considerations are made for most commodity polymers.⁹ Although 25 Mt of plastics additives were produced globally in 2015 (equivalent to the production of polystyrene),² only a small fraction of the over 10000 individual chemical substances have been widely studied with respect to their direct impact on chemical deconstruction of polymers.⁸ Yet, it has been widely speculated that both mechanical and chemical re/upcycling processes are significantly impacted by additives.^6,8,12,13

Experimental studies of valorization processes that treat specific plastic products (e.g., packaging, cables) highlight the challenges posed by additives.⁶ For example, brominated flame retardants in electronics waste diminish deconstruction activity by poisoning catalysts and contribute to lower product stream purity and value.¹⁴ In another instance, dissolution- or solvolysis-based approaches can concentrate certain toxic additives like phthalate plasticizers, precluding industrial adoption.^7,15 Although the impact of additives on valorization is widely accepted as fact or implied by experimental studies,^1,16–18 information about how common additives affect certain emergent chemical recycling strategies is needed to guide innovation in both plastics formulation and catalytic design. As scientific advances continue to address increasingly complex waste, novel processes that are tolerant to a wide range of additives are highly desirable.^1,4,19

Hydroconversion (i.e., hydrocracking and hydrogenolysis) has gained increasing attention as a strategy for the deconstruction of plastics.^1,18,20–22 This technology is particularly well suited to valorize polyolefins, which represent the largest percentage (by weight) of plastics consumed globally.² In hydroconversion processes, supported noble metal catalysts (e.g., platinum on tungstated zirconia, ruthenium on titanium dioxide) or solid acids (e.g., zeolites, tungstated zirconia) facilitate increased bond-scission activity and value-added product selectivity at mild operating conditions and short reaction times.^1,18,20–22 For example, our previously developed bi-functional hydrocracking catalyst, platinum on tungstated zirconium (Pt/WO₃/ZrO₂), exhibits high activity for polyolefin hydrocracking to alkane products. Furthermore, the selectivity of this catalyst toward desired molecular weights or degrees of branching can be easily controlled by alteration of the metal-acid balance [MAB, i.e., the molar ratio of Pt sites to Brønsted acid (BA) sites].^20,21 Unfortunately, the highly tunable nature of many hydroconversion catalysts is often countered by susceptibility to poisoning,^18,23,24 and this vulnerability is particularly important when considering hydroconversion approaches to treat real plastics waste.

In this work, we quantify the activity of our model catalyst, Pt/WO₃/ZrO₂, as impacted by common plastic additives. We examine the deconstruction of as-received HDPE (i.e., base resin) to demonstrate the high activity of Pt/WO₃/ZrO₂. We then identify the small-molecule additives present in the base resin and reveal the effect that typical additives have on product yield and distribution by deconstructing pure (i.e., stripped of additives) HDPE polymer. Next, we focus on the role of specific phenolic antioxidants in the reduction of activity and in the alteration of product distribution by adding three common molecules to HDPE. Our results indicate that both additive concentration and chemical structure contribute to changes to the effective MAB of Pt/WO₃/ZrO₂, and consequently, effect differences in product distributions. Finally, spectroscopic analysis is used to probe the chemical nature of the catalyst before and after poisoning by antioxidants. Our results reveal that the phenolic group of these small-molecule additives can generate phenoxy species that consume catalyst active sites; however, more severe poisoning can be caused by antioxidants that contain additional functional groups (e.g., acids, esters). These findings suggest that selection of antioxidants (and other additives) for plastic formulations with higher valorization potential should consider the potential effects of both additive loading and chemical structure.

Results and discussion

Initially, we quantified hydrocracking performance for as-received HDPE pellets (weight-average molecular weight, M_W, ∼90 kg mol⁻¹, melt flow index = 12 g/10 min, purchased from Sigma Aldrich), which contained a basic additive profile that facilitates processing during the fabrication of plastic products.^6,9,10 For all experiments, the catalyst to polymer mass ratio was fixed at 1:40 so that sufficient masses (typically >200 mg) of solid residues remained to be analyzed. This value was selected based on preliminary experiments, the results of which are given in Fig. S1–S3 and discussed in the ESI.† The detailed synthetic protocol and characterization of the catalyst are reported elsewhere.²¹ The overall activity measured in this study is higher than previously reported for LDPE at otherwise identical experimental conditions.²¹ The results in Fig. 1a show that Pt/WO₃/ZrO₂ cracked HDPE to extractable (i.e., gas and liquid) products with moderate yield (20%, on a moles of carbon basis as described in the ESI†). The product distributions obtained from gas chromatography with flame ionization detection (GC-FID) for the extractables (Fig. 1b) and high-temperature gel permeation chromatography (HT-GPC) for the solids (Fig. 1c) indicated the most abundant products, C_7–10 and C_∼30, were accompanied by a significant amount of unreacted HDPE.

To examine the effect of the native additives on hydrocracking performance, we compared the base HDPE resin results with those for a stripped HDPE. This purified polymer was recrystallized (by slowly cooling a 6 wt% solution in toluene from 130 °C) and extracted (via 24 h Soxhlet extraction in chloroform) such that the majority of additives were removed. A clear increase in activity was measured for the stripped HDPE sample, whereby extractable (i.e., gas and liquid) yield was enhanced from 20% (pellets) to 86% (stripped) (Fig. 1a). It is important to note that our extraction procedure resulted in HDPE particle diameters of ∼30 μm (i.e., a powder), whereas the original pellets were ∼4 mm. A more uniform distribution of the catalyst within the HDPE melt at the start of the reaction was achieved for the powder form, in comparison to the pellets, and likely was a key factor in the increase in measured catalyst activity (Fig. S4†). Therefore, stripped HDPE is more appropriately compared to a powdered HDPE, which for this study had the same particle size, because it was formed by recrystallization, but retained the additives in the as-received HDPE, as it was dried from solution without extraction. It should be noted that there is the potential for degradation of additives (particularly antioxidants) during the stripping process; however, it is likely that this reaction does not reduce the concentration of antioxidants significantly enough to affect our results. This powdered HDPE led to higher activity (44% extractables yield) than pellets (20%); however, the effect of additives remained pronounced with respect to the stripped HDPE (Fig. 1a).

Comparison of the product streams from the powdered and stripped HDPE samples suggests that the effect of the base additive profile leads to a dramatic change in the overall kinetics (or activity). The yields of individual components of the extractable products are shown in Fig. 1b. Although the distributions were similar, the ratio of selectivities toward carbon number 5 and 10 products was nearly double for the stripped HDPE (0.9) in comparison to both the pellet and powdered HDPE (0.48). This shift toward lower-molecular-weight alkanes suggested secondary cracking of the alkane products. The solid product distributions (Fig. 1c) demonstrated a similar trend. One key effect of additive removal was to increase the production of heavy (solid) alkanes, likely because of the occurrence of deeper cracking of intermediate waxes, and this cause is distinct from a change in conversion (e.g., from different catalyst loadings, see ESI†). Although hydrogen has been proposed to mitigate the impact of impurities on deconstruction,¹⁵ the results herein indicate that even base additive formulations can lower deconstruction reaction rates. Furthermore, mass transfer effects are likely important for both the polymer and small-molecule additives and are the subject of ongoing work.

To further examine the effects of base HDPE formulation on catalytic activity, we catalogued the specific additive chemistries extracted from our stripped HDPE. Four small-molecule additives were identified by gas chromatography-mass spectrometry (GC-MS): two primary, hindered-phenolic antioxidants and two fatty-amide slip agents (Fig. S5–S8 and Table S1†). It should be noted that larger-molecular-weight additives (i.e., phenolic phosphite secondary antioxidants) likely were present,¹⁰ but they could not be detected by GC-MS. Importantly, the inputs to most previously reported polyethylene-deconstruction processes have unwittingly contained additives from a (typically uncharacterized) base formulation that may be similar in composition to the HDPE in this work. Thus, many of these studies have drawn conclusions about polymer- or catalyst-specific effects but may be measuring the impact of additives to a substantial degree, and there exists a need to identify specific catalyst-additive interactions for common additives and catalyst. Given that most commercial plastics contain phenolic antioxidants,^9,10 fundamental quantification of the impact of these molecules is useful for assessing the performance of additive-tolerant valorization strategies.

Towards such an evaluation, we investigated three phenolic antioxidants (of different molecular sizes and that contain different functional groups) at two concentrations (0.5 wt% and 2 wt%) to mimic typical formulation conditions^10,25,26 and to emphasize the measured effects, respectively. The chemical structures of the antioxidants investigated in this work and overall product yields are summarized in Fig. 2. Note that we did not select the antioxidants found in the original HDPE for further study in order to extend our results to a wider range of potential plastics formulations.¹⁰ Hydrocracking was performed using the same method as above with a fixed 1:40 mass ratio of catalyst to polymer to confine the reaction to the low-conversion regime. The measured overall yield of extractable products varied strongly as a function of the antioxidant chemistry and concentration (Fig. 2b). In the absence of additives, the stripped HDPE was converted into mostly liquid products (81% yield); however, the HDPE samples that contained antioxidants always achieved less than 60% liquid yield. Less than 5% yield of gas was measured in all cases. With each different antioxidant chemistry, three distinct features were noted. For butylated hydroxytoluene (BHT), solid and liquid yields were comparable between the 0.5 wt% (43% and 40%) and 2 wt% (38% and 37%) loadings. This finding is a potential indicator of reversible poisoning (i.e., an equal, equilibrium concentration of BHT was adsorbed on the catalyst). For Irganox® 1010 (I-1010), deconstruction inhibition was dependent on antioxidant concentration. At 0.5 wt% I-1010, the highest yield of liquid products of all antioxidant-containing samples was measured (58%), but, at 2 wt% I-1010, the liquid yield decreased significantly to 9%. For Irganox® 3114 (I-3114), <2% extractable yield was achieved for both concentrations, which indicated severe, irreversible poisoning of the catalyst even at low concentration. These results suggest that antioxidants are a significant factor impacting catalytic activity of the base HDPE (Fig. 1) and that specialty antioxidants (e.g., I-3114) also poison hydrocracking catalysts.

The mechanism by which antioxidants interfere with hydrocracking was more complex than that measured for the base formulation of HDPE, as detailed by an analysis of product distributions obtained from GC-FID and HT-GPC analyses (Fig. 3 and S9†). The discrete carbon numbers of alkanes in the gaseous and liquid products were converted to approximate molecular weights, and the overall yields were combined with the continuous distribution (in terms of yield) of the solid products. The resultant summed distributions were calculated and plotted as a function of the molecular weight (see calculation details in the ESI†). Clear differences in the shape of measured distributions arose due to the production of distinct alkane/polyethylene populations that generated overlapping signals. The distributions were divided into six product groups on the basis of approximate bounds that correspond to light fuels (C₁–C₈), alkanes (C₈–C₁₃), diesel fuels (C₁₃–C₂₃), fuel and lubricant oil (C₂₃–C₇₀), heavy wax/low-molecular-weight polymer (C₇₀–C₇₀₀), and pristine polymer (Table S2†). A summary of the selectivities of these groups is given in Fig. 3c, d and S10.†

For stripped HDPE, groups I, II, IV, and V were the most prominent products, with high conversion indicated by significant group I and II yields (Fig. 3a, c and S11†). For 0.5 wt% BHT, the product distribution was shifted to higher molecular weight products (i.e., lower group I and II yields and higher amounts of unreacted polymer) due to lower activity. At 2 wt%, BHT led to similar product yields as 0.5 wt% except for group V products, which were diminished. A similar trend in selectivity was measured for the concentration dependence of I-1010 samples (Fig. 3b, d and S10†); however, at 2 wt% I-1010, conversion to groups I and II was eliminated in favor of group III, which indicated the catalyst was more severely poisoned at higher antioxidant levels. For I-3114, product distributions (Fig. S12†) showed only marginal yield of group III products in the solid fraction, which could be products generated at the early stages of the reaction before the antioxidant poisoned catalyst active sites. This result indicated that antioxidants reduce catalytic activity as cracking occurs, such that the conversion of waxy intermediates to small alkanes does not occur.

Comparison of the overall yield and composition of the solid residues further demonstrated the impact of antioxidant concentration and chemistry dependence on hydrocracking outcomes. A simple subtraction of a weighted pristine HDPE distribution from the measured product distribution (see details in the ESI†) was used to quantify unreacted HDPE in the solid products and to calculate a corrected selectivity of solid intermediates. This analysis demonstrated that overall solid yield is an insufficient indicator of conversion. Rather, quantification of the unreacted HDPE for antioxidant-containing samples (Fig. S13†) and different catalyst loadings (Fig. S14†) revealed that similar solid yields did not indicate the same conversion of HDPE. For example, as-received HDPE pellets, HDPE powder, and HDPE that contained 2 wt% BHT all yielded ∼40% solids (deconstructed at the same catalyst loading), but these residues contained ∼35%, ∼70%, and ∼30% unreacted polymer, respectively (see Fig. S13 and S14†). Although pristine Pt/WO₃/ZrO₂ led to a maximum yield of waxes (i.e., solid products) at 25%, in the presence of antioxidant, up to 40% yield of wax was achievable. A similar effect was measured in our previous study of LDPE deconstruction, wherein a series of Pt/WO₃/ZrO₂ catalysts of varied MAB produced similar solid yields but distinct molecular weight distributions for both the solid and extractable products.²¹ Therefore, the origin of the different product species distributions (Fig. 3) was likely related to a dynamic MAB induced by poisoning of catalyst active sites.

To validate the dynamic MAB hypothesis and determine the cause of chemistry-specific antioxidant effects, catalyst particles were impregnated with solutions of antioxidants to facilitate adsorption/poisoning and then heated under hydrogen flow at 250 °C to simulate hydrocracking conditions. Two spectroscopic probe molecules were used to quantify antioxidant-catalyst interactions via diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS): CO for Pt metal sites and pyridine for BA sites. The measured DRIFTS traces for pristine and poisoned catalyst exhibited antioxidant-mediated changes to the effective MAB. The metal sites of the pristine catalyst showed a strong carbonyl band at 2085 cm⁻¹ due to 8 to 9-fold coordinated Pt and a peak at 1855 cm⁻¹ caused by multiple sites bridged by CO (Fig. 4a).²⁷ After treatment with BHT and I-3114, well-coordinated Pt sites were eliminated, and a new band appeared at 2048 cm⁻¹, which indicated CO is bound to 6 to 7-fold (under) coordinated Pt sites.²⁷ Also, the band at 1855 cm⁻¹ completely disappeared, which suggested that Pt flat facets were blocked by adsorbed antioxidant, but corner and edge sites with lower coordination number were unaffected. Semi-quantitative analysis of DRIFTS data indicated that 30–35% of Pt sites remain accessible to CO. Strong attractions between π-electrons in the aromatic ring of phenol and metal atoms on the flat surface have been previously reported²⁸ and were likely responsible for blocking Pt sites.

The bi-functional nature of our hydrocracking catalyst is key to activity, which makes selectivity and conversion dependent on both metal and acid sites.²¹ Quantification of antioxidant-mediated changes to BA sites via pyridine chemisorption on Pt/WO₃/ZrO₂ showed formation of protonated pyridine (characteristic bands at 1639 and 1539 cm⁻¹) bound to BA sites and pyridine coordinated to Zr⁴⁺ LA sites (bands at 1609 and 1443 cm⁻¹) (Fig. 4b).²⁹ The pristine catalyst exhibited a BA site density that is comparable to our previously reported value (30 μmol g⁻¹).²⁰ Deposition of BHT (and an analogue, phenol) led to a four-fold decline in the peak areas of chemisorbed pyridine, which suggested ∼25% of the original BA sites remained accessible. In the case of the I-3114 treated catalyst, no signs of protonated or strongly coordinated pyridine were measured in the spectrum (Fig. 4b). Only bands at 1438 and 1596 cm⁻¹ arose due to weakly bonded pyridine. This result indicated that BA sites were completely blocked by I-3114, which effectively shifted the MAB toward infinity and the activity to zero. On the contrary, ∼25% of BA sites remained for BHT and I-1010, yielding finite, non-zero MAB and activity values.

The mechanism by which BA sites were depleted is related to the reaction of phenolic groups to form phenoxy species that were covalently bound to surface tungstate clusters. DRIFTS data for adsorbed phenol (Fig. S15 and S16a†) confirmed the formation of phenoxy groups. Hydrogenation of these phenoxy groups by Pt sites led to the formation of bound cyclohexanols, which subsequently desorbed or formed cyclohexene. Thus, some population of BA sites were regenerated and remained accessible to pyridine. BHT interaction with Pt/WO₃/ZrO₂ resembled that of phenol (Fig. S17 and S16b†); however, the t-butyl groups of BHT effectively shielded the phenolic hydroxyl group from interaction with BA sites such that there was less catalyst poisoning at low temperatures. In this context, it is noteworthy that I-1010 exhibited more severe poisoning at high concentration, in comparison to pure BHT. This difference was likely due to the central ester groups in I-1010. During BA site-catalyzed hydrolysis,³⁰ surface carboxylate groups were more likely to form than phenoxy groups, which led to more severely blocked BA sites via steric effects (Fig. S16c†). For Pt catalysts, hydrogenation of aromatic rings in phenoxy species occurs at a higher rate than hydrogenation of carboxylates.³¹ Thus, it is expected that hindered phenolic antioxidants with non-hydrocarbon functional groups (e.g., esters, amides) are likely to induce additional poisoning mechanisms and deactivate catalysts to greater degrees in comparison to BHT. An example of this effect was noted in the case of I-3114, which contained a central cyanuric acid group that was decomposed on Pt/WO₃/ZrO₂ at 250 °C as shown by a shift in the amidic carbonyl peak in the DRIFTS trace at increased temperatures (Fig. S18†). Furthermore, ∼30% of phenolic hydroxyls did not interact with BA sites because of steric hindrance, which suggests that the combined effects of cyanuric acid derivatives contributed significantly to poisoning.

Overall, antioxidants induced complex changes to both metal and acid sites, which led to differences in activity and product selectivity. A schematic of the effects of the extremes of antioxidant-catalyst interaction (BHT vs. I-3114) is provided in Fig. 5. The initial activity of the catalyst was largely determined by its MAB,²¹ the reaction conditions, and transport phenomena (discussed above). For the 1 wt% Pt and 25 wt% WO₃ on ZrO₂ catalyst used in this work, the native MAB (0.86 mol Pt per mol BA) promoted conversion of HDPE into light and mid-alkanes with moderate amounts of wax. In the presence of BHT, the catalyst lost significant activity because the number of total sites was reduced by blocking of metal sites or reaction with acid sites. The new effective MAB shifted product yields toward more mid-alkanes versus light alkanes, which equated to a general retardation in reaction kinetics. Higher exposure to poisonous additive caused by continuous throughput of plastic or lower catalyst loading likely would cause further damage, although regeneration via calcination in air could facilitate reuse of the catalyst. I-3114 blocked all Pt facet sites and leaves edge and corner sites intact, which yielded a catalyst with a similar number of metal sites as in the case of BHT poisoning; however, all BA sites were poisoned. Thus, the effective MAB increased towards infinity, and the activity was eliminated (because both metal and BA sites are required for hydrocracking). The resultant product consisted of only mid-alkanes, likely because transport of the antioxidant to the catalyst surface caused dynamic poisoning such that initial cracking products can accumulate to a small degree. Additionally, it should be expected that new antioxidant-derived chemical species can be generated during deconstruction because of the surface reactions of antioxidants with catalytic active sites, and these may lead to additional challenges to valorization processes (e.g., further poisoning, product separations, reactor fouling). Due to the low concentration and strong catalyst binding affinity, no derivative species were identified in the present study.

Conclusions

Additives are an intrinsic component of plastics, and therefore, plastics waste. In this work, we demonstrated the high activity of Pt/WO₃/ZrO₂ towards the deconstruction of HDPE. A comparison between as-received (i.e., containing antioxidants and slip agents) HDPE and stripped (i.e., pure) HDPE revealed that the small-molecule additives present in the base HDPE formulations were responsible for a 50% reduction in catalyst activity. Additionally, the specific role of phenolic antioxidants on the reduction of extractable yield and alteration of product distribution was examined, as these additives are a major constituent of commercial plastics. Both antioxidant concentration and chemical structure had complex effects on the hydrocracking process. For BHT, an equilibrium adsorption concentration was reached, and concentration played a lesser role on reducing catalyst activity (i.e., two-fold in comparison to stripped HDPE). For I-1010, the presence of ester groups was the probable cause of a more dramatic decrease in activity and a stronger concentration dependence (i.e., <2-fold and ∼8-fold reductions at 0.5 wt% and 2 wt%, respectively). In the case of I-3114, nearly no conversion occurred as the cyanurate moieties appeared to lead to complete poisoning.

Using HT-GPC and DRIFTS analyses, we connected antioxidant-mediated transformations of our bi-functional catalyst to complex changes to hydrocracking-product distributions. Specifically, the phenolic groups of the antioxidant adsorbed to Pt metal sites and left only undercoordinated edges accessible to HDPE. Additionally, both phenolic groups and other moieties formed complexes with acid sites to change MAB and, therefore, selectivity toward certain product groups. For I-1010, ester groups were easily cleaved to form complexes with and reduce the number of acid sites, yielding a catalyst that was more selective toward mid-alkanes. The cyanurate group of I-3114 underwent additional reactions over Pt/WO₃/ZrO₂ to generate poisonous compounds that completely consumed acid sites that were required for activity. Based on these results, it may be desirable in some cases to pretreat plastics waste to remove certain additives that negatively affect catalytic deconstruction processes. Overall, there are substantial opportunities to either design additives that have lower impact on deconstruction outcomes or to develop robust valorization strategies that are agnostic to a range of real-world plastic additives.

Author contributions

Z. R. H., T. H. E., and L. T. J. K. conceived the idea. Z. R. H., P. A. K., M. S., and B. C. V. performed experiments and analysis. Z. R. H. and P. A. K. co-wrote the initial draft. D. G. V., T. H. E., and L. T. J. K. supervised the work and obtained funding. All authors contributed to the writing and editing of the manuscript.

Conflicts of interest

The authors declare no conflicts of interest.

Acknowledgements

This work was supported as part of the Center for Plastics Innovation, an Energy Frontier Research Center funded by the U.S. Department of Energy, Office of Science, Basic Energy Sciences, under award DE-SC0021166.

References

1. Z. R. Hinton, M. R. Talley, P. A. Kots, A. V. Le, T. Zhang, M. E. Mackay, A. M. Kunjapur, P. Bai, D. G. Vlachos, M. P. Watson, M. C. Berg, T. H. Epps III and L. T. J. Korley, Annu. Rev. Mater. Res., 2022, 52, 249 CrossRef .
2. R. Geyer, J. R. Jambeck and K. L. Law, Sci. Adv., 2017, 3, e1700782 CrossRef PubMed .
3. T. H. Epps III, L. T. J. Korley, T. Yan, K. L. Beers and T. M. Burt, JACS Au, 2022, 2, 3 CrossRef CAS PubMed .
4. L. T. J. Korley, T. H. Epps III, B. A. Helms and A. J. Ryan, Science, 2021, 373, 66 CrossRef CAS PubMed .
5. J. N. Hahladakis and E. Iacovidou, Sci. Total Environ., 2018, 630, 1394 CrossRef CAS PubMed .
6. J. N. Hahladakis, C. A. Velis, R. Weber, E. Iacovidou and P. Purnell, J. Hazard. Mater., 2018, 344, 179 CrossRef CAS PubMed .
7. J.-P. Lange, ACS Sustainable Chem. Eng., 2021, 9, 15722 CrossRef CAS .
8. H. Wiesinger, Z. Wang and S. Hellweg, Environ. Sci. Technol., 2021, 55, 9339 CrossRef CAS .
9. Polymer Modifiers and Additives, ed. J. T. Lutz, Jr. and R. F. Grossman, CRC Press, Boca Raton, FL, 2018 Search PubMed .
10. M. Tolinski, Additives for Polyolefins: Getting the Most Out of Polypropylene, Polyethylene and TPO, William Andrew, Waltham, MA, 2nd edn, 2015 Search PubMed .
11. M. Bolgar, J. Hubball, J. Groeger and S. Meronek, Handbook for the Chemical Analysis of Plastic and Polymer Additives, CRC Press, Boca Raton, FL, 2008 Search PubMed .
12. F. P. La Mantia, Macromol. Symp., 1998, 135, 157 CrossRef CAS .
13. Recycling of Polymers: Methods, Characterization and Applications, ed. R. Francis, Wiley-VCH, Weinheim, Germany, 2017 Search PubMed .
14. C. Ma, J. Yu, B. Wang, Z. Song, J. Xiang, S. Hu, S. Su and L. Sun, Renewable Sustainable Energy Rev., 2016, 61, 433 CrossRef CAS .
15. I. Vollmer, M. J. F. Jenks, M. C. P. Roelands, R. J. White, T. van Harmelen, P. de Wild, G. P. van der Laan, F. Meirer, J. T. F. Keurentjes and B. M. Weckhuysen, Angew. Chem., Int. Ed., 2020, 59, 15402 CrossRef CAS .
16. L. D. Ellis, N. A. Rorrer, K. P. Sullivan, M. Otto, J. E. McGeehan, Y. Román-Leshkov, N. Wierckx and G. T. Beckham, Nat. Catal., 2021, 4, 539 CrossRef CAS .
17. G. W. Coates and Y. D. Y. L. Getzler, Nat. Rev. Mater., 2020, 5, 501 CrossRef CAS .
18. P. A. Kots, B. C. Vance and D. G. Vlachos, React. Chem. Eng., 2022, 7, 41 RSC .
19. A. J. Martín, C. Mondelli, S. D. Jaydev and J. Pérez-Ramírez, Chem, 2021, 7, 1487 Search PubMed .
20. S. Liu, P. A. Kots, B. C. Vance, A. Danielson and D. G. Vlachos, Sci. Adv., 2021, 7, eabf8283 CrossRef CAS PubMed .
21. B. C. Vance, P. A. Kots, C. Wang, Z. R. Hinton, C. M. Quinn, T. H. Epps III, L. T. J. Korley and D. G. Vlachos, Appl. Catal., B, 2021, 299, 120483 CrossRef CAS .
22. D. Munir, M. F. Irfan and M. R. Usman, Renewable Sustainable Energy Rev., 2018, 90, 490 CrossRef CAS .
23. E. G. Fuentes-Ordóñez, J. A. Salbidegoitia, M. P. González-Marcos and J. R. González-Velasco, Ind. Eng. Chem. Res., 2013, 52, 14798 CrossRef .
24. J. W. Gosselink and W. H. J. Stork, Ind. Eng. Chem. Res., 1997, 36, 3354 CrossRef CAS .
25. BASF, Irganox® 1010 Technical Information, 2015, https://dispersions-resins-products.basf.us/products/irganox-1010, (accessed April 2022) Search PubMed .
26. BASF, Irganox® 3114 Technical Data Sheet, 2019, https://documents.basf.com/7758655303a3460c5a14914b17d6cf15ffefc531/Irganox_3114_July_2019_R3_Adhesives.pdf, (accessed April 2022) Search PubMed .
27. M. J. Kale and P. Christopher, ACS Catal., 2016, 6, 5599 CrossRef CAS .
28. Y. Yoon, R. Rousseau, R. S. Weber, D. Mei and J. A. Lercher, J. Am. Chem. Soc., 2014, 136, 10287 CrossRef CAS PubMed .
29. T. Barzetti, E. Selli, D. Moscotti and L. Forni, J. Chem. Soc., Faraday Trans., 1996, 92, 1401 RSC .
30. Y. Wang, H. Ren, Y. Yan, S. He, S. Wu and Q. Zhao, Polym. Polym. Compos., 2020, 29, 1403 Search PubMed .
31. M. Guo, X. Kong, C. Li and Q. Yang, Commun. Chem., 2021, 4, 54 CrossRef CAS .

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d2gc02503e

‡ These authors contributed equally.

This journal is © The Royal Society of Chemistry 2022