Sustainable Design of Structural and Functional Polymers for a Circular Economy

3.1 Polyesters

Polyesters are well suited to highlight the connection between origin of raw materials, polymer structure, and EoL, which all have a direct impact on different sustainability dimensions of the life cycle. Structural polyesters can be divided into major categories based on the source of feedstock and their behavior in the biosphere (see Figure 7). (I) Durable and fossil polyesters: Many of the polyesters obtained on commercial production scale, such as polyethylene terephthalate (PET) and polybutylene terephthalate (PBT), belong to this category. These polymers are durable (i.e., non-biodegradable) and are derived from fossil feedstocks.⁴² (II) Durable and biobased polymers:⁴³ These are chemically identical, 1 : 1 biobased replacements of the fossil counterparts, e.g., PET, PBT, polytrimethylol terephthalate (PTT), or modified variants, e.g., polyethylene furanoate (PEF), polybutylene furanoate (PBF). The major property of durability remains and is not influenced by the biobased content, since biodegradability depends on the polymer structure and not on the origin of the carbon. (III) Biodegradable and fossil polymers:³⁰ These polymers e.g., polybutylene (adipic acid, terephthalic acid) co-polymer (PBAT), or polybutylene succinate (PBS) are biodegradable by design of their structure, and initial development in the 1990s relied on available fossil monomers.⁴⁴ With increasing introduction of biobased monomers like bio-1,4-butanediol and biobased dicarboxylic acids to the market, a shift has begun to the category of (IV) biodegradable and biobased polymers.³⁰ In addition to the biobased variants of PBAT-type polyester and PBS, polylactic acid (PLA) and polyhydroxyalkanoates (PHA) also belong to this category. These materials close the biosphere loop conceptually from biomass to biodegradation. Commercially, the use of biobased content (shift from category III to IV) depends on raw material costs and feedstock availability, which so far favor fossil origin. Change is therefore driven by sustainability commitments, consumer demand, and legislative requirements: in France, for example, 60 % biobased origin is targeted for biodegradable food and vegetable bags in 2025. Overall, the relevance of this category will increase significantly in the future. Biodegradability is the aligning property of materials for class III and IV and is determined by the material structure, abiotic factors connected to the environment (presence of water, oxygen, temperature etc.), and finally the biological factors (presence and types of microbes). Considerable progress has been made in recent years to understand this complex interaction.^45-47

On the raw-material side, carbon sources via the technical loop also become more important. At scale, (Ib) recycled durable polymers, mainly from mechanical recycling of PET so far (but, in future, also depolymerization²⁶ and emerging technical enzymatic recycling^{48, 49}) are examples with significant volumes. For recycled PET, quickly growing demand exceeds supply⁵⁰ and will spur further scale-up. Future polyesters could also be based on building blocks from CO₂,⁵¹ or use CO₂ directly as a reactant in polyester synthesis, as exemplified recently (Scheme 1) by Rapagnani et al.⁵² Their product poly-(EtVP) was shown to be biodegradable in aerobic aqueous environment (OECD-301B standard) and would be a rare, and still academic, example for (IIIb) recycled biodegradable polyesters.

In principle, all different combinations of feedstocks can and will play a future role in a circular economy that moves from incremental increases on a low percent level to system transformation and multiple cycle uses of carbon.⁵ Biobased carbon that enters a technical loop and is continuously recycled, for example, is a direct way of sequestering CO₂ from the atmosphere for the lifetime of the polymers.

Despite the significant difference in properties of durable (categories I and II) and biodegradable polyesters (categories III and IV), polymer design, processing, and production technologies can be remarkably similar. Biodegradability is a property that results from a convolution of all factors in the polymer design hierarchy (Figure 5), from molecular structure to microstructure and crystallinity. Looking at the structural polyesters in Figure 7, there is not a single factor like the presence of a specific monomer that dictates environmental fate. In the set of examples, the presence of rigid, aromatic diacids like terephthalic acid seems to play a role—but can be mitigated as shown by PBAT.⁴⁴ The durable polyester PBT was transformed into the certified compostable polyester PBAT^{53, 54} by “simply” introducing adipic acid as a co-monomer in the molecular structure. Additional commercially available dicarboxylic acids, such as succinic, azelaic, and sebacic acid, allow further tuning of material properties⁵⁵ and show the broad potential of the toolbox of chemical design.

On the other hand, modifications favoring biodegradability also impact material properties. Introducing longer-chain flexible linear diacids increases chain mobility and accessibility for microorganisms, but also influences other properties⁴² e.g. mechanical strength (Young's modulus of 2–3 GPa for PBT, and 0.14 GPa for PBAT), and glass transition temperature (T_g of 55–65 °C for PBT vs. −31 °C for PBAT). The materials therefore are suitable for different applications and will require specific design at the systems level.

The transformation to biobased variants often relies on the same basis molecules (e.g., bio-1,4-butanediol) and therefore can go hand in hand for durable and biobased polyesters as well as biodegradable and biobased polyesters. This close relationship between durables (PBT) and biodegradables (PBAT) is also reflected at the level of production technology. The first generation of PBAT by BASF, for example, was produced in a slightly modified PBT plant.⁵⁶ This consideration of the level of dependence between two polyesters in the categories of molecular design, properties, and production technologies resembles the application of the “phylogenetic tree” principle in biology.⁵⁷ The key property of industrial compostability of different polyesters can be achieved by vastly different synthesis approaches ranging from polycondensation (PBS, PBAT) and ring-opening polymerization (PLA) to direct fermentation (PHA) (Figure 8).

For functional polyesters, recycling is not an option, and the biosphere loop must be targeted instead. Esterases, such as lipases, are abundantly available in the environment and an integral part of activated sludge in wastewater treatment plants.⁵⁸ Therefore, the use of functional polyesters or the introduction of ester groups in functional polymers may be an interesting strategy. For the application of “classical” polyesters as soluble polymers in aqueous environment, one challenge is their intrinsic hydrophobicity. Different building blocks and side-chain functionalities are therefore needed in this realm of application.

In one recent example, a metal-catalyzed bulk polymerization process starting from Boc-Asp-dimethylester and PEG-diol produced a functional polyester;⁵⁹ removal of the Boc-protecting group then resulted in a water-soluble amine-functionalized polyester (Scheme 2). While the focus of this work was on the preparation of hydrogels for tissue engineering via subsequent cross-linking, the strategy creates options to make polyesters suitable for other applications.

Functional polyesters are established as soil-release polymers for laundry detergent formulations.⁶⁰ Copolymerization of terephthalic acid with poly(ethylene oxide) results in a water-dispersible polyester, which can be used in detergent formulations. The amphiphilic fossil polymer adheres to hydrophobic textile fibers, thereby making their surface more hydrophilic, which eases soil removal from the textiles.⁶¹ To the best of our knowledge, the biodegradation behavior of this polymer has not been disclosed in the literature. Depending on the length of the PET block,⁶² the polymer may have some potential to enter the biosphere loop due to the ester functionalities.

Unlike the introduction of amides, the introduction of ester groups in functional polymers has so far been limited. On the one hand, most polyesters have hydrophobic backbones and lack pendent groups that can be chemically functionalized. An example where this challenge seems to have been overcome is again in the field of soil-release polymers. Here, the use of sulfonated terephthalic acid as a comonomer, instead of unmodified terephthalic acid, resulted in an anionic water-soluble soil-release polyester.⁶³ On the other hand, polyesters have a high susceptibility to enzymatic and non-enzymatic hydrolysis in aqueous systems, releasing low-molecular-weight compounds. While non-enzymatic (i.e. acid/base) hydrolysis could support biodegradation of the functional material after use, the polymer is not allowed to (significantly) hydrolyze before or during use. Thus, formulation concepts are required on the systems design level to ensure the necessary stability and performance.

3.2 Polyamides

Like polyesters, polyamides have a broad variety of structures, properties, and targeted applications.⁴⁰ Based on the amide bond, they also represent an attractive toolbox for the development of sustainable polymers. Thanks to their high crystallinity and melting points deriving from the presence of strong hydrogen bonds, polyamides can exhibit superior stability and mechanical properties, as in the case of the most common representatives Nylon 6 (PA6) from caprolactam (as equivalent of ω-aminohexanoic acid) and Nylon 6,6 (PA6.6) from the corresponding α,ω-diamine and α,ω-diacid. These durable materials are produced at a scale of several millions of tons globally from fossil resources today (Category I in Figure 9), and are widely used as engineering plastics. Polyamides are already recycled today via the technical loop, i.e., via mechanical recycling or chemical recycling to monomers. Prerequisites for mechanical recycling are a relatively pure and uncontaminated waste stream. Often the use of additives, such as stabilizers and compatibilizers, or the addition of virgin polymer is necessary to achieve recyclability.⁶⁴ This has been shown e.g. for multilayer films⁶⁵ and injection molding compounds (Nypel®).⁶⁶ For waste streams such as PA6-containing textiles and fishing nets, the Italian Aquafil Group has shown that hydrolytic depolymerization under acidic conditions⁶⁷ is possible on an industrial scale to obtain the monomer caprolactam with virgin-like properties. No such industrial process exists for PA6.6 to date, as the addition of a small molecule (such as water) and the recovery of two monomers is more challenging than the caprolactam ring closure (see Section 3.4.2). Recycling processes for other polyamides in the technical loop are hindered by their low volumes and thus limited access to enriched waste streams of the respective polymers.

Biobased polyamides have recently been comprehensively reviewed by Winnacker and Rieger.⁶⁸ Several commercial products with partial or fully biobased content are present on the market, for example PA4.10 introduced by DSM under the brand name EcoPaXX® from biobased C₁₀-diacid and a fossil 1,4-butanediamine. So far, only few biogenic α,ω-diamines are available at scale.⁶⁹ Earlier in 2022, Covestro announced a partnership with Genomatica to produce biobased 1,6-hexamethylenediamine in ton amounts,⁷⁰ which if successfully scaled would also open a perspective for biobased PA6.6. By synthetic introduction of an amino group via halogenation, 11-aminoundecanoic acid can be obtained from biobased linoleic acid.⁶⁸ The polyamide resulting from this polymerization has been introduced as Rilsan® PA11 by Arkema.

Due to their chemical similarity to proteins built from amino acids, polyamides in principle possess bonds that could undergo enzymatic cleavage leading to metabolization by microbes. However, similarly to polyesters, the choice of monomers and comonomers strongly influences their biodegradability, by directly modulating reactivity of the bonds, or by secondary effects such as imparting high crystallinity or relatively high hydrophobicity, which are key performance criteria of the established structural polyamides. Organic recycling or environmental biodegradation of polyamides⁷¹ is therefore usually not possible, despite some reports about metabolization of some polymers or oligomers by single microorganisms. The most promising material in this respect is PA4, for which partial biodegradation was reported to a certain extent in sludge, soil,⁷² and sea water,⁷³ as well as its utilization by single microbes.⁷⁴

Overcoming this limitation, biodegradation could be achieved by further tuning the molecular design of polyamides through the introduction of additional monomers forming polyester bonds, and hence leveraging their established biodegradability. During the 1990s this approach was followed by Bayer AG through the introduction of BAK® polyesteramides.⁷⁵ The same monomers utilized for the synthesis of PA6 and PA6.6 are copolymerized with e.g. adipic acid and butanediol. The hybrid polymers were claimed to combine the mechanical properties of polyamides and biodegradability under composting conditions. The production of these materials, however, was stopped in early 2001.

What about functional polyamides? In nature, proteins, in addition to their structural functions, act as prime functional polymers. Intrinsically disordered proteins take on many functions in nature that synthetic functional polymers are capable of in their respective applications. Even highly charged polyelectrolytes are available in nature in the form of supercharged proteins and polypeptides.⁷⁶ For proteins, perfect sequence control is possible in synthesis, and biotechnology might in the future allow rational production at a price conducive to non-therapeutic material applications. Synthetic homo- and copolymers of amino acids are biobased functional polyamides which have received significant attention for their use in several consumer applications, such as home and personal care.³² As one example, biodegradable poly(aspartic acid) (PASP) has been evaluated as a promising alternative to non-biodegradable acrylic acid based polymers in home care applications. Following its discovery in 1926 by Wessely and co-workers,⁷⁷ polysarcosine (PSar) has regained attention recently as a potential alternative for poly(ethylene glycol) (PEG) in biomedicine.⁷⁸ Both polymers are non-ionic and are characterized by their excellent water-solubility, low affinity to protein, non-immunogenicity, biocompatibility, and non-cytotoxicity. The challenge with both PEG and homo-PSar is that they are not biodegradable at higher molecular weights up to 50 kg mol⁻¹. Although based on a naturally occurring amino acid (sarcosine), homo-PSar is reported not to be biodegradable, most likely due to the presence of the methylated amide bonds. Recently, Ling et al. described the random incorporation of alanine (Ala) into PSar via ring-opening polymerization of the respective amino acid N-carboxyanhydrides (NCAs)⁷⁹ and demonstrated its degradability by porcine pancreatic elastase within 50 days under mild conditions (pH 8 at 37 °C). Both the enzymolysis rate and water solubility of the copolymer depend on the content of Ala. While this study focuses on inducing biodegradation in the human body, the concept may be extended. Besides Ala, one could think of introducing other amino acids via NCA ring-opening polymerization to tune polymer properties and biodegradability behavior for a targeted application as technical functional polymer. PSar-containing copolymers have been investigated as antifouling surfaces for medical devices, food packaging, and in industrial and marine equipment.⁸⁰ To the best of our knowledge, no biodegradation studies have been published on these (co)polymers using OECD or ISO biodegradation methods, which would be necessary to assess their true potential as a circular, biodegradable functional polymer.

3.3 Polyurethanes

Polyurethanes (PU) are another general class of polymers that can serve both structural and functional applications. The urethane bond again is generally hydrolysable and potentially cleavable by enzymatic attack. A recent overview on polyurethanes from an industrial perspective has been given by Eling, Tomovic, and Schädler,⁸¹ while the use of biomass for PU has been reviewed and systematically organized by PhungHai et al.⁸² Owing to the modular building principle of PU from polyols and diisocyanates, the structural variety and versatility in applications is even broader than for the polymer classes discussed so far. Here, we only want to highlight some basic features that complement the earlier discussions in Sections 3.1 and 3.2, without aiming at being comprehensive.

Typical building blocks of polyurethanes today are, like for all polymers at scale, fossil based, and the resulting durable polymers are non-biodegradable. Typical PU products, like flexible or hard foams, and thermoplastic polyurethanes (TPU) are in the respective category I (Figure 10). Via the polyol component, PUs have a long history of incorporating biobased content: Multifunctional Hydroxy-decorated molecules, like sugars, are available directly as natural products. Common polyols are of the polyester-polyol type, leveraging the whole range of polyester chemistry discussed above in Section 3.1, together with the respective biobased diols and diacids. Similarly, polyether-polyols can be derived from biobased multifunctional alcohols. Using only bifunctional polyol and isocyanate, linear PUs are achieved, which as structural polymers yield thermoplastic polyurethanes (TPU) and are suitable for mechanical recycling. Using multifunctional building blocks, cross-linked systems including typical flexible and hard PU foams are created, which in the technical loop require chemical recycling approaches like glycolysis.²⁶ Algenesis, a startup out of UC San Diego, has started commercializing algae-based polyols⁸³ obtained by ozonolysis of unsaturated fatty acids from algae oils.⁸⁴ Biodegradable PUs,⁸⁵ for example, a foam system based on these algae-based polyols,⁸⁶ have also been described, although biodegradation studies proving complete mineralization are still missing. Recycled polyols are available from PU chemical recycling, e.g., foam glycolysis,²⁶ or via PET oligomers from recycling (e.g. TEROL® polyols from Huntsman⁸⁷). As one of the first examples of carbon capture and utilization (CCU) in a commercial polymer, Covestro offers a CO₂-based polyol (Cardyon®): CO₂ is reacted with propyleneoxide to yield a polycarbonate-based polyol, which has demonstrated benefits in a life cycle analysis (LCA)⁸⁸ and can be used in flexible foam applications.

Isocyanates, on the other hand, can be derived from respective biobased diamines (see also the discussion on polyamides in Section 3.2). Commercial examples are linear pentane diisocyanates (PDI) marketed as STABiO^TM PDI^TM by Mitsui Chemicals⁸⁹ and Covestro's Desmodur N7300 PDI-trimer as a hardener for coatings,⁹⁰ which is based on the phosgenation of pentane diamine derived from glucose.⁸² A wider variety of biobased diisocyanates has been demonstrated in a novel approach by the Burkart group, using flow-chemistry and Curtius rearrangement from renewable carboxylic acids to introduce the isocyanate functionality.⁹¹ For the two mainstay isocyanates TDI and MDI, on the other hand, variants have been introduced by Covestro⁹² and BASF⁹³ based on renewable raw materials from biomass, by a mass balance attribution (see Section 5.4). In the PU building block approach, combining biobased polyols and isocyanates paves the way to introduce biobased content in the final PU polymers.⁹⁴ One example is BASF's Elastollan® TPU which is available with biobased content.⁹⁵

PUs, due to their enormous structural variability, are also used as functional polymers: in adhesives,^{96, 97} in coatings and paints,⁹⁸ and in aqueous systems for hair styling.⁹⁹ Similarly, PU dispersions are employed as adhesives. Epotal® Eco by BASF additionally serves as an example of a PU application offering biodegradation and is used as a laminating adhesive for multilayer packaging¹⁰⁰ for fully compostable systems.

The multitude of different options based on PU molecular design again demonstrates the need for clear guidance on design that goes beyond chemical and polymer research. With these options, sustainable and circular design must be performed on the product level. Given the high reactivity of isocyanates, life cycle based SSbD (Figure 3) will factor in the potential hazard of these substances during production and processing, while considering the non-exposure of users to these intermediates in applications, and the benefits and safety of the final material. Additional paths towards polyurethanes are being explored that do not rely on isocyanate chemistry: Non-isocyanate PUs (NIPUs) have been reviewed.¹⁰¹ Chemical recycling via cleavage of ester polyols, as demonstrated by Magnin et al. using enzymes on polycaprolactone segments, would be another path towards reconstituted PUs in their subsequent life, reassembling the oligomeric products of the ester cleavage by polycondensation.¹⁰²

3.4 General Molecular Design Strategies for Circularity

The distinct structural features of the polymer classes discussed above are important for the circular potential of the respective polymers. As can be seen, the versatile behavior of ester bonds is already being leveraged, e.g., for polyesteramides and in the case of polyurethane building blocks. Generalizing these features that go beyond a single polymer class, molecular design strategies (Figure 11) can be derived and have been explored to improve the circular potential of materials, in the case of structural polymers, depending on their thermoplastic or thermoset nature. These strategies include (1) reducing material complexity on the systems level by interventions on the molecular level, (2) deliberately facilitating recycling via re-monomerization (breakdown into the corresponding monomeric building blocks), (3) transitioning a thermoset into a thermoplast, while still addressing the same applications, and (4) building on existing natural biogenic polymers by modification.

Strategies (2) and (3), remarkably similar to biodegradability, have in common the fact that a functional group in the polymer structure is required which is susceptible to an external stimulus (such as temperature and/or the addition of a solvent and a catalyst) and initiates the recycling process. As a last step, a chemical or physical workup process is required to isolate the monomer or polymer in such a way that it can be used again as starting material. The following section highlights some of the latest achievements and discusses challenges and opportunities of all four strategies.

3.4.2 Enabling Circularity via Re-monomerization and Other Strategies for Backbone Cleavage

Polycondensation polymers such as polyesters, polyamides, and polyurethanes have an intrinsic ability to depolymerize to monomers or oligomers, as the condensation reaction is reversible and depends on a thermodynamic equilibrium. The depolymerization occursby e.g., reaction with solvent such as water, alcohols, or amines present in excess (solvolysis). The back reaction to monomers is favored at high temperatures above the ceiling temperature due to the entropy increase.^{36, 105} The use of catalysts^106-108 and/or special solvents such as ionic liquids can increase reaction rates at lower temperatures.^{109, 110} While the process of re-monomerization has been extensively researched, especially for poly(ethylene terephthalate) (PET),^{26, 103, 106, 111, 112} purification of the obtained monomers is often a challenge and complicates the process for mixed waste streams.

For polyaddition polymers, such as the most widely used polyolefins as well as polystyrene and polyvinyl chloride (PVC), finding such a predefined pathway to monomers is more complicated, as in contrast to polycondensation polymers, C−C bond cleavage must take place. Due to their high ceiling temperature (e.g., 400 °C for ethylene¹⁰⁵) depolymerization only starts at much higher temperatures than for the polycondensation polymers, often yielding an undefined mixture of small molecules (pyrolysis oil) or, in the case of PVC, producing hydrogen chloride instead of the monomers. Even though the ceiling temperature can be influenced for example by adding solvent,¹¹³ the re-monomerization process is often energetically very costly for pure hydrocarbon backbones. Alternative chemical strategies seek to convert such waste into other useful chemicals by C−H functionalization,^114-116 but are likely limited to smaller fractions of the waste. Therefore, important research in the last years has focused on designing new polymers for which chemical recycling to defined monomers is possible. Coates and Getzler highlight in their review the possibilities of ring-opening polymerization (ROP).³⁶ ROP allows a near-ergoneutral polymerization reaction, which means the change in Gibbs free energy is small during polymerization and can therefore easily be reversed by ring-closing depolymerization (RCD). An example for such a near-ergoneutral system is given in Figure 12A for γ-butyrolactone. ROP has mainly been proposed for heterocycles,^{117, 118} where the main advantage over polycondensation polymers is the absence of a small molecule that needs to be added/removed from the process and thereby complicates the reaction. Many examples for three- to eight-membered ring monomers that can be used for closed-loop recyclable polymers are given in the review by Xu and Wang.¹¹⁸ In addition, cyclopentenes might present a possibility for alicyclic monomers to form an all-carbon backbone.^{119, 120}

A more detailed approach on how to replace polyolefins with a polymer that can be easily re-monomerized is given by Abel et al.¹²¹ They synthesize a polyacetal with comparable properties to polyolefins (regarding tensile strength and thermal stability) which could be used for packaging products. Synthesis relies on cationic ring-opening polymerization; re-monomerization of this polymer is possible using an acid catalyst at 140–150 °C and subsequent recovery of the pure 1,3-dioxolane monomers by distillation with a high yield (Figure 12B). Moreover, they tested their approach with a mixed and contaminated waste stream obtaining a remarkably high yield of 96 % of monomers without prior separation of the mixed waste.

Häußler et al. take a different approach to replacing polyolefins by introducing defined functional breaking points into the polymer chain.¹²² These breaking points, which are either polyester or polycarbonate functionalities, are introduced every 18 C atoms, thereby creating molecules with properties like polyethylene (Figure 12C). Recycling back to the monomer can be done by solvolysis at 120 °C in ethanol with a KOH catalyst. They obtain the monomers in high yields and purify them by recrystallization. A similar approach is the catalyst-mediated introduction of 1 % carbon monooxide (CO) into the polyethylene chain through the formation of keto bonds that promote degradation under UV exposure but do not affect the mechanical properties of PE.^{123, 124}

The introduction of functional groups in polyolefins while retaining their application performance might also open up a pathway towards biodegradation. In aqueous environments and towards biochemical transformations or enzymatic attack, the alkyl backbone itself is not easily accessible. Initial degradation of pristine polyethylene in the guts of wax worms, however, has been described.¹²⁵ While plastic uptake by the worms seems to be confirmed, digestion¹²⁶ and potential usefulness of the effect¹²⁷ are still subjects of debate. Utilization of polyethylene by microbes has also been reported,¹²⁸ but limited characterization of the investigated material, low biodegradation rates, and experimental conditions far from current biowaste processing technologies or environmental conditions, make these findings academically interesting, but in our view, unlikely to change the general picture. In many instances, partial biodegradation requires prior abiotic (via oxygen radicals or photochemical) backbone oxidation.¹²⁸ Impurities or deliberately added oxo-degradation catalysts can support these initial steps, but this dependence on prior oxidation limits the generality and leads to the danger of only partial biodegradation and subsequent microplastics formation in the environment. Consequently, oxo-degradable plastics have been banned in the EU.¹²⁹

On the other hand, chain-growth polymers with C−C - backbones, like polyacrylates in the case of many functional polymers, can be equipped with cleavable sites, e.g., ester bonds, in the backbone even via radical polymerization; radical ring-opening polymerization has been reviewed by Tardy et al. for that purpose,¹³⁰ including copolymerization¹³¹ of vinylic monomers with cyclic ketene acetals (Scheme 3).¹³⁰

All these publications present an impressive way of designing new polymers with a focus on both performance and recyclability or biodegradability. However, they do not consider a full life cycle analysis of the entire re-monomerization and re-polymerization process and neglect manufacturing costs, raw material costs, and the carbon footprint for the synthesis of the virgin material. These will be higher compared to the presently used polyolefins or other vinylic chain-growth polymers. However, in the case of monomer recovery, with each recycling cycle, the virgin material costs will play a diminished role. For the recycling process, costs and carbon footprint will be dominated by energy costs due to the reaction at higher temperature and the solvent recycling necessary. Additionally, a crucial factor to consider is the requirement to separately collect or identify and sort the specific polymers that were designed for the recycling process.

These considerations are also the focus of a publication by Vora et al.¹³² who focus on the chemistry of polydiketoenamines (PDKs, Figure 12D). Re-monomerization of the PDKs is achieved in sulfuric acid followed by ion exchange to recover the amine and by extraction to recover the triketone. The subsequent re-polymerization can be done without solvent at room temperature. Here, the production costs of the virgin material were estimated to be 30-fold higher (45 $/kg) than the recycled material, which is in the price range of commodity polymers such as HDPE (1.5 $/kg). Also, the carbon footprint for the virgin material is about 40 times higher than for the recycled material. A high collection rate and thus many recycling loops are therefore a decisive factor for such a circular solution. New business models including return of the materials and involving the whole supply chain will be necessary to implement such materials designed for circularity. The hurdles are specifically high for new polymers, starting out at low volumes, higher prices, and competition with the most abundantly produced polyolefins. Emerging markets in lower income regions globally will struggle even more to bear higher initial cost. Importantly, established bulk polymers such as PE and PP, as mentioned, will also improve along the value chain for more efficient mechanical recycling or conversion to hydrocarbon feedstock by pyrolysis. In any case, circularity will increase the value of EoL plastics, hopefully effectively reducing incentives for unsustainable waste exports and likewise encouraging quantitative collection everywhere on the globe.

3.4.3 Transitioning a Thermoset into a Thermoplast

Designing recyclability by introducing defined breaking points is not only a potential strategy for the inert backbones of polyolefins, as shown in the preceding section, but also decisive for the recyclability of cross-linked polymers (thermosets). Thermosets account for 10–15 % of total plastics production and are characterized by their high chemical, thermal, and mechanical resistance which is caused by the cross-linking between polymer chains.^{133, 134} Since these materials cannot be physically recycled as they are not re-processable by application of heat, design for recycling and/or processes for cleavage are necessary for achieving a circular economy. A viable strategy is to introduce dynamic or degradable cross-links that can be deconstructed by an external stimulus such as heat¹³⁵ or light.^{136, 137} Depending on the amount and position of the labile linkages, degradation products such as monomers, oligomers, or linear (thermoplastic) polymers are achieved and can be separated from additives, fillers, and reinforcements (such as fibers in structural applications) and be re-processed.¹³³ With respect to degradable cross-links, a common strategy is to introduce more thermally sensitive cleavable bonds such as esters^{138, 139} or bonds that can be easily cleaved in an acidic or basic environment such as acetals^{140, 141} or ketals (Figure 13A). A drawback of these degradable cross-links is that the matrix resin is often degraded into undefined fragments and the focus is usually on the recovery of structural fibers (glass or carbon), making closed-loop recycling challenging.¹³³ In contrast, labile bonds such as dynamic covalent networks can create re-processable thermosets. An important class are the vitrimers¹⁴² which show thermoplastic behavior at higher temperatures and behave as thermosets at lower temperatures. These were pioneered by Leibler et al. by introducing ester functionalities into epoxides which allow for exchange reactions at elevated temperatures making a rearrangement of the network structure possible¹⁴³ (Figure 13C). This strategy is also viable for polyurethanes where both transesterification and transamidation reactions,¹⁴⁴ or the introduction of disulfide bonds,¹⁴⁵ can be used to make the networks mechanically reprocessable. Another reversible exchange mechanism that can be used is the Diels–Alder reaction, the [4+2] cycloaddition of a diene and a dienophile¹⁴⁶ (Figure 13B). The adduct is thermally unstable and makes the thermoset recyclable at higher temperatures. An additional advantage of both vitrimers and Diels–Alder systems is their intrinsic self-healing ability, thus retaining mechanical properties after material damage,¹⁴⁷ and contributing to possibly extended lifetimes of products. A commercially available example of a recyclable thermoset is the material for a wind turbine blade generated by the cross-linking of the respective epoxy system with so-called Recyclamines® by Aditya Birla Advanced Materials. These are amine-based curing agents that enable cleavage of the thermoset at defined cross-linking points upon application of moderate heat and acid,^{148, 149} generating recyclable thermoplastic compounds. Using this technology, Siemens Gamesa have announced the production of the world's first recyclable wind turbine blade.¹⁵⁰ Alternatively, Arkema designed a methyl(meth)acrylate-based system which is radically polymerized into a thermoplastic resin called Elium®. Blades manufactured with this technology can also be fully recyclable through depolymerization of the thermoplast.^151-153

5.1 Standards and Legislation for Biodegradable Polymers

As previously mentioned in Section 2, polymers have so far been exempted from parts of chemical regulation due to their stability and high molecular weight. While in some applications biodegradable polymers are seen as a solution to replace persistent materials, regulations are still struggling in creating a consistent framework for the testing of these materials. During (bio-)degradation, the polymer is cleaved into smaller fragments and ideally metabolized into biomass, CO₂, and H₂O. The possible formation of intermediate (bio-)degradation products needs scrutiny,¹⁶⁴ but the development and standardization of specific testing schemes has helped address this issue.

An example shows the importance of standardized methods linked to a recognized legislative framework: in Italy the mandatory compostability for shopping and fruit and vegetable bags was introduced.¹⁶⁵ The development of a framework to test and certify bags suitable for biowaste recollection based on EN 13432 has created the premises for legislation for the mandatory introduction of compostable bags and the ban of PE bags. These measures helped reduce the probability that non-biodegradable bags will enter the composting stream, resulting in compost with substantially reduced plastics contamination and increased biowaste collection rates. Similar legislative measures are currently being followed by other countries.

5.2 Testing Methods for Biodegradability

Reliable testing therefore is key for acceptance, as well as defining design targets for scientists. Assessing polymer biodegradability is linked to several challenges. Polymers need to undergo initial cleavage to become bioavailable for metabolization, which is often linked to longer biodegradation times. This applies to dispersed functional polymers, but it is even more challenging for structural polymers in bulk. For these, the surface-to-volume ratio will influence the kinetics of biodegradation.¹⁶⁶ For this reason, plastic materials are typically assessed in a two-step process, including biodegradation of the matrix material on a micrometric scale, and disintegration of the final microscopic items. A second relevant aspect is linked to the chemical and structural diversity: polymer molecules can be chemically remarkably diverse and inhomogeneous in their properties, due to their polydispersity. Low-molecular-weight fractions are readily assimilated by microbes, and easily accessible partial domains of the polymer could wrongly lead to the classification of a whole material as biodegradable, if full biodegradation is not proven. Therefore, the use of biodegradation half-lives (Dt50) and the extrapolation of biodegradability from reduced testing times should be avoided.

In terms of standard methods to assess biodegradability of polymers, tests must rely on respirometric endpoints, such as tracking O₂ consumption and CO₂ evolution. Existing respirometric methods can be roughly divided into two classes: OECD testing guidelines, developed for the testing of small molecules, and standard methods developed for biodegradable plastics, such as ISO, CEN, and ASTM methods. Both categories have been acknowledged as suitable to assess the biodegradability of microplastic materials, and could work as well for other polymers, but validation and optimization is required. Existing methods to assess biodegradability of polymers have been recently reviewed by Albright¹⁶⁷ and within the ECETOC working group,¹⁶⁸ highlighting the need for systematic validation for different classes of polymers, and the lack of an overall framework to assess and regulate persistent and biodegradable polymers. Considering the urgent need for biodegradable alternatives to current existing materials, methods leading to shorter biodegradation testing times are needed. Studies investigating enzymatic cleavage of polymeric materials have been reported in the literature^{169, 170} and provide an excellent complement to respirometric methods. The use of enzymes allows rapid cleavage of the polymeric structure and bypasses the rate-limiting step in the biodegradation of polymers. Complete cleavage and identification of the oligomers requires thorough analysis by IR, Raman, NMR, GC-MS, LC-MS, or other methods. Challenges are the availability of a library of enzymes covering the most relevant reactions in nature, the different environmental compartments, and complex determination of the fragments formed. A universal, substrate-independent method is still missing.

For a polymer to be biodegradable, all its components, oligomers, fragments, and metabolites will have to be converted into biodegradation products. Methods to assess the biodegradability of these fragments can be based, beyond the testing of the final polymer, on a detailed analysis of the polymer building blocks and the breaking points introduced into the material. Identified oligomers could be either isolated after enzymatic or abiotic cleavage or synthesized in a targeted manner for further investigations. This approach should be favored over the use of labelling techniques on the polymers themselves, due to the high costs and the challenges in synthesizing polymers with the same properties as products on the market. Alternative, attractive methods enabling the identification of metabolites are microbial enrichments.⁴⁶ Microbes from environmental samples can be repeatedly exposed to the polymeric substrate as the only carbon substrate, to enrich exclusively the microbes involved in the biodegradation process, allowing testing with increased concentrations of polymer. This method has been proven to identify the different steps in the biodegradation process and has also provided deep insights into the metabolic pathways.

5.3 Modelling the Sustainable Molecular Design of Biodegradable Polymers

As seen above, biodegradation is a function of all hierarchical levels of polymer design, while testing for biodegradation is burdensome and requires time, making iterative development loops tedious. One of the most attractive emerging tools for the optimisation of the molecular design of polymers towards biodegradability in a certain environment, is molecular modelling. If the end-of-life behaviour of a structure can be accurately predicted before or during its development, this information can be used to guide design towards a more sustainable outcome. It of course also reduces the associated materials, energy, and investment required for such a development.

QSAR (Quantitative Structure–Activity Relationships) models are predictive models which analyse correlations between the properties of a chemical structure and its biological activity.¹⁷¹ These models have been discussed for use by REACH as an alternative method¹⁷² of creating data where none is available. Here, QSARs are recommended for prediction of physicochemical properties and for some environmental toxicity and fate properties, such as biodegradability; they are, however, not recommended for prediction of more complex systems.

For biodegradation, several QSAR models have already been developed which can predict the biodegradation of a molecular structure in a specific environment or dataset.^{173, 174} A number of these are even either freely available, or available through licencing to customers;^{175, 176} most recent developments include OASIS Catalogic¹⁷⁷ and EPI Suite BIOWIN (4, 5 & 6).¹⁷⁸ Some of the more recent tools, such as EnviPath, even use modelling of known degradation pathways to map chemical breakdown pathways. Most of these tools use open-source data from three frequently cited sources, although other databases have been described in the literature.¹⁷⁹ Since the implementation of the REACH Regulation in 2007, a large amount of biodegradation data is now available online through ECHA registration dossiers and is generally thought to be a very reliable dataset. Additionally, two databases are also freely available: the NITE database (based on MITI-I OECD 301C Tests, Chemical Inspection & Regulation Service, Japan) and BIODEG database (Syracuse Research Cooperation, USA).¹⁸⁰ However, prediction models for larger, more complex molecules—especially polymers—are unfortunately lacking, mainly due to the absence of large datasets and the complexity of calculations required for polymer modelling.

Although fewer in number, some additional investigations have been published on modelling the more complex relationships of polymer degradation.^181-184 Most recently, in 2020 Min et al. investigated the environmental degradation trends of 110 polymers to create a predictive model.¹⁸⁵ The degradation modelled was surface erosion, and so could not be described as true biodegradation in the sense of circularity. In most cases, literature describing biodegradability prediction of large molecules does not use microbial biodegradability as its output, but structural descriptions of degradation such as weight loss,¹⁸⁶ or “biodegradation” describing enzymatic or hydrolytic degradation in vivo.¹⁸⁷ This still is a clear deficit, and future research should consistently measure biodegradation in terms of e.g., CO₂ evolution.

With the ongoing digitalization of chemical research,¹⁸⁸ including materials characterization and biodegradation testing, one can expect the body of useful data to drastically increase, allowing further data-driven approaches, as has been demonstrated in other fields for accelerated discovery in multiobjective target settings.¹⁸⁹

5.4 Life Cycle Analyses and Mass Balancing

When designing sustainable polymers on a molecular or systems level, the result, i.e., how sustainable a certain design is in specific aspects, and finally all aspects of the life cycle need to be considered in a comprehensive life cycle analysis (LCA), in comparison to alternative designs. Life cycle CO₂ reduction by use of biomass, for example, needs to be considered jointly with other dimensions such as land or water use within the planetary boundaries.¹¹ Relevant examples e.g., for the introduction of biobased building blocks for polymers,¹⁹⁰ chemical recycling to feedstock,¹⁹¹ and recycling approaches for mixed waste streams,¹⁹² have been published. While complete LCAs are too complex and time-consuming to perform for every design consideration in early-stage research, the level of detail and quality of assessment must increase as a technology matures in the innovation chain. Furthermore, typical LCAs still have limitations applied to circular economy cases, such as representing the non-intended end of life (littering) as a CO₂ sink, which might lead to debatable conclusions. As a mitigation measure, microplastic formation as additional impact category has been suggested.¹⁹³

Another practical hurdle for the introduction of circular feedstock from biomass or recycling is the need for the corresponding production assets. The production plants needed might in many instances be identical to those in the established fossil value chain. Therefore, for a circular transformation to occur, the use of assets with mixed feedstock is unavoidable. Here, industry uses an allocation principle.¹⁹⁴ Recycled or renewable feedstock or raw material entering a chemical value chain at the beginning is allocated to downstream products, which are then marketed with a given renewable or recycled (circular) allocation, without direct physical traceability. The correct stoichiometric mass of circular feedstock at the beginning of the value chain (e.g., oil from plastics pyrolysis) must be in balance with the amount of claimed circular content in the downstream products (e.g., a polyamide grade). A prerequisite for allocation is a system boundary in which a physical connection exists between the production plants (Figure 16), enabling a physical flow that establishes a link between the feedstock and the product. For a transparent communication on mass balance, physical connectivity should be given.

Audits are performed to certify that the correct mass of products with circular content is claimed. Also here, the creation of standards is vital to create acceptance and trust for customers and consumers.

5.5 Digital Representation

Creating a physical, circular material world will require not only the right polymers and processes, but also a well-established digital infrastructure to successfully close the loops. Raw materials are turned to polymers, which are processed to (semi) finished goods and eventually built into a consumer product. Transparency is required along this value chain, e.g., to track and trace the recycled or renewable content in materials and products, as well as LCA or CO₂ emissions data. Digital product passports are required by the EU Green Deal, which contain sustainability-relevant data.¹⁹⁵ Catena-X,¹⁹⁶ for example, is a consortium led by the automotive industry to create a transparent digital value chain. At end of life, digital product passports can help to better sort and recycle certain polymers, if it is known which additives and polymers are contained in the materials. The complexity is high for a digital infrastructure for tracking sustainability data along complex value chains and more work will be required before the first commercial applications go live.