Technological watch

Dynamic polymer network points the way to truly recyclable plastics

The fact that most currently used plastics cannot be easily recycled has produced severe environmental problems, caused considerable losses to the global economy and depleted finite natural resources1,2. Of the most widely used plastics today, thermosets are of great interest for high-performance applications, but are particularly impractical for recycling because they cannot be reprocessed using heat or solvents. , Christensen et al.3 report thermosets formed using covalent links known as diketoenamines, which can be reorganized within the material’s polymer network. Remarkably, the diketoenamines allow the plastics to be recycled in an energy-efficient process to regenerate pristine monomers, which can then be used to make thermosets that are almost identical to the original material.

Read the paper: Closed-loop recycling of plastics enabled by dynamic covalent diketoenamine bonds

Conventional plastics are polymers that are designed to survive ambient onslaughts from light, water, heat and so on. But this resilience also makes them difficult to recycle. Most widely used plastics can be recycled in principle, but reprocessing is costly, energy-demanding and usually produces poor-quality materials. The performance of recycled plastics is therefore typically not as good as that of newly made polymers4.

Thermosets are particularly problematic because the polymer molecules are permanently crosslinked by covalent bonds. This crosslinking means that thermosets have much better solvent resistance and thermomechanical properties than do thermoplastics — a family of polymers that lack molecular crosslinks. Thermosets are therefore highly attractive materials for present and future use in high-temperature electronic devices or in automotive applications, and they currently account for 15–20% of global plastic production5. But the fact that they cannot be reshaped or recycled, either by melting or by processing in solution, is a major drawback6.

Researchers are therefore trying to make thermosets that can be recycled, reshaped in energy-efficient ways, or both, without compromising the materials’ excellent chemical and thermal resistance7. One idea is to use reorganizable polymer networks in thermosets8,9. Such plastics would have the desirable properties of thermosets under most conditions, but could be reshaped using certain stimuli.

For example, thermosets known as vitrimers have been intensively studied in the past decade. Vitrimers have glass-like fluidity: they flow when heated because the covalent bonds reorganize at high temperatures. This allows vitrimers to be thermally reshaped and eventually recycled, like thermoplastics10. But because high temperatures are often required during reprocessing, these dynamic thermosets tend to degrade with each usage cycle, which reduces their utility.

Some thermosets7 have been designed to depolymerize at low temperatures under acid conditions. This allows the plastics to be used in closed-loop cycles — in which monomers are recovered from the depolymerization and used to remake the plastic11,12. Nevertheless, more work is needed to develop polymers that can yield 100% monomer conversion rates, as well as to improve the sustainability of the chemistry and make it suitable for industrial-scale processes, and to find ways of depolymerizing thermosets in mixed plastic waste.

In their search for thermosets to use in closed-loop cycles, Christensen et al. designed polymers that form dynamic crosslinked networks based on diketoenamine bonds (Fig. 1). The authors’ networks are produced almost spontaneously from starting materials in a solvent-free, versatile process that could, in principle, be used to make many analogous materials. Ketoenamine groups — which differ from diketoenamines by having one carbonyl (C=O) group, rather than two — have previously been shown to reorganize within a polymer network to enable polymer reshaping13.

Figure 1 | Dynamic covalent bonds enable monomers to be recovered from a polymer network. Christensen et al.3 have synthesized a polymer network in which the molecules are crosslinked by diketoenamine bonds. The polymer is made by simply grinding together two monomers (amine and triketone compounds), and it belongs to a useful family of plastics known as thermosets. The crosslinks can reorganize within the polymer because amine (NH2) groups from different molecules react with the diketoenamine groups. When the polymer is treated with strong acid, the monomers can be recovered and used to make the polymer again.

Previous work14 had shown that ketoenamine-containing polymer networks are stable in acid, but Christensen et al. found that, unexpectedly, their diketoenamine network can be depolymerized by treating it with a strong acid. The process takes less than 12 hours and produces excellent monomer yields, with the difference in reactivity being associated with the presence of the extra C=O group in the diketoenamine crosslinks. The authors found that the monomers could be recovered using an operationally simple procedure, and that they could be reused to make thermosets that had nearly identical properties to the material that had been depolymerized.

Christensen et al. went on to carry out the depolymerization of their material in the presence of other widely used plastics and additives (such as fibreglass, colourants or flame-retardants) that are often found in plastic waste that has been collected for recycling. No contamination of the recovered monomers by these materials was observed, demonstrating that the reaction tolerates many different additives. This is a particularly interesting finding because it suggests that the depolymerization method could be used to recycle fibre-reinforced plastics — finding a way to recycle such composites is one of the biggest challenges in the field.

This latest work opens up avenues of research for the preparation of fully recyclable thermosets, but it also raises scientific and technological questions. For example, Christensen and colleagues carried out their polymerizations on a gram scale — can the reactions also be carried out on industrial scales for plastic manufacturing? Moreover, the authors needed to use large amounts of acidic and basic water to recycle and purify the monomers, so will this process be environmentally friendly on industrial scales?

To facilitate the translation of Christensen and colleagues’ strategy from academic experiments to commercial industrial processes, dynamic polymer networks should now be implemented in widely used thermosets such as polyurethane, polyester and epoxide resins. Academics and industrialists will also need to work together to define the costs and benefits of next-generation plastics compared with current polymers, and to carry out life-cycle assessments15. This will require chemists to work routinely with other scientists to evaluate the consequences of the entire life cycle of synthesized products.

The development of robust materials that combine excellent chemical and heat resistance with outstanding recycling capabilities could greatly assist the transition from the current linear model of plastics production and consumption — in which limited resources are used to make products that have a finite lifetime and are then disposed of — to a sustainable, circular economy that minimizes waste and maximizes resource use. Next-generation plastics will certainly need to be compatible with closed-loop life cycles to sustainably satisfy the expectations of the world’s growing population. In the meantime, work such as this from Christensen et al. takes us considerably closer to the synthesis of plastics that have minimal environmental impact.

Publication date: 22/04/2019

Reference: 10.1038/d41586-019-01209-3

Nature International weekly journal of science (News)

This project has been co-funded with the support of the LIFE financial instrument of the European Union [LIFE17 ENV/ES/000438] Life programme

The website reflects only the author's view. The Commission is not responsible for any use thay may be made of the information it contains.
Last update: 2019-05-09