Recycling Technologies
Mechanical, chemical, metallurgical, biological
Before we start
Every few months a headline promises that a new technology will finally solve plastic: enzymes that eat bottles, reactors that turn waste back into oil, "infinitely recyclable" plastic. Some of it is real science. Almost none of it is yet working at the scale the headlines imply. Learning to tell those two things apart — the science that works from the business that works — is the single most useful skill this module can give you, and it is what separates an engineer from a press release.
Recycling is not one process but a spectrum of technologies, each suited to different materials, costs, and quality targets. This module covers the four families and gives you a framework for choosing, and judging, between them.
By the end of this module you should be able to: (1) describe the four technology families and what each does well and badly; (2) match a waste stream to the most appropriate technology and justify it; (3) tell a proven technology from a promising one, using a live, current example; (4) read a "breakthrough" claim and ask the right questions about it.
Part A — The four families
Mechanical recycling: the workhorse
Most of what happens today. Sort by type, shred, wash, and reprocess (melt and pelletise for plastics, pulp for paper, remelt for metals and glass) without changing the chemistry. It is the cheapest, most energy-efficient route for clean, single-material streams, and it is proven at scale. Its weakness is quality loss: plastics degrade with each melt, so mechanical plastic recycling usually downcycles. It is the default; reach for something else only when it can't cope.
Chemical (molecular) recycling: powerful, mostly unproven at scale
These break materials down to their chemical building blocks to rebuild them. The main types: depolymerisation (break a polymer back to its monomers, then repolymerise; enzymatic PET is the headline case), pyrolysis (heat mixed plastics without oxygen into an oil or gas), and gasification (higher temperatures to a syngas).
The appeal is real: chemical routes can in principle handle mixed and contaminated plastics that mechanical recycling rejects, and depolymerisation can yield virgin-quality output. The honest caveats are equally real. These processes are energy-intensive and capital-heavy, many announced projects have slipped or stalled, and some "plastic-to-fuel" routes are really waste-to-energy in disguise, since burning the fuel re-emits the carbon. Treat chemical recycling as a complement for the hard fraction, not a replacement for mechanical recycling, and always ask for the life-cycle numbers and the operating record.
Metallurgical recycling: where metals and e-waste go
Metals are recovered in furnaces: steel in electric arc furnaces, aluminium in remelting furnaces, copper and precious metals through smelting and refining. For e-waste and batteries, the route combines mechanical pre-processing with pyrometallurgy (smelting) and hydrometallurgy (chemical leaching and refining). It handles complex mixtures well and recovers high-value metals, but it is energy-intensive and needs strict emissions control. This is the most powerful mature technology in the field, and the spotlight below shows it.
Biological recycling, and the last resort
Composting and anaerobic digestion return organic matter to use as compost or biogas. At the frontier, engineered enzymes break down some plastics (overlapping with chemical recycling). And waste-to-energy (incineration with energy recovery) is not recycling: it removes material from the loop permanently and should be used only for what genuinely cannot be recycled.
The part they don't teach in class: "the chemistry works" is not "the business works"
This is the judgement that will protect you for an entire career. A technology can be scientifically sound and still fail to become a real, affordable, operating plant. The live case is Carbios, in the spotlight, and the broader pattern is everywhere: across the world, advanced-recycling plants have been announced with great fanfare and then delayed, downsized, or quietly cancelled when the economics or the engineering proved harder than the press release. When you read that a technology "can recycle any plastic infinitely," your job is not to cheer or to sneer. It is to ask: at what scale, at what cost, with what energy, and is it doing it today or promising to?
Part B — Company & Tech Spotlight: two ends of the spectrum
Carbios — a frontier technology, told honestly
Carbios is a French company developing enzymatic depolymerisation of PET: engineered enzymes break the plastic into its monomers, which can be rebuilt into virgin-quality PET, in principle from coloured or mixed waste that mechanical recycling can't handle. The science is genuinely impressive. Its demonstration reactor in Clermont-Ferrand can break roughly two tonnes of PET — around 100,000 ground-up bottles — into its building blocks in a matter of hours (MIT Technology Review, 2021).
The part most summaries skip
In April 2024 Carbios held an opening ceremony for its first commercial plant in Longlaville, a 50,000-tonne-per-year facility it called the world's first PET biorecycling plant, then pointing to a 2026 start-up. Since then the project has stalled, mainly on financing. By early 2026 the company had secured ~€42.5m of public funding and pre-sales for roughly half the plant's capacity, but a portion of the financing was still not closed; it now targets closing that financing around Q3 2026 and starting production in H1 2028 — a clear slip from the original 2026 date. Alongside this it has faced management departures and legal disputes, and has pivoted toward licensing its technology abroad, including a project with Wankai in China.
So, is enzymatic recycling "real"?
The chemistry works. Whether it works at industrial scale, profitably, on schedule, is still being decided in front of us — and the timeline has already moved by about two years. That distinction is the most useful thing in this module. Treat any future "Carbios is now producing at scale" claim the same way: confirm it against the operating record, not the announcement. (Carbios's situation is volatile; re-verify its current status before teaching this section.)
Umicore — proven metallurgy and "urban mining"
At the proven end of the spectrum is Umicore. Its Hoboken site near Antwerp is an integrated smelter and refinery the company describes as the world's most advanced of its kind, recovering 17 precious and specialty metals from over 200 complex input streams, including electronic scrap.
Umicore also recycles batteries: a pyrometallurgical smelter yields a slag (carrying aluminium, manganese, lithium) and an alloy (cobalt, nickel, copper), with the alloy refined hydrometallurgically into materials for new batteries. And recall DESPRAY from Module 3: aerosol recycling is a third, specialised technology for a hazardous stream. Frontier biotech, mature metallurgy, niche purpose-built machines — the field is all three at once.
Who hires across this spectrum
Chemical and biotech recyclers hire process, chemical, and bioprocess engineers and scale-up specialists. Metallurgical recyclers hire metallurgists, process and environmental engineers, and emissions-control specialists. "Recycling engineer" is not one job; it is a family of them.
Part C — Do the work
Task 1 — Match the technology to the stream
For each, name the most appropriate technology family and justify it in one line (quality, energy, or maturity): (1) clean, single-colour PET bottles from a deposit system; (2) mixed, contaminated, multi-layer plastic film; (3) spent lithium-ion EV batteries; (4) household food and garden waste.
Task 2 — Proven or promising?
Write a short argument (6–8 sentences): should a European city today plan its plastic strategy around enzymatic recycling? Use the current Carbios situation as evidence on both sides. A strong answer separates "the technology is real" from "it is reliably available at scale now," notes that the timeline has already slipped about two years, and reaches a defensible conclusion.
Task 3 — Interrogate a claim
Find one recent news headline about a recycling "breakthrough" (any technology). Write the four questions you would ask before believing it (scale, cost, energy, operating record), and answer as many as the article actually lets you. This builds the habit that protects you from hype.
Deliverable. Your four technology matches with reasons, your "proven or promising" argument with a conclusion, and your interrogation of one real headline. Upload as text and/or a file.
Watch
One is enough here. Company and lab demonstration videos of chemical recycling are mostly not openly licensed, and I won't add one just to reach two.
A material scientist tours real advances, including enzymes that aim to make plastic recyclable many more times, while keeping the bigger picture that using less beats recycling more. A grounded look at where material technology is heading.
Open on TED ↗From the field
Sources
- Carbios — enzymatic PET depolymerisation; demonstration reactor in Clermont-Ferrand (~2 t/batch, ~100,000 bottles, 10–16 h) (MIT Technology Review, 2021). Longlaville commercial plant ~50,000 t/year, ~€230m, opening ceremony April 2024, originally targeting 2026 start-up; project stalled on financing; ~€42.5m public funding secured, ~50% pre-sales, ~€60m cash end-2025; financing close targeted ~Q3 2026, production targeted H1 2028; management departures and legal actions; licensing pivot incl. Wankai (China). carbios.com financial releases (Dec 2025, Mar–Apr 2026); labiotech.eu (Mar 2026); Yahoo Finance / just-style (Apr 2026). Volatile — re-verify before teaching.
- Umicore Hoboken — integrated smelter/refinery, 17 metals from 200+ input streams; ~300 g gold per tonne e-scrap vs ~5 g per tonne ore; battery pyro/hydrometallurgy. https://www.umicore.com/
- DESPRAY Environmental — aerosol recycling (see Module 3). https://www.despray.com/
- Andrew Dent, "To eliminate waste, we need to rediscover thrift," TED (2018). https://www.ted.com/ (Creative Commons).
All URLs accessed June 2026.
Submit your work for feedback
Each task is graded immediately by an AI tutor using the rubric below. An instructor can override the AI score and leave a final comment.
Match the technology to the stream
For each waste stream, name the most appropriate technology family and justify it in one line (quality, energy, or maturity).
Grading rubric
Credit for: (1) PET bottles → mechanical recycling (clean, single-material, proven, lowest energy); (2) mixed film → chemical recycling or — honestly — waste-to-energy where chemical isn't available; (3) EV batteries → metallurgical (pyro + hydro, Umicore-style); (4) food/garden → biological (composting or anaerobic digestion). One-line justification per choice. Score out of 10.
| Item | Technology family | Justification (1 line) |
|---|---|---|
| Clean, single-colour PET bottles from a deposit system | ||
| Mixed, contaminated, multi-layer plastic film | ||
| Spent lithium-ion EV batteries | ||
| Household food and garden waste |
Proven or promising? — the Carbios question
Should a European city today plan its plastic strategy around enzymatic recycling? Use the current Carbios situation to argue both sides, then state your conclusion.
Grading rubric
Credit for: (1) clear separation of science-works vs business-works; (2) reference to the Longlaville slip (originally 2026 → now H1 2028); (3) a defensible conclusion (don't bet today's strategy on it, but track it; or pilot at small scale) rather than fence-sitting. Score out of 10.
Interrogate a real "breakthrough" headline
Find a recent recycling 'breakthrough' headline. Answer the four screening questions — and mark honestly which ones the article does NOT answer.
Grading rubric
Credit for: (1) a real, linked headline from the last ~12 months; (2) all four questions clearly stated; (3) honest acknowledgement of which questions the article does NOT answer (this is the point of the exercise). Score out of 10.
Check your reading
- 1.
Which technology family is the cheapest, most energy-efficient default for clean, single-material streams — but tends to downcycle plastics with each pass?
- 2.
Which of these is NOT a form of chemical (molecular) recycling?
- 3.
Carbios's commercial Longlaville plant — originally targeting a 2026 start-up — has now slipped to which target?
- 4.
Roughly how much gold can Umicore Hoboken recover per tonne of electronic scrap — compared to ~5 g/tonne in natural gold ore?
- 5.
Why is waste-to-energy (incineration with energy recovery) NOT counted as recycling?
- 6.
When you read that a new technology "can recycle any plastic infinitely," what is the right engineering question to ask?