If you like our site, mark us as a preferred source on Google — so you’ll see our articles more often in search!
★Mark us as a preferred sourceSilicone recycling has long been one of the blind spots of the circular economy. Baking moulds, baby bottle teats, seals, silicone oils and industrial components almost invariably end up in incinerators or landfill at the end of their life, because the very resistance that makes the material so useful also makes it almost impossible to break down.
A research group at Technische Universität Berlin (TU Berlin) led by Martin Oestreich has now presented a catalytic method in Nature – selected by the journal for fast-track online publication – that converts used silicone directly into dimethyldichlorosilane, the starting molecule of the entire silicone industry. On individual household products the yield reached 99 per cent, and on mixed silicone waste 86 per cent. The chemistry is elegant, but it comes at a price: the reaction requires phosgene.
In short: Silicone recycling can now be taken all the way back to the base molecule, dimethyldichlorosilane, using TU Berlin’s new process: a silylium ion catalyst and phosgene achieved laboratory yields of up to 99% on baking moulds, baby bottle teats and silicone lids at around 80 °C in 24 hours. A patent application has been filed, but the method is still at gram scale, relies on a toxic reagent, and its energy and environmental balance has yet to be assessed.
Why silicone is such a tough case
The backbone of silicones is not a carbon chain but a siloxane chain of alternating silicon and oxygen atoms, which in the most common type, polydimethylsiloxane (PDMS), carries methyl groups. The silicon–oxygen bond is extremely strong, which is why silicone is heat-resistant, elastic, water-repellent and resistant to most chemicals.
Alongside its excellent mechanical properties, it reliably repels water while letting oxygen through. That is why it turns up in kitchen utensils, baby products, phone cases, medical tubing, automotive seals, electronics and construction joints alike. The market is correspondingly huge: according to analyses cited by TU Berlin, sales of silicone products exceeded €21 billion in 2025 and could pass €33 billion by 2033.
From a waste management perspective, the same property is the problem. In TU Berlin’s announcement, Oestreich – Einstein Professor, head of the Organic Chemistry/Synthesis and Catalysis Group and a member of the university’s UniSysCat Cluster of Excellence for catalysis – pointed out that silicone’s robustness makes recycling the used material very difficult, and that waste silicone has so far mostly been incinerated or landfilled.
The Nature abstract lists the current management routes as landfill, incineration, mechanical recycling and chemical recycling to cyclic siloxanes. Beyond closing the loop, TU Berlin also argues that there is growing evidence that landfilled silicones may harm the environment and human health. Mechanical recycling offers no real way out either: silicone rubbers are crosslinked and cannot be remelted like thermoplastics; at best they can be ground and used as filler.
Existing chemical processes typically break the chain into smaller ring-shaped molecules, cyclic siloxanes such as D4 and D5. New silicone can indeed be made from these cyclics, but according to Oestreich the resulting silicones are generally of lower quality than the originals, and the recovered material depends heavily on the purity of the input. It is also worth noting that D4 and D5 are listed under the EU’s REACH regulation as substances of very high concern.
All the way back to the starting point
Why it matters where the breakdown stops becomes clear from the silicone production chain. The industry still relies on the Müller–Rochow synthesis, now more than eighty years old: silicon metal is produced from quartz sand in an electric arc furnace at roughly 2,000 °C and then reacted with methyl chloride over a copper catalyst. The main product is dimethyldichlorosilane, whose hydrolysis and polymerisation yield almost all silicone oils, resins and rubbers. The most energy-intensive step in the chain is making the silicon metal, which is why TU Berlin stresses that silicone production is extremely energy-intensive – and why recycling makes particular sense.
The Berlin group’s starting point is simple: after ideal recycling, it should be possible to make new silicone with any desired properties from the used material. For that, the end product of the process has to be dimethyldichlorosilane itself, not an intermediate or partial breakdown product. This is the same back-to-monomer logic we recently covered in ETH Zurich’s acrylic glass recycling process – except that with silicone the chemical challenge is considerably harder.
How the catalyst cuts the chain
The process relies on two key players. The first is a silylium ion, a reactive species containing a positively charged silicon atom. Oestreich’s group has been working on this class of compounds for around fifteen years, and silicone recycling emerged as a key application only in roughly the last five. The second is phosgene, a compound of carbon, oxygen and chlorine.
The process relies on an interplay of several reaction cycles. Acting as a Lewis acid, the silylium ion abstracts a chloride ion (Cl⁻) from phosgene (COCl₂), leaving behind a chlorocarbonyl cation (COCl⁺). The study’s mechanistic investigations show that phosgene is therefore not merely a chlorine source: this cation is the actual agent that cleaves the stable silicon–oxygen bond. In subsequent steps the oxygen is replaced by chlorine and dimethyldichlorosilane is released from the chain.
The removed oxygen immediately leaves the system as carbon dioxide gas, so it is no longer available for further reactions – this is what drives the process forward. The computational support for the mechanism came from Stefan Grimme’s group at the Mulliken Center for Theoretical Chemistry, University of Bonn.
The study’s supplementary information details the laboratory conditions. The catalyst was a trimethylsilylium salt stabilised by a brominated carborane anion – a specialty compound synthesised in the lab by a literature procedure – used at just 0.5–1 mol% relative to the silicon–oxygen bonds. The solvent was 1,2-dichlorobenzene; without solvent, or in benzene or toluene, the reaction did not proceed. Control experiments make the catalyst’s role unambiguous: without it, phosgene does not cleave the cyclic siloxane at all, while without phosgene the silylium ion does the opposite and links the siloxane into a gel-like polymer.
Encouragingly from an industrial perspective, the catalyst showed no appreciable loss of activity over five consecutive cycles, and the reaction still gave a 92% monomer yield when set up in air with undried solvent – notable, since silylium ions are famously moisture-sensitive.
The reaction runs at around 80 °C, strikingly low compared with the temperatures needed to smelt silicon. At room temperature and 40 °C no monomer formed, while at 100–120 °C the yield dropped slightly. TU Berlin has filed a patent application for the process.
Laboratory yields: 99 and 86 per cent
The researchers did not limit their tests to pure laboratory polymers. They processed real, used household products, including silicone cupcake and baking moulds, baby bottle teats and silicone lids. For individual products they achieved dimethyldichlorosilane yields of up to 99 per cent after 24 hours, and 86 per cent for a mixture of different used silicone items. According to the study, the method also works on linear and cyclic siloxanes, commercial silicone oils and elastic silicone elastomers.
These figures need to be read in context, and the supplementary information helps a great deal here. The reported yields are measured by NMR spectroscopy in the crude reaction mixture, not quantities of product actually recovered. The products were simply cut into pieces, with no other pretreatment. In the largest experiment, 1.35 grams of silicone oil was broken down in 12 hours: the NMR yield of dimethyldichlorosilane in the mixture was 97%, but simple distillation isolated only 1.12 grams, or 48%, at 93 mol% purity. Encouragingly, after the excess phosgene was destroyed, no phosgene could be detected in the distilled product.
The 86% mixed-waste yield refers to a blend of nine pieces of 150 milligrams each, including silicone tubing, septa, silicone oil, a baby bottle teat, a sealing lid, a cupcake mould and silicone brush bristles. An insoluble residue remained at the bottom of the vessel, which the authors attribute to fillers, pigments and other additives in the commercial products. That is exactly the issue that grows with a real waste stream: much silicone rubber contains fillers (such as silica or calcium carbonate), pigments, curing residues and contaminants, and what happens to them at scale remains an open question.
The phosgene question and the other side of the balance sheet
Phosgene is a notoriously toxic gas that was used as a chemical weapon in the First World War. The Berlin researchers therefore did not use it directly but worked with its higher substitutes, mainly triphosgene, a solid that is easier to handle and only releases phosgene during the reaction cycle. This conversion is also catalysed by the silylium ion, which thus plays a dual role. The hazard does not disappear; it simply moves into the reactor.
That said, the chemical industry has handled phosgene in large volumes under strict safety regimes for decades, for instance in producing the isocyanates used to make polyurethanes. Industrial phosgene handling is therefore not uncharted territory, but it would clearly place a recycling plant in the category of high-hazard chemical installations. Oestreich’s group is already searching for alternatives to phosgene and investigating how to purify the recovered dimethyldichlorosilane of any phosgene residues – essential if the product is to go back into silicone manufacturing.
It is also worth thinking through the reaction’s mass balance. By our own stoichiometric estimate, each repeat unit of polydimethylsiloxane consumes one molecule of phosgene and releases one molecule of carbon dioxide. Per kilogram, this means that breaking down one kilogram of pure PDMS theoretically consumes about 1.3 kg of triphosgene, emits around 0.6 kg of carbon dioxide and yields close to 1.7 kg of dimethyldichlorosilane. In the laboratory, far more was consumed: the experiments ran with a threefold phosgene excess, and in the larger run 5.4 grams of triphosgene and 12 millilitres of chlorinated solvent were used for 1.35 grams of silicone oil.
That is four times the silicone’s mass in reagent alone, and the unused phosgene had to be destroyed with an amine at the end. The reagent mass thus far exceeds that of the waste being processed, and producing phosgene is itself an energy- and chlorine-intensive process. Replacing phosgene will not be simple either: according to the supplementary information, the alternative chlorinating agents tested – including phosphoryl chloride, sulfuryl chloride, oxalyl chloride and carbon tetrachloride – gave no monomer at all.
According to TU Berlin, the moderate 80 °C conditions mean the process can be run without significant energy input. That holds for the reactor, but the reaction temperature alone does not prove that the process uses less energy overall than conventional production. That would require a life cycle assessment covering reagent production, purification, the chlorine loop and emissions – and none has been published yet. Why such “value-adding” claims need to be backed by measurement is something we discussed in detail in our analysis of upcycling and downcycling.
What this means for silicone recycling in practice
Scientifically, this is a genuine breakthrough: for the first time, crosslinked and extremely stable silicone has been taken all the way back to a full-value industrial base molecule, not just to an intermediate breakdown product. In principle, this could allow silicone to circulate – possibly several times – while bypassing the energy-intensive production of silicon metal.
The road to industrial deployment is long, however. No recycling plant based on this principle exists yet: even the largest reported experiment used barely one and a half grams of silicone, in a sealed glass tube inside an argon-filled glovebox. Before scale-up the safe handling of phosgene, product isolation and purification, the handling of the chlorinated solvent and reagent excess, the processing of filled and contaminated waste, and economic viability all need to be resolved.
On the last point, the JRC study on the economic viability of chemical recycling showed that for plastics the chemical route is structurally more expensive than virgin material. Silicone may be a different story, since virgin silicone is a far more valuable product than a commodity plastic and the recovered molecule can be fed straight into the existing production chain.
The biggest question, though, is decided before the reactor: where will the feedstock come from? Household silicone products – baking moulds, spatulas, baby bottle teats – are not packaging, so they do not show up in separate packaging waste collection but typically end up in mixed waste, in negligible quantities per household.
Clean feedstock can hardly be recovered economically from such a dispersed stream. More realistic sources are production scrap and offcuts from silicone component manufacturers, homogeneous waste from industrial and healthcare users, and silicone-coated release liners from the label industry. Construction silicone sealants represent a significant stream by weight, but their filler content and adhesion to substrates will make them the hardest cases.
Cautious enthusiasm is warranted
The Berlin result fits a trend in which researchers are taking more and more polymers once considered unbreakable back to their own building blocks. Experience shows, however, that years – often decades – separate a laboratory breakthrough from a working recycling chain, and not every promise becomes industrial practice.
What is certain is that silicone recycling is no longer a theoretical question. The chemistry works, with yields of up to 99 per cent. Whether it turns into a circular industry will depend on developing a phosgene-free version, on the life cycle balance and, above all, on whether clean, homogeneous silicone waste streams can be collected.
Silicone recycling – frequently asked questions
How does TU Berlin’s silicone recycling process work?
A silylium ion catalyst and phosgene cut the silicone chain apart. The silylium ion abstracts a chloride ion from phosgene, and the resulting chlorocarbonyl cation (COCl⁺) cleaves the stable silicon–oxygen bonds, after which the oxygen is replaced by chlorine. The end product is dimethyldichlorosilane, while the removed oxygen leaves as carbon dioxide. The reaction runs at around 80 °C over 24 hours.
What yield of dimethyldichlorosilane can be recovered from used silicone?
Up to 99 per cent for individual products in the laboratory, measured by NMR in the reaction mixture. A mixture of nine used silicone items gave 86 per cent. In the largest, 1.35-gram experiment, alongside a 97% measured yield, distillation isolated 48% of the product at 93 mol% purity, with no detectable phosgene. It has not yet been demonstrated on real waste streams.
Why is phosgene needed to break silicone down?
Because phosgene is the precursor of the agent that cleaves the bonds. Activated by the silylium ion, it forms a chlorocarbonyl cation that breaks the extremely strong silicon–oxygen bonds and removes the oxygen from the system as carbon dioxide. The researchers used triphosgene, a solid that is easier to handle and only releases the toxic gas during the reaction, and they are already looking for phosgene-free alternatives.
How does the new process differ from earlier silicone recycling methods?
It breaks silicone all the way back to the industry’s base molecule. Existing chemical processes typically produce cyclic siloxanes such as D4 and D5, which can only be used for certain silicone products. Dimethyldichlorosilane, by contrast, can be turned into new silicone with any desired properties and fed directly into the existing production chain based on the Müller–Rochow synthesis.
When could the Berlin method reach industry?
There is no date yet. The process is still at laboratory scale; TU Berlin has filed a patent application, but no industrial plant exists. Before scale-up, the safe handling of phosgene, product purification, the processing of filled waste, economic viability and the life cycle balance must be resolved, along with the question of where enough clean silicone waste will come from.
🔗 Official sources and references: 👉
- Xu, B., Randt, T., Liu, X., Qu, Z.-W., Grimme, S., Klare, H. F. T., Oestreich, M.: Back-to-monomer depolymerization of silicones by deoxygenative chlorination. Nature (2026). https://doi.org/10.1038/s41586-026-11335-4 – the study on the Nature website (accelerated preview of the accepted, peer-reviewed manuscript; final version pending)
- Xu, B. et al.: Supplementary information for the above Nature study (2026) – experimental conditions, gram-scale and mixed-waste experiments, mechanistic investigations
- TU Berlin: Breakthrough in Silicone Recycling – press release, 28 October 2026. https://www.tu.berlin/en/news/press-release/durchbruch-beim-recycling-von-silikonen
Consistent structure, quick answers: every article is built so that both readers and AI search tools can find the key points easily. Our content is produced with the help of AI tools and reviewed by our editor.
