the zipper problem in textile recycling

Problem statement

Recycling old clothes back into new fibre — mechanical unravelling, cellulose pulping, or polyester depolymerisation — is the goal of the EU's textile strategy, and every one of those processes shares a precondition: the garment must arrive as textile only. Zippers, buttons, rivets, labels, elastic, coatings and prints are "disruptors" that jam shredders, contaminate pulp and poison catalysts, and the JRC's 2023 assessment states flatly that for the main recycling families "non-textile accessories such as zippers and buttons and generally also coatings must be removed." Today that removal — cutting out hard parts from millions of heterogeneous garments — is a manual step. Automated sorting has advanced (near-infrared machines can now assess fibre composition at 900–4,500 kg per hour versus 100–150 kg per person-hour by hand), but the JRC finds that "no automated technologies are currently available that could take over actions other than assessments of fibre and colors." The pre-treatment step, not the chemistry, caps how much collected textile can be recycled.

Why this matters

More than 8 million tonnes of post-consumer textile waste are incinerated or landfilled in the EU every year; separate collection is about 2.4 Mt/yr against a sorting capacity of roughly 1.8 Mt/yr, and Member States were required to set up separate collection for textiles by 1 January 2025, so volumes are rising into a system that cannot process them. The JRC estimates that only about 20% of separately collected textiles are homogeneous in fibre and have "no or removable disruptors," and that with current machinery at most ~25% of that fraction becomes high-quality spinnable fibre — so the disruptor bottleneck is compounded by yield loss. Because sorting and pre-processing are manual and labour-intensive, roughly half of collected European textiles are exported to third countries (about 1.8 Mt/yr, much of it attributed by the JRC to "the lower cost of the (manual) sorting"), where the JRC judges it "likely that a large share of the textiles is finally not re-usable," shifting the environmental burden abroad. Fibre-to-fibre recycling cannot scale in Europe unless the hand-labour step is engineered out.

What’s been tried and why it hasn’t worked

Mechanical recyclers producing wiping rags and insulation tolerate heterogeneous input and simply cut around hard parts, but that is downcycling; advanced mechanical, pulping and depolymerisation routes all require the pre-treatment. Design-for-disassembly is the main technical response: Wear2's microwave-triggered sewing thread and Resortecs' heat-dissolving thread let seams release in an oven or microwave in under a minute (Wear2 reached TRL 7 in the Circtex project, 2019–2022). Their limitation is structural — they only work on garments that were sewn with the special thread, so the installed base of everything already in wardrobes and bring-banks (and everything brands keep sewing conventionally) still needs cutting. Chemical pre-treatments for coatings and laminates (selective dissolution, triggerable polymers, reversible crosslinking) exist at varying TRL but do not remove metal and plastic hardware. Automated NIR sorters identify fibre and colour but cannot locate or remove trims, and NIR itself has limits — dark colours and chemically similar fibres — while elastane may hide undetected in fractions "classified as 'pure' … due to analytical limitations." The economics push the manual step offshore rather than automating it: the JRC expects manual processing "may continue to take place to a significant extent in countries with low labor costs, inside or outside the EU."

What would unlock progress

Two unlocks: an automated trim-removal station — vision or metal/X-ray detection of hardware and coated regions on flattened garments, followed by robotic or die cutting that maximizes recovered textile area — which is essentially the same problem as automated defect-cutting in leather and fabric manufacturing or foreign-object removal on food lines; and a design-side pathway (disassembly threads, minimal-trim design, digital product passports flagging hardware) that shrinks the problem for future garments. Data would help both: there is no published distribution of hardware types, positions and masses across the post-consumer stream that a machine designer could target.

Entry points for student teams

A team could build a benchtop detect-and-cut rig — camera plus metal detection over a flattened garment, a learned model that segments zippers, buttons, labels and prints, and a cutting path that maximizes usable fabric — and measure throughput, fabric yield and residual contamination against manual cutting on a sample of donated garments. A second team could characterize the disruptor content of a real sorted stream (types, mass, position, attachment method) to publish the first design dataset. An industrial-design team could prototype trim and seam specifications for a garment category that keep the item recyclable and test them with a local recycler. Skills: computer vision, mechatronics, textile/apparel engineering, industrial design, waste-stream sampling.