circular-economy · chemistry · food-safety · family: the missing yardstick
food-grade polyolefins have no rulebook
Recycled PET goes back into bottles every day, but post-consumer HDPE and PP almost never go back into food packaging — the polymer soaks up contaminants and nobody has agreed how clean is clean enough
Problem statement
Europe recycles PET beverage bottles back into food-contact bottles at scale because PET is a "low-diffusive" polymer: contaminants from misuse (someone storing chemicals in a bottle) are absorbed only at low concentrations and are removed by established super-clean processes, many of which EFSA has evaluated in process-specific opinions. Polyolefins — HDPE milk jugs and closures, PP tubs and trays — are "high-diffusive": in EFSA's words, "the diffusivity in HDPE is much higher than in PET," so the Panel expects misuse contamination in washed HDPE to be at least as high as in PET, and, as Welle puts it, traces left after decontamination migrate into the next food at much higher levels than from PET. When a process using Starlinger's recoSTAR HDPE technology was submitted (on behalf of PET to PET Recycling Österreich GmbH) to recycle HDPE bottle caps collected with PET beverage bottles from kerbside and deposit systems, the CEP Panel found that decontamination efficiency was above 99.5 percent for small surrogate contaminants but only 20.2 percent for the 547 Da surrogate, noted "the limited efficiency of the technology to remove contaminants with a molecular weight above 400 Da," and concluded the applicant "has not demonstrated ... that the recycling process is able to reduce the contamination of the HDPE flakes ... to a concentration that does not pose a risk to human health." As Fraunhofer IVV's Welle wrote in 2025, "While there are evaluation criteria for mechanical PET recycling processes, no such evaluation criteria have been published for recycled HDPE caps in food contact." Welle's same 2025 paper argues that for the deposit-return cap stream specifically, worst-case criteria derived from other polymers put the required cleaning efficiencies "within a range that is technically feasible for today's mechanical HDPE recycling processes" — so the gap is at least as much the absence of agreed criteria as it is the physics. The unsolved problem is thus doubled: the physics makes polyolefins harder to clean and more prone to re-release, and the evaluation criteria that would tell a recycler what "clean enough" means have been published for PET but not for recycled HDPE caps.
Why this matters
The EU Packaging and Packaging Waste Regulation (Regulation (EU) 2025/40) sets minimum recycled-content targets from 2030 for plastic packaging, including contact-sensitive packaging made of polymers other than PET; Welle notes that for high-diffusivity polymers such as polyolefins it will be much more challenging than for PET to reach a high recyclate content without compromising food safety. The polyolefin recycling processes named here that EFSA has assessed favourably take their input from "closed and controlled loops" (food crates, boxes, trays and pallets kept in a traceable circuit — the Morssinkhof Plastics and CAPEC crate processes); post-consumer HDPE caps must instead be submitted as a novel technology under Regulation (EU) 2022/1616. For post-consumer HDPE, Welle reports that EFSA's opinions on both bottle caps and milk bottles were inconclusive and that no industrial-scale cap-to-cap recycling has been established — the opposite of the circular loop PET already runs.
What’s been tried and why it hasn’t worked
The PET playbook has been applied to HDPE and it strains at every joint. EFSA's evaluation logic starts from a reference contamination level of 3 mg/kg for misused PET; the applicant used 0.5 mg/kg for closures, and the Panel did not accept it: because HDPE diffusivity is higher, it expected that the "highest misuse contamination levels in washed and dried HDPE" would be not less than those in PET (2,000–6,750 mg/kg) and that the reference level for closures would be not less than PET's 3 mg/kg. Migration modelling used the Piringer/Hinrichs polyolefin diffusion models recognised in EU legislation and, for the infant scenario, arrived at a migration criterion of 0.034 µg/kg food. The challenge test used surrogates only up to 547 Da (TEHTM) where the Panel asked for coverage to 800 Da; the largest surrogate was barely removed and its concentration actually rose after the decontamination step, unexplained; the input-contamination screening covered only "24 post-consumer samples" that the Panel found "not sufficiently well described to ensure their representativeness"; and up to 1 percent of the input closures were non-food, whose additives could contribute up to 10 mg/kg — none of which the applicant could resolve. Closed-loop crate approvals sidestep the problem by ruling out misuse through traceability rather than by cleaning, and so cannot be generalised to household waste. Welle's 2025 analysis argues that the required cleaning efficiencies for deposit-return caps "are within a range that is technically feasible for today's mechanical HDPE recycling processes" once evaluation criteria are derived from other polymers — but those criteria remain a proposal, and the applicant of record must still submit under the novel-technology route with no published benchmark.
What would unlock progress
Three things, in order of leverage: (1) published, polymer-specific evaluation criteria — surrogate sets, reference contamination levels, migration models and packaging-use scenarios — for HDPE and PP recyclate, so that recyclers know the target and regulators can compare processes; (2) decontamination steps that reach heavier contaminants (>400–800 Da) in a high-diffusivity matrix — longer solid-state residence at higher vacuum, supercritical or solvent-assisted extraction, or hybrid mechanical-solvent processes — proven at industrial not lab scale; and (3) input-stream segmentation, such as deposit-return caps or dedicated milk-jug collection, that lowers the realistic misuse-contamination assumption below the PET default. The adjacent precedent is the PET route itself, whose published criteria (a 3 mg/kg reference contamination level, a challenge test with surrogate contaminants and a fixed exposure threshold) underlie the many PET recycling processes EFSA has assessed; polyolefins need the same scaffolding, adjusted for their diffusivity, plus a compliant sensor-based screen (e.g., odour/VOC or spectroscopic) that can catch grossly contaminated flakes before they enter the melt.
Entry points for student teams
A chemistry/food-safety team could run a scaled-down challenge test on post-consumer HDPE caps: spike flakes with a surrogate ladder spanning 100–800 Da, apply a bench solid-state or extraction step, quantify residuals by GC-MS, and model migration into food simulants using published polyolefin diffusion coefficients — producing exactly the decontamination-vs-molecular-weight curve EFSA found missing. A regulatory-science team could draft the missing evaluation criteria for rHDPE closures by comparing EFSA's PET guidance, the Starlinger opinion and Welle 2025, then stress-test them against worst-case use scenarios (infants, fatty foods, long shelf life). An engineering team could survey what contaminants actually occur in deposit-return caps versus kerbside HDPE (screening 100+ samples), addressing the representativeness gap the Panel flagged. Relevant skills: analytical chemistry, polymer diffusion modelling, food-contact regulation, process engineering.
Genes
Structural cousins: stuck the same way, in other fields
Sources
EFSA Panel on Food Contact Materials, Enzymes and Processing Aids (CEP) (2022), "Safety assessment of the process Starlinger recoSTAR HDPE FC 1 – PET2PET used to recycle post-consumer HDPE closures into food contact closures," *EFSA Journal* 20(1):e07001, doi 10.2903/j.efsa.2022.7001, accessed 2026-08-18. Supplementary: Welle, F. (Fraunhofer IVV) (2025), "Recycling of Post-Consumer HDPE Bottle Caps into New Caps for Food Contact," *Recycling* 10(6):197, doi 10.3390/recycling10060197 (abstract via Crossref; full text not read), accessed 2026-08-18. go to source ↗
Working notes
Notes made while writing and checking this brief, shown as written. They use working shorthand.
All Starlinger-opinion details (decontamination efficiencies >99.5 percent for ≤270 Da vs 20.2 percent for TEHTM at 547 Da; the ">400 Da" statement; the 3 mg/kg vs 0.5 mg/kg reference-level dispute; 2,000–6,750 mg/kg misuse levels; the 0.034 µg/kg infant criterion; 24 samples; 1 percent non-food closures/10 mg/kg additive contribution; the 800 Da request; the conclusion quote) come from the EFSA Journal 2022;20(1):e07001 text read via PMC on 2026-08-18. The Welle 2025 quotes and claims are from the Crossref-indexed abstract (MDPI full text returned HTTP 403); note that the abstract concludes cap-to-cap recycling for DRS-collected caps "is realistic" under worst-case criteria and proposes submitting the process as a novel technology under Regulation (EU) 2022/1616 — the verifier added this nuance to the Problem Statement so the brief does not overstate the technical barrier for the cleanest stream. The Morssinkhof (EFSA Journal 2018;16(1):5117) and CAPEC crate approvals were confirmed via the Food Packaging Forum summary and PMC listings, not read in full. The PPWR 2030 recycled-content structure (30 percent for contact-sensitive PET, 10 percent for other contact-sensitive plastic packaging) was confirmed only through secondary summaries of Regulation (EU) 2025/40 Article 7 — the primary EUR-Lex text was not fetched; verifier should check the exact percentages and product categories. `failure:regulatory-mismatch` (no-approval-pathway / wrong-measurement-basis sub-patterns) is applied because the assessment framework was built for PET and no polyolefin-specific criteria exist; the technical decontamination limit is captured by `constraint:technical`. `temporal:window` is a deadline window (PPWR 2030 targets). `stakeholders:multi-institution` passes: EFSA/Commission own the criteria and approvals, deposit-return operators own clean feedstock, recyclers own process performance, converters own use scenarios. `constraint:coordination` rejected on filter (2): binding constraints are physical (diffusivity) and regulatory (absent criteria), not unwillingness to cooperate. Related collection briefs: `circular-recycled-plastics-pcr-spec-void` (a general PCR specification gap, not food-contact safety), `circular-plastic-recycling-quality-cost-barrier` (mechanical quality loss vs chemical cost); this brief is specifically the food-contact safety-assessment barrier for polyolefins.
Source type: Self-articulated (regulator's own scientific opinion identifying what could not be demonstrated; recycling-science lab naming the missing criteria)
Verified at intake 2026-08-18: gate (net) + adversarial source check + contested-tag second coding.
Related briefs (distinct sub-problems, cross-referenced 2026-08-18): `circular-recycled-plastics-pcr-spec-void`, `circular-plastic-recycling-quality-cost-barrier`.
Reconciliation 2026-10-01: Fact repair after reading the whole write-up against every source it names, all fetched on 2026-10-01. Read in full: the EFSA opinion on the Starlinger process (https://pmc.ncbi.nlm.nih.gov/articles/PMC8762564/, and the same text from Europe PMC); Welle 2025, which the intake could not open, from the Internet Archive copy of the publisher page (web.archive.org/web/20251022162852/www.mdpi.com/2313-4321/10/6/197), with its citation checked against the Crossref record; and Article 7 of Regulation (EU) 2025/40 from EUR-Lex (https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=CELEX:32025R0040). Also read in full, from the Europe PMC full-text records: the Morssinkhof Plastics opinion (EFSA Journal 2018, PMC7009420) and the CAPEC opinion (EFSA Journal 2022, PMC9240968). A second reader went over every repair against the same sources on the same day; two repairs were undone because the original text matched the source, and this note describes the final state. (1) The text said "Starlinger submitted" the process and spoke of "Starlinger's proposal." The opinion says the dossier was submitted on behalf of PET to PET Recycling Österreich GmbH; Starlinger recoSTAR is the technology. The text now says so. (2) The text said the caps were collected with deposit-return bottles and were "about the cleanest polyolefin stream that exists." The opinion says the input came from kerbside and deposit systems, and no source ranks the streams; the sentence now gives the opinion's description. (3) The text said polyolefins are "the largest plastic packaging stream by volume." No source says this; removed. (4) The misuse example "pesticide" is not in the sources (Welle says storage of chemicals), and "any contaminant a used container ever held migrates deep into the polymer" went beyond them; the sentence now gives the Panel's expectation about misuse contamination and Welle's statement that remaining traces migrate at much higher levels than from PET. (5) The text said the assessment framework "has never been written" for polyolefins. Welle says final criteria like those for PET have not been published for HDPE caps, and the opinion itself points to a 2015 EFSA opinion discussing criteria for recycled HDPE; the sentence now says the criteria have not been published for recycled HDPE caps. (6) Removed because no source contains them: that brands can meet the 2030 targets only with chemically recycled or mass-balance material, that this is far more expensive and energy-intensive, that non-food uses are saturated and prices depressed; the quoted phrase "one life in food, then downcycled"; that solvent-based purification adds cost kerbside polyolefins cannot bear; that the 0.034 µg/kg criterion is "at the top of what solid-state processes can deliver"; and that the PET criteria "took years to consolidate and then unlocked a steady stream of approvals." (7) The text said "only two kinds" of polyolefin recyclate have cleared EFSA and that the household stream "has no approved mechanical route at all." The sources do not make either count. The paragraph now says what they do say: the two crate processes take input from closed and controlled loops; caps must be submitted as a novel technology; and Welle reports that EFSA's opinions on HDPE caps and on HDPE milk bottles were both inconclusive and that no industrial-scale cap-to-cap recycling exists. (8) Two quotations did not match the source word for word. "highest misuse contamination levels in washed HDPE" is "in washed and dried HDPE" in the opinion. The word "only" was inside the quotation about the 24 samples and is not the Panel's word. Both now match. The text also described the closed-loop input as "crates and drums"; neither crate opinion mentions drums, and the Morssinkhof opinion lists crates, boxes, trays, pallets and containers, so the text now names those. (9) The text said the Panel "insisted on 3 mg/kg." The opinion says the Panel expects the reference contamination to be not less than 3 mg/kg; the sentence now says that. Confirmed in the sources and left as written: both citations on the Source line; the quoted phrase "closed and controlled loops," which is in the Morssinkhof opinion; "up to 10 mg/kg" for additives from non-food closures (the opinion says an additive "may reach ca. 0.1%, which would correspond to 10 mg/kg HDPE"); the efficiencies above 99.5 percent for surrogates up to 270 Da and 20.2 percent for the 547 Da surrogate; the 400 Da statement; the conclusion quotation; the 2,000–6,750 mg/kg range; the Piringer and Hinrichs models; the 0.034 µg/kg infant criterion; the request for surrogates up to about 800 Da; the unexplained rise of the largest surrogate; 24 samples; the non-food fraction of no more than 1 percent; all three Welle quotations; and the 2030 recycled-content targets (30 percent for contact-sensitive PET packaging, 10 percent for contact-sensitive packaging of other plastics). Not checked: the proposals in "What Would Unlock Progress" (longer solid-state residence, solvent-assisted extraction, sensor-based screening) are the write-up's own suggestions and are not statements from the sources; the Food Packaging Forum summary named above has no link and was not found.