Short answer
Recycled PET tends to yellow, darken and haze because each melt creates coloured degradation products and because contaminant particles scatter light; recycled polyolefins go greenish, grey or dark because coloured items are melted together. How far it goes depends on the polymer, its stabilisers, how many heat cycles it has seen and how well the feedstock was colour-sorted.
PET is dried and melted at high temperature every time it is reprocessed. Heat, oxygen and any residual moisture break the ester links in the chain, and the fragments react further to form chromophores — structures that absorb blue light and so look yellow or brown. Laboratory studies that re-extrude the same clean bottle-grade resin several times show lightness falling and total colour difference rising with every pass, in step with the drop in intrinsic viscosity. In practice rPET also carries traces of whatever the bottle met: residues, other polymers and inks. The result is that bottle-grade rPET usually has a slightly higher b* (more yellow) and lower L* than virgin resin, which is why rPET bottles can look faintly grey-yellow beside virgin ones.
Clarity is a separate problem from tint. Tiny insoluble particles — paper fibres, degraded polymer, fragments of other plastics — scatter light and make a bottle wall look milky. A Wageningen study that blew bottles at 25%, 50%, 75% and 100% recycled content from three rPET sources found haze rising linearly with particle count, and found that the origin of the rPET (a dedicated bottle collection versus mixed collection with other packaging) determined contamination and colour more than the percentage did. That is why deposit-return and dedicated bottle streams feed the clearest bottle-grade rPET, and why mixed-collection rPET more often goes into trays, strapping or fibre, where slight haze matters less.
Polyolefins yellow far less on reprocessing than PET when they are well stabilised: a study of a stabilised polypropylene copolymer without phenolic antioxidant found its yellowness index unchanged between the first and third processing cycles. Their colour problem is mixing. Natural HDPE milk bottles are clear-white, but their caps are pigmented; WRAP's milk-bottle research traced a green hue in recycled HDPE mainly to cap pigment and to labels that leached colour during washing. When many coloured bottles, tubs and closures are melted together, the pigments average towards a muddy grey, green-grey or brown, and the cheapest way to make that consistent is to add carbon black. That is why so much mixed polyolefin recyclate ends up as dark grey or black products.
Plastics carry antioxidants to protect them during processing, and the choice of antioxidant affects recycled colour. Studies of repeated extrusion of polyethylene have found that material with only a phosphite stabiliser stayed close to white, while a common hindered-phenol antioxidant left a slight yellowish-reddish tint from its breakdown products and some natural polyphenol stabilisers discoloured it strongly. So two batches of the same polymer with the same number of heat cycles can differ in colour purely because of what additive package they were first made with — a variable the recycler rarely knows.
| Extrusion cycles | L* | a* | b* | ΔE*ab vs cycle 0 | Yellowness index (ASTM E313) |
|---|---|---|---|---|---|
| 0 | 73.23 | -6.96 | 10.75 | — | 42.4 |
| 1 | 74.15 | -4.51 | 13.23 | 3.6 | 48.5 |
| 2 | 63.7 | -1.01 | 13.14 | 11.5 | 54.2 |
| 3 | 62.31 | -0.53 | 14.53 | 13.2 | 57.5 |
| 4 | 65.06 | -0.7 | 21.25 | 14.7 | 66.9 |
| 5 | 54.56 | 0.51 | 21.77 | 22.9 | 73.6 |
| 6 | 58.65 | -2.1 | 20.59 | 18.3 | 66.8 |
Why: Higher b* and lower L* from heat history plus particle haze in the recycled resin.
Fix: Specify rPET from bottle-only collection, agree a b* and haze limit with the supplier, or use a light blue tint that reads as intentional.
Why: Pigment from coloured closures carried into the natural HDPE stream.
Fix: Specify natural-only or colour-sorted flake, or add enough white to mask it and accept a greyer white.
Why: The mix of coloured items in the feedstock changes from lot to lot.
Fix: Choose a dark target colour that absorbs the swing, or buy from a supplier who blends large lots.
Each statement is labelled by kind — established fact, a standard’s requirement, observed market data, a convention, or Colourwise’s own interpretation or analysis — with the strength of the evidence behind it.
FactStrong evidence
Across six extrusion cycles of PLA, b* rose from 10.75 to around 21 and the ASTM E313 yellowness index from 42.4 to above 70, while L* fell, with a ΔE*ab of 3.6 after a single reprocessing cycle.
Caveat: PLA, not PET or polyolefin; one grade reprocessed under laboratory conditions.
FactModerate evidence
The collection route of rPET — mono-collection of bottles or co-collection with other packaging — was identified as the crucial factor for particle contamination and bottle colour.
Caveat: Based on three rPET sources in one study.
FactModerate evidence
In recycled HDPE from milk bottles, a green hue that limited recycled content was traced mainly to pigment in coloured caps and to over-laminated paper labels leaching colour in the wash.
FactModerate evidence
A stabilised polypropylene copolymer containing no phenolic antioxidant showed the same yellowness index (−15.7, DIN 6167) after the first and third processing cycles.
Caveat: One formulation designed for medical use; not representative of mixed post-consumer PP.
Colourwise interpretationLimited evidence
Mixed-colour polyolefin recyclate tends towards dark grey or black because averaging many pigments produces a muddy colour that is most easily made consistent by adding carbon black.
Based on: Follows from subtractive mixing of many pigments and from the HDPE colour-shift findings in the WRAP milk-bottle work; not a measured survey of recyclate colours.
Caveat: Colour-sorted polyolefin fractions (natural, white, single colours) can avoid this.
Reviewed 29 September 2026. Colourwise summarises its sources in its own words and does not reproduce standards text or proprietary colour data. Spotted an error? Tell us.