Regrind and Recycled Content in Extrusion: Ratios, Drying and Property Loss

Regrind is not free material

Every extrusion line produces scrap — start-up lumps, off-spec product, trim, purging. Grinding it and feeding it back looks like free money, and on a well-run line it is. On a line that treats regrind as a substitute for virgin resin, it is the beginning of a slow quality decline that nobody can trace.

The reason is that extrusion is a thermal process, and every pass through the extruder adds a heat history. Polymer chains break under heat and shear, and once broken they do not re-form. The result is lower molecular weight, which means lower viscosity, lower impact strength and lower elongation — while tensile strength and modulus barely move. That is why regrind problems show up as brittleness and surface defects rather than as weak parts.

What actually degrades

Table 1 — What regrind actually changes. The pattern — impact and elongation down, strength and stiffness unchanged — is the signature of chain scission.

The gel problem is worth separating. Gels are not degraded polymer; they are crosslinked or highly oxidised particles that do not melt at processing temperature. They come from hot spots in the extruder, from purging, or from contamination. Grinding does not remove them, and they pass through the screen pack if it is too coarse. In film and sheet they appear as visible defects; in profile they appear as surface roughness.

How much regrind each resin tolerates

Table 2 — Practical regrind ratios by resin. PVC's tolerance is real but it is tolerance of the process, not of the chemistry — HCl evolution still accumulates.

The PVC case is counterintuitive and worth understanding. Rigid PVC accepts 50% regrind in many profile applications because the compound is heavily stabilised and the processing temperature is low. But the stabiliser is consumed with each pass, and at some point the remaining stabiliser cannot protect the polymer. That point arrives suddenly, as a colour shift and then as degradation, and it is not predicted by the ratio alone — it is predicted by the accumulated heat history of the specific regrind batch.

How to feed it back properly

The mistake most lines make is feeding regrind by volume rather than by weight, and feeding it unmetered. Both cause variation that gets blamed on the resin.

Table 3 — Regrind feeding methods. Volumetric dosing fails because regrind bulk density varies with flake size, so the same volume is a different weight every batch.

Bulk density variation is the specific reason volumetric fails. Virgin pellets run at 550–650 kg/m³; regrind flake runs at 350–500 kg/m³ depending on how fine it was ground and how much dust it contains. A volumetric feeder set for 30% regrind by volume might deliver 20% by weight one hour and 38% the next, purely because the grinder was set differently.

The other requirement is consistent flake size. A grinder with a worn screen produces a mix of fines and large flakes, and the fines melt faster while the large flakes may not melt at all. Both cause surface defects. Screen size should match the application: 6–8 mm for profile, 4–6 mm for sheet, 3–5 mm for film.

Drying: the step that gets skipped

Regrind is more hygroscopic than virgin resin because it has far more surface area. A flake has roughly ten times the surface-to-volume ratio of a pellet, and moisture adsorbs onto that surface within hours.

Table 4 — Drying schedules for virgin versus regrind. The longer times for regrind are not conservative padding; the moisture is genuinely harder to remove from flake.

Wet regrind in a hygroscopic resin produces the classic hydrolysis failure: a drop in molecular weight that shows as brittleness, silver streaking and sometimes visible bubbles. It is often misdiagnosed as "bad regrind" when the regrind was fine and simply wet.

Tracking heat history

The ratio is a proxy for heat history, and a poor one. What actually determines whether regrind is safe is how many times the material has been through an extruder.

A closed-loop system where a part is ground and immediately fed back is one pass. A system where scrap from several products is collected, ground together, and fed back weeks later is an unknown number of passes — and if some of that scrap was itself made with regrind, it could be three or four.

Table 5 — Heat history by sourcing practice. Purchased regrind is the highest-risk category because the history is unknowable and the price rarely reflects it.

The practical control is a colour code or a label on every regrind bin recording the product, the date and whether the batch contains regrind. That sounds bureaucratic for a small plant and it is exactly what prevents the "why did last month's batch fail" investigation.

Where regrind is acceptable and where it is not

Table 6 — Where regrind belongs. The prohibitions in pipe and medical are certification constraints, and no amount of technical justification overrides them.

The certification point is worth stating plainly because it is often missed. Many pipe standards, and virtually all potable water and medical standards, prohibit or strictly limit regrind regardless of whether the material would perform. Using regrind in a certified product is not a quality risk — it is a compliance failure, and it invalidates the certification of the whole batch.