Every clear drinks bottle in the world is polyethylene terephthalate, and it is one of the most successfully recycled plastics on earth. That combination — abundant, familiar, recyclable — makes printing with it enormously appealing, and it is the reason bottle-to-filament projects keep appearing.
The reality is harder, and understanding why explains the existence of half the copolyester shelf.
Crystallisation is the whole difficulty
PET's chain is regular and symmetrical. Left to cool slowly it packs into crystalline regions, and crystals are denser than the disordered melt they came from, so a crystallising polymer shrinks substantially more than an amorphous one. Stored shrinkage runs from 0.4% right up to 1.2% — the widest band of any polyester here — and where in that band you land depends on how the part cooled rather than on anything you set in the slicer.
That is precisely what glycol modification was invented to prevent. Add a bulky co-monomer and the chain can no longer pack neatly, crystallisation is suppressed, and you get the dimensionally calm, forgiving material sold as PETG. PET is what PETG is a fix for.
Crystallinity also decides appearance. A bottle is clear because it was cooled fast enough to stay amorphous and then stretch-blown; PET cooled slowly on a print bed goes hazy and white. So a clear PET print is not simply a matter of buying clear filament.
Annealing brings the problem back after the fact: heat a PET part to relieve stress and it crystallises further and moves again, which makes the usual "anneal it for strength" advice actively dangerous here.
Water destroys it chemically, not just cosmetically
PET hydrolyses. At melt temperature, absorbed water attacks the ester links and cuts the chains, and a shorter chain means a weaker, more brittle part. Bottle-grade PET is dried to demanding specifications before moulding for exactly this reason, and there is no reason a printer should be more forgiving.
Dry at 70 °C for 6 to 8 hours, and treat it as non-negotiable rather than as best practice. The tell is a print that looks acceptable and then snaps under a load that a good PETG part shrugs off. No amount of temperature tuning recovers a chain that has already been cut.
This is also the honest problem with recycled bottle filament. Each thermal cycle — original moulding, shredding, extrusion into filament, printing — costs some chain length. The material can be excellent, but it depends entirely on how carefully it was dried at each stage, and that is not visible on the spool.
The densest common polyester
At 1.32 to 1.40 g/cm³ PET is heavier than every other filament on this site except the metal-filled grades. A 1 kg spool of 1.75 mm PET at 1.38 holds 301.27 m, occupying 724.64 cm³, against PETG's 327.36 m for the same mass. That is roughly 8% less filament for the same money, and the difference is entirely density rather than anything anyone is doing wrong. In practice that 8% shows up as a spool running out one print earlier than the profile predicted rather than as anything dramatic. Putting 1.38 and then PETG's 1.27 through the length and weight converter gives the difference in metres, which is easier to act on than a percentage.
Cool it hard, and keep the geometry compact
The bed and its adhesive layer matter more here than the temperatures do, because crystallisation shrink pulls corners hard and pulls them late in the print, when there is a tall part already attached to them. Work at 250 to 275 °C through the nozzle onto a 70 to 90 °C plate, with an enclosure if you have one.
Two practical adjustments distinguish PET from PETG:
- Cool it deliberately and evenly. Faster cooling suppresses crystallinity and keeps dimensions closer to nominal, which is the opposite of the low-fan advice that helps most high-temperature materials.
- Keep part geometry compact. Long flat spans give crystallisation shrink the leverage it needs. A part that would be fine in PETG at 200 mm may not be at all in PET.
What PET is genuinely bad at
- Dimensional predictability. Between the crystallisation range, the cooling dependence and the post-print movement, this is the least repeatable of the polyesters.
- Beginners. The process window is narrow and the failure modes are chemical rather than visual.
- Being clear. See above.
- Annealing.
- Availability. Very few brands sell unmodified PET filament, because PETG exists and is better in almost every way that matters to a printer.
- Damp storage of any kind.
Where it still makes sense
Genuine recycling projects, where turning bottles into objects is part of the point and the imperfections are acceptable. Parts that need PET's specific chemical resistance and its low gas permeability — it is a barrier polymer, which is why it holds carbonation. Applications where the material's food-packaging pedigree matters for regulatory paperwork rather than for the printed part itself.
On that last point, a caution worth stating flatly: PET resin's long history in food packaging says nothing about a printed part. Layer lines are a porous, uncleanable surface, most nozzles are not food-grade, and a recycled feedstock's history is unknown by definition. The resin's credentials do not transfer to the object you printed.
For everything else, the glycol-modified version is easier, cheaper, more available and dimensionally better behaved — which is what the PET against PETG comparison works through.