PET Sheet: Water Is Measured in Parts per Million and It Decides Everything

PET is a condensation polymer, and water runs it backwards

Polyethylene terephthalate is built by removing water. The ester bonds that make the chain long were formed by condensing an acid and an alcohol and driving off water, and the reaction is reversible. Put water back in at 280 C and the bonds break, the chain shortens, and the molecular weight falls. This is hydrolysis, and it happens fast enough to matter in the two minutes a pellet spends in a barrel.

The consequence is that PET has to be dried to a level that would be considered absurd for most polymers. Polyolefins are processed wet without complaint because they are not condensation polymers. PET at 800 ppm loses 0.55 dL/g of intrinsic viscosity in the extruder, which is the difference between a sheet that thermoforms well and one that sags, tears and cannot hold a corner.

The measurement unit matters here. Moisture specifications for PET are in parts per million, not percent, because the interesting range is between 20 and 200 ppm. A drier that reaches 0.02 percent is at 200 ppm, which is four times the specification and twice the level at which IV loss becomes visible.

Fifty ppm is not a conservative specification

Measured IV loss against pellet moisture: 0.02 dL/g at 50 ppm, 0.05 at 100, 0.12 at 200, 0.28 at 400 and 0.55 at 800. If the sheet specification allows 0.05 dL/g of loss, then 100 ppm is the limit, not a comfortable target. Fifty ppm is what gives the process room for a drier that drifts on a humid day.

The relationship is steeper than linear above 200 ppm, because hydrolysis produces carboxylic acid end groups that catalyse further hydrolysis. The reaction accelerates as it proceeds, which is why a wet batch fails suddenly rather than gradually: the first part of the run looks acceptable and the last part is unusable.

Drying itself has a trap. PET pellets absorb moisture and they also crystallise, and amorphous pellets stick together in a hopper drier at 160 C before they are dry. That is why drying is done at 160 to 180 C on crystallised pellets, and why a desiccant drier with a dew point of minus 40 C is the standard rather than a luxury: the drier is not removing water from the air, it is keeping the air from putting water back.

IV loss cannot be recovered downstream

Intrinsic viscosity is a measure of chain length, and chain length is destroyed by hydrolysis, by excessive shear and by thermal degradation. None of those are reversible in an extruder. A sheet that has lost 0.2 dL/g cannot be brought back by changing the temperature profile, and the only remedy is to start with higher-IV resin.

Table 1 — Starting intrinsic viscosity by application, and what it is buying

The practical approach is to buy IV headroom. A thermoforming sheet made from 0.80 dL/g resin that loses 0.06 in the extruder arrives at 0.74 and forms well. The same process on 0.76 resin arrives at 0.70 and produces sheet that sags in the oven and tears at the corners.

The chill roll decides whether the sheet can be formed

PET crystallises, and how much depends on how fast it is cooled. Quenched hard against a cold roll it stays amorphous and clear. Cooled slowly it crystallises and turns opaque, and opaque PET sheet does not thermoform into a clear part because the crystals do not disappear on reheating in the time available.

Measured crystallinity against chill roll temperature: 2 percent at 20 C, 5 at 40, 12 at 60, 25 at 80 and 38 at 100. The useful window for clear thermoforming sheet is below about 40 C, where crystallinity stays under 5 percent. Above 60 C the sheet starts to haze, and above 80 C it is visibly opaque.

The roll temperature is not the only variable. Line speed matters because it sets the contact time, and a roll at 30 C running twice as fast gives less cooling than the same roll running slowly. Water flow and roll internal design matter too, because a roll with poor internal flow has a surface temperature that varies across the width, which produces a sheet that is clear at the edges and hazy in the middle.

The two sides of a PET sheet are not the same material

The side that touches the chill roll cools fast and is smooth and amorphous. The side that faces the air cools slowly and is slightly rougher and slightly more crystalline. The difference is small in absolute terms and large in behaviour: the two sides have different gloss, different friction and different surface energy, and they form differently.

That asymmetry is why a formed PET part can be glossy on the outside and slightly matte on the inside from the same sheet, and why printing adhesion can differ between the two faces. It is also why a sheet that is wound the wrong way up produces parts with the appearance on the wrong side, which is a reject that nobody diagnoses correctly on the first attempt.

Table 2 — PET sheet defects and the setting that causes each

What to specify

A PET sheet specification has to carry the numbers that the downstream process depends on.

• Intrinsic viscosity of the sheet, not only of the resin, with the test method.
• Crystallinity or clarity, stated as a maximum, because it is a formability limit.
• Which face is the chill roll side, and which face is visible in the finished part.
• Thickness and thickness variation across the width, which affects heating uniformity.
• Residual moisture of the sheet if it will be stored before forming, since PET reabsorbs water.

The last item is the one most often left out. A roll of PET sheet left in a humid warehouse for a month picks up moisture through the surface, and a thermoformer who does not dry the sheet before forming will see hydrolysis in the oven rather than in the extruder. The sheet was in specification when it left and out of specification when it was used, and both statements are true.