Wet PET flake cannot be dried properly, and the reason is the shape of the polymer. PET is amorphous in the bottle wall, and amorphous PET softens around 70–80 °C. A dehumidifying dryer at 120 °C turns the outside of the flake soft first, and the piece sticks to itself and to the hopper wall before the moisture has left it. Crystallisation turns the flake rigid and stable so it can be heated hard without that happening.
Skip the crystalliser and you get a dryer that appears to work — the hopper is hot, the air is dry — and a flake that comes out at the wrong moisture, with the inside of the pellet still wet. That is what shows up downstream as hydrolysis: PET loses 0.01 dl/g of intrinsic viscosity per 100 ppm of moisture, and a food-contact line is specified at below 50 ppm after crystallisation.
The sequence, stage by stage
- Dewatering: centrifugal, from 30% to 2–5% surface moisture, and it costs little energy.
- Thermal drying: controlled hot air, below 0.5% residual moisture.
- Crystallisation: heated with agitation until the flake is rigid and stable.
- Dehumidified drying: dew point below −40 °C, to below 50 ppm, then extrusion and underwater cutting with IV loss below 0.03 dl/g.

The problem the crystalliser solves
PET flake coming out of a washing line is wet, and it is amorphous. Amorphous PET is soft, and soft PET flake cannot be conveyed reliably, fed evenly or melted predictably. If it goes into an extruder wet and cold, it hydrolyses.
Crystallisation is the step that prepares it. Heat is applied in a crystalliser with agitation, and the polymer chains reorganise into an ordered crystalline structure. The material becomes rigid, it holds its shape, it flows properly through the feed system, and — the point of the whole exercise — it can be dried at a temperature it survives.
This is why the order is not negotiable. Drying amorphous PET is difficult because the material sticks together before the moisture is gone, and clumps that have welded shut cannot be dried at all.
What happens if you skip it

The failure is quiet, which is what makes it expensive.
PET loses 0.01 dl/g of intrinsic viscosity for every 100 ppm of moisture it carries into the melt. If wet flake goes into the barrel, the hydrolysis continues inside the extruder, and the intrinsic viscosity of the finished pellet is lower than the intrinsic viscosity of the flake went in.
The visible symptoms are yellowing, bubbles and a loss of strength. The invisible symptom is that the material still passes most quality checks and simply cannot be used for bottle-to-bottle applications, which is where the value is.
And the cost of finding out late is high. A line that has been producing off-specification pellets for months has produced a mountain of material at a discounted price, and the customer relationship is the thing that is hardest to rebuild.
The specification that has to be met
For food-contact PET the published requirements are precise, and they are written into the material specification rather than left to operating practice.
Moisture below 50 ppm after crystallisation. That is the number, and it is a hard limit rather than a target.
Intrinsic viscosity loss below 0.03 dl/g. Our TL-PET platform is specified for exactly this, at 500–2,000 kg/h and US$120,000, and it is the only platform in the range where the crystallisation and dehumidified drying are part of the machine rather than something added beside it.
Food-contact material requirements. Published guidance is that the machine can be supplied with stainless contact parts, closed-loop water and documented process parameters — which is what certification bodies ask for — while actual food-contact approval also depends on feedstock traceability and, for PET, an SSP step. The approval is issued to the converter, not to the machine.
The feedstock traceability point is worth dwelling on. A machine can be perfect and the material still cannot be approved, because the certification follows the material, not the equipment.
The equipment sequence, stage by stage
The full path from wet flake to food-grade pellet runs like this, and each stage has a defined job.
1. Dewatering. A centrifugal dewaterer or a dewatering machine removes the free surface water mechanically. This costs almost no energy because it uses centrifugal force rather than heat, and it takes the flake from around 30% surface moisture down to roughly 2–5%.
2. Thermal drying. Controlled hot air evaporates the remaining bound moisture, bringing the flake below 0.5% — the level required before crystallising and extruding for food-grade use.
3. Crystallisation. The flake is heated with agitation so the polymer reorganises into a crystalline structure, making it rigid and stable.
4. Dehumidified drying. The dehumidifying dryer takes the crystallised material below 50 ppm with a controlled dew point below −40 °C.
5. Extrusion and pelletising. The now-dry flake goes into the extruder, and the underwater cutter produces the spherical pellet without the oxidation that air cooling would cause.
Steps 3 and 4 are the ones that are frequently combined into one package, and our TL-PET line does exactly that.
The PET drying sequence from wet flake to food-grade pellet
The sequence from wet flake to food-grade pellet, and what each stage costs in energy.
| Stage | Method | Result | Energy character |
|---|---|---|---|
| 1. Dewatering | Centrifugal dewaterer / dewatering machine | 30% → 2–5% surface moisture | Low — centrifugal, not heat |
| 2. Thermal drying | Controlled hot air | Below 0.5% residual moisture | Moderate — evaporative |
| 3. Crystallisation | Heated with agitation | Rigid, stable crystalline flake | Moderate — soak temperature |
| 4. Dehumidified drying | Desiccant rotor, dew point below −40 °C | Below 50 ppm | Low — closed loop, re-circulated air |
| 5. Pelletising | Extrusion and underwater cutting | Spherical pellets, IV loss below 0.03 dl/g | Standard extrusion duty |
Common questions, with short answers
Q1: Why does PET need a crystalliser before it can be dried?
Amorphous PET is soft and sticky, so clumps weld shut before the moisture is removed. Crystallisation makes the material rigid and stable, which is what allows it to be dried properly. The sequence is not negotiable — drying amorphous PET does not work.
Q2: How much intrinsic viscosity does wet PET lose?
PET loses 0.01 dl/g of intrinsic viscosity for every 100 ppm of moisture carried into the melt. At 300 ppm that is 0.03 dl/g, which is the entire budget for a food-contact line. The damage continues inside the extruder, so the loss is not stopped by drying the flake afterwards.
Q3: What is the maximum moisture for food-contact PET?
Below 50 ppm after crystallisation, with an intrinsic viscosity loss below 0.03 dl/g. Our TL-PET line is specified for exactly that, at 500–2,000 kg/h, with the crystallisation and dehumidified drying integrated into the machine.
Q4: Does the machine itself make the pellet food-contact approved?
No. The machine can be supplied with stainless contact parts, closed-loop water and documented process parameters, which is what certification bodies ask for. Actual food-contact approval also depends on feedstock traceability and, for PET, an SSP step, and it is issued to the converter rather than to the machine.
The pages that come up most often alongside this one: the 12-platform comparison.
The companion write-ups for this machine family are: Dehumidifying Dryer vs Hot Air Hopper for Plastic Granules, Dehumidifying Dryer Sizing: Hopper Volume and Airflow, Dehumidifying Dryer Maintenance: Desiccant and Filters and How to Buy a Dehumidifying Dryer Without Overpaying.
About the author
Zhou Gong (周工) builds recycling lines at TENGLONG Plastic Machinery, a factory in Linyi, Shandong, with 18,000 m² of floor area, 30-plus staff and 820-plus lines delivered to 46 export markets. The number I insist on seeing is the moisture of the flake going into the crystalliser. Above about 5% the crystalliser stage is being asked to do a job it cannot do, and it will show up as brown pellets months later. The numbers in this article come from test runs and from dated published sources, not from a sales sheet.
If you want a recommendation rather than a catalogue, send the material form, the contamination level and the target output. We return a model and a price range within one working day, and the test run on 10–20 kg of your own material is free before the balance is paid.