Polyethylene terephthalate (PET) selected for cast sheet production is defined less by melt flow index than by intrinsic viscosity, diethylene glycol content, and carboxyl end group concentration. A sheet-grade resin with an intrinsic viscosity of 0.72–0.84 dL/g, measured according to ASTM D4603-18, is dried in a desiccant dryer at 160–180 °C to a final moisture content below 50 ppm. At ambient relative humidity above 60%, hopper residence after drying should not exceed 30 min unless the vessel is closed-loop. The drying step is process-critical; residual water hydrolyzes the ester linkages during plastication, reducing molecular weight and shifting intrinsic viscosity downward by as much as 0.05 dL/g under severe conditions. The phrase Cast PET Sheet Line: Chill Roll, Thickness Control and Gel Inspection defines three interdependent process zones that determine whether the sheet can survive downstream plug-assisted thermoforming without cracking, optical haze, or localized thinning. The extrusion train is normally a single-screw barrier screw with an L/D ratio of 30:1 to 33:1, a melt pump, a flex-lip die, a vertical or near-vertical chill roll stack, and a scanning thickness frame. Melt temperature at the die entry is held between 270 °C and 285 °C; the flex-lip die temperature is controlled in zones with a tolerance of ±1.5 °C. Sheet thickness for thermoforming grades typically ranges from 0.20 mm to 1.50 mm. At the lower end of that range, a ±3% thickness deviation translates to ±6 µm, which is below the resolution of some older contact gauges, so non-contact scanning is required. The die exit melt is pinned electrostatically or by air knife to the first chill roll; pinning force determines surface replication and prevents air entrapment that causes micro-voids. The material must also comply with food-contact requirements under FDA 21 CFR 177.1630 and, for European applications, Regulation (EU) No 10/2011; these regulations constrain the migratory species that can be produced by thermal degradation and therefore make gel inspection a compliance-relevant measurement, not only a cosmetic test.

Cast PET sheet can be produced as amorphous APET or crystallized CPET. For APET, the chill roll temperature is low enough to quench the melt below its glass transition rapidly; for CPET, a subsequent crystallization oven is required after the cast line. The discussion that follows addresses APET sheet for packaging and display applications, where clarity and thermoformability dominate the acceptance criteria. Because PET is hygroscopic, the entire material handling path from dryer outlet to extruder feed throat is sealed and purged with dry air at a dew point of -40 °C or lower. Residual moisture above the target increases hydrolytic degradation, but excessive drying at 180 °C can oxidize the resin surface and raise gel counts. The balance between drying temperature, residence time, and air dew point is therefore an operational boundary that must be established for each regrind fraction.

Chill Roll Surface Replication, Water Flow, and Anti-Blocking Parameters

Chill roll temperature and surface roughness set the amorphous state of APET sheet and influence haze, gloss, and dimensional stability. For amorphous PET, the first chill roll is maintained at 18–25 °C; if the roll surface exceeds 70 °C, crystallization proceeds and the sheet develops spherulitic haze. The second and third rolls may be set at 20–45 °C to control sheet flatness. Roll surface finish is specified by roughness average Ra: gloss sheet requires 0.02–0.05 µm Ra, while matte or textured sheet uses 1.0–3.0 µm Ra depending on the embossed pattern. Water flow through the spiral channels of a 400 mm diameter roll is typically 60–100 L/min per roll; lower flow reduces heat transfer uniformity and creates machine-direction bands of haze. Nip pressure between the first and second rolls is maintained between 2 N/mm and 8 N/mm linear force, but the value is adjusted to avoid center-to-edge caliper variation. More critical is roll runout and bearing condition. A runout of 0.01 mm on the first roll can generate a repeating thickness variation equal to the roll circumference. In production, roll temperature setpoints are cascaded with the extruder output; when the line speed increases from 10 m/min to 30 m/min, the coolant flow is raised to maintain the same surface temperature. Failure to do so produces a heat-transfer-limited condition in which the sheet remains above its glass transition temperature after the first roll and develops blocking or flatness loss.

Anti-blocking additives such as silica at 0.1–0.3 wt% may be compounded into the resin, but their concentration must remain low because excessive silica reduces clarity and increases detector counts during optical inspection. The chill roll stack gap is also a control parameter. A gap of 0.25–0.45 mm is common for sheet from 0.20 mm to 0.80 mm; the gap must remain parallel across the width to prevent wedge-shaped sheet. A differential roll gap of 0.02 mm across the width cannot be corrected by die bolt movement alone. Roll change intervals are often determined by surface pitting or release coating degradation. A release coating failure on the first roll causes periodic sticking of the melt, visible as transverse chatter marks at the winder. Surface temperature uniformity is verified with an infrared line scanner; a variation of more than ±1.0 °C across the roll face is typically corrected by rebalancing the water circuit.

Why Does Thickness Deviation Persist Even With Closed-Loop Die Bolt Control?

Closed-loop control of cast PET sheet thickness is limited by dead time, die bolt interaction, and scanning frame sampling interval. The scanning gauge, typically a beta transmission or X-ray backscatter unit, traverses the web once every 20–30 s depending on sheet width; a 1.2 m wide sheet scanned at 300 mm/s yields a transverse profile every 4 s, but the control algorithm requires several consecutive scans before adjusting die bolts. The melt between the die lip and the scanning frame has a transport delay of 5–15 s at line speeds of 15–30 m/min. Consequently, a die bolt change made at time t affects the measured thickness at t plus the transport delay, and the controller must apply model-based dead-time compensation to avoid oscillation. Die bolt sensitivity is not identical across the die; edge bolts have a wider influence per 0.01 mm adjustment than center bolts because of the lower transverse constraint. Therefore, closed-loop systems use normalized gain matrices determined by bump tests at each die bolt. Melt pump stability is an essential precondition. A gear pump with speed control of ±0.05% reduces extruder pulsation to less than ±0.5 bar at the die entry; without this, the thickness scanner sees a high-frequency variation that cannot be corrected by slow die bolt movement.

The die lip gap is typically set at 0.8–1.2 times the target sheet thickness; because PET has a high draw ratio after the die, a 0.01 mm lip gap change produces a thickness change in the range of 5–10 µm, depending on line speed and roll gap. Thickness specification for cast PET sheet is often tested according to ISO 4593:1993 or ASTM D6988-21; the measurement should be taken at least 25 mm from the edge to exclude the neck-in zone. Profile tolerances are usually expressed as ±2–5% of nominal for sheet above 0.5 mm, while films below 0.25 mm may require ±5–8% because of the greater relative influence of die lip deflection. Automatic profile control is not a substitute for mechanical flatness. If the die body is heated unevenly, the lip opening changes due to thermal expansion, and the actuator may fight a moving baseline. The die bolt positions should therefore be mapped at operating temperature, not cold, and the die should be allowed to stabilize for at least 45 min after zone changes before profile data are accepted.

Inline gel inspection on a cast PET sheet line uses high-intensity transmitted light and a line-scan camera positioned after the chill roll stack, before the edge trim slitter or winder. Gels in PET are generally classified by size and origin: hydrolytic degradation produces crosslinked oxidized particles; insufficient filtration allows resin fines and oligomeric crystallites to pass; stagnant melt at die lip edges degrades and periodically releases black specks. The optical system is configured for the contrast between the gel and the clear amorphous sheet. A line-scan camera with 8K pixels across a 1.2 m web gives a nominal transverse resolution of 150 µm/pixel, but after sub-pixel processing the reported detection limit is typically 25–50 µm at a line speed of 30 m/min. Detection is calibrated using reference films with embedded spherical particles of known size, usually 25 µm, 50 µm, 100 µm, and 200 µm. The inspection algorithm records defect size, circularity, optical density, and machine-direction position. Because PET sheet edges are often trimmed, the system may be configured to suppress 10–20 mm edge bands to avoid false alarms from neck-in and edge beads. Acceptance criteria are typically derived from downstream thermoforming trials; a starting specification is ≤5 defects larger than 100 µm per 1 m² and no defect larger than 300 µm in food-contact APET. This is not purely cosmetic; a gel particle larger than about 150 µm can act as a stress concentrator during plug-assisted stretching and produce a pinhole in a deep-draw cup. The test standard ASTM D7310-20 provides a framework for optical defect detection and rating in plastic films and sheets, though inline systems often use customized algorithms calibrated against the standard's defect classes.

The line should also include a metal detector and an upstream screen pack sized to the gel specification; a 20–40 µm rated screen changer can remove larger particles but cannot compensate for degradation occurring after the filter, especially in the die lip region. PVC contamination from label waste is particularly damaging; as little as 50 ppm PVC in the flake stream can produce black specks and acidic degradation by-products. When recycled PET flake is used, gel counts rise with each heat history; validation runs are required because published data for a specific blend and line configuration is often limited. In addition, the inspection unit should be downstream of the first chill roll but upstream of any corona treater or silicone coating station, because surface treatments can alter the optical contrast and create false positives. The gel count is most useful when trended against extruder barrel temperature and melt residence time. A sudden increase in gels larger than 200 µm without a change in resin lot is usually traced to die lip stagnation or a failed screen pack gasket, not to the drying system.

When Downstream Thermoforming Rejects Uninspected Cast PET Sheet

Thermoforming exposes cast PET sheet to non-uniform stretching that amplifies both thickness variation and gel defects. In cup forming, a plug-assisted process with a draw ratio of 3:1 or higher takes a sheet of 0.50 mm average thickness and stretches the sidewall to below 0.15 mm; any pre-existing thickness deviation is multiplied by the local draw ratio. A 3% machine-direction low spot may become a 5–8% wall thinning at the transition radius because the plug applies less resistance to the thinner section. Cast PET sheet for thermoforming is therefore specified with both average thickness and profile tolerance. The scanner data must be synchronized with the winder length counter so that a defect map can be generated and compared with the downstream reject location. When a gel-rich section reaches the forming machine, the reject mechanism is often a pinhole or a crack at the gel boundary; the crack propagates because the gel has a higher modulus than the surrounding PET and does not deform ductilely. In food packaging, such pinholes create a microbial ingress risk. Under FDA 21 CFR Part 177.1630 and EC Regulation No 10/2011, the finished article must also have low total migration; oxidized gels are polar and can absorb moisture but are not necessarily high migratory species, yet their presence indicates thermal degradation that may also raise acetaldehyde. Acetaldehyde in bottle-grade PET is controlled below 1 ppm in the resin; in cast sheet for water-sensitive applications, the target may be below 5 ppm in the sheet.

A well-configured Cast PET Sheet Line: Chill Roll, Thickness Control and Gel Inspection operates as an integrated loop: roll temperature changes alter surface condition and can modify shrinkage; thickness control changes the local draw and can change gel visibility; gel inspection detects the product of upstream thermal exposure and therefore serves as a feedback variable for extrusion temperature and residence time. Without this integration, a line can produce sheet that meets average thickness but fails intermittently at the thermoformer due to defects that laboratory sampling misses. For trays with shallow draw ratios below 1.5:1, the tolerance to gels is higher, and the inspection system should be tuned to avoid rejecting usable sheet. For deep-draw cups and medical blisters, however, the specification must be stricter because the local stretch ratio at the corners can exceed 5:1. The relationship between gel size and thermoforming reject rate is not linear; a shift in gel size distribution from 50–100 µm to 100–200 µm may increase pinhole rejects by a factor of three, but the exact factor depends on sheet thickness, plug temperature, and forming speed.

Routine calibration of the thickness gauge and optical inspection unit is performed at shift change using certified reference shims and defect films. The beta gauge source, if used, is subject to isotope decay and requires air-density compensation; a temperature change of 5 °C in the scanning frame area can shift the reported thickness by more than 1 µm unless the air temperature sensor is active. Chill roll thermocouples are calibrated against a reference RTD with an uncertainty of ±0.2 °C; a bias of 2 °C on the first roll may not always be visible in the sheet haze, but it changes the effective cooling rate and therefore the amorphous orientation. The optical inspection system is verified with a defect film containing a known count of 50 µm and 100 µm particles; if the detected count deviates by more than 10%, the light source and camera gain are adjusted. Edge trim recycling adds heat history to the regrind stream, and the ratio of regrind is limited to 20–30% for food-contact APET unless the sheet is coextruded with a virgin skin layer. The roll gap is checked with a nip impression or lead-wire test after every roll change, because a differential gap of 0.02 mm across the width produces a wedge-shaped sheet that cannot be corrected by die bolt movement alone.