Polyethylene Terephthalate (PET): Thermoforming PET Sheet: Oven, Mould and Cycle Without Sag

The processing problem defined as “Thermoforming PET Sheet: Oven, Mould and Cycle Without Sag” is a combined heat-transfer and rheology constraint. Polyethylene terephthalate sheet is reheated from a semi-crystalline pellet-derived extrudate into an amorphous or partially crystallized web, then formed in a tool where sag is the visible result of pre-forming extensional flow. The window without sag requires that sheet surface temperature remain within a narrow forming range while the core reaches sufficient extensibility; that mould surface temperature removes heat at a controlled rate; and that the cycle does not advance until demould temperature and part crystallinity satisfy post-mould dimensional stability. The relevant material boundary is intrinsic viscosity. For packaging-grade amorphous PET sheet, intrinsic viscosity measured by ISO 1628-5:1998 is typically reported between 0.72 dL/g and 0.84 dL/g. At the lower end, the sheet has lower extensional viscosity and greater sag tendency, while at the higher end the sheet requires higher oven output to soften and may retain orientation. The glass-transition temperature is near 78 °C for dry amorphous PET; the crystalline melting endotherm measured by ISO 11357-3:2018 is generally observed in the range 245–255 °C. These values define the thermal boundaries that processors manipulate in the oven and the mould.

Moisture control is the first boundary for sag-free forming. Amorphous PET regains moisture from humid plant air; at relative humidity above 60%, sheet stored open for 1–2 h can reach a moisture content that causes hydrolytic chain scission during reheating. Desiccant drying at 150–170 °C with a dew point below -40 °C is specified in resin supplier technical literature to bring residual moisture below 0.003% by mass, measured by ISO 15512:2019. Residual moisture above 0.005% reduces intrinsic viscosity during the oven dwell and produces a melt-like sag that is not corrected by increasing vacuum. Batch-to-batch variation in regrind content is a common production-scale cause of sag: post-industrial trim absorbs moisture and carries degraded polymer chains into the sheet. If the regrind fraction exceeds 30 wt% without proportional drying capacity and moisture measurement, the intrinsic viscosity can drop below the threshold for a stable bubble. The same issue occurs in humid coating lines where the sheet is preheated before lidding-film application; preheating rewets the surface and creates a low-viscosity skin layer.

Why Does Oven Setpoint Fail to Predict Sheet Surface Temperature?

Oven setpoint is a machine parameter, not a sheet property. In zoned infrared ovens, the actual sheet surface temperature depends on heater wavelength, heater-to-sheet distance, air turbulence, and sheet thickness. Quartz heaters with a peak emission below 3 µm heat the surface efficiently, but thick sheet above 3 mm may develop a skin-core temperature difference in which the surface is at forming temperature while the core remains stiff. Ceramic or gas catalytic heaters emitting at longer wavelengths penetrate more evenly into the sheet and are specified for heavy-gauge sheet. Sheet surface temperature is the control variable that correlates with sag length and wall-thickness variation. On industrial rotary lines, a trailing pyrometer array after the final oven zone often records a surface temperature spread of 8–12 °C across a 700 mm web width, even when the oven setpoint is uniform. The centre band reaches a lower extensional viscosity sooner than the edge, sagging before the clamp frame transfers the sheet to the mould. The oven must therefore be zoned with a slightly lower centre-power output in the final heating zone to flatten the profile; a surface-temperature gradient of less than 5 °C across the indexed width is the target commonly noted in process records for stable bubble formation. Measurement by ASTM E1256-11 or a calibrated infrared pyrometer is required because oven air temperature alone does not account for reflected radiation and air convection differences.

Bubble pre-stretch is the first external resistance against sag. The sheet is clamped and inflated with low overpressure, typically 0.02–0.05 MPa, until a photoelectric sensor detects a target height. The bubble geometry redistributes material and compensates for small temperature differences. If the sheet is too hot, the bubble inflates unevenly and may touch the upper oven wall; if it is too cold, inflation is slow and the sheet sags near the centre. The correct bubble is achieved when the surface temperature is above the PET glass-transition temperature but below the crystallization acceleration range that begins to produce haze. For amorphous PET, the forming surface-temperature window is generally cited as 120–150 °C; for crystallizing PET tray stock, the surface temperature may be pushed to 130–160 °C because the mould will complete crystallization. These values are not universal; they depend on intrinsic viscosity, sheet thickness, and line speed.

When Mould Temperature Is Reduced to Shorten Cycle Time

Lowering mould temperature is not a self-contained cycle-shortening strategy. In amorphous PET forming, a mould surface temperature below 30 °C cools the part below glass transition quickly, but it also increases thermal shock in the sidewall and can produce stress whitening at draw ratios above 3:1. The cycle without sag is controlled more by the plug-assist stage than by final mould cooling. The plug must move the sheet into the cavity at a speed that prevents stick-slip and local thinning. Industrial plug materials for PET include syntactic foam, acetal, and polyetherimide; an aluminium plug is generally less suitable because its high thermal conductivity removes heat before the sheet has completed drawing. Plug temperature is often maintained in the range 100–120 °C, and plug speed is limited to below 250 mm/s for sheet thicker than 2 mm. If the plug enters too fast, the sheet is punctured at the centre; if too slow, the sheet sags away from the plug and becomes unavailable for the final vacuum distribution. Vacuum timing must be delayed until the plug has displaced at least 70% of the draw depth; premature vacuum seals the sheet against the plug and causes an uneven wall-thickness distribution. These parameters are measured by linear transducers and pressure sensors, not by visual judgement.

Rotary formers with 16–24 stations and linear shuttle machines differ in transfer time. In ambient plant conditions, sheet thinner than 1.5 mm cools quickly between the oven and the tool; a transfer time of 2–4 s is the typical upper limit before surface temperature falls below the lower forming temperature. Thicker sheet retains heat longer but still exhibits a stable-bubble loss if first-stage cooling occurs at the clamp frame. Clamp-frame force is set to accommodate thermal expansion; across a forming-temperature increase from 20 °C to 140 °C, PET sheet can expand approximately 0.008–0.010 mm/mm. Insufficient clamp force permits the sheet to pull into the cavity, producing sag-like edge movement; excessive force causes local yielding and edge tears. On twin-sheet forming lines, upper and lower sheets are heated on the inner surfaces to 120–140 °C while outer surfaces remain cooler to reduce sticking and maintain handling stiffness. Published data for the optimum clamp force on a specific tool is limited; the value is normally determined during first-article trials using force transducers and post-trim dimensional records.

Cooling Time, Shrinkage and Sag Recovery

Cooling time is the largest component of the total cycle for heavy-gauge PET sheet. A 5 mm thick amorphous PET part can require more than 60 s to reach a demould temperature below the glass-transition range if the cooling water circuit is not designed for turbulent flow. The heat transfer coefficient on the mould side must be higher than the heat transfer coefficient on the air side; otherwise, the part remains hot in the inside corners and continues to shrink after demould. Shrinkage after demould is a form of delayed sag: the part appears dimensionally stable in the tool but distorts outside the mould. Hot-air shrinkage testing by ASTM D1204-14 at 150 °C for 30 min is used to detect residual orientation in sheet and formed parts; values above 5% may indicate that the part was not quenched from the forming temperature quickly enough. In CPET production, the opposite condition applies: the mould is heated to 130–160 °C to induce crystallization after forming. The part remains in the heated mould until the necessary crystalline fraction is reached, after which it can be cooled. If the CPET part is ejected too early, it may pass through the glass-transition range under insufficient constraint and exhibit roundness loss or flange ripple. The cooling time is therefore not a single value but the sum of the time required for thermal stabilisation, crystallization, and dimensional freeze-in.

Crystallinity, measured by differential scanning calorimetry using ISO 11357-3:2018, provides the separation between amorphous and crystallizing PET sheet. In clear APET tray stock, crystallinity is intentionally kept below a visible haze threshold by maintaining mould temperature below 50 °C. In ovenable CPET tray stock, crystallinity is driven to a level that provides dimensional stability at reheating temperatures up to 220 °C; the sheet is formulated and processed so that heater dwell does not cause excessive crystallinity before forming. The rate of crystallization is slow enough that a sheet held too long in the oven below the melt temperature can develop spherulitic haze and a higher stiffness that resists sag but also resists uniform draw. This is a processing conflict: a longer oven dwell may appear to reduce sag because the sheet stiffens, but it also narrows the forming window by reducing drape. The solution is a timed infrared heating profile that raises the surface temperature quickly and then transfers the sheet before crystallinity exceeds the level at which the part can still be formed without stress cracking.

Mould Surface Chemistry and PET Hydrolytic Stability Boundaries

Mould release agents and cleaning chemicals alter the surface friction of PET sheet. Silicone-based external release agents can contaminate sealing and printing surfaces; fluoropolymer-based release coatings on plug surfaces reduce sticking but may not be suitable for all food-contact applications. For food-contact PET sheet, the final article must comply with FDA 21 CFR 177.1630 or Regulation (EU) No 10/2011, and process aids must be assessed under those frameworks. Hydrolytic stability is also limited by melt-temperature excursions; if the extruder melt temperature exceeds approximately 285 °C, acetaldehyde generation increases and intrinsic viscosity decreases. The thermoformer does not see the melt history, but the incoming sheet carries it as a reduction in extensional viscosity. Incoming sheet with lower intrinsic viscosity may form a stable bubble at lower oven temperature but will sag more readily if the line stops or if oven zones drift. For this reason, process control records should pair intrinsic viscosity with sheet surface temperature and mould temperature, not with oven setpoint alone.

Tensile properties of thermoformed PET parts are commonly measured using ASTM D638-14 or ISO 527-2:2012, but the reported values are influenced by wall-thickness distribution and residual orientation. A part with sag-induced variation will show higher elongation at break in the stretched sidewall and lower in the bottom, which may not appear in a single-point average. Oxygen permeability of PET packaging is tested by ASTM D3985-17, and moisture vapour transmission by ASTM F1249-20; these barrier values are affected by thickness non-uniformity caused by sag. The forming window without sag is therefore not only a defect-avoidance target but also a specification for barrier and mechanical consistency. Processors should record pre-form moisture, intrinsic viscosity, oven zone pyrometry, mould coolant temperature, plug speed, and vacuum timing as correlated data, because no single parameter is sufficient to hold a cycle without sag. When a sag defect recurs, the pattern should be read as a signal that the sheet temperature profile and the rheological condition of the incoming sheet have shifted relative to the tool timing.