Why Does HDPE Shrink After Thermoforming?

High-density polyethylene (HDPE) is a semicrystalline polyolefin whose typical density spans 0.941 to 0.967 g/cm³ when measured by ASTM D1505-18, and whose peak melting endotherm is generally reported between 130 and 137 °C by differential scanning calorimetry per ASTM D3418-15. Extrusion-grade sheet resins often exhibit a melt flow rate of 0.3–6.0 g/10 min at 190 °C under 2.16 kg load per ISO 1133-1:2011. Thermoforming requires heating HDPE sheet to 160–190 °C, above the crystalline melting range, at which point the sheet can be drawn into a cavity. The question—Why does HDPE shrink after thermoforming?—does not have a single cause. The dominant mechanisms are frozen-orientation recovery, thermal contraction, crystallinity-driven densification, and residual-stress redistribution after demolding. Each mechanism is sensitive to extrusion history, heating uniformity, draw ratio, mold temperature, and cooling time.

What Causes Post-Thermoforming Shrinkage in HDPE Sheet?

During sheet extrusion, HDPE melt is forced through a flat die and polished on a chill-roll stack. The melt is stretched in the machine direction between the die lips and the roll nip and is simultaneously cooled below the crystallization onset temperature. Because HDPE crystallizes rapidly, some chain segments are trapped in extended conformations rather than fully relaxed random-coil conformations. When the sheet is later reheated to forming temperature, these oriented chains recover entropy and the sheet shrinks in the machine direction before forming. Unrestrained reheating of anisotropic HDPE sheet commonly produces greater dimensional loss in the extrusion direction than in the transverse direction.

The forming step itself introduces a second orientation state. When the heated sheet is drawn into the cavity by vacuum or plug-assisted pressure, biaxial or uniaxial extension aligns chains along the mold surfaces. The outermost layers quench first against the tool, preserving this orientation before complete relaxation. After demolding, the inner layers continue to cool and crystallize, generating a through-thickness stress profile. Over hours to days, chain relaxation at ambient temperature reduces the residual stress and produces measurable post-forming shrinkage. Published data for this specific configuration is limited because shrinkage is strongly grade-, draw-ratio-, and tooling-dependent.

On a roll-fed pressure-forming line running HDPE sheet at 1.2–2.5 mm gauge, sheet surface temperature is typically maintained at 160–175 °C using quartz panel heaters with infrared feedback. Plug-assisted forming with a syntactic foam plug and an aluminum water-cooled tool at 65–80 °C produces parts that are demolded at surface temperatures often near 70 °C. Immediate post-demold dimensions are transient. A dimensional inspection at 2 h after forming may record 0.8–1.5% linear loss in the machine direction and 0.4–1.0% in the transverse direction relative to the cold tool dimensions. The machine-direction excess is attributed both to residual extrusion orientation and to the directional draw that accompanies roll-fed indexing. These values are representative production observations rather than specification-grade values for every HDPE grade.

Cooling-rate asymmetry is a further source of post-shrinkage. The formed part remains in contact with the female cavity face longer than with the plug side, so the cavity-side skin is quenched more rapidly and develops lower crystallinity than the air-side surface. Subsequent crystallization of the undercooled skin increases density and reduces part dimension. This effect is especially visible in deep-draw parts with a plug displacement greater than 50% of the cavity depth.

Crystallinity, Cooling Rate, and Frozen-In Stress in HDPE Semi-Finished Sheet

In HDPE, the crystalline unit cell is denser than the amorphous phase. Density values of 0.945 to 0.960 g/cm³ correspond to crystallinity fractions roughly between 60% and 80% when referenced to 1.000 g/cm³ for the crystalline phase and 0.855 g/cm³ for the amorphous phase. A one-percent change in crystallinity therefore produces an observable volumetric change. During sheet extrusion, fast chilling suppresses spherulitic perfection and lowers sheet density relative to slow-cooled HDPE. When the sheet is reheated for thermoforming, partial melting destroys residual crystalline order. On cooling in the mold, new crystallites form and the material densifies. If the part is demolded before this crystallization is complete, continued secondary crystallization in the first 24–72 h after forming contributes additional shrinkage.

Cooling rate and sheet thickness jointly govern the crystallinity gradient. For sheet of 2.0 mm gauge cooled on a roll stack with surface temperature near 85 °C, the core cools more slowly than the skins, so core crystallinity and spherulite size can be higher. After forming, the thermal history is reversed: the tool-contact surface cools faster than the free surface. Parts show anisotropic shrinkage because the two surfaces have different crystallinity and orientation. This is a production-scale phenomenon regularly observed on water-cooled aluminum tools maintained at 60–80 °C, where the cooling rate at the cavity interface can be several times greater than that at the plug-side free surface.

Thermal contraction from forming temperature to ambient is also linear. Using a representative coefficient of linear thermal expansion of 120–220 × 10⁻⁶ m/(m·K) for HDPE by thermomechanical analysis per ISO 11359-2, a nominal temperature drop of 140 K from the forming temperature to 23 °C predicts a linear thermal contraction of 1.7–3.1% for an unrestrained state. In practice, partial crystallinity and mechanical constraint during cooling reduce this value, but the calculation demonstrates the magnitude of the thermal contribution. The actual measured shrinkage is the sum of recoverable strain, thermal contraction, and crystallization-induced densification, less any restraint from mold surfaces until demolding.

Tooling compensation for HDPE cannot rely on a single published shrinkage factor. A cold-tool dimension of 100.0 mm in a shallow tray may correspond to a formed part dimension of 99.0–99.5 mm at 24 h, while a deep-draw container with a draw ratio above 2:1 may require different compensation in the top flange, sidewall, and bottom. Measurement at 24 h under 23±2 °C and 50±5 %RH is common practice because initial post-demold shrinkage is not complete. Parts are not fixture-checked until the post-forming dimensional change has stabilized. This practice aligns with the conditioning requirements used in dimensional test methods for plastics.

When Tool Temperature Falls Below the Effective Recrystallization Threshold

Tool temperature exerts a threshold effect. If the aluminum water-cooled tool is held below roughly 65 °C, the contacting surface can quench to below the crystallization onset temperature before full spherulitic development occurs. The skin becomes more amorphous and more orientation-bearing. After demolding, this undercooled skin undergoes enthalpy relaxation and secondary crystallization, generating tensile skin stresses and concave or convex warp depending on part geometry. If the tool is held above 85 °C, cycle time increases and the part may stick or distort during ejection because the modulus remains low. For many HDPE grades, the practical forming tool temperature window is 65–85 °C, with a target of 70–80 °C used on pressure formers to balance shrinkage control and cycle-time economics.

The frozen-in stress profile is not uniform across the wall. In plug-assisted forming, the plug side undergoes additional contact and reheating due to frictional heat at the plug surface. The cavity side receives the earliest quench. The resulting differential crystallinity and residual stress frequently produce inward bowing of shallow lids or outward bowing of deep-draw containers. This is most pronounced when demolding occurs before the average part temperature falls below 60 °C. On production equipment, insufficient cooling time is a common root cause of post-thermoforming dimensional drift, and it is corrected by extending the in-mold cooling phase or by reducing tool temperature within the specified window.

Measuring Shrinkage: ASTM D2732 and ISO 11501 in Production Quality Control

Unrestrained linear thermal shrinkage of HDPE sheet is characterized by methods such as ASTM D2732-18 and ISO 11501. These methods use flat sheet specimens and expose them to a controlled elevated temperature for a specified time. They are most useful for evaluating incoming sheet orientation because they isolate the sheet's frozen-in strain from tooling effects. For formed parts, dimensional change is measured directly against the tool drawing using coordinate measuring machines or dial indicators. The distinction is important: a sheet may pass an unrestrained shrink test at 150 °C and still produce parts with unacceptable post-forming shrinkage because the forming step has superimposed new orientation.

The interpretation of these tests must account for the difference between unrestrained sheet shrinkage and restrained in-mold shrinkage. A sheet with high machine-direction orientation may show elevated machine-direction versus transverse-direction shrinkage in an unrestrained test, and this anisotropy will persist into the formed part as differential edge contraction. Production control therefore tracks both the absolute shrinkage and the machine-direction/transverse-direction ratio rather than relying on a single value.

Rheology, Regrind Fraction, and Batch-to-Batch Shrinkage Variation in Thermoforming

Melt flow rate variation within the 0.3–6.0 g/10 min range changes draw behavior and orientation retention. Lower-melt-flow-rate HDPE grades exhibit higher extensional viscosity and can retain more molecular orientation after forming, which may increase post-forming shrinkage in the draw direction. Higher-melt-flow-rate grades flow more easily and often produce more uniform wall thickness, but they may sag during heating and can show greater variability in plug-assist consistency. The relationship between rheology and shrinkage is not monotonic because it is mediated by sheet extrusion conditions and tool temperature.

Regrind addition further complicates shrinkage control. Thermoforming skeletons, edge trim, and rejected parts are ground and blended back into sheet at levels commonly between 10% and 50% by weight. Repeated thermal history shifts molecular weight distribution, reduces extensional viscosity slightly, and can alter crystallization kinetics. A production lot containing 30% regrind may therefore show different post-forming shrinkage than a virgin-only sheet of the same nominal grade. Batch-to-batch variation in pigment masterbatch, nucleating agents, or catalyst residuals can also shift the crystallization onset temperature enough to require tool-temperature or cooling-time adjustment.

On roll-fed cup lines running at 60–100 cycles/min, the practical consequence of uncontrolled post-shrinkage is frequently detected as inward rim movement after trimming. If immediate post-trim dimensions are used for tooling qualification, lid fitment failures can occur after 24–48 h warehouse aging. Tooling is therefore adjusted using aged part measurements rather than immediate measurements. Documented process records for sheet surface temperature, mold temperature, plug displacement, regrind percentage, and cooling time provide the traceability needed when HDPE parts exhibit field-dimensional deviation.