In production of 55–80 L passenger car fuel tanks, a 6-layer coextruded parison is generated from a multilayer accumulator blow moulding machine in which separate extruders supply high-molecular-weight bimodal HDPE outer and inner skins, a regrind HDPE layer, two maleic anhydride-grafted tie-resin layers, and a central EVOH barrier layer. The parison is extruded vertically through an annular die with a diameter programmed between 80 mm and 150 mm, cut to a controlled length, transferred to a mould, and inflated at 0.7–1.2 MPa blow pressure. During the interval between die exit and mould closure, which commonly lasts from 2 s to 6 s for a 4.5 kg shot, gravitational stress causes unsupported parison elongation and non-uniform wall thinning known as sag. The resistance of the extruded parison to this sag is strongly influenced by melt strength, a property measured as the tensile force exerted by a melt strand during continuous uniaxial drawing. Melt strength values obtained on a Göttfert Rheotens 71.97 at 190 °C and 2.4 mm/s² draw acceleration for automotive fuel tank HDPE grades are typically reported in supplier technical literature within a range of 0.15 N to 0.35 N, although exact values vary with regrind content and high-molecular-weight fraction. Automotive fuel tank designs are validated under UN/ECE R34, and evaporative emission compliance under CARB LEV III and Euro 6d depends on continuous EVOH barrier layer coverage after blow moulding, which in turn depends on parison wall thickness uniformity. Parison sag therefore has a direct regulatory consequence in addition to its effect on pinch-off weld quality, minimum wall thickness at tank corners, and barrier layer thinning.

How Does Strain Hardening in Bimodal HDPE Limit Gravitational Parison Elongation?

Melt strength is not an intrinsic material property but an engineering response measured under non-isothermal uniaxial extension. In fuel tank blow moulding, the relevant deformation occurs at low Hencky strain rates, often between 0.1 s⁻¹ and 2.0 s⁻¹, as the parison extends under its own weight. The sag rate is controlled by the transient extensional viscosity at these strain rates and by the zero-shear viscosity of the melt at the local melt temperature. Bimodal HDPE grades used for fuel tank skins typically contain a high-molecular-weight fraction that raises zero-shear viscosity above 5×10⁴ Pa·s at 190 °C, while retaining a 21.6 kg melt flow rate from 4 g/10 min to 8 g/10 min as determined by ISO 1133-1:2022. The low-molecular-weight fraction maintains processability under die-land shear rates that can reach 100–1000 s⁻¹. Capillary rheometry per ISO 11443:2021 on a Rosand RH7 instrument can quantify the shear viscosity decrease associated with shear thinning; for bimodal fuel tank HDPE, the Carreau-Yasuda power-law index is usually between 0.30 and 0.50. Strain hardening is less pronounced in linear HDPE than in LDPE, but the long relaxation time of the high-molecular-weight fraction provides a transient resistance to sag during the short open-parison interval. A higher high-molecular-weight fraction increases Rheotens melt strength and reduces sag, but it also raises die pressure and shear heating. In production-scale accumulator machines, switching regrind-rich layers to virgin HDPE at constant die gap produces die-pressure variations of 3–6 MPa and melt-temperature increases of 2–4 °C, which must be compensated by barrel temperature adjustments. Published peer-reviewed data that directly correlate Rheotens melt strength to six-layer parison sag at production shot masses of 4–6 kg is limited; the operational relationship is therefore derived from material supplier technical data and validated on individual machines using parison length monitoring.

Regrind is introduced into the fuel tank structure as a discrete layer because post-industrial flash and rejected tanks contain EVOH barrier material. The melt strength of regrind-containing HDPE is lower than that of virgin skin HDPE due to thermo-oxidative degradation during reprocessing and the presence of dispersed EVOH gel particles. In a typical six-layer automobile fuel tank, the regrind layer can occupy 20–40 wt% of total wall thickness, and its melt strength may be 10–25% lower than virgin HDPE at the same extrusion temperature. The position of the regrind layer directly beneath the outer skin means its reduced melt strength can create localized thinning at the transition zones adjacent to the mould pinch-off. A commercial accumulator blow moulding line running a 120 mm grooved-feed extruder with a 24:1 L/D barrier screw and a 35 kg accumulator head has demonstrated that increasing regrind content from 20 wt% to 40 wt% forced a die-gap reduction of 0.5–1.0 mm to maintain a consistent parison length at a 4.5 kg shot weight. The same line also showed that parison sag, measured as the percentage increase in parison length during a 4 s open interval, increased from 8% to 14% when the high-molecular-weight fraction of the skin HDPE was replaced by a lower-melt-strength extrusion grade. These values are representative of plant data but are not universal; published data for this specific equipment configuration is limited, and each line must be characterized with its own parison sag measurement system.

The extrusion of EVOH in the centre layer imposes a strict thermal window. EVOH copolymers used for fuel tank barriers, typically containing 27–32 mol% ethylene, are processed at 210–230 °C. Above 240 °C, thermal degradation produces acetic acid, crosslinked gels, and brown specks, which reduce barrier performance and disrupt layer continuity. When the skin HDPE melt strength is increased by raising barrel temperatures to reduce viscosity, the interfacial heat transfer from the HDPE melt to the EVOH layer through the tie resin is increased. If the EVOH layer reaches 235–240 °C in the accumulator head due to shear heating or extended hold time, the resulting gel formation can force line stoppages for die cleaning. A 0.05 N increase in Rheotens melt strength produced by a switch to a higher high-molecular-weight HDPE grade has been observed to correlate with a 2–3 °C increase in the EVOH layer temperature at a 60 kg/h total throughput, although the magnitude depends on the accumulator head design and layer ratio. This is the central processing conflict: the melt strength required to control sag also increases heat generation in the barrier layer, narrowing the processing window to approximately ±5 °C around the EVOH setpoint.

When EVOH and tie-layer viscosity mismatch creates interfacial waviness

Multilayer parison stability is not determined solely by the skin layer melt strength. The EVOH barrier layer and the maleic anhydride-grafted tie layers have different shear viscosity–shear rate relationships and different thermal sensitivities. At the shear rates generated in the die land, which may range from 50 s⁻¹ to 500 s⁻¹, the viscosity of a 29 mol% ethylene EVOH at 220 °C can be higher than that of HDPE at 230 °C, but the difference varies with shear rate and temperature because EVOH has a stronger Arrhenius temperature dependence. Viscosity mismatch at the layer interfaces can produce interfacial waviness, layer thickness oscillation, and local barrier layer thinning during parison sag. A continuous EVOH layer is required to meet permeation values; if sag-induced wall thinning causes the EVOH layer thickness to fall below 2–3 µm at a tank corner, the oxygen and hydrocarbon barrier function is compromised. The tie layers, typically maleic anhydride-grafted HDPE with a graft level of 0.15–0.30 wt%, act as stress-transfer layers between the HDPE skins and EVOH. Their melt strength lies between that of the skin and the EVOH, and their thickness is often 4–8% of the total wall. When the outer skin has high melt strength but the tie layer does not, the parison can sag internally with the outer skin retaining shape while the barrier layer moves downward, producing a continuous outer appearance but non-uniform internal layer distribution. On a six-layer spiral mandrel accumulator head, interfacial waviness can be detected by sectioning moulded tanks and measuring layer thicknesses with a digital optical microscope at 20× magnification. Production audits under SAE J1737 or CARB TP-901 have shown that barrier-layer waviness greater than ±15% of nominal thickness correlates with increased permeation variability. The processing remedy is to match tie-layer viscosity to the skin layer by adjusting tie-resin grade or by modifying the die-temperature profile, but this adjustment can reduce overall melt strength if the tie resin is less viscous.

Parison programming on accumulator blow moulding machines compensates for sag by varying the annular die gap during extrusion. A 64-point parison programmer controls a hydraulic servo valve that moves the die mandrel to create thicker sections where the parison will sag or where pinch-off material must support the tank seam. The programmed wall thickness profile is based on the sag behaviour of the specific compound and the mould geometry. For a fuel tank with a 1200 mm parison length and a 4.5 kg shot, sag compensation can require a die gap that opens from 2.0 mm at the top of the parison to 8.0 mm near the bottom, but these values depend on layer structure and melt strength. If melt strength is increased, the same sag compensation curve becomes flatter and less bottom-heavy wall thickness is required. However, high melt strength can also reduce pinch-off weld quality because the high-molecular-weight HDPE skin layers resist interdiffusion at the pinch line. Charpy notched impact tests on weld seams per ISO 179-1:2010 at -40 °C can reveal a drop in impact energy when a melt-strength grade is used without re-optimizing pinch-off pressure and mould temperature. On one production line, a change from a 0.18 N Rheotens melt strength HDPE to a 0.28 N grade improved parison sag from 12% to 7% but reduced weld-line impact strength by 15–20%, requiring an increase in pinch-off pressure from 0.8 MPa to 1.0 MPa to restore weld performance. This trade-off demonstrates that melt strength cannot be evaluated independently of weld-line mechanics.

Rheotens melt strength and zero-shear viscosity limits for long parison dwell

Long parison dwell occurs when large tanks, complex moulds, or slow clamp movements extend the open time beyond 6 s. At these dwell times, sag is controlled primarily by zero-shear viscosity and the time-dependent extensional viscosity. Bimodal HDPE with a zero-shear viscosity above 8×10⁴ Pa·s at 190 °C can support a 4.5 kg parison for 6–8 s with acceptable sag, but such a high zero-shear viscosity may exceed the torque capability of smaller extruders. The practical upper bound is set by the extruder drive and the melt temperature limit of the inner EVOH layer. A 100 mm extruder with a 24:1 L/D barrier screw can process a 21.6 kg MFR of 4 g/10 min HDPE at 60 kg/h, but a grade with 2 g/10 min may require reducing output to 40–50 kg/h to keep melt temperature below 245 °C. The lower throughput increases residence time in the accumulator head, which exacerbates EVOH thermal exposure. Thus a melt strength strategy that uses lower MFR can become self-defeating for multilayer fuel tanks because the barrier layer residence time rises. The preferred method is to use a bimodal HDPE with a low shear viscosity but high extensional viscosity, achieved by a high-molecular-weight fraction that contributes to strain hardening without excessively raising die-land pressure. Such grades are designed with a shear-thinning index near 0.30 and a melt strength above 0.25 N at 190 °C. A production-scale accumulator machine with a 150 mm annular die and 6-layer feedblock can maintain a parison sag below 10% at a 5 s open time when the skin HDPE reports a Rheotens melt strength of 0.25 N or greater and the EVOH layer is held at 225 °C. If the melt strength falls below 0.18 N, sag compensation may fail at the lower parison regions, producing tank corners with wall thickness below specified minimums. The relationship is not linear; a change from 0.15 N to 0.20 N may reduce sag by 20–30%, whereas a change from 0.25 N to 0.30 N may yield only 5–10% further improvement, depending on parison length and layer ratio.