In accumulator-head extrusion blow molding of 220 L open-head high-density polyethylene drums, the relationship between parison programming, mold heat transfer, and post-cooling fixture dwell determines whether the finished container remains within UN 1H2 design-type dimensional and stress-cracking limits. A typical machine configuration uses a 90 mm to 120 mm barrier-flighted extruder with a 30:1 L/D ratio, delivering a melt temperature of 190 °C to 210 °C into an accumulator head with shot capacity between 12 kg and 20 kg. The accumulator piston compresses the melt reservoir and forces it through a diverging annular die gap, while a 100-point or 200-point parison programmer varies the die gap axially to compensate for swell, hang time, and variable blow-up ratio. For a 9 kg to 11 kg shot, the programmed wall thickness profile typically ranges from 2.5 mm in the body to 6.0 mm or more at the top chime and bottom wear-ring regions. Cooling uniformity is therefore not a single-mold condition but a sequential problem: partial heat removal occurs during blow molding, followed by constrained thermal contraction in a post-cooling fixture that contacts the drum lip, sidewall, and bottom chime. Published data for this specific configuration is limited, but production-scale observation indicates that the mold alone removes between 40% and 55% of the total enthalpy, with the remainder removed by internal air circulation, water-cooled contact plugs, or ambient convection after demolding. The open-head design introduces an additional asymmetry because the top opening is unsupported by a closed dome and is free to ovalize during cooling unless a temperature-controlled sizing plug is inserted within the post-cooling station.
The high-density polyethylene grade selected for 220 L open-head drums is usually a bimodal copolymer with a density of 0.945 g/cm³ to 0.955 g/cm³, a melt mass-flow rate of 0.2 g/10 min to 0.5 g/10 min at 190 °C/2.16 kg, and a carbon black content of 2.0% to 2.5% for ultraviolet stabilization in outdoor storage. The bimodal molecular weight distribution provides a high-molecular-weight fraction for melt strength and environmental stress crack resistance and a low-molecular-weight fraction for processability. The zero-shear viscosity of these grades at 190 °C is typically between 50,000 Pa·s and 150,000 Pa·s, which is why accumulator heads are required rather than continuous reciprocating-screw blow molders for large shot weights. The high viscosity also means that self-heating during parison extrusion can add 5 °C to 10 °C to the melt temperature if the die gap is too small or the accumulator piston speed is too high, and this temperature rise directly reduces parison melt strength and increases sag. Cooling uniformity is therefore influenced by the rheological consistency of the melt before it enters the mold, not only by the mold and fixture heat transfer.
The limiting thermal resistances are, in descending order, the polymer-mold interface conductance after shrinkage, the conductive path through the mold shell, and the convective boundary layer in the cooling channels. For high-density polyethylene with a density of 0.945 g/cm³ to 0.955 g/cm³, thermal conductivity is commonly reported between 0.38 W/m·K and 0.52 W/m·K, while thermal diffusivity ranges from 0.13 mm²/s to 0.18 mm²/s. Theoretical conduction estimates for a 4 mm wall using the one-dimensional plane-slab solution yield substantially shorter cooling times than are observed on production lines because the equation assumes perfect thermal contact and neglects the latent heat of crystallization, which for HDPE is approximately 180 J/g to 220 J/g. Once the part shrinks away from the mold cavity, the interfacial heat transfer coefficient can fall from an initial 1,500 W/m²·K or higher to below 500 W/m²·K, depending on mold surface texture, cavity pressure, and local wall thickness. The consequence is that thick chime sections continue to cool long after the adjacent body wall has solidified, creating thermal gradients that drive anisotropic shrinkage, sink marks, and residual stress. In the upper open-head region, the absence of a blow pin or dome means that the lip and cover-seating area are in contact with the mold only on the outer surface unless an internal sizing plug is applied post-mold. The lower chime, by contrast, is often the thickest zone because it must resist side-impact and fork-tine damage, and it sits in a mold corner where coolant channel access is geometrically restricted. Therefore, cooling channel design in the chime corners is the primary process conflict: the need for maximum metal removal for drilled or conformal channels must be balanced against the structural stiffness of the mold insert and the part ejection force. Turbulent cooling water at a Reynolds number above 4,000 is generally required, with channel diameters of 12 mm to 16 mm and circuit pressure drops of 1 bar to 3 bar. Chilled water at 8 °C to 15 °C is typical, but lower temperatures below the local dew point can produce mold surface condensation that disrupts parison adhesion and creates surface defects. Published data for specific mold configurations is limited; however, production-scale evidence from large-format accumulator lines indicates that reducing the cooling water supply temperature by 5 °C does not proportionally reduce cycle time if the polymer-mold contact resistance is controlling.
Accumulator-head tooling with a diverging die gap and axial parison programming produces a wall-thickness distribution that is intentionally non-uniform to compensate for local blow-up ratio and mechanical performance requirements. The programmed die gap is typically smallest at the parison tail for the top lip and largest at the base and upper chime segments, but the exact profile depends on accumulator piston speed, melt viscosity, tooling diameter, and hang time. A tooling diameter between 350 mm and 450 mm is common for 220 L drum bodies, with a die gap range of 0.8 mm to 3.0 mm. The die land length is usually 10:1 to 15:1 relative to the gap at the smallest opening, which increases backpressure and homogenizes the melt temperature. The piston speed profile is adjusted so that the shear rate at the die exit stays within a narrow band, preventing melt fracture and controlling die swell. Weight variation from shot to shot is a direct input to cooling uniformity because a heavier shot increases local wall thickness and total enthalpy, while a lighter shot reduces the contact pressure for the mold and can shift the onset of shrinkage-gap formation. Modern accumulator-head controls hold shot weight within approximately ±0.5% to ±1.0% of target on a well-maintained machine, but this still corresponds to 45 g to 110 g of HDPE per 10 kg shot. When a post-cooling fixture is indexed from the mold to a calibration station, the drum transfers from a cavity at 10 °C to 15 °C into an environment where internal air at 10 °C to 15 °C is circulated at 3 m³/min to 5 m³/min per cavity. The fixture may include a water-cooled aluminum plug inserted into the open top, a bottom plate with cooling channels positioned under the wear ring, and pneumatic side fingers that maintain roundness. The contact plates are usually maintained at 10 °C to 16 °C and exert a clamping pressure of 0.3 MPa to 0.6 MPa on the external wall. The plug must be inserted while the top lip is still above the solidification temperature but not so early that the material is displaced or the cover gasket seating surface is damaged. This window is typically a 10 s to 20 s interval after mold opening, but published data for specific fixture configurations is limited.
The post-cooling fixture functions as a constrained heat exchanger and dimensional fixture, not as a simple support. Its geometry must be matched to the shrinkage anisotropy of the molded drum, which can exceed 1.5% to 3.0% in linear dimensions depending on wall thickness and cooling rate. A typical station includes a water-cooled top plug, a bottom plate, and two or four segmented side shells. The top plug is often machined from 6061 aluminum or 7075 aluminum with internal drilled channels arranged in a spiral or manifold pattern, and it enters the drum opening with a radial clearance of 0.5 mm to 1.5 mm per side when the part is at demolding temperature. The bottom plate is designed to contact the wear-ring area with a high interface pressure, which not only removes heat but also sets the bottom chime flatness. Side fixture segments are contoured to the body profile and apply uniform radial restraint to prevent ovalization, which is critical because the open-head lip can display a diameter variation of 2 mm to 4 mm if unrestrained. Water circuits in the top plug and bottom plate use turbulent flow to maintain a heat transfer coefficient above 2,000 W/m²·K on the water side; the limiting resistance remains the polymer-metal contact, which can be improved by applying a 0.2 MPa to 0.4 MPa surface pressure. Common circuit layouts include parallel branched manifolds with balance valves to prevent flow starvation in the lower chime. Flow rates of 20 L/min to 40 L/min per cooling circuit are typical, with a water temperature rise of 2 °C to 4 °C across the fixture. The post-cooling dwell time is usually between 60 s and 120 s for a 10 kg drum, but the actual value is determined by the target ejection temperature, which must be below 90 °C to avoid post-demolding shrinkage and above the dew point to avoid condensation. In multi-station rotary fixtures, each station is supplied from a common chiller loop, and balancing valves are set to maintain a temperature spread of no more than ±1 °C across all cavities. If one fixture circuit receives insufficient flow, the affected drum leaves the station with a hotter bottom chime, and the resulting thermal contraction after release can produce a convex bottom or a lip diameter that falls outside the cover-seal tolerance. Published data for specific fixture designs is limited, but the production-scale failure mode is well documented: unbalanced post-cooling produces intermittent dimensional defects that cannot be corrected by mold adjustments alone.
| Drum zone | Typical wall thickness | Cooling mode | Channel diameter | Reynolds number | Polymer-side heat transfer coefficient | Observed defect if mismatch |
|---|---|---|---|---|---|---|
| Top lip and cover seat | 5.0–7.0 mm | Mold outer surface plus post-cooling plug | 10–12 mm | 4,000–8,000 | 1,000–2,500 W/m²·K | Ovality, cover seal leakage |
| Body sidewall | 2.5–4.5 mm | Mold cavity channels | 12–15 mm | 4,500–10,000 | 800–1,800 W/m²·K | Sink marks, wall thinning, barrel distortion |
| Bottom wear ring | 5.0–8.0 mm | Mold corner plus bottom plate | 12–16 mm | 3,500–7,500 | 900–2,200 W/m²·K | Flatness error, fork-impact failure |
| Pinch-off weld | 4.0–8.0 mm | Flash pocket and secondary contact pad | 10–14 mm | 3,000–6,500 | 700–1,600 W/m²·K | Weld-line cracking, ESCR reduction |
Mold temperature control in the top chime and bottom wear-ring regions depends not only on coolant temperature but also on channel spacing, mold insert material, and the presence of high-conductivity inserts. Many large blow molds for 220 L drums are produced from wrought 7075 aluminum or cast aluminum alloys, with thermal conductivity values roughly 130 W/m·K to 170 W/m·K, compared with 40 W/m·K to 50 W/m·K for P20 tool steel. The higher conductivity accelerates initial heat removal but also makes the mold surface more responsive to coolant temperature fluctuations, which can create transient thermal gradients across the cavity. Channel spacing in the sidewall is often set at 1.5 to 2.0 times the channel diameter, with a distance of 25 mm to 35 mm from the cavity surface. In chime areas, space constraints force tighter channel placement and may require drilled plugs or beryllium-copper inserts with thermal conductivity above 300 W/m·K to move heat from the thickest sections. The use of conformal cooling has increased but remains limited by the cost of additively manufactured inserts; when conformal channels are used, they usually follow the top chime profile with a spiral or helical path to maintain a constant distance to the cavity. The transition from the molded drum to the post-cooling fixture occurs at a surface temperature that is highly non-uniform. Infrared thermography of production drums typically shows the top lip and bottom wear ring at 90 °C to 110 °C while body walls may already be at 50 °C to 70 °C. The core thermal asymmetry is that the post-cooling fixture must remove the most heat from the zones that also have the tightest dimensional tolerances. If the fixture contact is initiated too late, the thick sections have already passed through crystallization and the fixture can only function as a dimensional constraint, not a heat exchanger. If contact is too early, the part is too soft and the top plug can introduce surface deformation. The control strategy therefore requires mold-open temperature monitoring or a fixed transfer delay based on cavity temperature feedback, not simply a timer. Published data for specific mold configurations is limited.
When a multi-cavity post-cooling carousel is supplied from a central chilled water loop, the uniformity of the final drum dimension depends on the hydraulic balance of the circuit and the thermal load sequencing. A typical four-station fixture may draw 60 L/min to 120 L/min total chilled water at 10 °C, with return water at 14 °C to 16 °C. If the supply pressure fluctuates by 0.2 bar or more during indexing, the flow rate through high-resistance branch circuits can drop enough to reduce the water-side heat transfer coefficient by 20% to 30%, even though the overall chiller outlet temperature remains stable. This is because the polymer-side interface resistance is not the only control in the post-cooling stage: once the part has contracted from the mold, the fixture contact pressure becomes the dominant factor. A reduction in water flow in the top plug circuit may not change the chiller setpoint but it allows the plug surface temperature to rise transiently, and the next drum in that cavity is therefore cooled at a lower effective temperature difference. Because the open-head lip is thin relative to the lower chime, it responds quickly to fixture temperature changes, while the lower chime integrates the thermal history over a longer dwell. The result is a batch-to-batch diameter variation where the top lip moves by 0.5 mm to 1.0 mm while the bottom wear ring moves by 1.0 mm to 2.0 mm, producing conical or barrel-shaped sidewall profiles. Process capability studies on large-format blow molding lines indicate that a process capability index Cp of 1.33 for top-lip diameter can degrade to below 1.00 if the water balancing is not checked after every tooling change. The cooling-water supply should be filtered to 50 µm or finer because mineral scale and biofilm in small-diameter fixture channels are a known cause of progressive flow imbalance. A supply temperature below the ambient dew point can also cause condensation on the fixture, which transfers water to the drum surface and can interfere with ultrasonic cover welding or in-line leak testing. Published data for this specific configuration is limited, but the mechanical failure pattern is consistent with flow starvation rather than chiller capacity loss.
During post-cooling, crystallization of high-density polyethylene is the additional thermal load that distinguishes practical cycle times from idealized calculations. The enthalpy of crystallization is approximately 180 J/g to 220 J/g, which is a significant fraction of the total heat removed from the melt and can create a plateau in the cooling curve near 120 °C to 125 °C depending on comonomer content and cooling rate. In thick chime sections that cool slowly, the crystallization plateau extends the time during which the polymer is semicrystalline but dimensionally active. Differential cooling between the outer skin and the core also produces residual stress, which is measurable as a change in the stress-crack resistance of the molded article. The open-head lip is particularly susceptible because it experiences rapid quenching against the mold and then re-heating from the residual heat of the adjacent top chime when the mold opens. This re-heating can raise the lip surface temperature by 10 °C to 20 °C after demolding, just as the post-cooling plug is inserted, causing a transient expansion and a change in the effective plug clearance. If the plug clearance was set on a fully cooled sample, the interference fit during production may be too tight, producing surface scoring or lip deformation. The reverse is true if the plug clearance is too loose: the lip remains unsupported and free to ovalize. A common production compromise is a segmented plug with spring-loaded segments that maintain a constant radial force of 0.1 MPa to 0.3 MPa on the lip while allowing thermal expansion. The lower chime is often cooled by a water-cooled bottom plate with a raised ring that matches the wear-ring groove, and the contact pressure is supplied by the drum weight plus a pneumatic cylinder applied to the top plug. The applied force is typically 500 N to 1,200 N, which is low enough to avoid column buckling but high enough to ensure contact. Published data for specific fixture configurations is limited; however, production experience indicates that force-controlled contact is more repeatable than position-controlled contact because the part shrinks during the post-cooling cycle and a fixed position cannot maintain uniform pressure.
The pinch-off zone at the base of an accumulator-head molded drum is a thermal singularity. The mold pinch-off compresses the parison tail and forms a weld line by squeezing out flash, and the local wall thickness after deflashing can be 4 mm to 8 mm, but the cooling channel pattern around the pinch-off is often interrupted by the mold parting line and the flash pocket. The result is a hot stripe that extends from the pinch-off toward the bottom wear ring, while the open-head lip at the opposite end is cooled primarily by the outer mold surface and, after transfer, by the post-cooling plug. During the first 30 s after mold close, the body wall solidifies from the outside inward, but the pinch-off weld remains hotter than the adjacent wall because it is thicker and because it is subjected to compressive heat from the flash. This thermal gradient is reversed during post-cooling, when the top lip is actively cooled by the plug while the pinch-off area may only be cooled by the bottom plate if the wear ring contact is sufficient. If the bottom plate does not contact the pinch-off directly, the pinch-off can leave the post-cooling fixture at 90 °C to 100 °C while the adjacent body wall is below 60 °C. The resulting residual stress concentration at the weld line can reduce impact strength and environmental stress crack resistance, particularly when the drum is later exposed to aggressive liquid cargoes. ASTM D1693-15 ESCR testing of samples cut from the pinch-off weld versus the sidewall may show failure times that differ by an order of magnitude. For UN 1H2 open-head drums, the design-type approval includes drop tests and stacking tests, but the long-term chemical compatibility of the pinch-off zone with actual filling goods is not fully captured by short-term hydrostatic tests. A post-cooling fixture that extends a water-cooled contact pad into the pinch-off recess can reduce this thermal asymmetry, but the pad must be designed to avoid trapping flash fragments and to allow ejection without damaging the bottom chime. Published data for this specific configuration is limited; the industrial practice is to use mold temperature and fixture temperature as indirect indicators of weld-line residual stress.
| Standard/code | Test | Cooling-sensitive response | Typical production acceptance band |
|---|---|---|---|
| UN 1H2 | Open-head design-type qualification | Top lip ovality, bottom chime flatness, drop/stack integrity | No leakage, no rupture; dimensional tolerances per design type |
| ASTM D1693-15 | Environmental stress crack resistance | Residual stress from pinch-off and chime cooling gradients | Specified by material grade; often 500 h to 1,000 h for notched HDPE drum grades under Condition B |
| ISO 1133-1:2022 | Melt mass-flow rate | Parison drawdown, wall thickness distribution, cooling time | 0.2 g/10 min to 0.5 g/10 min at 190 °C/2.16 kg for typical drum HDPE |
| ASTM D2659-16 | Column crush properties | Buckling resistance as affected by non-uniform wall and residual stress | Peak load and deflection per design type; no fractured sidewall before specified deflection |
| ISO 12048:1994 | Package compression testing | Top-load stability under warehouse stacking | Determined by design-type approval; product-specific stacking factor and load duration |
| ASTM D638-14 | Tensile properties | Weld-line and pinch-off strength relative to sidewall | Weld factor 0.7 to 0.9 of sidewall yield strength for optimized cooling |
Blow air pressure, air temperature, and internal cooling duration are coupled to the post-cooling fixture because they determine the initial internal surface temperature and the residual air temperature at mold opening. Typical blow air pressure for 220 L drums ranges from 0.7 MPa to 1.0 MPa, with a first-stage pre-blow at 0.2 MPa to 0.4 MPa to distribute the parison before full inflation. The internal gas temperature can be reduced by using a chilled-air or nitrogen supply at 5 °C to 15 °C instead of plant compressed air at 25 °C to 30 °C. A lower internal gas temperature increases the heat flux from the inner surface and can reduce the post-cooling dwell by 10 s to 25 s for a 10 kg drum, provided that the mold is not below the dew point and that condensate is removed by a desiccant dryer. Internal cooling is particularly effective for the body wall but less effective for the thick top and bottom chimes, which are shielded by the mold pinch-off and the flash pocket. If post-cooling fixture contact is the limiting resistance, lowering the blow air temperature may simply produce a colder body while the chime remains hot, increasing the thermal gradient and the risk of lip ovality. Published data for this specific configuration is limited.
In production-scale process control, top-lip diameter, bottom wear-ring diameter, shot weight, and post-cooling fixture outlet temperature serve as the primary variables for cooling uniformity. Laser scanning systems measure top lip diameter at 8 to 16 angular positions and feed the data into a statistical process control routine, with a control limit of ±1.5 mm to ±2.0 mm on lip diameter and ±2.0 mm to ±3.0 mm on bottom chime diameter depending on the UN design type. These dimensions are not independent of cooling; a top lip that is measured in-spec at the fixture but released at a high internal temperature can shrink asymmetrically and fall outside the tolerance after 24 h of stabilization at 23 °C and 50% relative humidity. The use of a post-cooling dwell that is too short therefore displaces the dimensional adjustment to the warehouse, where pallet loads can imprint ovality on drums that had passed the in-line check. Production facilities may add a conditioning period of 24 h before final dimensional approval, but continuous lines often rely on a temperature-compensated acceptance model that correlates fixture outlet temperature with subsequent ambient shrinkage. The open-head cover seating surface is also tested with a plug gauge or a coordinate measuring machine trace, and the roundness tolerance is typically 1.0 mm to 1.5 mm on the seating diameter. Published data for this specific configuration is limited, but the mechanical failure data generated by cover seal leak tests indicates that top lip ovality is the dominant cooling-related defect in open-head drums.