Melt Pressure Control in CO2 Physical Foaming Sheet to Annular Die Conversion

Melt pressure in a continuous carbon-dioxide physical foaming line is not a single machine parameter but a distributed pressure envelope extending from the blowing-agent injection point through the static mixer, melt temperature control zone, gear pump, screen pack, adapter, and die. When a flat sheet die is removed and an annular die is installed, the terminal pressure profile changes even though the upstream extruder, gas dosing system, and melt pump remain unchanged. Carbon dioxide dissolves in the polymer melt only when the local pressure exceeds the equilibrium solubility pressure at the local melt temperature. Because the die adapter pressure in an annular configuration can be lower than the pressure in a sheet die at the same throughput, a conversion made without additional pressure compensation can place the die entry below the bubble-point pressure. The result is premature phase separation, uncontrolled nucleation in the die channels, coarse foam morphology, and melt-pressure oscillations that propagate backward toward the gear pump. The conversion therefore requires the pressure control strategy to be re-established from first principles rather than carried over from the sheet configuration.

In a typical tandem extrusion foaming line, the first-stage extruder plasticates the polymer, a high-pressure positive-displacement metering pump injects CO₂ after a melt seal, and a downstream cooling or mixing section homogenizes the gas-laden melt before a melt gear pump generates the discharge pressure. Flush-mounted pressure transducers are normally placed at the gas injection block, upstream and downstream of the melt cooler, at the gear pump discharge, and in the die adapter. The gas injection port pressure is held above the CO₂ solubility pressure to avoid gas cavity formation in the barrel. The gear pump discharge pressure provides the main process setpoint, while the die adapter pressure serves as the critical terminal measurement. In sheet extrusion, the coat-hanger or fishtail manifold and the parallel land provide a predictable pressure drop that helps maintain a high die pressure. In annular extrusion, the spiral mandrel, side-fed distributor, and annular lip land generate a different relationship between throughput, viscosity, geometry, and pressure drop. Because the annular die often offers a larger total flow circumference than the sheet width, the pressure drop at a given throughput can be lower, and the terminal pressure may fall below the safe window unless the gear pump speed, screen-pack resistance, or die gap is modified.

The melt pressure control problem is compounded by the fact that dissolved CO₂ reduces the melt viscosity, so the pressure loss through the annular die is not equal to the pressure loss of the same polymer without gas. The viscosity reduction is caused by free-volume plasticization and changes the non-Newtonian flow behavior of the melt; apparent shear thinning may be shifted to different shear-rate ranges. A pressure control loop that assumes the neat polymer pressure-drop curve will overestimate the die pressure and may fail to recognize that the melt is running too low. During conversion, the operator should record the injection pressure, gear-pump suction and discharge pressures, die-adapter pressure, melt temperature at the die, and die-head pressure ripple. These records allow the new die to be operated with a pressure boundary that preserves the single-phase solution state until the melt exits the land.

What Melt Pressure Window Prevents Premature Cell Nucleation Upstream of the Die?

For a gas-laden polymer melt, the lower bound of the operating window is the bubble-point pressure: the pressure at which dissolved CO₂ begins to form the first gas phase at local temperature. Above this pressure, the polymer–gas mixture remains a single phase; below it, bubbles nucleate in the melt, and the pressure signal becomes noisy as compressible gas pockets enter the melt pump or die. The equilibrium concentration of dissolved gas can be represented as c = K_H(T) p, where c is gas concentration in cm³(STP) g⁻¹, K_H(T) is the temperature-dependent Henry solubility coefficient in cm³(STP) g⁻¹ MPa⁻¹, and p is absolute pressure in MPa. Henry solubility coefficients for CO₂ in polyolefins and styrenic polymers generally decline with increasing temperature. Consequently, a hotter melt requires a higher pressure to retain the same CO₂ loading, and cooling the melt to the die temperature raises the solubility at a given pressure. This thermosolubility relationship is why tandem cooling-stage design and melt temperature control are inseparable from pressure control.

Above the bubble point, the pressure window is limited by mechanical and rheological constraints. Excessively high pressure at the gear-pump discharge accelerates wear on thrust bearings, increases seal leakage, and can exceed the maximum pressure rating of the transducer and die body. Industrial foam extrusion equipment may have pressure ratings of 25–40 MPa on the melt pump and 15–30 MPa on the die body, but the actual control band is usually much narrower. For CO₂-loaded melts, common operating practice is to maintain a pressure margin of 1.5–3.0 MPa above the estimated bubble-point pressure at the die entrance, with the margin adjusted upward if the transducer calibration error, pressure ripple, or melt temperature variability is high. The margin should not be interpreted as a fixed universal value because the bubble-point pressure depends on the exact CO₂ concentration, resin grade, melt temperature, and the presence of nucleating agents or other additives. Processors should measure the pressure upstream and downstream of the die adapter because the pressure at the die entry, not the pressure at the pump, determines whether phase separation occurs inside the die channels.

Pressure affects the viscosity of the gas-laden melt rather than acting only as a solubility gate. Published studies on CO₂-plasticized polypropylene, polystyrene, and low-density polyethylene have reported viscosity reductions from 20% to 60% relative to the neat polymer depending on gas loading, temperature, and shear rate. This reduction lowers the pressure drop across the die and can move the terminal pressure below the bubble point even when the pump discharge remains constant. The rheological consequence is that pressure control during annular die conversion cannot copy a pressure-versus-throughput curve generated with the unfoamed resin. The operating window must be re-derived from the gas-laden melt curve, with an explicit lower alarm set at the bubble point plus the selected margin, and an upper alarm set by the equipment mechanical limit or the threshold above which die deflection disturbs the radial gap uniformity.

From the gear pump discharge to the die land, the total pressure head is partitioned among the discharge piping, static mixer, melt cooler, screen pack, adapter, die manifold, and die land. Of these, the screen pack is usually the most practical pressure-adjustment element because changing mesh count or adding breaker-plate layers alters the pressure drop without changing the metering section. A clean screen pack composed of 40/60/80 or 40/80/100 mesh layers may generate a pressure loss of 2–8 MPa at industrial throughputs, depending on the open area, melt viscosity, and flow rate. In sheet-to-annular conversion, adding a screen layer or selecting a finer final mesh can offset the lower pressure drop of the annular die and maintain a safe die-entry pressure. The screen pack also removes contaminants, but its principal pressure function on a foaming line is to create a controlled downstream restriction. The pressure loss across the screen pack should be monitored with two transducers or a differential pressure instrument so that the operator can distinguish plugging from a true pressure increase. A plugged screen pack produces an increasing differential pressure, while a normal screen pack produces a stable differential pressure that depends on throughput and melt temperature.

The static mixer and melt cooler also contribute to the pressure profile, usually with a lower pressure drop than the screen pack. Static mixers with 6–12 elements are often installed downstream of the injection point to homogenize the CO₂ distribution before the melt enters the cooler and pump. The melt cooler, often a static or dynamic cooling device, adds pressure loss because the colder melt has a higher viscosity near the wall. The transducer configuration should include a sensor before the cooler and another after the cooler; a rising pressure drop across the cooler may indicate cooling-channel fouling, partial freezing, or reduced melt flow in some channels. In annular die operation, this pressure drop is no longer available to keep the die pressure high because it occurs before the gear pump, not after it. The gear pump is the pressure-isolating element: its discharge pressure is controlled independently, and the pressure upstream of the pump must only exceed the bubble point at that location. For CO₂ physical foaming, the gear-pump suction pressure should also be held above the local bubble point; otherwise, gas cavitation inside the pump reduces volumetric efficiency, creates pressure pulsations, and can damage the wear plates.

The adapter between the gear pump and the annular die is a critical pressure-holding zone. It is often a tapered or straight pipe section with a side entry to the annular distribution system. The pressure drop across the adapter should be measured because it may account for 1–4 MPa of the terminal pressure drop, and any abrupt diameter change can create a low-pressure region that triggers nucleation. The inner diameter of the adapter should be designed so that the melt velocity remains below the threshold at which shear heating and extensional flow promote premature foaming. If the adapter diameter is too large, the melt residence time increases and the pressure may sag in the centerline; if too small, the pressure drop rises and the melt temperature increases. During conversion, adapter geometry is usually fixed, so the pressure must be controlled upstream by the gear pump and screen pack instead of by modifying the adapter.

When Annular Gap Geometry Replaces a Flat Sheet Manifold

When an annular die replaces a flat sheet die, the terminal geometry changes from a planar coat-hanger or fishtail manifold to a radial or side-fed spiral mandrel that distributes the melt around a central mandrel before it enters an annular land. For an isothermal Newtonian fluid in a narrow slit, the pressure drop can be approximated as ΔP = 12 Q η L / (W h³), where Q is volumetric flow rate, η is viscosity, L is land length, W is slit width, and h is die gap. In an annular land, W is replaced by the mean circumference π D_m, where D_m is the mean diameter. If a 750 mm sheet die is compared with an annular die of 250 mm mean diameter, the annular mean circumference is approximately 785 mm, so the two geometries can have comparable pressure drops at the same gap and land length. However, annular foam dies frequently use a larger radial gap or shorter land length to reduce shear rate and avoid premature cell growth in the lip; those changes can reduce the die pressure drop significantly. The exact pressure-loss coefficient must be calculated for the actual spiral-channel geometry, because spiral mandrels create additional pressure drops at the entry ports, channel cross-overs, and the convergent section before the lip.

The annular die also differs from a flat sheet die in the way pressure and residence time vary across the exit circumference. In a coat-hanger sheet die, the manifold is designed to deliver a uniform transverse flow at the land entry. In a spiral mandrel annular die, the melt enters through multiple spiral grooves that gradually leak melt over the land entry region, and the pressure drop along each spiral depends on the number of starts, channel cross-section, and melt flow index. If the die pressure is too low, gas can come out of solution in the spiral channels before the melt reaches the land, creating a two-phase mixture that produces isolated bubbles, weld lines from the spider legs, and lumps on the tube surface. The annular die may also have a lower pressure drop because the radial gap can be adjusted during operation; opening the die gap to reduce the density of the foamed tube can lower the die pressure further and worsen premature nucleation. A pressure-controlled annular die conversion therefore requires keeping the die inlet pressure above the bubble point when the die gap is moved to produce the target wall thickness and foam density.

Because published pressure-drop data for CO₂-laden polymer melts in specific annular die geometries are limited, the reliable method for a line conversion is to install temporary pressure taps in the adapter and die body, or to use existing die-pressure ports, and to map the pressure profile across a range of gear-pump speeds and die-gap settings. The pressure-drop coefficient can be determined by measuring the adhesive pressure loss as a function of throughput and melt temperature, then comparing the results with the sheet-die baseline. If the annular die entry pressure is lower than the sheet-die baseline, the operator should increase the gear-pump discharge pressure, install a finer screen pack, reduce the die temperature, reduce the CO₂ loading, or partially close the die gap. Each of these actions changes the foam density, cell size, or output rate, so the pressure-control response must be coordinated with the product specification. The annular die should not be operated with a die pressure below the bubble point simply because the upstream pump discharge pressure appears adequate; the critical measurement is the pressure at the die entrance, not the pressure upstream of the restriction.

Pressure Transducer Diagnostics and Closed-Loop Control on Annular Foam Lines

Pressure transducers for CO₂ physical foaming service should be flush-mounted, corrosion-resistant, and compatible with temperatures up to at least 250 °C. The transducer full-scale range should be at least 1.25 times the maximum expected pressure, and the combined accuracy should be better than 0.5% of full scale. If the pressure range is too wide, the resolution near the bubble point may be insufficient; if the range is too narrow, an upset can exceed the transducer limit and cause a wrong signal or diaphragm damage. The primary control loop should use the gear-pump discharge pressure as the process variable, with a fast inner current or speed loop on the pump drive. A cascade loop may be used: the outer loop compares the die-adapter pressure to the lower limit, and the inner loop adjusts the gear-pump speed or extruder screw speed to keep the pump discharge pressure on target. The die-adapter pressure should be wired as a hard interlock rather than as a soft alarm only. If the die pressure falls below the configured lower limit, the control system should reduce the extruder speed, increase the gear-pump speed, or clamp the die gap, depending on which action can restore the pressure without exceeding the maximum melt temperature.

Pressure ripple in a gas-laden melt system is a diagnostic signal that often precedes visible foam defects. A stable single-phase melt produces a relatively smooth pressure trace at the die, with minor pulsations from the gear pump or screw. When the die pressure oscillates with a peak-to-peak amplitude above 0.3–0.5 MPa at a frequency of 0.5–2 Hz, the cause is often intermittent gas release in the adapter or die manifold. This can occur when the die pressure sits just above the bubble point and small pressure dips allow microscopic bubble formation. The corrective action should not be to filter the signal; the pressure ripple should be treated as a direct indicator of an insufficient pressure margin. Increasing the die pressure until the ripple amplitude falls below the threshold usually restores the single-phase condition. The pressure transducer sampling rate should be fast enough to resolve this ripple; a 10–100 ms update interval is normal for PLC-based extrusion control, but a dedicated signal conditioner or data acquisition system may be required for high-resolution diagnosis.

Closed-loop control should include the melt temperature at the die because the bubble point moves with temperature. A temperature change of 5 °C can alter the solubility enough to shift the required minimum pressure by a measurable amount. If the die temperature is lowered, the melt viscosity increases, the pressure drop rises, and the die pressure may rise naturally; the upper pressure limit must be respected because excessive die pressure can deflect the annular lip and create radial thickness non-uniformity. The die gap should be measured at multiple circumferential positions with feeler gauges or an electronic gap sensor, and the lip adjustment bolts should be moved in small increments with the die at operating temperature. ISO 1133-1:2022 provides a melt mass-flow rate determination that can be used to compare the incoming resin lot to the previous lot; a shift in melt flow rate of more than 10–15% may require a new pressure setpoint. For foamed-tube density and cell structure, ASTM D792-20 and ASTM D3576-20 provide standard test methods that can be used to quantify whether the pressure-control strategy is delivering the intended product.

A pressure-controlled annular die conversion is judged in practice by the absence of die-lip pre-foaming, the stability of the tube diameter, and the uniformity of the cell structure at the target density. Three failure modes are particularly diagnostic. First, if the die entry pressure is too low, gas separates in the spiral channels and produces a coarse, irregular cell structure with visible weld lines along the spider or mandrel support zones. Second, if the die gap is too small and the shear rate in the lip exceeds the critical shear rate for the gas-laden melt, sharkskin or melt fracture appears on the tube surface. Third, if the die pressure is too high, the die lips can be forced apart by internal pressure, increasing the gap and producing a heavy, poorly foamed tube with circumferential thickness variation. The annular die should be protected by an upper pressure limit that accounts for the die-body mechanical rating and the lip-adjustment system. The lower limit should be set to the bubble point plus the selected safety margin, and the transducer alarm should be interlocked with the gear-pump drive.

Operational boundaries are specific to the polymer, CO₂ concentration, melt temperature, and die geometry. Pre-drying of hygroscopic resins is generally required before extrusion if the resin moisture content exceeds 0.05 wt%, because water acts as a secondary volatile that can prematurely nucleate bubbles even when the CO₂ pressure is above its own bubble point. Additives that lower the melt surface tension, such as certain fluoropolymer processing aids or nucleating agents, can reduce the energy barrier for nucleation and make the die pressure margin more sensitive. Conversely, amine-based additives that could interact with acid species or other compounding ingredients should be evaluated for compatibility; they are not intrinsically required in CO₂ foaming and may create degradation products that alter the pressure signal. The operator should also recognize that the pressure readings at the die may be affected by transducer drift, diaphragm coating, or a partially solidified layer near the pressure port. Any sudden change in the die pressure that is not accompanied by a change in throughput, melt temperature, or die gap should be investigated with a portable calibration unit before altering the process.

Pressure control during annular die conversion is only one part of the terminal processing envelope. The melt temperature at the die should be selected so that the bubble-point pressure at the desired CO₂ loading is below the safe die pressure but above the pressure that would cause excessive shear heating in the lip. The CO₂ metering accuracy must be maintained because a variation of 0.2–0.5 wt% in gas loading can shift the required minimum pressure enough to move the die entry out of the safe window. Gas dosing pumps should be calibrated against a reference flow meter or mass-balance method, and the gas concentration in the melt should be inferred from the injection rate and the extruder throughput. If the line is operated with a return or recycle stream, the melt pressure at the return valve must also be held above the bubble point to avoid phase separation in the return piping. No single setpoint can guarantee stable annular foam production across all resin lots and die gaps; the pressure window must be re-estimated whenever the resin, CO₂ concentration, die gap, or melt temperature is changed.