Polyolefin elastomer (POE) grades used in footwear midsoles are predominantly ethylene-1-octene copolymers synthesized via single-site or constrained-geometry catalysis. The narrow molecular weight distribution and low residual crystallinity distinguish POE from conventional EPDM or ethylene-propylene rubber in microcellular applications. Commercial foam grades are supplied with densities of 0.857 g/cm³ to 0.885 g/cm³ per ISO 1183-1:2019, melt indices of 0.5 g/10 min to 30 g/10 min at 190 °C/2.16 kg per ISO 1133-1:2022, and peak melting endotherms between 55 °C and 95 °C per ISO 11357-3:2022. In midsole formulations, POE is compounded with ethylene-vinyl acetate, azodicarbonamide, dicumyl peroxide, zinc oxide, and fillers to produce foamed densities from 0.15 g/cm³ to 0.30 g/cm³. The primary manufacturing conflict is the simultaneous control of crosslink density, gas solubility, and cell wall stability within a temperature window that can be as narrow as ±5 °C in peroxide-cured systems.

Why Does POE Content Reshape the Foaming Window in Supercritical Fluid Molding?

In supercritical N₂ or CO₂ injection foam processes, POE lowers compound viscosity and increases gas solubility relative to high-VAc EVA. Capillary and plate-plate rheometry according to ISO 6721-10:2015 indicates that replacing 20–35 wt% of EVA with POE reduces complex viscosity at 120 °C and 100 rad/s by 15–35%. This viscosity reduction permits lower melt temperatures of 105–135 °C but also accelerates pressure-driven cell coalescence when gas dosage exceeds 0.6 wt%. Cell nucleation density is governed by the pressure drop rate across the mold gate. Published industrial processing data for POE-modified microcellular polyolefins show stable cell densities of 1×10⁸–1×10⁹ cells/cm³ only when pressure drop rate is maintained above 1×10⁷ Pa/s; below that threshold, cell size increases beyond 150 µm and split-tear strength falls below the 5 N/mm target in ASTM D624-00(2020) die C.

The solubility of supercritical N₂ in POE is slightly higher than in semicrystalline EVA, but the gas is less effective at plasticizing high-octene POE. This creates a processing asymmetry: POE-rich compounds require 0.2–0.4 wt% lower gas dosage but tolerate 5–10 °C lower melt temperature. Conversely, high-VAc EVA compounds need higher dosage to achieve equivalent cell density but are less prone to cell coalescence because of their higher melt viscosity. Closed-cell content should remain above 90% by ASTM D6226-15 for acceptable rebound energy return.

Melt Rheology and Pressure Drop Requirements

Melt strength in POE/EVA foam compounds is not a single-point property but a function of extensional rheology and strain hardening. Oscillatory shear data at 1 Hz and 120 °C do not fully predict sagging or cell wall failure during foaming. Extensional viscosity measurements using a filament stretching rheometer show that POE-rich compounds with 30 wt% POE exhibit lower Hencky strain at break than EVA-dominant compounds, especially at strain rates above 1 s⁻¹. To compensate, processors increase gas dosage or reduce mold temperature to raise the elongational viscosity at the bubble wall. Pressure drop rate across the gate must be in the range of 1×10⁷–1×10⁹ Pa/s; below this range cell density is insufficient, while above it cell rupture and foam anisotropy become severe. The actual gate geometry, typically 0.8–1.5 mm thickness, determines whether a given injection speed produces the necessary pressure drop. Mold trials on 200 t clamp machines show that a nozzle shut-off valve delay of 0.1–0.5 s is required to retain dissolved gas before injection; delays longer than 0.5 s allow gas to accumulate in the check ring and produce short shots.

What Cell Size Distributions Are Obtained With Activated Azodicarbonamide Versus Supercritical N₂?

Activated azodicarbonamide in compression molding typically produces broader cell size distributions than supercritical N₂ injection foaming. Scanning electron microscopy on midsole cross-sections shows that chemical foaming yields mean cell diameters of 80–150 µm with coefficient of variation between 25% and 40%, while supercritical N₂ can produce mean diameters of 20–80 µm with coefficient of variation below 20%. Azodicarbonamide decomposition releases approximately 220 mL/g of gas at STP. The activated decomposition onset can be shifted to 150–175 °C with zinc oxide and urea derivatives. In comparison, supercritical N₂ foaming does not generate chemical residues, but the narrow cell distribution is achieved only when gas saturation pressure is held at 12–20 MPa and the pressure drop occurs across a restricted gate. Published data for direct comparison of identical POE/EVA formulations under chemical and physical foaming is limited; industrial mold trials indicate that supercritical N₂ requires tighter cavity temperature control because nucleation is less forgiving of heat-transfer variability.

For compression-molded midsoles, POE is typically introduced at 20–40 wt% of total polymer. Azodicarbonamide is activated with zinc oxide and urea derivatives to decompose between 150 °C and 190 °C, while dicumyl peroxide at 0.6–1.2 phr generates crosslinks. The mold cavity is heated to 160–175 °C under hydraulic pressure of 15–25 MPa for 8–12 min. The required crosslinking degree, measured as torque increase in a rotorless curemeter according to ASTM D5289-19a, is typically MH−ML ≥ 10 dN·m for dimensionally stable foams. If crosslinking occurs before gas evolution, expansion is restricted and density remains above 0.25 g/cm³; if blowing agent decomposition precedes sufficient crosslinking, cell walls rupture and surface collapse appears. Batch-to-batch variance in POE melt index of ±0.5 g/10 min shifts the optimum cure time by ±30 s, making fixed-cycle compression molding the most sensitive operation in midsole manufacturing.

Trials on multi-cavity compression presses with cavity-to-cavity temperature spread above 3 °C produce visible density streaking from inconsistent cell nucleation. The same formulation can produce densities of 0.18 g/cm³ in the center cavity and 0.24 g/cm³ in the edge cavity. This is not a mixing failure but a heat-transfer problem: edge cavities lose heat through the mold frame and require cartridge heater trim or insulated chase blocks. Phenolic antioxidants at 0.1–0.3 phr are preferred over amine-based antioxidants in these POE/EVA peroxide-cured systems; amine additives scavenge free radicals and can increase t_c90 by 30–60 s.

If Peroxide Cure Kinetics Overlap With Azodicarbonamide Decomposition

The processing window becomes critical when dicumyl peroxide half-life at 170 °C is approximately 35–45 s and azodicarbonamide maximum gas evolution occurs within the same interval. Differential scanning calorimetry and cure rheology show that a ±5 °C mold temperature deviation shifts the ratio of crosslink density to gas-generation rate enough to alter foam density by 0.02–0.04 g/cm³ and Asker C hardness by 3–5 points. On production lines with multi-cavity tools, cavity-to-cavity temperature differences above 3 °C produce visible density streaking from inconsistent cell nucleation. The use of POE with higher octene content increases molecular mobility but reduces plateau modulus, so crosslink density must be raised by 0.2–0.4 phr dicumyl peroxide to maintain equivalent dimensional stability.

When the compound is heated rapidly, the measured onset of azodicarbonamide decomposition shifts upward, but cure torque development also accelerates. The practical method is to run a coupled rheometer-DSC analysis and select a mold temperature where 90% torque development occurs no more than 15 s before gas release peaks. Published data for POE-rich configurations above 50 wt% POE with this specific activator package is limited; high-POE compounds often exhibit lower gas pressure retention and require increased crosslink density to prevent open-cell collapse.

Hardness, Rebound, Compression Set, and Density Mapping

Midsole performance specifications are segmented by use category because the mechanical demands of running, court, and casual footwear differ. Hardness is measured with an Asker C spring durometer per JIS K7312-1996, rebound with a falling ball tester per ISO 8307:2018, and compression set per ISO 1856:2018 method A at 50% compression for 22 h at 70 °C. Density is determined by ISO 845:2009 on die-cut samples 25 mm × 25 mm × 10 mm. Tensile strength and elongation follow ISO 1798:2008, while tear resistance uses ASTM D624-00(2020) die C. Abrasion loss is measured with a rotating drum per DIN ISO 4649:2020.

The density-hardness-rebound relationship is nonlinear. Reducing density from 0.25 g/cm³ to 0.18 g/cm³ in a POE/EVA midsole lowers Asker C hardness by approximately 8–12 points and can reduce rebound by 5–10% if cell walls become excessively thin. For running midsoles, the specification conflict is between low density for energy return and sufficient cell wall strength to survive repeated compression. Compounds with POE content above 35 wt% often fail the ≤25% compression set target unless crosslink density is increased; the resulting hardness increase then violates the 45–55 Asker C limit. This is the primary formulation boundary for POE in running midsoles.

On twin-screw compounding lines with L/D 32–40, POE pellets are fed downstream after EVA melting to limit shear heating. Screw speed is held at 300–500 rpm for dispersive mixing, but melt temperature at the die must remain below 130 °C to prevent premature azodicarbonamide decomposition when the blowing agent masterbatch is included. Strand pelletizing of POE-rich compounds above 45 wt% POE shows die-face smearing unless underwater pelletizing water temperature is kept below 15 °C. If production shifts to supercritical injection molding, the compounded pellets require 2 h drying at 60 °C when ambient relative humidity exceeds 60%, because moisture accelerates hydrolysis of vinyl acetate sequences and increases surface blisters.

Zinc stearate, a common release agent and blowing activator, migrates to the skin of POE/EVA midsoles during storage at 40 °C for 7 days. Surface concentration increases to levels that reduce ink adhesion and require plasma or corona treatment before printing. This migration is more rapid in POE-rich compounds because the POE phase has lower crystallinity and provides less tortuosity for additive diffusion.

Injection Molding Equipment Exhibits Shot-to-Shot Density Variation Below 0.03 g/cm³ Only When Gas Dosing Is Closed-Loop

Open-loop gas dosing results in shot mass variation of 1.5–3.0% and foam density variation above 0.05 g/cm³ across 500-cycle runs. Closed-loop mass flow control with no-loss weighers reduces gas delivery deviation to ±0.5%, stabilizing cell size at 20–80 µm measured by scanning electron microscopy. Screw recovery time must be held within ±10% because extended residence time above 120 °C degrades POE chain segments and lowers melt strength. On 180–250 t clamp force injection molding machines, cushion position is maintained at 8–12 mm to prevent gas from accumulating in the check ring; gas accumulation causes short shots and density spikes in the heel area.

Mold trials on running midsole tools with 12–16 cavities show that balancing hot runner temperature to ±2 °C is necessary to keep cavity-to-cavity density within 0.02 g/cm³. If the hot runner manifold is poorly insulated, edge cavities solidify before full foaming expansion and produce hardness shifts of 5–7 Asker C. In injection foam, the combination of POE and supercritical N₂ reduces melt viscosity enough that lower clamping force can be used, but the check ring must be replaced after 50,000–100,000 cycles because wear increases backflow and density inconsistency.