Menu
In a co-rotating intermeshing twin-screw extruder configured with a segmented barrel of 40:1 L/D and screw diameters between 25 mm and 77 mm, polypropylene impact modification with 20 wt% ethylene-propylene-diene monomer rubber is carried out under conditions that simultaneously disperse the elastomer phase and initiate peroxide-driven crosslinking reactions within the diene-bearing EPDM domains. The EPDM grade selected for such reactive blending normally exhibits an ethylene content of 55 wt% to 70 wt%, an ethylidene norbornene diene content of 4.5 wt% to 8.0 wt%, and a Mooney viscosity ML (1+4) at 125 °C between 20 MU and 70 MU. The objective of the 20 wt% loading is not simply to add soft domains but to shift the ductile-brittle transition of the polypropylene matrix toward lower service temperatures while limiting the loss in tensile stiffness. Peroxide addition is maintained in the range of 0.05 wt% to 0.25 wt% of the total formulation, with the exact level adjusted to compensate for filler surface area, stabilizer package, and EPDM diene content. The screw configuration typically places two or three right-handed kneading blocks before the injection point for the peroxide, followed by a restrictive left-handed element that creates a melt seal, and then a second dispersive section. Barrel temperatures are profiled from 170 °C in the feed zone to 220 °C at the die, though the actual melt temperature in the reactive zone can be 10 °C to 20 °C higher because of viscous dissipation. The process is monitored by specific mechanical energy, die pressure, melt temperature, and post-extrusion pellet melt flow rate measured according to ISO 1133-1:2022 at 230 °C and 2.16 kg. Published data for this specific configuration is limited, but production-scale experience indicates that stable operation requires tight control over feed rate and peroxide injection to avoid rapid shifts in melt viscosity and pressure fluctuations.
One constraint arises because the dialkyl peroxide must decompose after EPDM droplets have been reduced to a critical morphology but before the polypropylene phase has accumulated excessive chain scission. The selected peroxide, commonly 2,5-dimethyl-2,5-di(tert-butylperoxy)hexane, has a supplier-listed one-minute half-life temperature in the range of 175 °C to 185 °C. If the melt temperature in the first mixing zone remains below 190 °C, only a small fraction of the peroxide is converted into alkoxy radicals during the available residence time, and the EPDM remains insufficiently crosslinked. Under-crosslinked EPDM domains are less resistant to coalescence, and the extrudate can exhibit a mottled surface, low melt strength, and poor pellet integrity. Conversely, when the melt temperature exceeds 230 °C, the polypropylene molecular weight falls by β-scission, the melt flow rate increases, and the compound loses impact resistance because the matrix itself becomes a low-viscosity, lower-toughness phase. The allowable processing window is frequently no wider than 5 °C to 10 °C through the reactive mixing zone; this is the deepest process conflict in the system. Temperature control is made more difficult because peroxide decomposition and EPDM crosslinking are exothermic, while the kneading blocks generate localized shear heating. Shell-and-tube or water-circulated barrel zones with PID control are normally specified, and melt temperature probes are positioned downstream of the reactive zone to trigger feed-rate reduction when the melt temperature exceeds the upper bound. Published data for this specific configuration is limited, but compounders operating 40:1 L/D lines often record melt temperature excursions of 15 °C within a single barrel diameter when screw elements are worn.
The process conflict is further amplified by the drive system. At fixed screw speed, an increase in torque indicates thickening or crosslinking, while a decrease indicates polypropylene chain scission or upstream feed instability. A drive operating at 50% to 85% of rated torque provides enough reserve to absorb viscosity fluctuations but does not guarantee stable morphology. Die pressure fluctuations above 0.3 MPa in a 25 mm line are often traced to phase inversion or to gel particles from over-cured EPDM. The gel particles act as stress concentrators in finished parts, and their presence is detected as surface roughness or as fisheyes in film and sheet. Therefore, the practical window is defined not only by barrel set points but also by torque and die-pressure signals.
Because the peroxide is typically injected in the liquid state after the first intensive kneading block, the location of the injection port becomes a critical scale-up variable. In a 40:1 L/D barrel, injection at zone 5 of 10 zones allows EPDM to be pre-dispersed in the polypropylene melt before free radicals are introduced. If the peroxide is injected too early, the matrix may undergo chain scission before the EPDM domains have developed a sufficiently small particle size. If injected too late, crosslinking occurs only in the outer skin of the extrudate and the core remains under-cured. The downstream mixing section must generate repeated extensional and shear flow rather than simple laminar flow; forward conveying elements with 45° pitch, neutral kneading blocks at 90°, and left-handed elements are combined to increase filled length. The barrel is vented at -0.08 MPa to -0.09 MPa relative to atmospheric pressure to remove decomposition by-products and low-molecular-weight volatiles. Inadequately vented systems show porosity in pelletized strands and a density outside the range of 0.90 g/cm³ to 0.92 g/cm³. The pellet moisture target before extrusion is below 0.05 wt%; when relative humidity in the feed area exceeds 60%, pre-drying of the EPDM and polypropylene is required for at least 2 h at 80 °C. These operational boundaries are derived from production-scale equipment behavior rather than from neat laboratory mixing.
When screw speed is increased while feed rate is held constant, the residence time distribution narrows but the average residence time falls, and the specific mechanical energy rises because more shear heat is dissipated in the kneading blocks. In a 25 mm co-rotating twin-screw extruder, raising screw speed from 300 rpm to 600 rpm at a fixed throughput can reduce mean residence time by 30% to 50% while increasing specific mechanical energy by 20% to 40%. The consequence is that peroxide decomposition may remain incomplete if the melt leaves the reactive zone too quickly, while the polypropylene phase may still undergo chain scission because local melt temperature increases. This decoupling produces a compound with low crosslink density but high melt flow rate, a combination that fails in impact-critical applications. If screw speed is reduced while feed rate is increased, the filled length in the restrictive element rises, torque increases, and the residence time distribution broadens. Some EPDM domains remain in the reactor long enough to form gel particles, while others pass through with minimal deformation. The resulting morphology is bimodal, with large EPDM domains coexisting with highly crosslinked microgel. Bimodal morphology is detrimental to low-temperature notched impact behavior measured according to ASTM D256-10(2018) because the large domains initiate unstable fracture. Screw speed and feed rate therefore cannot be optimized independently; they are coupled by the required specific energy input and the peroxide half-life. The stable operating envelope on a 40:1 L/D line is often bounded by 350 rpm and 550 rpm for a 25 mm screw, with throughput adjusted to maintain a specific energy of 0.25 kWh/kg to 0.40 kWh/kg.
Torque measurements provide early fault detection in this operating window. A gradual torque decline at fixed screw speed and feed rate often indicates that the polypropylene phase is undergoing excessive chain scission, while a sharp torque increase suggests that the EPDM is over-curing or that an upstream feed restriction has developed. In either case, the corrective action is not simply to adjust barrel temperature because the local melt temperature in the kneading blocks is weakly coupled to the barrel set point. The preferred response is to adjust feed rate, peroxide injection rate, or screw speed in small steps while monitoring die pressure and pellet melt flow rate. Production-scale lines commonly use automatic alarms tied to a torque deviation of 5% to 10% from the established set point, but such alarms only detect drift after the morphology has begun to change. Real-time rheometric monitoring at the die provides earlier indication of viscosity shifts, but published data for this specific configuration is limited.
Process data from production-scale compounding lines indicate that batch-to-batch variance in the EPDM bale viscosity is a more frequent cause of excursion than peroxide concentration drift. EPDM bales that are cold from storage or that have absorbed surface moisture feed irregularly and cause momentary torque dips. Loss-in-weight feeders with an accuracy of ±0.5% of set rate are normally assigned to the polypropylene and EPDM streams, and the EPDM is often preblended with a portion of the polypropylene to create a uniform dry blend before the main feed port. The peroxide is metered by a mass-flow-controlled liquid injection pump with a check valve at the injection port to prevent melt backflow. If the check valve sticks, peroxide delivery becomes intermittent, and the pellets show alternating light and dark gel regions.
The reactive modification is controlled by the concentration of pendant diene units in the EPDM and by the radical stoichiometry at the interface between EPDM and polypropylene. At a fixed 20 wt% EPDM loading, an EPDM with 8.0 wt% ethylidene norbornene generates a denser crosslink network than an EPDM with 4.5 wt% ethylidene norbornene at the same peroxide content. The crosslink density is commonly estimated by solvent swell or by moving-die rheometer torque, and the cure rate is described by vulcanization kinetics rather than by simple peroxide decomposition. Coagents are employed to increase the efficiency of free-radical crosslinking and to reduce the formation of low-molecular-weight fragments. Trimethylolpropane trimethacrylate and triallyl cyanurate are the most frequently used in polyolefin-based reactive blending; the coagent level is generally maintained between 0.5 phr and 3.0 phr of EPDM, with the exact level tied to the EPDM diene content. The coagent functions by scavenging alkyl radicals and forming stable multifunctional bridges, but excess coagent can accelerate gel particle formation and produce a compound that cannot be pelletized. Aromatic amine antioxidants and hindered amine light stabilizers are generally avoided because amine-based systems interfere with the peroxide-derived radical pool and can produce inconsistent crosslink density; suppliers of the EPDM grade and the peroxide ordinarily provide incompatibility lists for stabilizers. The boundary between acceptable and unacceptable crosslink density is narrow: an over-cured compound blocks the die, while an under-cured compound lacks pellet integrity and shows poor impact retention. Published data for this specific configuration is limited, but production experience indicates that the coagent to peroxide ratio is a more sensitive control variable than the total peroxide addition alone.
Mechanical property test results for reactively blended 20 wt% EPDM-modified polypropylene compounds are governed by the competition between elastomer domain content, crosslink density, and matrix molecular weight. Tensile yield stress and flexural modulus decrease as the EPDM content increases, while notched impact strength improves only when the EPDM domains are sufficiently small and sufficiently bound to the matrix. The property measurements are typically conducted after injection molding test bars and conditioning for 40 h at 23 °C and 50% relative humidity. Tensile tests according to ASTM D638-14 are run at 50 mm/min on Type I specimens; flexural tests according to ISO 178:2019 are run at 2 mm/min; and notched Izod tests according to ASTM D256-10(2018) are run on specimens of 63.5 mm × 12.7 mm × 3.2 mm. The notched Izod values for unmodified polypropylene are typically below 5 kJ/m² at 23 °C, while reactively modified compounds with 20 wt% EPDM can exceed 20 kJ/m² or 30 kJ/m² depending on the peroxide and coagent balance, but published data for this specific configuration is limited.
| Measurement | Test method | Specimen condition |
|---|---|---|
| Notched Izod impact | ASTM D256-10(2018) | 63.5 mm × 12.7 mm × 3.2 mm, 23 °C, notch radius 0.25 mm |
| Charpy notched impact | ISO 179-1:2010 | 80 mm × 10 mm × 4 mm, 23 °C |
| Tensile yield stress | ASTM D638-14 | Type I, 50 mm/min |
| Tensile modulus | ISO 527-2:2012 | 1A, 1 mm/min |
| Flexural modulus | ISO 178:2019 | 80 mm × 10 mm × 4 mm, 2 mm/min |
| Melt mass-flow rate | ISO 1133-1:2022 | 230 °C, 2.16 kg |
| Density | ISO 1183-1:2019 | 23 °C, immersed in water |
| Hardness Shore D | ASTM D2240-15 | 15 s delay, 3 mm plaque |
The exact values of tensile yield stress and flexural modulus depend on the EPDM viscosity ratio relative to the polypropylene matrix and on the degree of interfacial grafting. When the EPDM is lightly crosslinked, the dispersed domains deform and coalesce during injection molding, producing anisotropic shrinkage and weld-line weakness. When crosslink density is high, the domains retain their shape during molding but can act as stress concentrations if the domain size exceeds 2 µm. Transmission electron microscopy of stained thin sections typically shows a dispersed EPDM domain size in the 0.5 µm to 2.0 µm range for optimized impact modification. Dynamic mechanical analysis shows a secondary loss peak near the EPDM glass transition, usually between -50 °C and -40 °C, and the intensity of that peak correlates with the EPDM content. The melt mass-flow rate of the final compound according to ISO 1133-1:2022 at 230 °C and 2.16 kg is used as a lot-release criterion because it is sensitive to both polypropylene chain scission and EPDM crosslinking. A lot with a melt flow rate outside the agreed range is rejected before molding because it indicates that the reactive balance shifted during compounding.
The final properties are not governed by a single structural variable; particle size and crosslink density interact. A low crosslink density permits EPDM droplets to coalesce into larger domains during melt processing, while a very high crosslink density can prevent the domains from cavitating under impact and thus reduces toughness. The optimum crosslink density for impact modification is generally lower than that required for compression-set resistance in a fully vulcanized thermoplastic vulcanizate. Solvent extraction or gel content measurement can be used to estimate the extent of crosslinking, but gel content alone does not reveal whether the crosslinked phase is located in the desired submicron domains or in agglomerates. Imaging by scanning electron microscopy after cryofracture and solvent etching shows that the best impact retention occurs when the EPDM phase is present as discrete domains with a number-average diameter of 0.5 µm to 1.5 µm and a narrow distribution. When the domain size distribution is bimodal, the notched Izod energy at -30 °C drops because the largest domains initiate cracks before the smaller domains can stabilize propagation. Crosslink density influences particle formation by changing the viscosity ratio between the EPDM phase and the polypropylene matrix; as crosslinking proceeds, the dispersed phase viscosity increases and the droplets become less deformable. If crosslinking occurs too early in the mixing section, the EPDM domains cannot be reduced to the required size, and the final morphology contains large gel particles. The relationship between gel content and impact strength is therefore non-monotonic. Published data for this specific configuration is limited, but morphological studies of similar polymer systems support the conclusion that the process must first achieve the correct domain size and only then freeze that morphology by controlled crosslinking.
When the compounded pellets are fed to a downstream injection-molding cell with a clamp force between 1,200 kN and 3,000 kN, the melt temperature is normally limited to 210 °C to 230 °C to avoid further degradation of the polypropylene matrix and to preserve the EPDM crosslink structure. The injection speed is set high enough to fill thin-wall sections before the melt solidifies, but excessive shear in the gate can cause the crosslinked EPDM domains to separate from the matrix. Weld lines in impact-modified parts are weaker than in the bulk because the elastomer domains align perpendicular to the flow front, and this effect is magnified when the EPDM content reaches 20 wt%. Mold shrinkage compared with unmodified polypropylene is lower and more anisotropic, with typical shrinkage values in the 1.0% to 1.5% range, but the exact value depends on the EPDM content, crosslink density, and glass-fiber content if present. The compound is dried to a pellet moisture content below 0.05 wt% before molding when the storage environment exceeds 60% relative humidity; otherwise surface silver streaks and internal porosity appear. Amine-containing colorants and amine-based stabilizer masterbatches are avoided because they alter the residual peroxide chemistry and can cause lot-to-lot impact variability. These restrictions define the downstream processing boundary for the 20 wt% EPDM reactively blended material.
