High-Load POM: Selecting by Creep Limit, Not by Tensile Strength begins from an operational contradiction. A datasheet tensile yield of 60–70 MPa at 23°C for unfilled polyoxymethylene homopolymer, or 55–65 MPa for many unfilled copolymers, is generated under ISO 527-2:2012 at a relatively high test speed, typically 50 mm/min for yield determination. A gear tooth, valve actuator shaft, snap-fit arm, or pump rotor does not operate at that strain rate. It operates under continuous or intermittently sustained load for 5,000 h, 10,000 h, or longer. Polyoxymethylene is a semicrystalline acetal resin with crystallinity commonly between 60% and 80%. The crystalline phase provides short-term stiffness and yield strength, but the amorphous phase retains time-dependent chain mobility. Under continuous load, that mobility appears as tensile creep, and the strain accumulates even when the applied stress is far below the published yield stress. Material selection based on a tensile strength divided by a safety factor therefore misses the governing failure mode: creep-induced dimensional change, loss of interference fit, and tooth flank shift.

The creep limit is not a fixed mechanical property. It is the tensile stress that produces a specified total strain after a specified time at a specified temperature. Creep limit is read from isochronous stress-strain curves constructed from tensile creep tests, not from a short-term stress-strain curve. Where tensile yield is measured under ISO 527-2:2012, tensile creep is measured under ISO 899-1:2017 or ASTM D2990-17. In the creep test, a constant load is applied to a specimen at controlled temperature, and strain is recorded at logarithmic time intervals. The resulting isochronous curves for 1 h, 10 h, 100 h, 1,000 h, and 10,000 h allow the designer to select a stress that limits total strain to an acceptable value, often 0.5% to 2.0%. For a high-load POM part, that value is the design criterion. The tensile yield remains a lot-to-lot quality check, not a service load limit.

Why Does a 65 MPa Yield Stress Permit Failure Below 25 MPa Continuous Load?

The apparent contradiction is explained by the difference between a constant strain-rate test and a constant stress test. In the ISO 527-2:2012 tensile test, the specimen is pulled to failure in minutes. The measured yield stress is therefore dominated by the crystalline phase and by fast chain-segment orientation. In a creep test under ISO 899-1:2017, the same grade may show a creep strain of 1.0% or more at a stress of 18–25 MPa after 10,000 h at 60°C. The stress is below yield by a factor of roughly 2.5 to 3.5, yet the part is already outside a typical draw-free or functional-clearance limit. The failure mode is not brittle fracture or ductile rupture; it is progressive strain accumulation in the amorphous phase, followed by stress redistribution, relaxation, and eventual loss of load-bearing geometry.

Creep modulus is more useful than tensile modulus for such applications. Creep modulus is defined as the applied stress divided by the time-dependent strain. At 23°C and 10 MPa, an unfilled POM copolymer may have a creep modulus at 1,000 h of approximately 1,000–1,400 MPa. The same resin can show a short-term tensile modulus of 2,600–3,000 MPa. The reduction is large. At 60°C, creep modulus falls further because chain mobility increases near the glass transition of the amorphous phase and because the crystalline lamellae begin to permit slip. A design equation based on short-term modulus therefore overestimates real stiffness. High-load POM selection should be based on the isochronous stress-strain curve at the service temperature and the required service life. The allowed service stress is then read directly from the curve at the maximum permitted strain. This is the creep limit, and it replaces tensile yield in the design calculation.

An injection moulded gear carrier for an electric window drive illustrates the difference. The component is produced on a 1,500 kN hydraulic clamping force machine with a 40 mm three-zone screw and a hot-runner valve gate. The material is a medium-viscosity unfilled POM copolymer with a melt flow rate of 9 g/10 min at 190°C under 2.16 kg, measured to ISO 1133-1:2022. Barrel temperatures from feed to nozzle are maintained at 170°C, 190°C, 200°C, and 205°C. The mould temperature is held at 85–95°C to promote crystallinity at the gear flank and reduce post-mould shrinkage. Melt residence time is limited to 15 min because thermal degradation of POM begins to generate formaldehyde at temperatures above 230°C. A vented barrel or controlled insert degassing is used where moisture exposure before moulding exceeds 0.20% surface moisture. Pre-drying at 80°C for 3–4 h is applied only when surface condensation is observed; POM absorbs low moisture, but surface water causes splay and is not removed by a general-purpose screw without vacuum venting.

In the gear carrier, gate freeze is confirmed by cavity pressure decay. The gate pressure during packing is held at 70–85 MPa, and gate-seal time is kept between 6 s and 8 s. Overpacking is avoided because excess packing pressure raises frozen-in stress near the gate and can reduce creep resistance later. The service load at the tooth root is estimated at 18–22 MPa peak at 60°C during motor stall. The material certificate may list tensile yield at 61 MPa. A short-term calculation suggests an apparent safety factor above 2.5. However, tensile creep data generated to ISO 899-1:2017 show that the 1% creep limit at 10,000 h and 60°C can be below 15 MPa for unfilled copolymer. The gear carrier therefore operates above its creep limit despite being far below yield. The result is an increase in backlash and a shift in center distance, not a sudden tooth fracture. This is why the acceptance criterion for the part is a maximum creep strain under load, not a minimum tensile yield.

Test Methods for Creep-Limited POM Selection

Selection requires test data that captures the time-temperature response of the specific grade, not a generic datasheet value. The primary method is tensile creep under ISO 899-1:2017. The test applies a constant tensile stress and records strain over time. From the recorded data, the laboratory constructs isochronous stress-strain curves for the required design time and temperature. The creep limit is then defined as the stress at which the curve intersects the permitted strain limit. For many high-load mechanical parts in POM, a total strain limit of 1.0% after 10,000 h is used when gear backlash, press-fit interference, or actuator stroke is critical. ASTM D2990-17 provides a comparable framework for tensile, compressive, and flexural creep and creep-rupture. The standard is frequently used in North American specifications where tensile creep is not the only load mode.

Extrapolation of creep data beyond the measured time should be avoided unless the long-term creep curve is linear on a logarithmic time scale and the semi-crystalline morphology is stable. POM can exhibit delayed physical ageing and crystallisation under load, so accelerated elevated-temperature data are not a direct substitute for long-term data at the service temperature. If a supplier provides only 100 h creep data, the designer may use the 100 h isochronous curve for a preliminary sort but cannot claim a validated 10,000 h creep limit. The design validation should include a specimen conditioned at the service temperature, loaded at the highest service stress, and measured for total strain at the required life. The acceptance limit is the specified strain, not the fraction of tensile yield.

A glass-fiber-reinforced POM pump impeller in a hot-water circulation loop shows the material-selection trade-off. The impeller experiences a hoop stress of 6–10 MPa at 80°C. An unfilled POM grade would be near or above its long-term creep limit in this condition. A 25% glass-fiber-reinforced grade raises short-term tensile yield to roughly 90–110 MPa, but the more important effect is the reduction in creep strain. Glass fibers constrain the amorphous phase and increase the creep modulus at 1,000 h by approximately 1.5–2.0 times relative to unfilled resin at the same stress. The impeller retains its outer diameter and volute clearance over time. However, the grade change carries processing and design penalties. The filled compound requires a wear-resistant screw and barrel because glass fiber is abrasive. Back pressure is limited to 0.5–1.0 MPa hydraulic during plastication to avoid excessive fiber fracture. The gate must be larger than for unfilled POM to prevent premature freeze before packing. The notched impact strength of the glass-filled grade is lower than that of unfilled POM, and the filled material can abrade an unfilled pump housing. Published creep data for glass-filled POM in 80°C circulating water are more limited than dry-air data; validation under continuous circulation is required before release.

The final acceptance of a high-load latch plate for an industrial valve is based on a creep-limit test. The plate is moulded from a POM homopolymer with a tensile yield of 68 MPa at 23°C. The calculated static service stress is 22 MPa. A short-term stress check gives an apparent safety factor of 3.1. The creep-limit curve at 10,000 h may intersect the 1.0% strain limit at 16–20 MPa depending on grade and internal stress. The part therefore has little or no long-term margin. The tensile strength value from the material certificate is recorded for traceability, but it is not the acceptance number. The release criterion is a creep strain below 1.0% after 10,000 h at the maximum service temperature under a stress of 22 MPa. Only that measured result places the component inside its allowable dimensional and functional window.