When POM Gears Wear Out Early: Friction, Heat and the Limits of Self-Lubrication

Polyoxymethylene (POM) homopolymer and copolymer gear blanks are typically injection molded with melt flow indices of 2.5–9.0 g/10 min at 190°C, 2.16 kg according to ISO 1133-1:2022. The material is selected for small power-transmission stages because dry sliding against hardened steel can exhibit a steady-state coefficient of friction of 0.25–0.35 under ASTM D1894 conditions. That observation, however, is not an unconditioned material property. The low friction requires a stable transfer film and a narrow envelope of counterface roughness, contact pressure, sliding speed, and heat removal. When a gear pair operates outside that envelope, the apparent self-lubrication disappears and the flank enters a wear-out regime before the design projection.

What Changes a Wear-Resistant POM Grade into a Thermally Softened Failure?

The transition from stable sliding to early tooth loss is not governed by bending stress alone. In VDI 2736 design practice for thermoplastic gears, flank wear and thermal softening limits are evaluated separately from tooth-root stress. Unfilled POM gears tested against acetal worm wheels show abrasive scoring when counterface roughness exceeds 0.4 µm Ra; below 0.05 µm Ra, adhesion and stick-slip can raise local heat flux. The accepted working window for dry POM-steel contacts is approximately 0.10–0.30 µm Ra, measured according to ISO 4287. Within that band, the transfer film remains thin enough to limit abrasion but thick enough to prevent metal-to-polymer adhesion. Pin-on-disc data collected under ASTM D3702 are often generated at a counterface roughness of 0.13 µm Ra and 23°C, but a molded gear flank is not a flat coupon. The real contact at the pitch line is a line contact with cyclic stress, sliding reversal, and heat accumulation. The limiting pressure-velocity value for unfilled POM against steel is commonly placed near 0.10–0.15 MPa·m/s for continuous dry operation. Above this value, surface deformation, transferred debris, and thermal expansion combine to accelerate wear.

In injection molded gear blanks, POM is semicrystalline; homopolymer crystallinity typically reaches 55–65% as measured by ISO 11357-3. A mold temperature below 80°C quenches the surface and produces smaller spherulites, but the skin can be less wear-resistant than a slowly crystallized layer. At mold temperatures of 80–100°C for copolymer, the crystallized layer is more stable and less likely to pull away during early wear. Higher mold temperatures, however, increase cycle time and can create warpage in thin webs. The limiting factor is not the core crystallinity but the surface morphology at the tooth flank. A high-gloss mold surface does not guarantee low wear; the polymer must also replicate the tool roughness. For dry running, a cavity polished to 0.10–0.30 µm Ra is often more effective than a super-mirror finish below 0.05 µm Ra, because the polymer surface replicates the roughness and needs a controlled texture to retain transfer film. This interaction is specified in tooling standards and supplier design guides; the final tribological behavior must be confirmed on the molded gear rather than on a polished plaque.

A gear cavity filled too slowly against a mold temperature below 80°C can create a highly oriented, brittle skin that spalls before a stable film forms. On a 40:1 L/D co-rotating twin-screw extrusion line compounding unfilled POM, barrel set temperatures are typically held at 190–210°C; screw speeds above 300 rpm can add viscous dissipation and raise melt temperature by more than 5°C. During injection molding, melt stock temperature above 220°C for homopolymer or 230°C for stabilized copolymer begins to release formaldehyde by chain unzipping. The resulting gas defects on the tooth flank appear as silver streaks or microporosity. A part that passes a short-term tensile check under ISO 527-2 can still fail tribologically because the degraded surface layer is removed during early wear. Supplier processing guidelines for POM therefore set maximum melt residence times; at 230°C, residence time beyond 30 min is generally excluded for unprotected grades, while actual hot-runner cycles should remain far shorter. This processing boundary is tighter than the general instruction to avoid high temperature suggests, and it becomes a critical threshold when the same hot runner feeds an eight-cavity mold with different gate freeze times.

Barrel temperature profile, screw back pressure, and cushion size modify the melt history. A cushion below 3 mm on a 25 mm screw can introduce unmelted or partially melted granules into the cavity, producing low-strength weld lines at the gear hub. In multi-cavity runner systems, gate freeze-off time must exceed packing time; otherwise the tooth flank sinks and contact ratio decreases. For POM homopolymer, holding pressure is typically maintained until gate freeze, with pack times of 2–5 s per 1 mm wall thickness. These parameters are set on the molding machine and verified by short-shots and pressure-drop studies. A process shift of 10°C in the rear barrel zone can alter the melt viscosity enough to change the fill pattern and leave visible flow lines on the tooth flank. Those flow lines are not merely cosmetic; they can act as initiation sites for abrasive wear.

When the PV Limit Is Crossed by a Small Margin in Dry Gear Stages

The consequence of crossing the PV limit is nonlinear. POM has a thermal conductivity near 0.31 W/m·K for unfilled homopolymer and heat capacity in the range 1.46–1.50 J/g·K. Frictional heat is therefore concentrated in a shallow flank layer. A 10% increase in sliding speed over a gear tooth does not simply raise thermal input by 10%; it raises the flash temperature because the thermal penetration depth per contact pass decreases. In instrumented single-tooth rigs, a pitch-line temperature shift from 60°C to 85°C is sufficient to reduce the dynamic coefficient of friction but also to soften the surface. The reduced stiffness allows the contact width to increase, raising the real contact area and generating more heat in a feedback loop. At flank temperatures above 90°C, the copolymer still retains geometry, but the surface layer loses load-carrying capacity and the wear rate can increase by an order of magnitude. VDI 2736 addresses this behavior through material load factors that decline as pitch-line velocity increases. The exact transition depends on tooth flank geometry, backlash, lubrication, and counterpart material; published data for grease-lubricated plastic gears under intermittent motion are limited.

The transient heat problem can be understood through the thermal penetration depth. For unfilled POM, thermal diffusivity is approximately 0.15 mm²/s at 20°C. At a contact-pass duration of 1 ms, the thermal penetration depth is on the order of 20–30 µm. Frictional heat is therefore trapped in the outermost 20–30 µm of the tooth flank. A load increase raises the flash temperature before the bulk gear temperature changes. This is why thermocouples inserted 2 mm below the flank may show only a 5°C rise while the actual surface reaches 85°C or more. The thermal softening of the outer layer is not captured by bulk heat deflection temperature measurements under ISO 75-2. Equipment for this measurement includes surface thermocouples and infrared spot pyrometers with a 0.5 mm measurement field. The limitation is that the emissivity of POM changes as the transfer film forms, so the pyrometer must be calibrated against a blackened reference surface.

Self-lubrication in POM is conditional, not intrinsic. Unlike PTFE, which retains a low-shear surface over a wider temperature range, unfilled POM forms a transfer layer only when the interface operates inside the roughness, pressure, and temperature window. Once the surface temperature reaches the softening range, the film is replaced by a viscoelastic layer that increases the real contact area. The result is not just higher wear but a change in failure signature. Progressive flank wear under VDI 2736 is identified by a loss of tooth thickness that shifts the effective backlash and torque signature; this wear mode should be tracked separately from tooth-root fracture. Optical profile scanning against a master gear provides a traceable dimensional record, and a change of more than 0.05 mm in flank profile over a service interval is often treated as a condition for replacement in actuator maintenance. Published data for a specific gear stage may vary, so the threshold should be established from a run-in curve on the actual assembly.

On production-scale actuator benches, the transition from stable transfer film to early flank wear is more easily detected by debris morphology than by dimensional checks. A small POM spur gear running against a hardened steel pinion at 1,500 rpm may show a flank temperature in the 60–80°C range at moderate torque; a 25% torque increase can push the pitch-line temperature above 85°C under the same ambient conditions because of the nonlinear thermal feedback. The wear debris changes from fine powder to rolled flakes when the surface layer softens. Those rolled flakes are a field indicator that the local temperature has exceeded the continuous use range. Nanoindentation at 10 µm from the flank surface often shows a hardness loss of 20–40% relative to the core. Published gear-specific data for this exact configuration is limited, but the trend is consistent with the measured flank temperature and the thermal conductivity of the grade. The contact zone where sliding and rolling components reverse direction is particularly vulnerable. Inspection under incident-light microscopy typically shows adhesive transfer near the pitch line followed by abrasive scoring toward the tip. The core remains below the softening threshold, so the gear does not collapse immediately even though the flank is being consumed.

Moisture Uptake Changes the Wear Map at Low Sliding Speed

POM absorbs moisture slowly, with equilibrium uptake around 0.2% at 23°C and 50% RH for unfilled copolymer; water absorption after 24 h immersion under ISO 62 is typically 0.5–0.7%. That small uptake is sufficient to alter dimensions by 0.2–0.4% in a thin-walled gear web. In a low-speed, high-torque actuator, a moisture-equalized gear may exhibit lower dry friction because water acts as a boundary lubricant. However, the same moisture reduces dimensional stability and can increase backlash loss. At 85% RH, a molded gear can expand enough to alter the contact ratio by shifting the effective center distance. The wear pattern migrates from the central flank toward the addendum. In practice, parts are conditioned to 23°C and 50% RH for 48 hours before dimensional checks under ISO 291 or ASTM D618 conditioning standards. POM housings and gears in office machines can be exposed to 5–10% RH indoors during winter heating, which changes the clearance and torque signature. The tribological effect is not monotonic: at very low humidity, the transfer film becomes thicker and more stable but the coefficient can rise; at high humidity, friction drops but so does surface strength. This trade-off is a limitation of the term “self-lubricating” when applied to POM in uncontrolled environments.

When unfilled POM cannot meet the PV requirement, a common modification is the addition of 20 wt% glass fiber or a combination of internal silicone and PTFE. The glass-filled grade raises stiffness and heat deflection temperature but increases counterface abrasion. A 20 wt% glass-filled POM running against a hardened steel gear may show a wear factor that is higher than unfilled POM unless the steel is carburized and polished below 0.2 µm Ra. The counterface must also be hardened to at least 58 HRC; soft stainless steel shafts, sometimes selected for corrosion resistance, can wear faster than the POM gear. Internally lubricated POM grades containing PTFE or silicone can reduce dry friction, but they do not solve the fundamental thermal limit; they shift the PV curve upward by perhaps 20–40% depending on counterface and speed. These modifications also change mechanical properties: PTFE-filled POM often shows lower tensile strength under ISO 527-2 and lower impact toughness under ISO 179-1/1eU. The selection therefore requires a trade-off between tribological and structural requirements. Published data for a specific 20 wt% glass-filled POM gear stage is limited, so the designer should obtain supplier PV curves measured on the actual counterface material rather than relying on generic plastic bearing data.

Transfer-Film Stability and Grease Interruption in Low-Backlash Actuators

A POM gear pair that runs dry and stable can fail soon after relubrication. When a lithium soap grease is introduced into a dry POM stage, the grease initially lowers friction to 0.05–0.10 but also lifts the transfer film from the flank. If the grease later dries or is ejected from the contact, the polymer surface is rougher than the original molded skin. Restart torque rises and the wear rate becomes worse than a part that was never greased. This incompatibility is documented in maintenance records of small office-automation gearboxes, where relubricating after noise leads to a shorter remaining life than continued dry operation. The design response is not to treat all POM gears as dry-running, but to select a lubricant whose base oil and additives are compatible with acetal. Greases containing solid particles or certain extreme-pressure additives are not universally suitable for POM. Compatibility should be evaluated by immersion testing under ISO 175 and by functional torque testing on the actual motion profile. A blocked-rotor torque test on a gearmotor with a torque transducer can detect the rise in friction that precedes visible flank damage. If the torque rises by more than 20% from the post-run-in baseline, the gear should be inspected for transfer-film loss rather than simply being relubricated.