Polyoxymethylene (POM) copolymer and homopolymer grades are specified for precision motion-transfer parts because of their crystallinity, low moisture uptake, and predictable frictional behaviour. The addition of short glass fibre at 10 wt%, 20 wt%, or 30 wt% changes the material from a ductile thermoplastic with high elongation into a stiff, strongly anisotropic composite. That shift is accompanied by a measurable reduction in notched impact strength and flatness control. For components such as helical gears, fuel system pump housings, window regulator carriers, and electrical switch brackets, the decision to specify glass-filled POM is a trade between tensile modulus gain and the reduction in multiaxial energy absorption. Compounding lines running 25 mm to 75 mm co-rotating twin-screw extruders with 40:1 L/D and side stuffing for glass fibre commonly produce pellets with mean fibre length of 150–300 µm; barrel temperatures are held at 170–200°C to avoid formaldehyde evolution. The glass sizing chemistry, fibre diameter, and pellet fibre length distribution are co-variables with the nominal glass content.

At Which Glass Loadings Do Notched and Multiaxial Impact Properties Fall Below Unfilled Copolymer Levels?

Under ISO 179-1/1eA notched Charpy testing at 23°C, unfilled POM copolymer typically exhibits 6–8 kJ/m². A 10 wt% short-glass POM may retain 5–7 kJ/m² if the glass sizing is optimised for acetal, but bulk elongation measured by ISO 527-1/2 drops from 15–30% for unfilled resin to 2.5–4.0%. At 20 wt% loading, notched Charpy frequently falls to 4–6 kJ/m²; at 30 wt%, values of 3–5 kJ/m² are typical. The transition is not linear with fibre content because fibre end density and matrix ligament thickness control crack initiation. Under ISO 6603-2 instrumented multiaxial impact, total penetration energy for 25 wt% glass-filled POM can be 30–50% lower than unfilled POM at the same plaque thickness. Fractographic examination of impacted plaques shows that cracks initiate at fibre ends, propagate along the fibre-matrix interface, and link through the thin matrix layers between adjacent fibres. This failure mode is particularly visible in unnotched falling-weight tests of 2.0 mm plaques, where the reinforcement suppresses large-scale shear yielding but does not provide a continuous fibre network capable of absorbing multiaxial energy. Unnotched Charpy testing under ISO 179-1/1eU also shifts from a ductile no-break response in many unfilled POM grades to a brittle break in glass-filled grades; the exact non-break threshold depends on fibre length and specimen thickness.

These values are typical supplier datasheet ranges, not specification minima. Sizing chemistry, fibre diameter, and compounding intensity can shift the impact response by more than 1 kJ/m². For specification purposes, the moulding must follow ISO 294-1 or ASTM D3641 and conditioning according to ISO 291 at 23°C and 50% RH before mechanical testing.

Stiffness gain is accompanied by changes in density and thermal expansion. Density measured by ISO 1183 increases from 1.41 g/cm³ for unfilled POM copolymer to 1.55–1.63 g/cm³ at 30 wt% glass; this increases part mass by 10–15% for the same cavity volume. The coefficient of linear thermal expansion measured by ISO 11359-2 falls from 110–130 ppm/K for unfilled POM to 40–60 ppm/K in the flow direction and 80–100 ppm/K transverse for 25–30 wt% glass. That thermal expansion anisotropy adds a second distortion mechanism when parts are exposed to under-hood or pump housing temperatures: thermal expansion differentials across flow boundaries can produce warpage even if the part is moulded flat.

Fibre orientation in a glass-filled POM moulding is not uniform through thickness. The skin layer, typically 100–200 µm thick, contains fibres aligned predominantly in the flow direction; the core may contain transverse or randomised orientation depending on the shear rate profile and fountain flow. Because POM crystallises rapidly in a cold tool, the oriented skin layer shrinks less in the flow direction than the unreinforced matrix or the core; the resulting in-plane shrinkage anisotropy drives out-of-plane bowing. In a 2.0 mm-thick plaque moulded with a single edge gate, a 20 wt% glass-filled POM grade typically shows flow-direction shrinkage near 0.6% and transverse shrinkage of 1.2–1.4%; an unfilled POM plaque shows 1.8–2.0% in both axes. The imbalance produces concave or convex deflection that scales with flow length, packing pressure decay, and wall thickness. This distortion is commonly misclassified as post-mould warpage, but in glass-filled acetal it is primarily orientation-shrinkage anisotropy that is locked in before ejection. Moisture absorption is not the dominant contributor below 0.4% water content at 23°C and 50% RH, unlike polyamide 66 or polybutylene terephthalate, so dimensional variation is largely thermal and orientation-driven.

When Gate Location Overrides Wall Thickness in Warp Control

Because the orientation pattern is formed by the melt front, gate placement is often more influential than nominal wall thickness in controlling out-of-plane distortion. A single centre gate on a rectangular cover generates radial flow and a more homogeneous fibre orientation distribution than an edge gate, but it may create a weld line at the last fill location. Moving the gate to a long edge produces a unidirectional flow front that aligns fibres parallel to the length; the resulting cross-flow shrinkage differential can exceed 0.8–1.0% on a 150 mm flow length and produce measurable bow after ejection. Packing pressure above 80 MPa hydraulic pressure with a holding time sufficient to seal the gate reduces the amplitude but does not eliminate anisotropy. In production, flatness specifications below 0.10 mm total indicator reading on a 100 mm diameter part require a combination of multiple gates, reduced glass content, or a redesign of the gating to balance the orientation tensor. Mould temperature also participates: increasing tool temperature from 80°C to 120°C reduces the frozen-in orientation gradient but extends cycle time and can increase oxidative surface effects if the melt temperature exceeds 230°C. Gate size is adjusted upward to 0.8–1.2 mm for a 2.0 mm wall because the more viscous glass-filled melt freezes faster than unfilled POM. Injection speed in the range 100–200 mm/s screw advance is typical; excessive speed causes jetting and turbulent flow, while very low speed increases fibre orientation in the frozen skin.

Snap-fit and press-fit designs that were proven in unfilled POM cannot be transferred directly to glass-filled grades. A cantilever snap arm designed for 2.0–3.0% outer-fibre strain in unfilled POM will fail in a 25 wt% glass-filled grade with tensile elongation at break below 2.5% under ISO 527-1/2. The flexural modulus ratio between the two materials may be 2.5–3.0×, so the same deflection requires much higher insertion force and generates higher stress at the root. In field returns from automotive window regulator carriers and furniture adjustment mechanisms, root cracking is observed when the nominal strain calculation used the unfilled POM modulus; the actual peak strain at a sharp root radius may exceed the glass-filled material’s failure strain by a factor of 1.5–2.0×. Ribs and roots should be radiused above 0.8 mm, and the snap arm thickness should be reduced to keep outer-fibre strain below 1.0% for long-term assembly reliability.

Weld-Line Strength and Boss Loading in Glass-Filled Acetal Components

Glass fibres act as flow markers around cores, bosses, and holes. At the melt front, fibres orient tangentially to the flow and create a visible knit line; in glass-filled POM, weld-line tensile strength is commonly 40–60% of the parent material, compared with 70–85% for unfilled POM. This reduction is measured under ISO 527-1/2 on double-gate injection moulded tensile bars or on plaques with insert-moulded holes. A hole placed in a load-bearing web in 20 wt% glass-filled POM can reduce static failure load by more than a notch sensitivity factor of 2.0 when the load acts transverse to the weld line. Boss pull-out strength also becomes anisotropic: screw withdrawal force parallel to fibre orientation is higher than transverse, but the surrounding material may split at lower hoop strain because the fibres provide little transverse reinforcement. For threaded metal inserts, the differential shrinkage around the insert combined with low transverse elongation can generate radial cracks during thermal cycling between -40°C and 120°C. This failure mode is frequently observed in under-hood brackets and pump housings when the insert is located near a weld line. If a glass-filled POM boss must accept a self-tapping screw, the boss outer diameter should be increased to at least 2.0–2.5× the screw nominal diameter, and the gate should be placed so that flow wraps around the core rather than splitting against it.

Gear and sliding-contact applications do not show uniform benefit from glass fibre. The improved tensile modulus raises tooth bending stiffness and reduces tooth deflection at operating temperature, but exposed fibre ends can act as aggressive abrasion points. In a worm-and-wheel pair where one member is 20 wt% glass-filled POM and the other is unfilled POM, the unreinforced member typically becomes the primary wear site. Wear testing according to ASTM G133 or ASTM D3702 is required to establish a usable PV limit for the selected pair. Published data for specific gear geometries is limited; actual wear rates depend on pitch-line velocity, contact pressure, lubrication regime, and surface roughness. If the part operates in dry or grease-lubricated intermittent motion, a 10 wt% glass fibre grade or an internally lubricated POM with lower fibre loading may provide better wear life than 30 wt% glass reinforcement. Fibre ends at the tooth flank also increase the surface roughness after initial break-in; a post-mould honing or controlled wear-in cycle is often used to reduce abrasion on the counterpart.

Fatigue response is similarly anisotropic. In uniaxial tensile fatigue at 10 Hz and R=0.1, a 20 wt% glass-filled POM may show higher endurance stress at 10⁶ cycles than unfilled POM because the higher modulus lowers cyclic strain for a given load. However, in impact fatigue or repeated drop-impact loading, the lower crack initiation energy dominates and failure can occur after fewer events. Published S-N data for glass-filled POM is limited across all grades; designers should specify fibre-characterised fatigue data generated under ASTM D7791 or an equivalent laboratory procedure before using the material in vibrating brackets or pressure-pulsating fuel system housings.

Processing Restrictions Are Set by Formaldehyde Depolymerization and Fibre Attrition

Glass-filled POM must be processed within the same thermal limits as unfilled POM, but the presence of glass fibres increases viscous dissipation and local melt temperature at the screw tip. Melt temperatures above 230–240°C accelerate formaldehyde depolymerization; excessive gas evolution, nozzle drool, and surface silver streaks develop quickly. Barrel settings for 20 wt% glass-filled POM are typically 170/180/185/190°C from rear to nozzle, with actual melt temperature verified at 200–220°C by an insertion thermocouple. Mould temperature is held at 90–110°C for dimensional stability. If the mould is too cold, rapid solidification intensifies fibre orientation and increases warpage; if too hot, cycle time increases and plate-out may form on the cavity surface. Drying is required only when moisture content exceeds 0.10%; at 80°C for 3–4 h in a desiccant dryer to -30°C dew point is sufficient for most glass-filled grades. Injection units with 18:1–22:1 L/D and low-compression barrier screws are preferred; high-shear mixing sections are avoided because they reduce mean fibre length from the pellet value to below 100 µm and further depress impact resistance. Residence time at melt temperature above 200°C should not exceed 10 min; in hot-runner systems, dead spots and long tips must be individually heated and purged. A stall above 15 min at 220°C in a hot runner produces a noticeable formaldehyde odour and can degrade local impact performance. Tooling wear is a process cost: glass fibre abrasion demands mould steel hardness of 50–54 HRC with nitrided or hard-chrome plating for production runs exceeding 500,000 cycles; the barrel and check ring must be hardened for glass-filled resins above 20 wt%.

Pump impellers, valve bodies, and high-pressure fuel system components in 30 wt% glass-filled POM gain creep modulus retention at 80°C but may exhibit notched Charpy values below 4 kJ/m² in the as-moulded condition. If the impeller hub contains a metal insert, the combination of mould shrinkage anisotropy and the 6–10× difference in thermal expansion between acetal and steel can generate hoop stress sufficient to initiate radial cracking at the insert interface during thermal cycling from -40°C to 120°C. For such parts, lower glass fibre loadings or mineral-filled POM are evaluated, but published data for this specific configuration is limited and component testing under thermal shock protocols is necessary before release.