Polyoxymethylene (POM) is specified for high-stiffness, low-wear parts because its semicrystalline acetal backbone provides a short-term tensile modulus near 2,800 MPa and a dry coefficient of linear thermal expansion close to 110 × 10⁻⁶ K⁻¹. Field returns for loose fits in POM gears, snap-fits, bushings, and threaded bosses are frequently traced not to initial molding defects but to combined creep and moisture uptake. Ignoring POM Creep and Moisture: The Failure Mode Behind Loose Fits is a progressive loss of interference caused by stress relaxation in the amorphous phase and volumetric swelling at the part surface. The problem becomes most severe in parts molded with low internal stress but then exposed to cyclic humidity or continuous load at ambient temperatures above 40 °C.

Moisture absorption in unfilled POM follows Fickian diffusion. At 23 °C and 50 % relative humidity, equilibrium water uptake is approximately 0.2 % by mass, while saturation in liquid water at 23 °C reaches about 0.8 % for many copolymer grades. The resulting linear expansion is not isotropic in molded parts because orientation from the frozen-in flow layer and weld-line anisotropy produce differential swelling. A local thickness increase of 0.5 % may reduce a hole diameter and raise hoop stress in a press-fit bushing, but it can also unlock a mating steel shaft if the surrounding boss is constrained and swells outward.

In injection-molding production, pre-drying is mandatory when ambient relative humidity exceeds 60 %. A desiccant dryer with a dew point below −40 °C is used to reduce pellet moisture to below 0.1 % by mass before the resin enters a 25 mm or 35 mm reciprocating screw. Hydrolysis at melt temperatures above 210 °C cleaves acetal linkages and releases formaldehyde; the result is surface splay, loss of molecular weight, and reduced elongation at break measured by ASTM D638-14. Homopolymer grades have a narrow processing window of approximately ±5 °C around 205 °C because formaldehyde evolution accelerates above 210 °C; copolymer grades tolerate up to 230 °C melt temperature. Acidic purge compounds and additive systems that liberate halogen acids are incompatible because acid-catalyzed depolymerization can reduce molecular weight even when the measured pellet moisture is low.

Why Moisture Uptake Converts an Interference Fit into a Clearance Fit

Interference fits in POM are generally designed with diametral interference of 0.2 % to 0.6 % of the shaft diameter. The apparent modulus used for assembly force is not the short-term tensile modulus; it is the creep modulus. For a POM bushing pressed onto a steel shaft, hoop stress is proportional to the effective modulus and the initial interference. When service temperature fluctuates between 23 °C and 60 °C, the creep modulus after 1,000 h can fall to approximately 50 % of the initial value. At the same time, moisture swelling at the bore increases contact pressure initially but then accelerates stress relaxation because the plasticized amorphous phase has a higher molecular mobility and a lower local glass transition. The result is that radial retention force after 500 h can fall below the threshold required to prevent axial migration of the shaft under vibration.

On a 1,100 kN all-electric injection machine with a 35 mm screw and 22:1 L/D ratio, POM bushings are molded at 215 °C melt temperature and 90 °C mold temperature. Cavity pressure transducers show that packing at 80 MPa reduces sink but also freezes in residual stress. After ejection, the part continues to shrink and condition to ambient humidity. Bushings measured after 48 h show bore closure of 0.05 mm for a 20 mm nominal bore, but measured values vary depending on gate location, flow-front temperature, and packing time. The dimensional check alone does not predict the loss of fit because creep and moisture act together after assembly.

Creep Modulus, Crystallinity, and the ISO 899-1 Data Set

Creep data measured by ISO 899-1 record strain after 1 h, 10 h, 100 h, and 1,000 h under constant load at controlled temperature and humidity. POM creep is nonlinear above 10 MPa stress; at 20 MPa and 23 °C, published data for a medium-viscosity copolymer show tensile creep strain exceeding 1.5 % at 1,000 h, while the same grade under 10 MPa remains below 0.8 %. The amorphous phase contributes most of the time-dependent deformation because the crystalline lamellae restrict chain slip. Service temperatures above 60 °C shift the creep curve upward; at 80 °C the 1,000 h creep strain under 10 MPa can exceed 1.2 %. For interference fits, this means stored elastic energy dissipates as viscoelastic flow, and contact pressure decays even if the part dimensions appear stable.

Published data for this specific configuration is limited, and the values should be confirmed with supplier long-term creep data for the specific POM grade, filler content, and weld-line condition. The important comparison is not the initial modulus but the ratio of creep modulus at the expected service temperature and moisture state to the design modulus used in the tolerance stack. Failure analysis with digital image correlation during fit-retention testing shows strain localization near gates and weld lines, where oriented skin layers are thinner and moisture uptake is faster.

Snap-fit retention in POM electronics housings follows the same mechanism. A cantilever snap arm is deflected during assembly and must maintain a residual force after mating. If the arm is loaded to a strain of 2.0 % during insertion, the initial retention force may be 45 N. After 1,000 h at 40 °C and 70 % R.H., stress relaxation and moisture plasticization reduce the residual force by 30 % to 50 %. The failure mode is not sudden fracture but a loose fit that allows audible rattling, intermittent electrical contact, or accidental opening. Snap-fit designs using POM should limit flexural strain to below 1.0 % when continuous load is expected, or use metallic spring reinforcements.

When POM Components Are Molded Below −10 °C Dew Point

Processing with a dehumidifying dryer at a dew point below −10 °C does not remove the need for service moisture management. Dryer performance controls pellet moisture before plastication; it cannot prevent post-molding absorption from ambient air or condensation in cold-start conditions. A POM housing taken from a 110 °C mold to a 25 °C assembly area at 65 % R.H. begins to absorb moisture immediately. Dimensional movement of 0.1 % can be reached within 24 h depending on wall thickness and crystallinity. For a 50 mm long housing, that represents 0.05 mm of linear growth, enough to alter a snap-fit latch engagement by several tenths of a millimeter.

Moisture analysis by ISO 15512 Karl Fischer titration or parallel ISO 1110 conditioning confirms the part moisture state but does not predict local moisture gradients. In a hot runner mold, the gate region cools more slowly and develops higher crystallinity; the thicker crystalline zone absorbs less moisture than the adjacent oriented skin. This gradient creates transient internal stress that can temporarily tighten a fit and then release as the part equilibrates. The effect is observed on production lines when a POM coupling passes incoming inspection after 4 h from molding but fails mating-force testing after 72 h because the bore has shifted by 0.02 mm.

Threaded metal inserts in POM bosses fail differently. The plastic surrounding the insert creeps under the radial stress generated by the insert itself. A brass insert with an outer diameter of 6 mm installed into a POM boss imposes a hoop strain that relaxes over time. When the assembly is exposed to 80 % R.H., the POM swells and initially tightens around the insert, but the plasticized polymer then loses its load-bearing capacity. Pull-out force measured on a production test fixture with a 10 mm/min crosshead speed can decrease from approximately 1,200 N dry to 700 N after 500 h at 40 °C and 90 % R.H. The insert remains visually seated, but the loose fit appears when the screw torque drops below the specified 0.8 N·m seating torque.

Fastener Retention in POM Threaded Bosses Without Metal Inserts

Direct-threaded POM bosses are limited by thread stress relaxation. A self-tapping screw forms a thread in the polymer; the hoop stress generated during thread forming is high, and the creep modulus of POM controls the residual clamping force. At 23 °C dry conditions, a thread-forming screw can provide a clamp load of 800 N. After moisture saturation and 1,000 h at 60 °C, the residual clamp load can drop below 300 N. The joint remains assembled but becomes loose under vibration. Torque-to-loosen measurements show initial breakaway torque of 1.2 N·m falling to 0.3 N·m. The failure is mechanical relaxation, not chemical attack, and repeated retorquing accelerates thread damage because the bore has already opened.

In a fan drive coupling molded from glass-fiber reinforced POM, the failure mode is observed on a production line after 2,000 h of duty. The coupling bore is designed with an interference of 0.1 mm on a steel motor shaft. The glass fibers reduce moisture swelling to about 0.2 % linear, but they also reduce the creep benefit at weld lines because fiber orientation across the weld limits load transfer. The failure surface shows circumferential cracking and fiber pull-out; the bore has opened by 0.04 mm. The shaft can rotate relative to the coupling under torque reversals, producing a knocking sound. Compliance with FDA 21 CFR 177.2470 or RoHS does not alter this physical aging response. The operational boundary for this geometry is therefore continuous hoop stress below 10 MPa and service relative humidity below 60 % unless creep and moisture swelling are explicitly included in the tolerance stack.