Annealing Polycarbonate Parts: Time, Temperature and Why Skipping It Costs More

Thermal conditioning of polycarbonate (PC) is a stress-relief operation placed between primary forming and downstream finishing. The process exposes molded, extruded, or machined workpieces to controlled oven residence below the glass transition. For unfilled grades, the upper soak temperature is generally limited to 135°C. The glass transition of amorphous PC is measured at 147–150°C by differential scanning calorimetry according to ISO 11357-2. Heat deflection temperature under 1.8 MPa load is 125°C when determined by ISO 75-2. Annealing at 121°C for thickness-dependent hold periods relaxes frozen-in chain orientation generated during injection molding, extrusion, or cold machining. Skipping the step transfers cost to later field failure modes: solvent crazing, dimensional drift after secondary finishing, or crack initiation at press-fit inserts. The relationship among time, temperature, and wall thickness is the central control variable.

Residual stress in transparent PC can be screened qualitatively with polarized light. High fringe order near gates, weld lines, or inserts indicates molecular orientation that was frozen before relaxation could occur. Birefringence fringes are not a direct stress value, but their reduction after annealing is routinely used on production lines as a batch release indicator. Quantitative stress measurement is possible with hole-drilling strain gauges or layer removal, but these methods are slower and geometry-dependent. Published data for specific stress magnitude reduction in molded PC is limited because the distribution depends on gate geometry, packing pressure, mold temperature, and part thickness. The practical role of annealing is therefore validated most often by comparative solvent exposure and by dimensional stability after conditioning.

What Is the Measurable Result of Skipping Thermal Stress Relief?

In an injection-molded PC component, flow-induced orientation and differential cooling generate a residual stress profile. The skin solidifies against a mold surface that is typically held between 80°C and 120°C for unfilled PC. If the mold surface is below 80°C, freeze-off occurs earlier and the core remains molten longer. The resulting stress profile places the surface in compression and the core in tension. When a secondary operation introduces a sharp feature—such as a threaded brass insert, a drilled hole, or a snap-fit groove—the local stress concentration can exceed the craze initiation limit at room temperature. A stress-related failure may not appear at the molding machine. It appears later under contact with lubricants, cleaning fluids, adhesive solvents, or plasticizing oils. Annealing at 121°C reduces the stress gradient by allowing segmental motion below the glass transition. The measurable result is that a part that would craze in a solvent screening test before annealing remains uncrazed after a correctly executed cycle. The effect is verified by solvent exposure tests such as ASTM D543 or by environmental stress cracking evaluation under ISO 22088-1.

On horizontal injection molding machines with clamp forces from 600 kN to 4,000 kN, packing pressure and hold time can be adjusted to reduce residual stress at the source. A cold mold at 60°C may fill faster but produces higher stress. Raising mold temperature to 100°C reduces stress but increases cycle time. Annealing becomes necessary when cycle-time constraints prevent high mold temperatures. The process therefore functions as a downstream compensation, not a full replacement for correct molding. Time-temperature superposition in PC below the glass transition also means that longer exposure at 121°C is not indefinitely beneficial. Stress relaxation reaches a practical plateau after the part has achieved uniform temperature plus the required relaxation period. Continuing for many additional hours produces marginal further stress reduction while increasing energy consumption and reducing throughput.

Because heating is conduction-limited, the required dwell time scales with the square of the thickest cross-section. A thin wall reaches oven temperature quickly; a thick boss or rib root lags behind. The following representative values for unfilled PC assume a circulating air oven at 121°C with part-to-air temperature difference less than 3°C at thermocouple checkpoints and a cooling ramp not exceeding 20°C/h to 80°C before removal.

Above 6 mm, the ramp-up rate becomes as important as the hold time. A thermal gradient across a thick PC bar can itself generate stress if the surface heats too quickly. Production ovens should be profiled with thermocouples embedded in sacrificial parts at the thickest section, the gate area, and the thinnest wall. A uniform air temperature reading is not sufficient; part-core lag can exceed 20°C during the first hour for dense glass-filled grades. The one-hour-per-millimetre rule is not a material constant; it is a heat-transfer approximation. For a semi-infinite plate, the time to thermal equilibrium scales with the square of thickness. A 2 mm wall therefore reaches core temperature roughly four times faster than a 4 mm wall. Complex gate and rib geometry departs from a flat plate approximation, so direct thermocouple mapping is the only reliable validation. Published data for specific glass-filled and flame-retardant PC configurations is limited; grade-specific supplier bulletins should override generic settings.

When Oven Airflow Creates a Temperature Spread Greater Than ±3°C

Forced-circulation ovens with horizontal airflow at 1.5–2.0 m/s provide more uniform heat transfer to PC batches than gravity convection ovens. If airflow is uneven or parts are stacked directly on a solid shelf, the local temperature spread can exceed ±3°C. In such a case, thinner sections may reach 135°C while thicker sections remain near 115°C. The thin sections can distort under self-weight because 135°C is above the heat deflection temperature of 125°C under load. The thick sections, by contrast, are under-annealed and retain residual stress. The result is a batch with non-uniform properties. Fixturing must therefore support parts without imposing additional load. Stacking should be avoided unless the part geometry is self-supporting. Perforated aluminum racks, not solid trays, are preferred because they allow airflow to the underside. Thermocouple mapping should be repeated after any change in rack material or load density.

Cooling from annealing temperature should be slow enough to avoid regenerating the stress gradient. If the part is removed from the oven at 121°C and exposed to room-temperature air, the surface cools faster than the core. That creates a new stress profile that can erase much of the benefit. The cooling segment is therefore as important as the hold segment. Industrial practice is to leave parts in the closed oven and allow cooling to 80°C at a rate not exceeding 20°C/h. Faster cooling may be acceptable for very thin parts, but only if validated by post-anneal solvent testing. Polycarbonate is also incompatible with strong alkalies, amines, and certain solvent blends. Oven atmospheres should be free of these vapors. Annealing in an oven shared with polyvinyl chloride stock is inadvisable because hydrogen chloride release can attack the PC surface.

Solvent Craze Resistance After Thermal Conditioning

Polycarbonate is amorphous and notch-sensitive, with moderate chemical resistance. Solvents such as toluene, acetone, methyl ethyl ketone, and some phthalate plasticizers can induce crazing when tensile stress is present. Annealing does not make PC chemically resistant; it removes a component of stress required for craze formation. The distinction is critical. A properly annealed PC part can still be attacked by strong solvents, but it resists the early cracking that occurs when residual molded-in stress combines with a mild cleaning fluid. Chemical compatibility should be verified separately by ASTM D543 immersion testing or by supplier chemical resistance data. The stress-relief cycle should be added after any machining operation that cuts more than 0.5 mm from the surface, because fresh machining can reintroduce localized stress at the cut line.

Extended exposure above 140°C can reduce impact strength and shift elongation at break, as measured by ISO 527-2. The issue is not immediate melting but cumulative oxidative degradation. The practical upper limit for unfilled PC annealing is therefore set below 140°C, with 121°C preferred to preserve toughness. Some flame-retardant or glass-filled PC grades have different thermal stabilization packages. Their published supplier data should replace generic settings when available. Published data for these specific formulations is limited without grade-specific bulletins. Moisture interaction adds a further boundary. Polycarbonate is hygroscopic, and heating a damp machined part to 121°C can drive off absorbed water. That can reduce surface haze in subsequent welding, but rapid heating of damp constrained sections may promote local surface marking. Pre-drying at 80°C for 2–4 h in a desiccant dryer is recommended when a part has been stored above 60% relative humidity before annealing.

The failure cost of an unannealed PC part appears in several areas. A snap-fit cover that crazes after contact with cleaning solvent fails during assembly or in field use. A gear or impeller that drifts after absorbing oil can bind. An injection-molded manifold that carries a metal thread insert may develop radial cracks during insertion if the hoop stress from molding remains high. These failure modes are diagnostic: multiple radial cracks around a metal insert in a solvent-free environment often indicate high molded-in stress rather than chemical attack. The immediate production saving of skipping a 4-hour oven dwell for a 3 mm part is offset by higher scrap rates at next operations, customer-return quarantine, and repeated validation runs. Cost accounting varies by plant, but the technical trade-off is that annealing functions as a process control point rather than an optional surface finish.

After annealing, dimensional inspection should be performed after the part has cooled to 23°C and conditioned at 50% relative humidity according to ISO 291. Annealed parts may show small anisotropic relaxation, especially in the flow direction of a molded part. Parts intended for precision assemblies should be checked with go/no-go gauges only after this conditioning step. If dimensions shift beyond drawing tolerance, the molding parameters or gate size should be corrected before the anneal cycle is altered. The anneal step is not intended to fix gross molding defects such as sink marks, voids, or severe warpage; those require process changes at the press. If critical dimensions hold after conditioning and solvent exposure, the oven cycle can be considered technically valid for that part family.