Polycarbonate (PC) is an amorphous engineering thermoplastic whose bisphenol A carbonate backbone gives a glass transition temperature of approximately 147 °C and saturated water uptake of 0.15–0.35% by mass per ASTM D570-98. In fused filament fabrication, the material sits inside a narrow processing band. The melt must be hot enough for layer fusion but dry enough to prevent hydrolysis, while the surrounding chamber must be warm enough to limit shrinkage stress but cool enough to avoid softening the printer structure or degrading stepper motors. Supplier datasheets and operational field reports indicate that unfilled PC filament generally requires an all-metal hot end holding 270–310 °C, a heated enclosure at 70–90 °C, and filament dried to below 0.02% moisture by mass. Published round-robin data for PC extrusion on FFF machines is limited, so the values cited below represent manufacturer technical bulletins, ISO 1133-1:2022-based viscosity references, and observed production-scale practice rather than a single universal standard.

How Does a Heated Enclosure Shift Warp and Interlayer Diffusion in PC?

The enclosure is not a convenience; it changes the residual stress state of an amorphous deposition. The linear coefficient of thermal expansion for unfilled PC is typically 65–70 ppm/°C when measured by ISO 11359-2. A melt deposited at 290 °C onto a bed at 110 °C in a 25 °C room cools through the glass transition much faster at the top surface than at the base. The resulting asymmetric contraction creates tensile stress near the perimeter and can exceed first-layer adhesion, producing corner lift, delamination, or mid-build fracture. Raising the chamber setpoint to 70–90 °C lowers the vertical thermal gradient and keeps the just-deposited layer above approximately 65–75 °C for several minutes. That residence time above the glass transition onset is required for polymer chain diffusion across the layer interface. For injection molded PC, ISO 294-4 shrinkage values of 0.5–0.7% are representative, but FFF shrinkage is anisotropic and currently has no equivalent single ISO standard. Passive insulated enclosures heated only by the bed often plateau at 45–60 °C; this may be sufficient for small PC coupons but is inadequate for unfilled PC structures longer than 100 mm in the X/Y plane. Active chamber heating with a recirculating fan, as found on industrial heated-chamber FFF systems, typically uses a setpoint of 80 °C with a bed setpoint of 100–120 °C. Below 70 °C, warp suppression becomes strongly geometry-dependent and should not be assumed for tall, thin-walled parts.

Chamber temperature uniformity is a separate control variable. A single convection source can create a vertical gradient of 10–15 °C between the build plate and the chamber ceiling. On a tall part, the upper layers then see a colder environment than the chamber thermistor indicates, which reintroduces edge stress. Production-scale heated-chamber machines therefore use multiple thermistors and forced airflow, with the control sensor placed away from the heated bed. Door seals and enclosure insulation also affect the gradient. Stepper motor temperature limits introduce a practical upper bound: motors rated for 80 °C can lose torque in a chamber held at 90 °C for extended runs, so active motor cooling or a lower chamber setpoint is required. The build surface must also be selected for the combined bed and chamber temperature; PEI or polycarbonate sheets are common, but bed temperatures above 120 °C can degrade some adhesives. Supplier data for heated-chamber PC prints indicate that Z-direction tensile strength can fall below 50% of XY strength when the chamber is below 60 °C, while a chamber at 80 °C typically raises the Z/XY ratio above 0.60. Published print-to-print variance data across multiple printer models is limited, but the directional trend is consistent with polymer diffusion theory.

For first-layer adhesion, a bed temperature of 110 °C on PEI or polycarbonate sheet is typical, with first-layer height of 0.2–0.3 mm and speed reduced to 15–20 mm/s. First-layer extrusion width is often set to 120–140% of nozzle diameter to increase contact area. If chamber temperature is below 60 °C, edge lift may still occur despite these settings, because the first layer remains bonded while upper layers contract. Some production prints use an adhesion raft or brim of 8–12 mm, but these features add post-processing steps and do not replace chamber heating.

Carbon-fiber-reinforced PC changes the enclosure requirement. The filler lowers the coefficient of thermal expansion to approximately 30–40 ppm/°C, which reduces warp at a given chamber temperature. However, filled PC does not eliminate the need for a warm chamber because the polycarbonate matrix still undergoes glass transition and the interlayer weld remains temperature-dependent. The hot end constraints discussed below become more severe for filled grades because of abrasive wear and higher thermal conductivity.

Hot End Metallurgy and the Necessity of All-Metal Thermal Paths

PTFE-lined hot ends are incompatible with polycarbonate. PTFE loses mechanical integrity above 260 °C, and the PC melt zone must operate at 270–310 °C. The correct configuration uses an all-metal heat break with a sharp thermal transition, typically a bimetallic or titanium alloy tube with a wall thickness of 0.2–0.4 mm. The heater block should hold the nozzle setpoint within ±2 °C, which requires a cartridge heater and a PID controller with autotune. A 40 W or 60 W heater is common for 0.4 mm nozzles. Melt volume rate for FFF-grade PC typically ranges from 10 g/10 min to 25 g/10 min at 300 °C/1.2 kg per ISO 1133-1:2022. Lower-MVR resins require nozzle setpoints near 300–310 °C, while higher-MVR grades can be processed closer to 270–285 °C. The bed is maintained at 100–120 °C to maintain first-layer adhesion and reduce early cooling. Layer heights of 0.15–0.25 mm with extrusion widths of 0.4–0.6 mm are typical for 0.4 mm or 0.6 mm nozzles.

Extruder architecture determines feed reliability. Polycarbonate filament is stiff and can buckle or fracture in long Bowden tubes. Direct-drive dual-gear extruders are preferred, and retraction distance should be minimized to 0.5–1.5 mm to prevent pulling molten PC into the cold side of the heat break. A part cooling fan is generally disabled for PC, or limited to 10–20% duty for small features, because rapid quenching increases delamination risk. A silicone sock over the heat block reduces temperature ripple from residual airflow. Brass nozzles are acceptable for unfilled PC, but carbon-fiber-filled PC requires a hardened steel, ruby, or tungsten carbide nozzle. Brass orifice wear in filled grades can shift the actual line width by more than 0.05 mm and create under-extrusion artifacts long before the nozzle fails structurally.

Volumetric flow limits exist because PC is a high-viscosity melt. Standard 0.4 mm nozzles on a 40 W hot end may begin to skip filament at volumetric rates above approximately 10 mm³/s; the exact limit depends on heater power, melt zone length, and barrel temperature. PC exhibits shear-thinning behavior, so viscosity decreases at typical extrusion shear rates, but the melt remains more viscous than PLA. High melt pressure at the nozzle tip can contribute to back pressure and filament buckling if extruder idler tension is too high. Polished drive teeth or excessive tension can create PC dust, which accumulates in the extruder and causes slip. Thermal runaway protection should be configured for the specific heater cartridge and thermistor type, and the hot end should be assembled with a torque wrench according to the nozzle tightening procedure because cyclic thermal expansion at 300 °C can loosen a nozzle and create internal leakage.

At nozzle temperatures above 320 °C, thermal degradation becomes measurable even in dry PC. The melt darkens, molecular weight decreases, and impact strength falls disproportionately compared with tensile modulus. If a print is paused, the hot end should be purged with fresh dry PC or a high-temperature purge compound before resuming; extended static residence at 300 °C can cause chain scission in the stagnant boundary layer near the nozzle wall.

Drying of spooled PC filament is a diffusion-limited operation, not a surface evaporation problem. Water absorbed in the amorphous matrix cannot be removed by a brief preheat or by passing the filament through a heated tube at print time. The accepted melt-processing endpoint is below 0.02% moisture by mass, with many PC filament suppliers specifying 0.01–0.02%. A spool of 1.75 mm PC filament exposed to 50% RH at 23 °C can reach approximately 0.2% moisture at the outer wind layers within 24–48 h, while inner wind layers remain drier because of diffusion resistance. This nonuniform distribution causes intermittent extrusion defects: the first layers of the spool may print acceptably, while deeper windings begin to foam or produce splay. The following drying methods are used for spooled PC filament.

Forced-air convection is the most practical method for single spools, but the oven must have active air circulation and sufficient clearance around the spool flanges. Vacuum drying shortens the cycle by lowering water partial pressure, but oil-sealed vacuum pumps require protection from evolved volatiles. Active filament dryers work only if the dried air is fed to the hot end through a sealed PTFE tube; any open-air path between the dryer outlet and the extruder inlet allows moisture reabsorption. After drying, spools should be transferred immediately to a sealed container with molecular sieve 13X or activated alumina and held at <10% RH. In facilities where ambient relative humidity exceeds 60%, a single drying cycle is insufficient for prints longer than 12 h; continuous dry feeding or periodic re-drying is required.

Filament diameter influences drying time. Because water removal scales with diffusion path length, 2.85 mm PC filament can require roughly 1.5–2× the time of 1.75 mm filament at the same oven temperature. Winding tension also matters: tightly wound spools block airflow to the inner layers, and the hub section can remain above moisture specification after a nominal cycle. Field data from spool drying operations indicates that wide-flange spools with loose windings dry faster than dense spools with small flanges at the same temperature, but published standardized data for PC filament spool drying is limited. Molecular sieve 13X regenerates at 200–250 °C for 4 h in a ventilated oven; indicator desiccant color must not be the sole determinant because some indicators lose sensitivity after repeated regeneration. A dew point meter inserted into the dry box provides a direct measurement rather than an inference from color change.

When Ambient Relative Humidity Exceeds 40%, Hydrolysis Becomes Transport-Limited

Water-induced degradation in PC is rapid once the melt reaches processing temperature. Absorbed water attacks the carbonate linkage, generating chain scission, carbon dioxide, and bisphenol A. A residual moisture content of 0.05% by mass is sufficient to produce visible splay and internal voids at 300 °C, while 0.15% moisture can reduce molecular weight and notch toughness dramatically. Notched Izod impact measured by ASTM D256-10 can fall from the dry range of 600–900 J/m to a low-toughness failure mode, but the exact reduction depends on resin grade and print porosity. The degradation rate increases with temperature, which means that upper-range nozzle settings are less tolerant of residual moisture. Once chain scission has occurred, the damage is permanent; re-drying or annealing cannot restore the original molecular weight.

The operational boundary is explicit: at ambient relative humidity above 40%, an open spool can re-absorb enough moisture within 12–24 h to create intermittent extrusion popping and reduced interlayer toughness. Above 60% RH, continuous sealed dry feed with a desiccant bed and a PTFE outlet tube is required for prints longer than 2 h. Partially used spools should be sealed in vacuum bags with desiccant sachets immediately after removal from the printer. A dry box with a dew point of -40 °C is the only reliable method for long-duration printing in uncontrolled humidity, because desiccant alone cannot keep up with repeated opening of the container. Validation of dryness on the print line can be performed with a moisture analyzer or Karl Fischer titration of a sacrificial sample; infrared analyzers require calibration for PC. Operators should log drying time, oven temperature, ambient RH, and dry box dew point so that intermittent extrusion defects can be traced to moisture excursions rather than attributed to hot end or chamber failure. The interaction of moisture, nozzle temperature, and enclosure temperature makes PC a system-level material: the hot end and chamber cannot correct for wet feedstock, and drying cannot compensate for a cold chamber or an unstable thermal path.