Anionic polymerization of ε-caprolactam in a closed steel casting cell is initiated by sodium caprolactamate, typically at 0.4–0.8 mol% relative to monomer, and activated by acyl caprolactam species such as N-acetylcaprolactam at 0.4–0.6 mol%. The activator provides an acyl imide initiating site, while the lactamate anion propagates by ring-opening of additional ε-caprolactam at the 90–110 °C casting temperature. The monomer melts at 69 °C but is held under nitrogen before catalyst addition to prevent solidification and to control viscosity. Residual water is the principal termination agent; each water molecule hydrolyzes a lactam anion to form free caprolactam and sodium hydroxide, thereby reducing the number of propagating centers. Industrial monomer-casting specifications require residual moisture at or below 0.03 wt% before catalyst injection. When moisture rises to 0.05 wt%, a 50 mm thick casting can exhibit a delayed exotherm onset from 5–8 min to 15–20 min and a peak exotherm reduction of 10–20 °C, below the 180–200 °C range observed in properly dried monomer. Free ε-caprolactam content rises above 3 wt%, and the part develops a weak, waxy core that cannot be demolded cleanly.

Water enters the monomer during storage, transfer, and melt preparation. Caprolactam flakes equilibrate with warehouse relative humidity; at 60% RH and 20 °C, surface moisture can exceed 0.1 wt% within 24 h. Production-scale vacuum dryers with 500 kg charge capacity are operated at 80 °C and 20–30 mbar absolute pressure for 4–6 h to achieve 0.03 wt% residual moisture. Field data from polytetrafluoroethylene-lined double-cone dryers show moisture variability of 0.01–0.03 wt% across a single batch when vacuum pump performance degrades or when the condenser outlet temperature exceeds 10 °C; the condition is compensated by extending drying time to 8 h or reducing charge depth. In-line near-infrared moisture analyzers calibrated against coulometric Karl Fischer titration according to ISO 15512:2019 are installed at the dryer outlet; probe fouling from caprolactam sublimate underreports moisture by 0.01–0.02 wt%, so calibration must be verified at least once per shift.

Coulometric Karl Fischer titration with a vaporizer set at 160 °C is preferred over volumetric titration because caprolactam interferes with direct injection and releases water slowly. The method specified in ISO 15512:2019 requires validation of recovery using hydrated caprolactam check samples; recovery below 90% indicates incomplete water release and requires higher vaporizer temperature or longer purge time. In-line capacitance sensors are not reliable for caprolactam flakes because the dielectric constant changes with temperature and packing density; production lines therefore rely on periodic grab sampling. Liquid caprolactam transfer lines must be heated and purged with dry nitrogen with a dew point below -40 °C. Open handling at 60% RH can raise the moisture content of the melt surface by 0.05–0.10 wt% within 30–60 min, and production scheduling must therefore sequence drying immediately before casting.

What Moisture Threshold Governs Sodium Caprolactamate Catalyst Activity?

The inhibition effect of water is stoichiometric rather than catalytic. A residual moisture of 0.03 wt% corresponds to approximately 1.67 mmol of water per 100 g of caprolactam. Because the active sodium caprolactamate concentration in monomer-casting formulations is also in the millimolar range, the water-to-catalyst ratio is sufficiently large to depress the number of active centers without being visually detectable. At 0.10 wt% moisture, the water content is approximately 5.56 mmol per 100 g, which may exceed the total catalyst charge in some formulations. The threshold is therefore not a single concentration but a function of catalyst and activator molar balance; formulations with higher activator levels can tolerate slightly higher moisture, but only at the cost of accelerated reaction after the induction period and increased risk of exothermic runaway in thick sections exceeding 80 mm. Published data for exact kinetic constants of water termination in sodium caprolactamate systems is limited; however, the empirical threshold of 0.03 wt% is consistently observed across monomer suppliers for cast PA6 grades.

Amine-based additives that regenerate free lactam or consume acyl activator are incompatible with sodium caprolactamate casting systems because they alter the catalyst/activator ratio. Even a 0.2 mol% addition of a primary amine can consume the acyl activator at an equimolar ratio, causing premature gelling or incomplete conversion. Material substitutions should therefore be evaluated by thermal analysis according to ISO 11357-3:2018 and residual monomer extraction before production use.

Above the threshold, polymerization does not simply slow; it produces a heterogeneous structure. The skin of the casting, in contact with a heated mold at 120–140 °C, polymerizes first because water in the boundary layer is driven into the bulk or vaporized. The core remains molten and moisture-rich, so the reaction front propagates inward slowly. Differential scanning calorimetry according to ISO 11357-3:2018 reveals a broad melt endotherm with shoulder peaks at 200–210 °C and a reduced degree of crystallinity below 30%, compared with 35–45% for dry-monomer cast stock. This skin-core mismatch increases residual stress and leads to post-demolding warpage and center-line porosity.

Values in the table are representative of steel-mold casting cells at 120 °C mold temperature and are not transferable to aluminum molds or uncontrolled ambient conditions; verification by thermocouple and residual monomer analysis is required for each new mold geometry.

Demolding Force, Shrinkage Compensation, and Cooling Curve Control

Demolding of cast nylon is governed by three coupled variables: core temperature, mold surface temperature, and the coefficient of thermal contraction. For a 50 mm thick plate cast in a steel mold held at 130 °C, the exotherm peak occurs at 15–20 min after catalyst addition, and the part reaches a core temperature of 180–200 °C. Demolding before the surface temperature falls below 90 °C produces visible ejector pin penetration and localized tearing because the material remains above its heat deflection temperature. The heat deflection temperature of dry cast PA6 under 1.8 MPa load, measured according to ISO 75-2:2013, is approximately 75–85 °C; therefore, a practical demolding criterion is to wait until the part surface thermocouple reads 60–80 °C. At this condition, the linear mold shrinkage of unrestrained cast PA6 is in the range of 1.5–2.5%, measured by dimensional comparison according to ISO 294-4:2018. If the mold is opened too early, the part may shrink at different rates across its section, producing bowing that exceeds 0.3 mm per 100 mm of section length. Bowing is measured with a calibrated granite surface plate and feeler gauges; a limit of 0.3 mm per 100 mm is applied for gear blanks before machining.

On production-scale vertical press casting cells with hydraulic ejectors, demolding force is measured by load cells mounted in the ejector plate. In a 100 mm diameter sleeve mold, ejection force can exceed 5 kN when surface temperature is above 90 °C; after cooling to 60 °C, force drops below 2 kN because the part has contracted away from the core and developed sufficient modulus. Published data for this specific configuration is limited, and the force values are representative of equipment load-cell readings rather than universal design limits. Mold release agents, typically silicone-based, reduce stick-slip but can contaminate surfaces intended for post-molding machining or bonding; a release agent that migrates into the part surface by 10–20 μm can reduce lap shear adhesion in subsequent polyurethane bonding, measured according to ISO 4587:2003, by 30%. For applications requiring adhesive assembly, polytetrafluoroethylene-coated mold surfaces and dry demolding are preferred, but they require a cooler surface temperature of 50–60 °C to avoid sticking. Mold surface roughness also interacts with moisture. A cast iron mold with average roughness Ra of 0.8 μm can show higher demolding force than a polished steel mold with Ra of 0.2 μm because caprolactam exudate wets the rougher surface. The effect is amplified when residual moisture exceeds 0.03 wt% because the exudate contains more free caprolactam. Surface preparation procedures therefore specify polishing to Ra ≤ 0.4 μm and application of sacrificial zinc stearate layers for first cycles.

Cooling time scales with the square of section thickness. The time for a slab to cool from exotherm peak temperature to a centerline temperature below 100 °C can be approximated by the Fourier number method; for a 50 mm plate, the cooling time is 0.5–1 h, while for a 100 mm plate it increases to 2–4 h in a 130 °C mold. Production scheduling uses thickness-dependent demolding windows rather than fixed cycle times; a 200 mm block cast in the same mold may require overnight cooling. Published data for exact cooling times in complex geometries is limited, and thermocouple verification is mandatory for new part designs.

When Residual Moisture Exceeds 0.04 wt% in Thick-Section Castings

Moisture overshoot above 0.04 wt% in sections thicker than 80 mm creates a characteristic failure sequence that is distinct from simple rate suppression. The casting skin polymerizes at the mold wall because the wall heat drives water migration away from the reaction zone; the inner core retains water and unreacted caprolactam for an extended period. After the mold is opened, atmospheric moisture and oxygen accelerate hydrolysis and discoloration, producing yellow-brown centers. The residual ε-caprolactam content measured by methanol extraction followed by gas chromatography, using the procedures outlined in ISO 6427:2013, can exceed 5 wt% in the core while the skin remains below 2 wt%. Such parts exhibit hardness below 70 Shore D measured by ISO 868:2003, compared with 78–82 Shore D for dry-monomer cast PA6, and tensile strength below 50 MPa measured by ISO 527-2:2012, compared with 70–80 MPa for sound material. The part is not recoverable by post-curing because the catalyst has been consumed and the remaining monomer lacks active centers to continue polymerization.

Thick-section demolding failures from moisture overshoot are often misclassified as mold release problems. In a 120 mm thick bearing block cast in a 130 °C mold, the surface thermocouple may read 70 °C while the core remains at 140 °C because the thermal diffusivity of cast PA6 is approximately 1.2 × 10⁻⁷ m²/s. Demolding under these conditions releases a part with a soft core; post-molding contraction of the core then causes sink marks and internal voids. Process controls therefore require at least three thermocouples: one embedded in the mold wall, one placed 10 mm from the part surface, and one at the part centerline. Demolding is permitted only when the centerline temperature falls below 100 °C and the surface temperature is below 60 °C. For a 120 mm section, this window occurs 3–4 h after peak exotherm in a 130 °C mold; published data for exact times in other mold geometries is limited.

Quality-control standards applied to cast PA6 moisture and demolding validation are listed in the matrix below.

Cast nylon wheels, rollers, gears, and wear pads are produced in thicknesses from 20 mm to 200 mm. A 50 mm thick gear blank cast in a 120 °C steel mold using sodium caprolactamate at 0.6 mol% and N-acetylcaprolactam at 0.5 mol% reaches a surface demolding temperature of 60 °C at approximately 25–35 min after exotherm peak. A 200 mm thick block requires 4–6 h because the core retains heat. The demolding window also depends on mold release chemistry and part geometry; undercuts, deep hubs, and metal inserts reduce the permissible demolding force and narrow the window by 5–10 °C. Parts with pressed-in metal inserts must be cooled to 40–50 °C before demolding to avoid differential contraction that exceeds the interference fit tolerance of H7/p6 specified in ISO 286-1:2010. If moisture exceeds 0.03 wt%, the demolding window is not simply shifted; it can disappear entirely because the core never develops sufficient green strength.

After demolding, cast PA6 absorbs atmospheric moisture; conditioned parts at 50% RH and 23 °C according to ISO 291:2008 reach approximately 2.5 wt% water, and dimensions grow by 0.2–0.3% per 1 wt% water uptake. Machining allowances for cast nylon components must therefore be based on the moisture service condition rather than the dry as-cast state. Residual caprolactam above 3 wt% acts as a plasticizer and accelerates moisture absorption; extraction in boiling water reduces residual monomer to below 1 wt% but adds a post-machining conditioning step. Demolding windows documented for dry-monomer casting are not transferable to high-moisture batches, and each batch must be verified by oven-dry weight loss or Karl Fischer analysis before catalyst injection.