Moisture absorption in PA6 under conditions of 23 °C and 50% relative humidity reaches approximately 2.7–3.0 wt% when measured according to ISO 62:2008; immersion in water at 23 °C raises saturation uptake to 9.0–9.5 wt%. In underhood components, this water is chemically reactive rather than merely plasticising. Hydrolysis of the amide linkage proceeds by random chain scission that is autocatalytic in the presence of carboxylic acid end groups. The rate depends on local water activity, temperature, and the concentration of proton-donating species. For unstabilized PA6 with an acid end-group content of 35–45 meq/kg, viscosity number measured per ISO 307:2019 declines from an initial 145–155 mL/g to 80–100 mL/g after 1,000 h immersion in 50:50 ethylene glycol/water at 125 °C. This corresponds to molecular weight loss into the brittle range, with tensile strength retention measured per ISO 527-2:2012 falling to 35–45%. Comparative North American tensile data generated according to ASTM D638-14 show the same trend. Production-scale experience on a 2,500-kN clamp force injection moulding machine with a 40 L/D twin-screw compounding stage showed that pellet residual moisture above 0.10 wt% produced visible splay and reduced notched Charpy impact strength by 15–25% after only 500 h of coolant exposure at 120 °C. The impact measure is more sensitive than tensile strength because surface degradation and weld-line attack do not require bulk molecular weight collapse. Post-moulding conditioning at 23 °C and 50% relative humidity for 500 h before service shifted the failure mode from ductile to brittle when residual caprolactam exceeded 0.3 wt%. The same effect was not observed in parts dry-aged at 80 °C for the same duration, indicating that water uptake and residual monomer interact to reduce the effective entanglement density. Pre-drying of PA6 pellets at 80 °C to a residual moisture content below 0.10 wt% is required when ambient relative humidity exceeds 60%; otherwise hydrolytic degradation during melt processing reduces viscosity number before the part enters service.

Why Do Acid End Groups and Residual Caprolactam Set the Autocatalytic Baseline?

During hydrolytic ageing of PA6, each chain scission consumes one water molecule and generates one carboxylic acid end group and one amine end group. The newly formed carboxylic acid protonates the amide oxygen and increases the local hydrolysis rate, creating an induction period followed by an accelerating molecular weight decline. Acid end-group concentration greater than 40 meq/kg after melt processing shortens the induction period in hot aqueous glycol. Residual caprolactam, typically 0.1–0.5 wt% as determined by gas chromatographic methods aligned with ISO 11337:2023, plasticises the amorphous phase and raises water diffusivity. Some automotive material specifications limit residual caprolactam to 0.3 wt% for underhood PA6. Reactive extrusion with vacuum venting at −0.095 MPa and melt temperature of 270–280 °C can reduce residual monomer, but melt temperatures above 290 °C degrade amine ends and alter the acid/amine balance. In controlled compounding trials, PA6 with an acid end-group concentration of 15 meq/kg retained Charpy impact strength above 70% after 1,000 h at 130 °C in 50:50 aqueous ethylene glycol, while the same base resin with 50 meq/kg retained less than 40% under identical conditions. The impact tests followed ISO 179-1:2023. The incompatibility observed with amine-based additives is not universal; some high-molecular-weight hindered amine stabilisers provide thermal oxidation resistance but can shift the end-group equilibrium in ways that increase equilibrium moisture uptake. Therefore, formulation screening for underhood PA6 requires titration of acid and amine end groups after compounding rather than relying on base resin data.

Coolant exposure in a pressurised loop generates local temperatures from 105 °C to 135 °C in normal operation, with hot-spot excursions to 160 °C near cylinder head outlets. The hydrolysis rate constant for PA6 approximately doubles for every 10 °C increase, consistent with an Arrhenius activation energy near 90 kJ/mol. Ethylene glycol in a 50:50 volume mixture lowers water activity compared with pure water, which slightly suppresses hydrolytic attack at the same temperature. However, glycol oxidation products accumulate over time and include glycolic acid, formic acid, and oxalic acid. When coolant pH drops from an initial range of 8.0–9.5 to below 7.0, the rate of PA6 chain scission accelerates because the acid end groups remain protonated. Chloride contamination above 50 ppm is recognised as a threshold above which acid-catalysed chain scission and stress cracking accelerate in PA6 underhood components. Copper ions released from brass fittings can catalyse oxidative chain scission at temperatures above 120 °C. The operational boundaries therefore include coolant replacement intervals and exclusion of chloride-containing water sources beyond the engine manufacturer's tolerance. Long-term immersion in fresh organic acid technology coolant with pH buffered between 8.3 and 9.0 produces less hydrolytic degradation than the same coolant after inhibitor depletion, but quantitative retention varies with glass loading and stabiliser package.

Glass-Fibre Orientation, Weld Lines, and the Hydrolytic Front

Hydrolytic degradation in moulded PA6 components follows water ingress pathways created by fibre orientation and weld lines. In 35% glass-fibre-reinforced PA6 radiator end tanks, filling analysis produced a weld line along the hose nipple. Cross-sections after hydrolytic exposure showed a 150–200 µm deep degraded layer at the weld line after 2,000 h at 120 °C in 50:50 aqueous ethylene glycol, while adjacent oriented regions showed 50–80 µm penetration. Viscosity number in the weld-line skin dropped from an original 145 mL/g to below 90 mL/g, whereas the bulk retained above 110 mL/g. This localised reduction in molecular weight causes cracking at the hose nipple under clamp loads. The significance of the weld line is quantified by comparing tensile bars with and without a deliberate weld line. Hydrolysis-stabilized glass-filled PA6 retains 60–70% of unwelded tensile strength after 1,000 h at 130 °C, but weld-line specimens retain only 30–40%. Glass-fibre sizing chemistry further controls hydrolytic stability. A standard amino-silane size supports a dry tensile strength of 180–220 MPa in 30% glass-fibre-reinforced PA6. After 1,000 h at 125 °C in coolant, interfacial shear strength measured by microdroplet pull-out decreases by 25–50% depending on the size chemistry. Capillary transport at the fibre-matrix interface is up to 10 times faster than bulk diffusion, so the hydrolysis front advances along the glass surface. Processing adjustments mitigate this effect. A mould temperature of 80 °C raises the crystallinity of the skin layer from 30% to 40% as measured by differential scanning calorimetry, reducing water permeability by approximately 25% in oriented regions. Low shear screw designs and hot runners with short residence times preserve the fibre length and the silane bond.

When Underhood Thermal Spikes Exceed 140°C Adjacent to Exhaust System Components

Components located near exhaust manifolds, turbocharger housings, or EGR coolers experience simultaneous thermal oxidation and hydrolytic attack. At surface temperatures above 140 °C, heat-stabilized PA6 grades lose hydrolytic stabilisers through volatilisation and leaching by acidic coolant condensate. Accelerated ageing in an autoclave at 135 °C and 100% relative humidity following dry thermal preconditioning at 150 °C for 500 h reveals a synergistic loss beyond either exposure alone. In a validation programme for a PA6 air cleaner resonator bracket positioned 120 mm from an exhaust heat shield, the measured peak surface temperature exceeded 155 °C. Standard heat-stabilized 30% glass-fibre-reinforced PA6 retained only 45% of original tensile strength after 1,500 h in the combined duty cycle, while the same material in coolant immersion at 120 °C retained 75%. A hydrolysis-resistant PA6 grade with a copper halide stabiliser system retained 70% under the same combined cycle. The test matrix used ASTM D3045-21 for dry thermal ageing and an internal autoclave procedure with 50:50 ethylene glycol/water. Failure was dominated by surface embrittlement rather than bulk molecular weight loss. The degraded skin was 250–350 µm thick over a ductile core, permitting crack initiation at surface flaws and propagation under pressure pulses. Validation for these positions also requires thermal cycling between −40 °C and 150 °C because the moisture-plasticised skin undergoes freeze-thaw stress. Substituting PA66 does not automatically resolve the problem; although PA66 has lower equilibrium moisture uptake, its hydrolysis rate at temperatures above 130 °C can approach that of PA6. The selection must be based on comparative immersion data under the actual coolant and temperature profile.

Accelerated hydrothermal testing of underhood PA6 parts commonly uses a pressure vessel containing 50:50 coolant concentrate and deionised water at 120–130 °C for 1,000–2,000 h. Acceptance criteria in vehicle manufacturer specifications typically require tensile strength retention above 70% after 1,000 h and notched Charpy impact retention above 50%. Test fixtures apply a strain condition equivalent to 0.3–0.5% outer-fibre strain because tensile stress increases water sorption and accelerates crack formation. Weight gain curves for 30% glass-fibre-reinforced PA6 in coolant at 125 °C show an initial diffusion-controlled increase up to 500 h followed by saturation near 4–5 wt%. Degradation continues after saturation because hydrolysis proceeds in the plasticised matrix. Viscosity number decreases linearly with reciprocal molecular weight, and a fall from 145 mL/g to 100 mL/g corresponds to an approximate Mn decline from 21,000 g/mol to 13,000 g/mol; at this point notched impact strength may fall by more than half. Published failure analyses also identify engine oil and road salt as accelerants not captured by single-fluid immersion. Sulfur-containing additives in motor oil can generate acidic species at elevated temperature, while sodium chloride from winter road de-icing concentrates at junctions and increases ionic ingress. Combined-fluid testing that alternates between coolant, engine oil mist, and 5 wt% sodium chloride solution has been used to reproduce field cracking in radiator end tanks. Published data for all combinations of coolant additive packages, glass loadings, and external contamination remains limited.

The test matrix applied to underhood PA6 is compiled from standardised conditioning, mechanical, thermal, and chemical methods. No single test captures the combined effects of stress, coolant composition, thermal oxidation, and contamination.

Copper halide stabiliser systems for PA6 function by redox reactions that deactivate hydroperoxides and slow radical chain scission, but they also influence hydrolysis because the halide component can generate acidic species at high temperature. Potassium bromide with copper iodide is effective in glass-filled PA6 at use temperatures below 130 °C. Above 150 °C, bromide may convert to volatile or water-soluble species that reduce stabiliser content and create surface pitting. Alternative hydrolysis stabilisers include carbodiimide additives that react with free acid end groups, lowering the acid concentration from 40 meq/kg to below 15 meq/kg during compounding. In one twin-screw compounding study using a 40 L/D extruder with melt temperature 275 °C, a polycarbodiimide masterbatch reduced acid end groups by 55% and extended time to 50% tensile strength loss in coolant at 125 °C by 400 h. However, carbodiimide must be added late in the process to avoid crosslinking reactions that raise melt viscosity and reduce flow length. The compounding window is narrow; too little additive fails to cap acid ends, while excess carbodiimide increases viscosity number above 160 mL/g and requires higher injection pressure. Processing at melt temperatures above 290 °C with carbodiimide can produce gel particles and surface defects. These interactions explain why stabilisation decisions cannot be separated from moulding parameters. Validation must include measurement of acid end groups, viscosity number, and tensile properties after ageing, not merely initial mechanical data.