When PA66 Went Short: A 90-Day PBT Switchover for a Customer Programme

Polybutylene terephthalate (PBT) was introduced as a candidate replacement when polyamide 66 (PA66) supply became constrained. The customer programme comprised injection-moulded connectors and sensor housings for automotive under-hood positions, and the changeover had to fit within a 90-day requalification window. Existing cavity steel had been dimensioned for PA66-GF30, so the substitution had to account for differences in melt rheology, shrinkage anisotropy, moisture response, hydrolytic stability, and flame-retardant additive behaviour. Screening compared unreinforced PBT and PBT-GF30 against the incumbent PA66-GF30 using ISO 527-2, ISO 178, ISO 75-1/-2, ISO 1133-1:2022, ISO 62, IEC 60112, and UL 94. The following technical boundaries emerged from processing trials and published datasheet comparisons.

Hydrolysis Stability Is Not a Secondary Concern in Hot-Wet Connectors

Water absorption after 24 h at 23°C under ISO 62 is substantially lower for PBT-GF30 than for PA66-GF30, typically 0.07 % to 0.15 % versus 1.0 % to 1.7 %. That lower equilibrium moisture uptake does not make PBT insensitive to water. The ester group in the PBT backbone is hydrolytically cleaved at melt-processing temperatures when hopper moisture exceeds 0.02 % by weight. Melt viscosity falls, tensile strength drops, and volatiles increase. For hot-wet service conditions such as water-glycol contact above 90°C, unreinforced PBT or standard PBT-GF30 can undergo progressive hydrolysis; a hydrolysis-stabilised grade is required unless component testing proves otherwise.

Melt-processing of PBT-GF30 therefore began with desiccant drying at 120°C for 4 h using a closed-loop dryer with a dew point of -40°C or lower. Return-air dew-point monitoring was used because hopper residence time and ambient humidity shift the final moisture condition. At moulding shop relative humidity above 60 %, open hoppers and unsealed vacuum loaders were not acceptable. The hopper was sized for 4 h residence time at maximum shot weight, and the system was purged with dry air during any extended stoppage. Pellets that accumulated in the machine throat without dry-air purge were discarded because they could exceed the 0.02 % moisture limit quickly at 23°C and 80 % relative humidity.

What Drying Boundaries Apply to a PBT-GF30 Switchover?

Drying must reduce moisture to 0.02 % or less before the melt phase. A desiccant dryer with a dew point of -40°C and a set point of 120°C for 4 h is the recognised starting point for glass-reinforced PBT. Hot-air tray dryers are generally insufficient because they cannot reach the target moisture under normal ambient conditions. In-line moisture analysis is preferred when lot-to-lot variation matters; melt quality on the press is better controlled by monitoring screw recovery time and cushion consistency than by appearance alone.

The melt-processing window for PBT-GF30 is bounded by 240°C and 270°C. Barrel zones were kept at 240°C to 255°C, with nozzle temperature 250°C to 260°C. Temperatures above 270°C cause thermal degradation of the ester backbone and increase the risk of volatile by-products and surface defects. Screw L/D ratio was 18:1 to 22:1 with a low-compression screw of 2:1 to 2.5:1. Mould temperature was maintained at 80°C to 90°C for close-tolerance connector parts because lower mould temperatures reduce crystallinity and produce post-mould dimensional drift. The 0.8 mm wall sections required high injection speed to avoid early gate freeze.

Injection moulding trials on a 32-cavity valve-gated hot-runner mould revealed the central processing conflict. The PA66 process had used a hot-runner manifold temperature of 285°C to 295°C; applying that thermal profile to PBT-GF30 produced valve-gate stringing, nozzle drool, and surface defects. The manifold set point was reduced to 255°C to 265°C, and the nozzle tips were trimmed to 250°C to 260°C. The mould was held at 80°C. Fill speed was kept high because the thin 0.8 mm sections froze quickly; slow fill generated short shots at the last cavity, while fast fill without adequate venting raised flash risk. Hold pressure was reduced below the PA66 level because PBT solidifies rapidly and gate freeze occurs sooner. Transfer position was adjusted to begin packing before the flow front decelerated in the runner. Published data for this exact hot-runner geometry is limited, so valve-pin sequencing required iterative adjustment rather than direct substitution of the PA66 parameter set.

The dimensional effect of the lower melt viscosity was most visible at the parting line. With the PA66 manifold temperature, PBT-GF30 flashed in areas where vent depth had been set for PA66. The solution was not an increase in clamp force alone; vent depth and hot-runner tip temperature had to be rebalanced. A 1,200 kN hydraulic machine was used for the trial, and clamp-force margin was maintained above 80 % of available capacity to avoid mould breathing. Injection pressure was sufficient to fill all cavities without exceeding 120 MPa hydraulic pressure. Screw rotation speed was kept below 100 min⁻¹ to limit shear heating and maintain melt homogeneity. Back pressure was set at 0.5 MPa to 1.0 MPa; higher back pressure increased melt temperature and caused resin yellowing at the screw tip.

Shrinkage Anisotropy and Comparative Tracking Index

Mould shrinkage for PBT-GF30 is anisotropic. Under ISO 294-4, flow-direction shrinkage typically ranges from 0.2 % to 0.4 % at 0.8 mm to 3.0 mm thickness, while transverse shrinkage is generally 0.5 % to 1.0 %. The incumbent PA66-GF30 exhibits a similarly anisotropic pattern, but the absolute values and the response to hold pressure differ. Because the existing cavity dimensions were fixed, the initial PBT-GF30 parts showed bore diameters close to the lower specification limit in connector pin openings. That condition was corrected by increasing mould temperature to 85°C and holding pack pressure long enough to prevent premature gate freeze; however, excessive pack pressure increased flash risk in the runner area.

Post-mould dimensional stability was monitored by thermal cycling from -40°C to 140°C under dry conditions. Post-mould annealing at 120°C for 2 h reduced residual shrinkage during subsequent exposure to 140°C. Comparative tracking index under IEC 60112 was a critical requirement for connector creepage distance. PBT-GF30 with a non-halogen flame-retardant package can maintain a CTI of 500 V to 600 V or higher, depending on glass content and additive composition. This is an advantage over some PA66 grades that fall in the 400 V to 500 V range when reinforced and flame-retarded. Surface resistivity under IEC 60093 and volume resistivity were recorded at 23°C and 50 % relative humidity because PBT has low moisture uptake and stable electrical behaviour in dry and normal service conditions. For high-voltage connector spacing, the CTI class determined minimum creepage distances; the PBT-GF30 candidate was not accepted solely on datasheet values but was verified at the moulded thickness of 0.8 mm after thermal ageing.

When UL 94 V-0 at 0.8 mm Is Required in a Non-Halogen Programme

A non-halogen phosphorus-based flame-retardant PBT-GF30 was screened because the customer programme prohibited polybrominated diphenyl ethers and antimony trioxide combinations. The selected grade met UL 94 V-0 at 0.8 mm after conditioning under UL 94 requirements. The processing window narrowed: barrel temperatures above 255°C produced visible depolymerisation odour and mould deposit, and residence time above 5 min darkened the melt. A flat barrel profile of 235°C to 245°C was used, with nozzle temperature 245°C to 250°C. Screw speed was held below 100 min⁻¹. Short cycle time and reduced shot capacity were preferred; using a barrel size where the shot was only 20 % to 30 % of capacity increased residence time and led to degradation, so a machine with shot size closer to 40 % to 60 % was selected.

Purging required special attention. Residual polyamide from previous production runs or amine-containing purging compounds should not be introduced into the PBT melt. Amine residues can accelerate ester-chain breakdown and reduce molecular weight. The changeover protocol used a PBT-compatible purging compound, followed by clean PBT-GF30 purge until melt temperature and screw recovery stabilised. The material handling system was isolated to prevent PA66 fines from contaminating the PBT feed. Regrind use was restricted after spiral-flow testing and only when the regrind fraction was dried to the same 0.02 % moisture specification. High regrind levels decreased tensile elongation at break and increased lot-to-lot variability.

The remaining boundary is service environment. PBT-GF30 exhibits good dimensional stability in dry thermal cycling, but continuous hot-water or water-glycol exposure above 90°C is not covered by standard datasheet values. Hydrolysis-stabilised grades and component-level testing according to the applicable coolant exposure standard are required before production release. Published long-term data for this specific connector geometry under hot-wet conditions is limited, and the 90-day requalification did not replace full vehicle validation.