Tracking Failure in the Field: When Polybutylene Terephthalate Insulation Carbonises

Field returns from automotive under-hood connectors, industrial terminal blocks, and power electronic housings have identified a recurrent failure mode in which polybutylene terephthalate insulation darkens, loses surface resistivity, and forms a conductive char. The visible carbonised track is not the result of a single-pulse dielectric breakdown; it is the endpoint of a progressive surface degradation sequence involving contamination, moisture film formation, leakage current, Joule heating, pyrolysis of the polyester matrix, and carbonaceous consolidation. Electrical stress is often below the short-term dielectric strength of the moulded part. Standard withstand-voltage screening therefore may pass parts that later fail in service after weeks or months of intermittent condensation and voltage hold, particularly where connectors are exposed to salt spray, flux residues, or ionic fingerprints.

What Converts Surface Leakage into Conductive Carbon in PBT Insulation?

At the molecular level, PBT degradation during tracking is initiated by thermally driven ester scission. When a surface moisture film becomes conductive through dissolved ions, the resulting leakage current raises local temperature at points of highest current density. Dry bands form because film evaporation reduces conductivity; the electric field across the dry gap then initiates small arc discharges. These arcs deliver temperatures far above the decomposition threshold of PBT, which is commonly reported between 380 °C and 410 °C under nitrogen in thermogravimetric analysis. The polyester backbone undergoes random chain scission, releasing tetrahydrofuran, butadiene, carbon monoxide, carbon dioxide, and aromatic fragments. Once carbonaceous char forms, surface resistivity drops from an initial insulation range typically above 10¹² Ω per ASTM D257 into a semiconducting or conductive regime. Further current flow concentrates at the char edge, extending the track through repeated pyrolysis.

Arc discharge temperatures are estimated to exceed 1000 °C at the root. This is sufficient to convert carbonaceous polymer into conductive domains with graphitic character. The process is self-accelerating: conductivity increases locally, leakage current increases, dry-band arcing intensifies, and new carbon is deposited. In field parts, the track path often follows a boundary between moulded surfaces and metallic inserts or along weld lines where the surface is less dense and more susceptible to carbon retention.

Under IEC 60112, comparative tracking index is measured using 0.1 % ammonium chloride solution with droplets of approximately 20 mm³ applied every 30 s between platinum electrodes separated by 4 mm. The objective is to determine the highest voltage at which no tracking occurs over 50 drops. The method uses clean, moulded specimens; it does not simulate aged surfaces, industrial film contamination, thermally degraded contact regions, or persistent contaminant flow.

On injection-moulding lines producing PBT terminal blocks and connector housings, the processing window itself can create surface conditions that reduce tracking resistance. Barrel temperatures are typically set between 240 °C and 270 °C, with mould temperatures held at 60–100 °C depending on crystallinity targets and dimensional tolerances. Melt residence times above 8–10 min at the upper end of this range produce yellowing and local polymer decomposition; degraded surface films have a lower comparative tracking index than the bulk. High injection speeds and cold mould surfaces freeze in flow-front skins that may be resin-rich or glass-depleted. In glass-fibre-reinforced PBT moulded on hot-runner systems, the orientation of fibres at the surface can create capillary paths for moisture and ionic contamination. Weld-line regions in multi-gate tools show higher fibre exposure and lower surface smoothness, and field failures are frequently observed to initiate along these boundaries.

Pre-drying of PBT pellets before injection moulding is required at 120 °C for 3–4 h to reduce moisture below 0.02 %. Hydrolytic molecular-weight reduction during compounding and moulding lowers melt viscosity and creates carboxylic acid end groups that persist in the moulded part. A shift in melt flow index measured to ISO 1133-1 is an early indication of processing-induced degradation. Parts moulded from poorly dried material may show surface defects, but more importantly they carry a degraded surface layer with reduced tracking resistance.

When Thermal Aging Precedes Electrical Stress

Long-term heat aging changes the tracking response of PBT because oxidation lowers the activation energy for surface pyrolysis and creates microcracks that retain contamination. Laboratory samples exposed to air oven aging at 140 °C for 1000 h often show surface carbonyl formation and a measurable reduction in molecular weight. The embrittled skin is more easily abraded during connector mating; abraded polymer particles then mix with moisture and salts to form deposits at the base of contacts. If the compound is not hydrolytically stabilised, high-humidity exposure accelerates ester hydrolysis. Aging under 85 °C/85 % RH conditions, commonly specified in automotive qualification programmes, can produce substantial strength loss in unprotected PBT grades. Representative supplier data indicate that a hydrolytically stabilised grade may retain more than 70 % of initial tensile strength after 2000 h, while an unstabilised grade can fall below 50 % depending on wall thickness and carboxylic acid end-group concentration.

Filler geometry influences carbon track propagation. Glass-fibre bundles oriented in the flow direction create continuous interfaces that guide leakage current. The tracking path in a 30 wt% glass-reinforced PBT returning from an engine bay environment is often observed to follow exposed fibre ends at the surface rather than traversing the polymer matrix directly. Mineral-filled PBT grades, especially those containing talc or mica, may show shorter discontinuous tracks because the mineral platelets interrupt phase continuity, but water absorption increases if filler-matrix adhesion is poor. Carbon black pigmented PBT grades present an additional diagnostic problem because the filler itself is conductive; surface electrical readings cannot always distinguish carbon black from pyrolytic char. For high-voltage insulation applications, carbon black levels are minimised or replaced with non-conductive organic pigments to preserve surface resistance.

Flame-Retardant Package Effects on Tracking Resistance

The choice of flame-retardant system determines whether a PBT compound retains high tracking resistance after moulding. Unfilled PBT without flame retardant can achieve 600 V comparative tracking index under IEC 60112, placing it in PLC 0 per UL 746A and material group I per IEC 60664-1. Glass-fibre reinforcement alone reduces this performance in many formulations because the fibre-matrix interface creates higher surface roughness and can wick electrolyte. Brominated flame retardants combined with antimony trioxide are widely used to meet UL 94 V-0 at 0.8 mm, but these packages often lower CTI into the 175–249 V range. The presence of antimony halides and halogen acid species under discharge conditions increases ionic conductivity and promotes carbon formation. Phosphinate-based flame-retardant systems, by contrast, are reported to maintain CTI at 600 V in certain reinforced PBT compounds. The trade-off is not purely electrical: brominated systems may offer lower compound density and lower cost, while phosphinate systems may require higher loadings and different processing conditions to avoid screw wear and mould deposit formation.

Designers applying PBT insulators in power electronic assemblies verify clearance and creepage distances using IEC 60664-1. If a reinforced flame-retarded PBT is assigned to PLC 2 or PLC 3, the required creepage distance for an 800 V dc system under pollution degree 3 is larger than for an unfilled PLC 0 grade. Carbonisation failure in field returns is therefore not necessarily a resin defect; it can be the result of selecting a flame-retarded PBT grade with lower comparative tracking index while retaining the same package geometry as a higher-CTI material. For creepage-limited connector designs, changing to a phosphinate-stabilised grade with CTI ≥ 600 V can reduce the required creepage path depending on the insulation coordination table selected.

Mould Release Agents, Ionic Contamination, and Track Initiation

External mould releases containing silicone or metallic stearates leave insulating or hygroscopic films that can mask surface defects during production and later participate in leakage-current development. In field-return analysis, tracking paths on PBT connectors have been associated with residues containing silicon, sodium, chloride, and sulphur. These residues absorb moisture and form an electrolyte film at relative humidity levels above 60 % or during dew-point transitions. The resulting surface resistivity can fall from above 10¹² Ω to below 10⁶ Ω, even before arcing begins. Cleaning processes for high-voltage PBT insulators therefore require verification by ion chromatography or surface extract conductivity tests. Solvent wiping with isopropanol does not reliably remove ionic contamination from glass-filled surfaces because the fibre-matrix interface retains salts below the apparent top surface.

Hydrolytic degradation produces carboxylic acid end groups that increase surface acidity and attract moisture. This accelerates the ion transport that precedes dry-band arcing. In PBT exposed to water-glycol mixtures or aqueous urea solutions in under-hood systems, the ester linkage is attacked, generating terephthalic acid and tetramethylene glycol fragments. The resulting oligomeric species can migrate to the surface and form a residue that traps dust and salts. The relationship between hydrolysis and tracking is particularly severe in high-voltage connectors used in battery packs, where coolant leakage can expose the polymer to glycol-containing mixtures at elevated temperature. In such cases, the CTI measured on a new, dry part is a weak predictor of service performance.

At 800 V dc charging stations, field replacements of PBT connector housings have shown carbonised tracks radiating from contact retention slots and from the edges of seal interfaces. The conductive path width is typically between 0.2 mm and 1.0 mm, with char extending into the bulk to depths of 0.1–0.4 mm depending on arc energy. Optical microscopy of cross-sectioned failures reveals aligned porosity and exposed glass fibres in the char zone. Scanning electron microscopy with energy-dispersive X-ray analysis identifies carbon, oxygen, and frequently chlorine, sodium, or sulphur species embedded in the blackened layer. The carbon is not uniformly graphitic; Raman spectroscopy often shows disordered carbon with D-band and G-band maxima near 1350 cm⁻¹ and 1590 cm⁻¹. These features are consistent with pyrolytic degradation rather than metallic dendrite growth.

Which Test Methods Distinguish Surface Tracking from Carbonisation?

Tracking resistance is assessed by two categories of standardised electrical tests. The first category measures low-voltage proof under simulated contamination droplets; the second category measures high-voltage inclined-plane tracking and erosion under continuous contaminant flow. Field carbonisation of PBT insulation is better represented by inclined-plane methods because the contaminant film is replenished and dry-band arcing persists for longer durations. However, most material specifications for PBT connectors still rely on comparative tracking index as the primary metric.

Failure-analysis laboratories typically measure the comparative tracking index of the returned moulding, but the service-aged CTI is usually lower than the virgin datasheet value. The aged surface is therefore sectioned and examined under reflected-light microscopy at magnifications from 50× to 500×. Fourier-transform infrared spectroscopy of the char zone reveals loss of the ester carbonyl near 1715 cm⁻¹ and the appearance of aromatic and unsaturated carbon bands. Differential scanning calorimetry of failed parts often shows a shifted melting peak and broadened crystallinity distribution due to surface reheating. All findings are compared against control parts from the same production lot. The root cause is frequently assigned to a combination of reduced creepage distance below the IEC 60664-1 design minimum for the material group, ionic contamination from installation or moulding, and a flame-retardant package that lowered tracking resistance.