Electrical tracking resistance in thermoplastic components for high-voltage battery modules is not determined solely by the base polymer dielectric strength. The as-moulded surface of a cell separator, busbar carrier, or terminal insulator is exposed to a combined electrical and chemical stress that is reproduced by the comparative tracking index test defined in IEC 60112:2003 + A1:2009 and ASTM D3638-21. In a traction battery module operating at 400 V DC to 800 V DC, with high-voltage busbar potentials and occasional switching transients, surface contamination from condensation, electrolyte residues, metal wear debris, and glycol from cooling-line leaks creates a weakly conductive electrolyte film. The resulting leakage current concentrates at electrode gaps, heats the polymer surface by Joule dissipation, decomposes the matrix, and leaves a carbonaceous residue with higher conductivity than the original resin. Once a partial track forms, current density increases and the track propagates along creepage paths, often across a busbar retainer rib or around a threaded insert boss, until a phase-to-phase or phase-to-ground fault occurs. The CTI value obtained under the standard is therefore not a bulk dielectric strength measurement but a system-dependent pass/fail voltage with a specific electrolyte and electrode geometry. In battery module design, creepage distances are selected from IEC 60664-1:2020 using the material group derived from CTI, so any loss of tracking resistance after moulding, ageing, contamination, or physical modification of the surface directly changes the required minimum creepage distance. A material that is specified only by its supplier datasheet CTI, without verification on the moulded component surface and after thermal cycling, can comply with the initial design but fail in service at a level below the dimensional safety margin.

The standard method places two platinum chisel electrodes, 4.0 mm apart, on a moulded or cut plaque with a 1.00 N contact force. A 20 mm³ droplet of electrolyte solution A, 0.1 % ammonium chloride in deionized water with a resistivity of 3.95 Ω·m at 23 °C, is dispensed between the electrodes every 30 s. The test voltage is held at the selected level until either 50 drops have been applied without tracking or a persistent overcurrent condition occurs. The maximum voltage at which the material withstands 50 drops without evidence of tracking is reported as the comparative tracking index. Because the test outcome depends on surface finish, moisture content, filler distribution, flame retardant package, and plate-out, values measured on polished plaques may overestimate the performance of a textured moulded rib. The material groups defined by IEC 60664-1 and the performance level categories of UL 746A convert the CTI result into a creepage-distance selection input. Table 1 lists the accepted CTI intervals for the four IEC material groups and the corresponding UL performance level categories.

IEC 60664-1 material group	CTI interval	UL 746A PLC category
I	≥ 600 V	PLC 0
II	400 V to < 600 V	PLC 1
IIIa	175 V to < 400 V	PLC 2 and PLC 3
IIIb	100 V to < 175 V	PLC 4

How Does a Conductive Track Emerge on a Contaminated Thermoplastic Surface?

The initiation of electrical tracking on a thermoplastic surface proceeds through several coupled regimes. First, the electrolyte film provides ionic conduction, which is temperature-dependent and increases as the surface becomes wetter. The applied field between busbar and ground or between adjacent busbars drives the electrolyte to carry current across the narrow insulating gap. Second, Joule heating occurs in the film and at microscopic high-current-density spots where the electrolyte layer is thinnest. The polymer surface temperature may exceed the onset of thermal decomposition, typically above 250 °C for aliphatic polyamides and above 300 °C for partially aromatic grades, depending on heating rate and atmosphere. Third, decomposition products from the polymer, glass sizing, pigments, heat stabilizers, and flame retardants form a carbon-rich residue. Aromatic structures and unsaturated backbone segments cyclize and condense into graphitic or turbostratic carbon, which has an electrical resistivity orders of magnitude lower than the original polymer. Fourth, the dry band region around the carbonized spot concentrates the electric field, and the conductive residue acts as an electrode extension. The process repeats in discrete steps as the electrolyte re-wets the surface, the conductive track lengthens, and the creepage distance effectively shrinks. In glass-fibre-reinforced grades, the exposed glass strand acts as a capillary wick and a local mechanical discontinuity, while the silane sizing decomposition products can accelerate matrix breakdown. Tracking failure therefore does not require the bulk material to carbonize uniformly; a thin surface layer on a rib or weld line is sufficient to create a permanent low-resistance path. The thermal conductivity of the base polymer and filler system affects the local temperature rise, which is why mineral-filled grades may show higher tracking resistance at the same CTI value than thin unfilled sections under identical test conditions.

The CTI test uses an alternating current potential, whereas the service potential in a high-voltage battery module is direct current. Direct current tracking can involve electrochemical degradation mechanisms such as anodic oxidation, electrolytic metal dendrite formation from dissolved busbar metal or plating, and slower thermal runaway. Published data for direct-current tracking under battery-specific contamination are limited, but the IEC method remains the accepted screening test because battery management system switching transients and inverter-fed voltages introduce alternating or pulsating components. The test surface is normally a moulded plaque 3 mm thick; if a component creepage path crosses a weld line, a mating surface, or a gate vestige, the standard plaque does not reproduce these discontinuities. For that reason, internal verification on production mouldings with the same surface texture and insert geometry is required before final creepage distances are frozen. A moulded surface with texture depth above 25 µm, measured according to ISO 21920, may retain more electrolyte and may produce lower CTI than a smooth machined surface, depending on the wetting angle of the resin.

Flame Retardant Chemistry Determines Residue Conductivity More Than Base Polymer Electrical Strength

A material comparative tracking index is frequently dominated by the flame retardant package, not by the resin dielectric strength. Brominated polystyrene and other brominated flame retardants combined with antimony trioxide are effective at limiting ignition in polyamides and polyesters, but they tend to produce dense, carbonaceous decomposition products during surface arcing. As a result, PBT GF30 grades containing brominated flame retardants typically exhibit CTI values between 175 V and 250 V under IEC 60112, placing them in material group IIIa. In contrast, halogen-free PA66 GF30 formulations using melamine polyphosphate and organic metal phosphinate stabilizers can retain CTI values above 500 V, and some grades reach 600 V. The difference is not simply the presence of halogens; it is the char conductivity and the extent of intumescent or gas-phase activity during decomposition. A flame retardant that promotes a strong, insulating, expanded char can disrupt the tracking path, whereas a flame retardant that forms a dense aromatic char with low volatile content can create a semiconductive bridge. The effect of filler content is non-linear: adding glass fibre generally lowers CTI relative to the unfilled resin, but the magnitude depends on the flame retardant system and the glass sizing. A 30 wt% glass-fibre addition to an unfilled PA66 with CTI 600 V can reduce the value to 475 V or 400 V, while the same glass-fibre addition to a brominated PBT system may produce less relative change because the matrix already fails at a lower tracking voltage.

Mineral and glass bead fillers affect tracking resistance differently. Calcium carbonate and talc can create high-surface-energy interfaces and can absorb stabilizer components, potentially reducing CTI; however, high-purity talc in polypropylene often retains CTI values of 600 V because the resin decomposes to volatile aliphatic fragments rather than aromatic conductive char. Carbon black, conductive graphite, and metallic pigments must be excluded from creepage zones because they lower surface resistivity and can reduce CTI below 100 V. Even small quantities of carbon black used for colouring can shift a material group from I to IIIa if the dispersion is poor or the loading exceeds the percolation threshold of the particular polymer system. Mould release agents and slip additives present another risk: an additive that migrates to the surface may increase initial hydrophobicity and surface resistivity, but subsequent thermal cycling can re-distribute the additive into island-like films that increase electrolyte wetting heterogeneity. The resulting non-uniform film allows dry-band formation and concentrates leakage current in smaller areas, accelerating tracking initiation.

Across commercially available injection moulding grades used in high-voltage battery modules, the comparative tracking index varies widely. Unfilled, heat-stabilized PA66 typically achieves CTI 600 V and is classified as material group I, but moisture absorption and hydrolytic ageing require molecular weight retention. PA66 GF30 with halogen-free flame retardant may exhibit CTI between 475 V and 600 V, while contact with 85 °C and 85 % relative humidity for 1,000 h can reduce the value to 400 V or below depending on the stabilizer system. PPA GF30 and PA6T/6I GF30 can combine similar CTI values with lower moisture uptake and higher glass transition temperatures, making them suitable for busbar carriers that must survive 150 °C hot-spot temperatures. PBT GF30 is used extensively for terminal insulators and connectors because of its dimensional stability, low moisture uptake, and fast crystallization, but its halogenated versions often sit in material group IIIa; halogen-free PBT grades can reach 400 V but may sacrifice mechanical strength. PPS GF40 provides high chemical resistance and high continuous use temperature, but its CTI is typically 125 V to 175 V, placing it in material group IIIb. PEI GF30 and PES can exhibit CTI values between 150 V and 175 V, limiting their use in live parts unless creepage distances are increased and field strengths are reduced. LCP offers thin-wall flow and low coefficient of thermal expansion, but its as-moulded surface is highly anisotropic; CTI values are typically 175 V to 225 V, and supplier datasheets often report different values parallel and perpendicular to melt flow. Table 2 summarizes representative CTI ranges for unaged injection moulded specimens under IEC 60112 solution A, compiled from public material supplier datasheets.

Polymer system	Filler/flame-retardant	CTI range	IEC material group / UL PLC
Unfilled heat-stabilized PA66	None	600 V	Group I / PLC 0
PA66 GF30	Halogen-free FR	475–600 V	Group II to I / PLC 1 to PLC 0
PPA GF30	Halogen-free FR	475–600 V	Group II to I / PLC 1 to PLC 0
PBT GF30	Brominated FR	175–250 V	Group IIIa / PLC 3 to PLC 2
PBT GF30	Halogen-free FR	250–400 V	Group IIIa to II / PLC 2 to PLC 1
PPS GF40	Glass-filled	125–175 V	Group IIIb / PLC 4
PEI GF30	Glass-filled	150–175 V	Group IIIb / PLC 4
LCP GF30	Glass/mineral	175–225 V	Group IIIa / PLC 3
PP GF20 mineral	Talc/mineral	500–600 V	Group II to I / PLC 1 to PLC 0

These ranges are not design limits; part geometry, pigment, stabilizer, and moulding parameters shift the actual value on a production component.

If the Moulding Process Alters the Surface Skin, Then Tracking Test Values From Plaques Will Misrepresent the Part

If the injection moulding process changes the outer skin composition, then CTI values measured on flat plaques can misrepresent the component. During injection moulding of glass-fibre-reinforced thermoplastics, the resin-rich surface forms when low-viscosity melt slips against the cavity wall. At high injection velocities above 300 mm/s or under high shear, glass fibres align in the flow direction and can break through the surface skin, especially at the end of fill and at weld lines. The exposed or shallow glass strands create local capillaries for the electrolyte, and the silane sizing decomposes under heating to form a lower-viscosity degraded layer. Weld lines formed downstream of core pins or around insert bosses contain elevated glass-fibre concentration at the meeting front. The CTI value at a weld line can be more than 100 V lower than the surrounding moulded surface, but the standard IEC 60112 plaque without an intentional weld line does not capture this loss.

Mould release agents, anti-slip additives, and external lubricants extend the discrepancy between plaque and component. Fatty acid ester or amide release additives migrate to the surface during cooling and can create a monolayer that increases the water contact angle and artificially raises the initial CTI; after thermal cycling or solvent exposure, the film breaks into isolated islands and the CTI falls. Silicone-based external mould release agents are particularly persistent and can produce a high initial surface resistivity, but they also create non-uniform wetting and local dry bands. A production controller may observe batch-to-batch CTI shifts of 50 V to 100 V when mould release agent concentration, purge cycles, or screw residence time change. For PA66, pre-drying is mandatory when relative humidity exceeds 60 %; residual moisture above 0.15 wt% at 280 °C melt temperature leads to chain scission during plastication, reducing molecular weight and ageing resistance. A twin-screw extruder with an L/D ratio of 40:1 and a side feeder for glass fibre is commonly used for compounding; if the screw speed is too high, local melt temperature can exceed 330 °C in PA66, and the resulting degradation products lower CTI.

After moulding, workers do not normally clean every busbar carrier with solvents, but if cleaning is applied, the solvent type matters. Isopropyl alcohol or 2-propanol is preferred for removing ionic contamination, while petroleum-derived hydrocarbon cleaners can leave low-volatility residues that reduce tracking resistance. Ionic contamination from machining, cutting, or handling should be limited to below 5 µg NaCl equivalent/cm² in high-voltage creepage zones, based on IPC-TM-650 method 2.3.25, because tramp ions increase electrolyte film conductivity. Laser marking can carbonize thin surface layers, and the resulting black marking should be excluded from the shortest creepage path by at least 2.0 mm unless the marking is verified not to reduce CTI.

Creepage Dimensioning Inputs for High-Voltage Busbar Carriers and Cell Separators

Creepage distances in a battery module are dimensioned according to IEC 60664-1:2020, using the working voltage, pollution degree, installation altitude, and material group. The material group is derived from CTI but is not a direct multiplier of voltage. A design that selects material group I because the as-moulded CTI is 600 V may become non-compliant if the same component loses CTI to 400 V after moisture ageing. The upper voltage boundary of the battery module, typically 800 V DC for high-performance systems, must be treated as the root mean square value for AC transients when creepage distances are selected under IEC 60664-1. Battery modules mounted at altitudes above 2,000 m require derating because the reduced air density lowers clearance and creepage performance. Pollution degree 2 applies when only non-conductive dust or occasional condensation can occur; many sealed modules are designed to pollution degree 2, but external connectors and vents may require pollution degree 3. The difference in required creepage distance between material group I and material group IIIa is significant at 800 V under pollution degree 3, and the standard table should be consulted rather than applying a fixed clearance-to-voltage ratio.

On a busbar carrier, the shortest creepage path is often not the straight-line distance between two copper busbars. The path may wrap around a rib, follow a snap-fit groove, cross a heat-stake weld line, or travel through a clearance hole for a threaded insert. A three-dimensional CAD creepage analysis must include all possible surface paths shorter than the intended clearance, especially where the component sits against another insulating surface or a metal cooling plate. The use of conformal coatings or potting can modify the effective surface because the coating material has its own CTI and may crack during thermal cycling. Conformal coating applied only to flat areas may leave uncoated rib edges that become the new creepage path. If a component is overmoulded with a low-CTI thermoplastic elastomer for vibration damping, the overmould layer must be included in the creepage path calculation because it reduces the effective insulating surface. Busbar edge radius and terminal geometry also affect local field stress; sharp edges increase the electric field within the electrolyte film and can reduce the voltage at which tracking initiates.

In a prismatic cell module, the cell separator must perform mechanical compression, electrical isolation between adjacent cells, and thermal insulation. In modules with cell surface potentials close to 4.0 V per cell and series strings up to 800 V, the separator is not subjected to the full pack voltage across a single gap but must withstand local voltage differences between adjacent cells and the cooling plate. Material choices such as PA66 GF30 and PPA GF30 are common because they can meet CTI 600 V and 475 V to 600 V, respectively. However, the separator ribs and weld lines are formed around alignment pins and may create low-tracking paths. The injection mould should be gated so that the rib bottom, not the rib tip, is the filling path; otherwise, the rib tip may contain high glass concentration and exposed glass ends. If a separator is ultrasonically welded or heat-staked, the locally remelted flash and squeeze-out can have lower molecular weight and altered filler orientation, reducing CTI at the stake location. The weld flash should be removed or the creepage path should be routed away from the stake area by at least 3.0 mm.

Terminal insulators and high-voltage interlock connectors are moulded in smaller wall thicknesses, often 0.8 mm to 1.2 mm, with press-fit or overmoulded brass inserts. The combination of high clamp force and thin wall creates local stress concentrations that can craze under thermal cycling, and crazes act as electrolyte wicking paths. In connectors exposed to external environment, the conditions may reach pollution degree 3 and condensation with road de-icing salt, which is significantly more conductive than the 0.1 % ammonium chloride solution used in IEC 60112. The presence of sodium chloride and calcium chloride can accelerate electrochemical degradation and produce chlorine-containing decomposition species. A terminal insulator in PBT GF30 with CTI 250 V may be acceptable for a 400 V system with generous creepage, but for an 800 V system, a PPA GF30 with CTI 475 V to 600 V is preferred if the module is not hermetically sealed.

For overmoulded busbars and sensor carriers, the thermoplastic must adhere to copper or aluminium conductors and survive thermal expansion mismatch. The interface between rigid busbar and polymer creates a microgap during thermal cycling, and the gap can fill with condensation. The resulting narrow electrolyte channel concentrates the field and creates a tracking path along the conductor surface. In such designs, adhesion promoters and surface roughening alone do not eliminate the microgap; the creepage distance should be based on the assumption that the interface is wetted. Silane coupling agents can improve initial adhesion but may hydrolyze after 1,000 h at 85 °C and 85 % relative humidity. Published data for this specific configuration are limited, but the failure mode has been observed in production in the form of blackened tracks along overmoulded copper busbar edges after environmental cycling.

What Operational Boundaries Apply to Hydrolytically Stabilised PA66-GF30 in 800 V Modules?

Hydrolytically stabilized PA66 GF30 can be used in 800 V modules, but operational boundaries must be enforced. The material must be pre-dried to a residual moisture level below 0.10 wt% to 0.15 wt% before processing, and melt residence time at 290 °C should be kept below 10 min to avoid thermal degradation. The maximum continuous use temperature in air is typically 120 °C to 130 °C for heat-stabilized grades, but hot-spot temperatures against busbars can reach 150 °C for short periods. Above 150 °C, oxidative embrittlement and loss of tracking resistance can occur rapidly. The material is not recommended for continuous exposure to glycol-based coolants because ethylene glycol can plasticize the polyamide and reduce surface resistivity; when a cooling-line leak wets a PA66 live part, the creepage path becomes more conductive and the CTI is no longer representative. Combinations with amine-based lubricants or certain amine-based heat stabilizers can cause premature crosslinking and discolouration during processing; the formulation supplier should be consulted if an external lubricant is added. If a module design requires a material with CTI 600 V after 2,000 h at 85 °C and 85 % relative humidity, PA66 GF30 may be insufficient unless the grade is specifically formulated for hydrolytic stability and the creepage distance is derived from aged CTI. In that case, PPA GF30 or a polyphthalamide blend is the next material step.

In service, PA66 GF30 should not be exposed to alkaline electrolyte leakage from a cell. Lithium-ion cells can vent electrolyte containing lithium hexafluorophosphate that hydrolyzes to hydrofluoric acid in humid air; this acid attacks glass fibre and polyamide, and the resulting surface deposits reduce tracking resistance. If a cell vent path can direct electrolyte mist onto a busbar carrier, the carrier should be made from a fluoropolymer-lined or higher-chemical-resistance polymer such as PPS or LCP, despite their lower CTI values. The design must then increase creepage distance to compensate. No material combines maximum CTI, maximum thermal resistance, and maximum chemical resistance in a single injection-mouldable grade; the selection must be made against the actual vent flow, condensation behaviour, and thermal load in the module.