For round copper conductors with nominal diameter in the range 0.100 mm to 5.000 mm, the thermal class assigned to a trimellitic anhydride-based polyamideimide enamel layer is not determined by a single chemical degradation threshold but by a system-level interaction of cure conversion, solvent retention, interfacial oxide chemistry, film thickness distribution, and the standardized ageing environment. The polymer backbone formed by the condensation of trimellitic anhydride with aromatic diisocyanates and diamines contains imide and amide segments that give the cured film a room-temperature elongation of 25% to 40% on 0.400 mm diameter wire, yet this initial flexibility does not guarantee long-term thermal endurance. According to the procedures of ASTM D2307, thermal class means the temperature at which the extrapolated time to failure reaches 20,000 h for a selected retention of dielectric strength or flexibility after twisted-pair ageing. For commercial TMA-PAI enamel on round copper, values of 200 °C and 220 °C are attainable only when the degree of imidization exceeds 92%, the residual N-methyl-2-pyrrolidone content is kept below 0.5 wt%, and the maximum eccentricity ratio is held below 1.3. In practice, deviations in any of these parameters shift the ageing curve downward in a manner that is more severe than a simple Arrhenius translation, because the film undergoes multiple simultaneous degradation mechanisms: thermo-oxidative amide scission at the copper interface, plasticization by retained solvent, differential thermal expansion cracking, and partial discharge erosion at voids formed by overbaking. The result is that a wire lot may pass initial dielectric breakdown at room temperature and still fail thermal endurance testing at 200 °C before 5,000 h if the underbake condition remains undetected.

When Partial Cure Leaves Residual NMP in the Enamel Microstructure

When partial cure leaves residual N-methyl-2-pyrrolidone in the enamel microstructure, the limiting factor shifts from backbone decomposition to plasticization and vapour pressure accumulation at the copper-enamel boundary. NMP has a normal boiling point of 202 °C at 101.3 kPa; at 220 °C its vapour pressure is approximately 130 kPa, while a fully imidized TMA-PAI film typically develops a copper adhesion strength between 8 MPa and 12 MPa. A retained solvent content above 0.8 wt% therefore generates an interfacial tensile stress that exceeds the adhesion limit during the first high-temperature thermal shock cycle, producing blistering and local delamination. Thermal gravimetric analysis of underbaked enamelled wire frequently shows a two-step mass loss: the first between 200 °C and 300 °C corresponds to residual solvent and water volatilization, while the second above 450 °C corresponds to polymer decomposition; when the first mass loss exceeds 0.7%, the sample seldom survives ageing at 220 °C beyond 5,000 h. In addition, residual NMP acts as a plasticizer that reduces the glass transition temperature from the fully cured range of 260 °C to 280 °C to as low as 225 °C, which means the enamel enters its rubbery plateau precisely in the intended continuous operating range and permits carbon dust or metallic wear debris to embed more readily in the softened surface. The cure window for preventing this condition is narrow: zone temperatures below 340 °C leave too much amic acid and solvent, while zone temperatures above 430 °C on copper wire below 0.300 mm initiate surface oxidation and copper-catalysed decomposition before imidization is complete. A further complication arises from the fact that the outer surface of an uncured pass can densify before the inner region, trapping solvent in a low-permeability shell that is not removed by later passes.

Because the copper-enamel interphase is not a passive boundary, migration of copper oxide and ionic copper into the polymer matrix during high-temperature ageing accelerates thermo-oxidative scission of the amide segments more rapidly than bulk pyrolysis of the imide ring. On a 1.000 mm round conductor with a 45 µm single-coat TMA-PAI build, energy-dispersive X-ray mapping after 2,000 h at 220 °C in air frequently shows copper enrichment of 2 wt% to 4 wt% in the first 5 µm above the substrate, while the outer surface remains at bulk composition; published data for the exact enrichment gradient is limited because it depends on the oxide layer condition, enamelling speed, and pass schedule. This diffusion gradient creates a two-layer dielectric system in which the inner layer has a higher dielectric constant and lower volume resistivity, shifting electrical stress toward the outer intact region and reducing partial discharge inception voltage under inverter-type waveforms. The presence of copper at the interface also lowers the activation energy of thermo-oxidative degradation from approximately 110 kJ/mol for the unaged polymer to 80 kJ/mol for the metal-contaminated inner layer, which means that the limiting thermal class measured on twisted pairs may be dominated by failures that initiate in the first few micrometres rather than by bulk film thinning. Suppression of this interfacial mechanism requires a primer or an oxide conditioning step that produces a compact cuprous oxide layer less than 0.5 µm thick; if the oxide is too thick or loosely adherent, the aged enamel exhibits longitudinal cracks that follow the wire drawing direction and cause dielectric failure at bend test radii below 2d. The interphase effect is particularly severe in round conductors with high surface roughness, because asperities create local areas of high copper diffusion flux and reduce the effective contact area for enamel adhesion.

What Limits Thermal Shock Resistance After 20,000 Hours of Ageing?

Thermal shock resistance after long-term thermal exposure is limited not by the initial elongation of the enamel but by the loss of elongation caused by secondary crystallization, oxidative crosslinking, and chain scission at the copper-enamel interface. A round copper conductor has a coefficient of linear thermal expansion of approximately 17 µm/(m·K), while the cured TMA-PAI film has a coefficient between 30 µm/(m·K) and 40 µm/(m·K); this mismatch produces interfacial shear stresses that are partially reversible at short times but become locked in as the film modulus rises from 3.0 GPa to 4.5 GPa after 20,000 h at 200 °C. Aged film failure under thermal shock is typically observed as radial or longitudinal cracks when the specimen is subjected to a rapid temperature change from 220 °C to 20 °C followed by mandrel winding; the failure location correlates with maximum film thickness or with pre-existing wire guide abrasion sites. The thermal shock acceptance criterion in IEC 60317-26 uses a mandrel diameter based on the conductor diameter, but for aged wire the more severe discriminator is the retention of elongation, which should exceed 50% of the initial value after ageing if the wire is intended for high-speed automated winding. If the enamel passes through its glass transition during the thermal shock, the stress cannot relax quickly enough and crazing initiates at surface defects where the local coating thickness is below 15 µm. Thus, thermal shock resistance after ageing is actually a function of both the chemical stability of the polymer and the geometric uniformity of the enamel layer; a high average enamel build cannot compensate for a low minimum build at the point of stress concentration. In practice, aged TMA-PAI films that retain more than 70% of their original ultimate elongation at 200 °C are more likely to withstand a 1d mandrel bend after thermal shock, while those that drop below 40% fail even at 4d.

Material blending before application is straightforward; the mill base is dispersed in NMP and filtered to remove particles above 5 µm.

Oven Zone Residence Time, Solvent Flash-Off, and Dielectric Loss Tangent

Oven zone residence time, solvent flash-off, and dielectric loss tangent jointly limit the maximum production speed for TMA-PAI enamel on round copper because underbaking leaves a measurable dip in glass transition temperature while overbaking produces surface oxidation that interferes with inter-coat adhesion. Horizontal enamel ovens with zone temperatures from 320 °C to 520 °C and wire speeds between 80 m/min and 350 m/min impose a thermal history on each pass that lasts only 1.5 s to 7.0 s, during which solvent must evaporate, amic acid groups must close to imide, and the film must densify without trapping volatile reaction by-products. The dielectric loss tangent at 1 MHz is a sensitive production-line indicator of undercure because it increases from a fully cured baseline of 0.008 to 0.025 when residual polar amic acid and NMP remain above their threshold concentrations; this measurement can be performed on representative straightened samples with a micrometer electrode according to ASTM D150. In high-speed coating of wires below 0.200 mm, the short residence time means that solvent flash-off can dominate over imidization, so the first pass may contain 3 wt% to 5 wt% residual solvent even when the zone air temperature is 380 °C; subsequent passes then trap this solvent beneath a denser outer skin, creating the conditions for mid-life blistering. Conversely, for wires above 2.000 mm, the large thermal mass of the copper slows the heat-up of the enamel-substrate interface, making it possible to overbake the outer surface while the inner layer remains underbaked. The resulting film has a hard, oxidized outer surface with a low coefficient of friction but poor inter-coat adhesion, and a soft inner region with high dielectric loss; such a construction is not acceptable for thermal class 220 °C because the gradient in mechanical properties produces delamination during thermal cycling. The practical consequence is that each conductor diameter requires a dedicated oven temperature profile and line speed envelope, and changing either without recalibrating the imidization state of the final film invalidates the thermal class determination.

Since eccentric enamel build on round copper produces a non-uniform thermal and dielectric cross-section, the thermal class of a production lot can be invalidated by a maximum film thickness that exceeds the design thickness on one side while the opposite side approaches bare copper. For a 0.800 mm grade 2 wire with a minimum diameter increase of 0.080 mm, a concentricity ratio of 1.5 creates a one-side enamel build of approximately 48 µm and an opposite-side build of approximately 32 µm; this distribution places the thinnest film at risk for partial discharge and the thickest film at risk for local overheating because the thermal conductivity of cured TMA-PAI is only about 0.20 W/(m·K). During locked-rotor or stalled-rotor tests, the copper-adjacent temperature at the thick-side film surface can be 10 K to 15 K higher than the average winding temperature, which accelerates interfacial copper diffusion and reduces the time to dielectric failure. Eccentricity also changes the distribution of mechanical bending strain; the thinner side elongates less per unit bending stress while the thicker side develops cracks earlier when the conductor is wound on a small radius. Production lots intended for thermal class 200 °C and above are therefore screened by enamel build measurement using optical cross-sections at magnifications of 400× to 1000×, with acceptance limits typically requiring a minimum build on any point of the circumference to be at least 70% of the maximum build. The concentricity requirement becomes more difficult as the conductor diameter decreases, because wire vibration during coating and variation in die alignment produce larger relative deviations in enamel thickness on fine wire below 0.200 mm. Even when the average build meets the dimensional specification, a low minimum build location can become the primary failure site under partial discharge screening with rise times below 500 ns.

Thermal Endurance Classification Under IEC 60216 Requires Multipoint Arrhenius Extrapolation

Thermal endurance classification under IEC 60216 requires multipoint Arrhenius extrapolation rather than single-temperature exposure because the dominant degradation mechanism can change with temperature and the resulting data set may contain non-linearities that invalidate a simple logarithmic extrapolation. For TMA-PAI enamel on round copper, ageing temperatures of 180 °C, 200 °C, 220 °C, and 240 °C are commonly used, with end-of-life defined as the time at which the dielectric strength on twisted pairs falls below 1 kV or the mandrel bend fails at 2d. At 180 °C the failure mode is often gradual oxidative embrittlement, whereas at 240 °C the failure mode may shift to rapid copper-catalysed interfacial degradation, producing a break in the Arrhenius slope and making the 220 °C extrapolation sensitive to the selection of data points. The compliance matrix in Table 1 identifies the standard methods and limiting variables that are evaluated when determining whether a specific TMA-PAI enamel formulation can be labelled as class 200 °C or class 220 °C on round copper.

In high-speed automated winding of 0.315 mm TMA-PAI enamelled round copper for automotive alternator stators, the limiting factor is often not bulk thermal degradation but localised film damage arising from the combination of residual solvent and the mechanical action of needle winders running at 1,200 strokes/min to 1,800 strokes/min. At this winding speed, the enamel surface can reach 120 °C from frictional heating, and if residual NMP exceeds 0.5 wt% the surface softens sufficiently to allow wire guide debris to embed and create a defect that fails after 500 h at 180 °C in subsequent thermal endurance testing. The higher thermal class of the polymer therefore becomes irrelevant when the enamelling process leaves a residual solvent level that is only 0.3 wt% above the threshold, because the failure mechanism is mechanical defect propagation rather than thermo-oxidative chain scission. Similarly, in hermetic compressor motor windings, the interaction between TMA-PAI enamel and phosphate ester lubricants can reduce the dielectric strength of the film at 200 °C by promoting hydrolysis of unreacted amic acid segments, although published data for this specific configuration is limited and batch-to-batch variance in compressor charge chemistry makes reliable numerical limits difficult to establish without sealed-tube compatibility testing on the final wound unit.