(138d) Embedded Sphere Method for Measuring Dielectric Breakdown in Polymers and Polymer Composites

This work describes a new approach, the embedded sphere method, for measuring the dielectric breakdown field strength of pure polymers and polymer composites. In this method, a polymer film is coated onto a polished planar electrode by spin casting or melt pressing. After the film is dried or cooled, a spring presses a polished metal sphere onto the air side of the polymer film. When the assembly is heated above the polymer's glass transition temperature, the spring pushes the sphere into the polymer, resulting in partial embedment of the sphere. The degree of embedment can be varied by controlling the initial film thickness and the heating time and temperature. After cooling, we ramp the voltage across the electrodes and measure pre-breakdown leakage current and the breakdown voltage. After breakdown, we remove the sphere, confirm that breakdown occurred at the center of the indentation left by the sphere, and measure the minimum thickness at the center of the indentation. The breakdown field strength equals the breakdown voltage divided by the minimum indent thickness. Results are presented in the form of two-parameter Weibull plots. In this presentation, we show results for spin cast polystyrene films and composites. We find that the embedded sphere method minimizes the deleterious effects of point contact between the measuring electrode and the specimen, especially premature breakdown due to air near the point of contact. We confirm that the indentation depth can be controlled, is highly reproducible, and is much greater than the average surface roughness of the film. Our results show that embedding the sphere in the polymer leads to a greater apparent breakdown field strength for pure polystyrene with much less scatter in the data. We present results, for PS and various composites, showing the variation of breakdown field strength with degree of embedment, minimum indentation thickness, and filler weight loading. We also present results showing the effect of increasing embedment temperature, from below the glass transition to above the melt temperature, to investigate the possibility that air-filled micro-voids at the metal-polymer interface may affect the measurements.