Filtern
Dokumenttyp
- Vortrag (4)
Sprache
- Englisch (4) (entfernen)
Referierte Publikation
- nein (4) (entfernen)
Schlagworte
- Dielectric strength (4) (entfernen)
Organisationseinheit der BAM
Eingeladener Vortrag
- nein (4)
Dielectric strength is a critical property for materials used as electrical insulators. The measurement of dielectric strength is well established and straightforward, but the values determined in a measurement are strongly dependent on the measurement setup and the specimen characteristics. For example, the size of the electrodes has a significant influence on the results. ASTM D149 covers a range of electrode sizes and does not unambiguously prescribe the quality of the electrodes. Thus, different test setups and procedures are used in the field. Consequently, a comparison of dielectric strength values of different origins, for example material supplier and customer, is often not meaningful.
To quantify the influence of the test procedure on the dielectric strength values, a batch of industrial thick-film substrates has been tested with different electrode configurations under AC conditions. Opposing cylinders with diameters of 6.4 mm and 25 mm, and a thick-film metallization with a diameter of 25 mm were used in the study. At least 20 specimens were tested with each type of electrodes. The results range from 21.7 ± 0.7 kV/mm measured with printed electrodes to 26.7 ±1.2 kV/mm measured with reused 6.4 mm cylinders. This means a difference of 23 %. Measurements performed with 6.4 mm cylinder electrodes produce significantly lower values (ANOVA, α = 0.01) if a new set of electrodes is used for each measurement instead of reusing the same pair of electrodes for the entire batch. The dielectric strength measured with new 25 mm cylinders is 11.2 % lower than the values determined with new 6.4 mm cylinders. No significant difference (ANOVA, α = 0.01) was found for measurements with printed electrodes and opposing 25 mm cylinders. Weibull evaluation of the data showed that all tested electrode configurations result in a similar reliability of the specimen (Weibull modulus) but in a significantly different characteristic dielectric strength (scale factor, Bonferroni, α = 0.05).
This study emphasizes the importance of a thorough and comprehensive documentation and communication of the test procedure for dielectric strength measurements. It further helps to evaluate the significance of differences in dielectric strength data provided by different sources.
The dielectric breakdown strength of ceramics strongly depends on the test conditions. Thus, standardized test procedures and thorough documentation are indispensable. However, during dielectric strength testing the breakdown often occurs near the electrode edge or even outside the specified electrode area. This behavior is similarly observed for printed and cylindrical electrodes. The aim of the presented study was to calculate the electric field strength distribution in a ball-on-plate testing setup for metallized samples and to correlate the field distribution with the observed breakdown locations. Small misalignments in the test setup were also considered in the simulations. Furthermore, the field strength at the breakdown Location should be compared to the experimentally determined dielectric strength. Therefore, Finite Element Models of several test conditions with varying printed electrode areas and sample thicknesses were created and electrostatic calculations of the electric field Distribution were performed. The simulation results were compared to experimental data. Alumina (96 %) was used as test material. The calculations show that the electric field strength maxima match the experimentally observed locations of breakdown. Without any fitting of the model, the maximum calculated field strength is in reasonable agreement with the experimental dielectric strength. The FE analysis is a helpful tool to understand the observations in experimental dielectric strength testing.
Dielectric breakdown of ceramics is widely believed to originate from microstructural defects. Still, there is no commonly accepted model for the origin and process of dielectric failure that covers all observed phenomena and dependencies. In analogy to mechanical strength, the Weibull distribution is commonly used to evaluate dielectric strength data. This works well for a given group of specimens with constant geometry. But unlike mechanical strength, dielectric strength scales with the inverse square root of sample thickness. This cannot be explained by the classic Weibull concept. The Griffith type energy release rate model of dielectric breakdown proposed by Schneider is based on space charge injection and conducting filaments from the sample surface. This model incorporates the distinct thickness dependence and the pronounced influence of surface defects. Based on this model and the classic Weibull probability of failure, Schneider’s group theoretically derived a probability of breakdown that predicts an increase of failure probability with increasing electrode area. In our study we tested this model with dielectric strength data measured on dense alumina samples using different electrode areas. Weibull modulus and characteristic dielectric strength (scale parameter) were determined for a set of measurements using small electrodes. These values were used to calculate the failure probability under large electrodes according to the model. The calculated data excellently fits the measured values. Thus, our experiments substantiate the assumptions made in the breakdown model and the significance of surface defects for dielectric failure.
Dielectric strength testing of ceramics is simple and yet challenging. The execution of a breakthrough voltage measurement of a given sample is fast and straightforward. ASTM D149 describes the standardized procedure. But, there are versatile effects of test conditions and sample properties that affect the result of such a measurement. As one example, ASTM D149 allows different sizes of test electrodes and does not unambiguously prescribe the condition of the electrodes. Thus, different electrode configurations are used in the field. We conducted several test series on alumina samples to comprehensively quantify the effect of test conditions and sample properties on dielectric strength results. In our study, testing of alumina substrates using different electrode configurations resulted in differences of mean values of up to 20%. Further test series on alumina focused on the effect of voltage ramp rate. The results are complemented by calculations of failure probability at different voltage levels and corresponding withstand voltage tests. We conclude that a communication and comparison of single dielectric strength values is insufficient and may be misleading. A meaningful comparison of dielectric strength studies from different sources requires a thorough consideration of test conditions.