Ultrasonic characterization and image analysis of Ag-Cu-Zr active braze joints between Al{sub}2O{sub}3 and kovar

摘要

An active brazing process is being developed using Ag-Cu-Zr (97Ag-1Cu-2Zr and 98Ag-2Zr) active braze alloys (ABA) to join alumina to Kovar (Carpenter Technoloy Corp.), a low expansion Fe-Ni-Co alloy. The braze joints must be hermetic and free of excessive porosity and other defects. A design of experiments (DOEx) approach was taken to identify the main effects variables in the brazing process. For the DOEx study, it was important to provide reliable inputs as a basis for the statistical interpretation of the response data. In the past, a subjective rating system was used, for example, with a scale of 1 to 5. A rating of 1 corresponded to "poor" and a rating of 5 represented "excellent" joint quality. In this work, a new characterization scheme is described which employs ultrasonic (UT) imaging and quantitative image analysis (QIA). In particular, the UT system produced "false-color" images and the QIA software employed color thresholding to identify regions of porosity and other defects. These characterization techniques allow for the quantitative measurement of braze defects such as porosity and excessive braze flow (braze "runout"). Correlations were developed between the UT/QIA results and the hermeticity of the braze joints. In addition, the UT imaging technique was compared to traditional x-ray radiography and metallographic cross-sectioning techniques for defect analysis. The UT imaging technique avoided x-ray radiography artifacts, such as shadowing and distortion, and the color QIA procedure provided enhanced ability to charactrize defects compared to traditional grayscale imaging. The objective of the study was to provide reliable, quantitative inputs for statistical analysis of the DOEx data and identification of the main effects variables.