Characterization of microstructure of lead free solder by different image processing algorithms

The microstructure of lead free solder joints was investigated by three different image processing techniques. Tin (Sn) was extracted from the cross sectioned samples by selective etching technique. Samples were generated by laser soldering. The cellular network of the Sn-Ag intermetallic compound remained intact. The images were taken by SEM and the compounds were identified by EDX measurements. A special Elastic String Image Processing (ESIP) algorithm was developed in order to measure the average gap size of this network. The results of the algorithm were compared to results prepared by basic thresholding and watershed techniques. The experiments show us that due to the fibrillose structure the thresholding methods does not fulfil the requirements, while using either the watershed, or our new algorithm, the average grain size of the microstructure was determined. ESIP algorithm results carry some further information about the surface energy of the crystallite grains, so technological parameters could be concluded only by the observation of the microstructure of the solder.