Computer analysis of the cemented carbides' microstructure

Cemented carbides have been known for about a hundred years and are now widely used in mining, civil and road construction, and mechanical engineering. Their improvement requires the development of novel technologies based on fundamentally new approaches. Here the method for microstructure analysis using computer processing with elements of machine learning and artificial intelligence is proposed. It has been applied to the analysis of micrographs obtained by the scanning electron microscopy of the WC-Co cemented carbides. The geometric parameters of the WC/Co interphase boundaries are extracted using mathematical methods for processing digital pictures. This method is applied to the micrographs of three samples with different WC mean grain sizes. It has been found that the distribution of the contact angles of WC/Co interphase boundaries has a pronounced bimodal structure, and the values of the peak angles are practically the same for the samples of the fine-, medium-, and coarse-grain cemented carbide grades. The distributions of the semiaxes for the isolated areas of the cobalt binder are also obtained. The probability of finding a particular value of long and short semiaxis decreases exponentially with increasing semiaxis length. Contrary to the contact angles, the exponents are different for samples with different WC grains sizes. The obtained results are discussed using the ideas of wetting of grain boundaries and faceting-roughening of interphase boundaries.