Assessing glass‐ceramic homogeneity and nucleation self‐correlation by crystallization statistics

In this work, we implemented and tested a statistical method to evaluate the microstructural uniformity of partially crystallized glasses employing the aggregation index (R‐index), a parameter derived from the Poisson distribution function, which has mainly been used in Ecology studies. Since the crystal nucleation rate strongly depends on the chemical composition, the spatial crystal distribution in glass‐ceramics can be used to infer their chemical homogeneity. We also tested the hypothesis, advanced by some authors, of preferential secondary nucleation close to preexisting crystals (even in chemically homogeneous samples). To this end, we conducted a nearest‐neighbor statistical analysis of the spatial crystal distribution in partially crystallized Li2Si2O5 and Ba5Si8O21 glasses, used as model materials, by inspecting optical micrographs obtained at different magnifications. The resultant R‐indexes indicate a very high degree of homogeneity of the crystal number distribution, reflecting the uniform distribution of the chemical elements in the parent glasses. Moreover, the results for both glasses refute the suggestion that crystal nucleation is self‐correlated. These outcomes allow us to suggest the R‐index as a valuable and easily implemented tool to evaluate the chemical homogeneity of glasses that undergo internal nucleation, such as those used for glass‐ceramics.