Quantitative analysis of the uniformity of CVD diamond nucleation via Kullback-Leibler divergence

Chemical vapor deposition (CVD) diamonds are gaining traction in multiple sectors, yet effective quantization methods for evaluating nucleation uniformity have been limited. This limitation results in obscure when evaluating nucleation processes and patterns, potentially impeding the preparation of higher quality and larger area diamonds. To address this issue, this study leverages the Kullback-Leibler (K-L) divergence, a fundamental algorithm in artificial intelligence, to establish a novel quantification algorithm. The algorithm maps the space distribution of diamonds on the substrate to a probability distribution on a two-dimensional plane and calculates its K-L divergence from a uniform distribution. Comparative analyses of CVD diamond nuclei on silicon substrates with various pre-treatments reveal that our quantitative algorithm outperforms traditional algorithms like nucleation density and variance in terms of reliability, validity, and robustness. Additionally, this novel algorithm not only enhances the assessment of diamond nucleation quality but also holds broad potential for research involving particle distribution.