New Quasicrystal Approximant in the Sc–Pd System: From Topological Data Mining to the Bench

Intermetallics contribute significantly to our current demand for high-performance functional materials. However, understanding their chemistry is still an open and debated topic, especially for complex compounds such as approximants and quasicrystals. In this work, targeted topological data mining succeeded in (i) selecting all known Mackay-type approximants, (ii) uncovering the most important geometrical and chemical factors involved in their formation, and (iii) guiding the experimental work to obtain a new binary Sc–Pd 1/1 approximant for icosahedral quasicrystals containing the desired cluster. Single-crystal X-ray diffraction data analysis supplemented by electron density reconstruction using the maximum entropy method, showed fine structural peculiarities, that is, smeared electron densities in correspondence to some crystallographic sites. These characteristics have been studied through a comprehensive density functional theory modeling based on the combination of point defects such as vacancies and substitutions. It was confirmed that the structural disorder occurs in the shell enveloping the classical Mackay cluster, so that the real structure can be viewed as an assemblage of slightly different, locally ordered, four shell nanoclusters. Results obtained here open up broader perspectives for machine learning with the aim of designing novel materials in the fruitful field of quasicrystals and their approximants. This might become an alternative and/or complementary way to the electronic pseudogap tuning, often used before explorative synthesis.