Rational design of binary halide scintillators via data mining

We introduce a new search strategy for the development of novel inorganic scintillators. For designing new scintillation host media having the improved properties, the potential candidate materials were chosen by using a chemical selection scheme based on a multi-dimensional similarity metric. For the quantitative assessment of the chosen materials, predictive models based on informatics were built by correlating a set of key parameters which reflect the features of the host materials with the performance of inorganic scintillators. The resulting design rules generated from the relationships serve as a guide to identify HfI4 and TaI5 as two new host lattices with high light yield. The method we have outlined here serves as a new computational template based statistical learning method to search for new inorganic scintillators with targeted properties.