Non-destructive X-ray based characterization of materials assisted by multivariate methods of data analysis: from theory to application

X-ray based non-destructive techniques are an increasingly important tool in many fields, ranging from industry to fine arts, from medicine to basic research. Over the last century, the study of the physical phenomena underlying the interaction between X-rays and matter has led to the development of many different techniques suitable for morphological, textural, elementary, and compositional analysis. Furthermore, with the development of the hardware technology and its automation thanks to IT advancements, enormous progress has been made also from the point of view of data collection and nowadays it is possible to carry out measurement campaigns by collecting many GigaBytes of data in a few hours. Already huge data sets are further enlarged when samples are analyzed with a multi-technique approach and/or at in situ conditions with time, space, temperature, and concentration becoming additional variables. In the present work, new data collection and analysis methods are presented along with applicative studies in which innovative materials have been developed and characterized. These materials are currently of high application interest and involve composites for radiation protection, ultralight magnesium alloys and eutectic mixtures. The new approaches have been grown up from an instrumental viewpoint and with regard to the analysis of the data obtained, for which the use and development of multivariate methods was central. In this context, extensive use has been made of principal component analysis and experimental design methods. One prominent topic of the study involved the development of in situ analysis methods of evolving samples as a response to different types of gradients. In fact, while in large structures such as synchrotrons carrying out analyzes under variable conditions is now consolidated practice, on a laboratory scale this type of experiments is still relatively young and the methods of data analysis of data sets evolving systems have large perspectives for development especially, if integrated by multivariate methods.