μLIBS imaging for the characterization of heterogeneous archaeological materials
Résumé
In archaeology, studying of building materials such as mortar or architectural ceramic provides an important insight into the technological choices and skill, as well as the supply dynamics of construction sites and the construction phasing. This is especially their heterogeneity and complexity in phases, mainly in regard to the mineral filler, that are rich in information. The study of those heterogeneous materials is commonly lead by observation with binocular magnifier, but for broader interpretations, mineral identification and fine analysis, an observation under polarising light microscope is necessary. Preparation of thin sections for such study is time consuming and classic petrography is limited with regard of data treatment. In this work, we will show that μLIBS imaging brings a new dimension to the study of this kind of materials, both in terms of sample needs and phases identification. The technique allows to conduct analysis on a simple polished section and to carry out geochemical identifications on phases present in heterogeneous materials, even for the smallest grains thanks to a high resolution of about 15 μm. The possibility to work on large samples is also an advantage that show its benefits in granulometric and morphological approaches with LIBS images. The elemental data provided by LIBS ( especially minor and trace elements) can thus provide an effective means of comparing phases between different samples. Furthermore, the use of an artificial intelligence system (artificial neural network) for the classification of spectral data is an effective way to identify the components of these archaeological materials, while allowing for fast and automated analysis.