XRF data-based statistical clustering of ancient ceramic artifacts and analysing provenance with thermo-mineralogical analysis

Abstract

The chemical composition of a ceramic assemblage 48 sherds from the Museum of Asian Art (MoAA), Universiti Malaya, was determined using X-Ray Fluorescence (XRF) and the statistical analysis of the results was compared with the classification made by the museum curator on the basis of typology. The major objective of this study was to establish a link between the components used to produce Chinese porcelain from Jingdezhen, South China, throughout the different dynasties. A statistical analysis is applied based on 16 major and minor oxides employing hierarchical cluster analysis (HCA), principal component analysis (PCA), biplots, boxplots, and descriptive statistics. Fourier Transform Infrared Spectroscopy (FTIR), X-Ray Diffraction (XRD), and Thermomechanical Analysis (TMA) techniques were used to analyse the material characterization of sherds for classification, resulting in conclusions consistent with provenance and dating. The HCA analysis separated 48 sherds into 5 clusters. Each cluster except cluster 5, lacked of anorthite, which typically forms above temperature 850 °C, indicating that the firing temperature was not exceed 850 °C for clusters 1–4. Similarly, the deformation point for clusters 1–4 were low temperature and high temperature for cluster 5. The observed chemical composition and SiO₂/Al₂O₃ ratio (3.76–4.18) for clusters 1–4 and cluster 5 ratio of 2.92 indicates the Ming and Qing dynasty respectively with origin of Jingdezhen, South China of the examined ceramics. The HCA classification of sherds was based on XRF raw data without combining all experimental techniques like XRD, TMA, and FTIR data. The most representative sherds of each HCA cluster were analyzed thermo-mineralogically. A combination of strong evidence due to material characterization can therefore overcome the hindrance of unreliable sources that cause ambiguity of the sherd's originality and classification.