A Multivariate Statistical Approach To Detecting Adulteration Of Cement, Using Gamma Spectroscopy And Instrumental Nuclear Activation Analysis (INAA) Data

This study is considered as an attempt for employing multivariate statistical analysis, such as multiple linear correlations, principal component analysis and cluster analysis, to infer the spatial distribution of elements in cement samples through which we can detect possible adulteration of the samples based on measured radioactivity and elemental levels. It was estimated on the bases of data from five types of cements produced and or used in Ghana. The study results showed that the adapted classification rule leads to the classification of all CM I, CM II and CM III, about 60% of the total cements sampled, to the group which they belong. From the classification of the remaining 40%, it is highly probable that CM IV and CM V are from the same product but packaged under different names, whilst the CM IV and CM V from Accra and Kumasi are adulterated forms of this product. The analysis adopted was found to be effective in classifying the cement samples into their true statutes. This shows that multivariate statistical analysis of data from gamma spectroscopy and INAA may be an efficient and useful tool for predicting adulteration of cement.