Application of multivariate statistics in assessment of green analytical chemistry parameters of analytical methodologies

摘要

The study offers a multivariate statistical analysis of a dataset, including the major metrological, "greenness" and methodological parameters of 43 analytical methodologies applied foraldrin determination (a frequently analyzed organic compound) in water samples. The variables (parameters) chosen were as follows: metrological (LOD, recovery, RSD), describing the "greenness" (amount of the solvent used, amount of waste generated) and general methodological parameters (sample volume, time of analysis, injection volume) and scores of greenness assessment with NEMI and eco-scale. The results of the study show that all analytical methodologies have been grouped into three clusters. The first one consisted of "non-green" LLE and SPE methodologies and the other two consisted of solventless or virtually solvent-less methodologies. The NEMI and eco-scale scores are well correlated, which indicates the similarity between these two assessment scales. A self-organizing maps technique is not feasible for easy and quick labeling of analytical methodologies in terms of their greenness. However, the multivariate analysis of analytical methodologies can give information about clustering of methodologies to "green" or "non-green" groups and some extra information about relations between objects inside clusters of interest.