Accurate classification of the molecular characteristics of soluble portions from various lignites: Joint analysis of thermal dissolution experiments and data mining methods

Gas chromatography-mass spectrometry (GC-MS) was used to analyze the soluble portions (SPs) of eight lignites which were obtained by extracting using cyclohexane, methanol and ethanol, respectively, as the solvent. For thermal dissolution at 300oC using cyclohexane, the intermolecular interactions between mobile phase and macromolecules in coal were weakened. Consequently, non- and low-polar compounds such as alkanes and aromatics were well dissolved. Methanol and ethanol were more likely to closely interact with oxygen-containing functional groups and promoted the breakage of hydrogen bonds and weak covalent bonds. Hierarchical cluster analysis classified all the 24 SPs, and the SPs from cyclohexane and ethanol solvents exhibited more similarity. Partial least squares discriminant analysis was established to reproduce the classification results of SPs entirely and accurately. The SPs were further classified by combining a small number of geochemical characteristic compounds rather than massive GC-MS data.