ABSTRACT
Lignin is considered a promising renewable source of valuable chemical compounds and a feedstock for the production of various materials. Its suitability for certain directions of processing is determined by the chemical structure of its macromolecules. Its formation depends on botanical origin, isolation procedure and other factors. Due to the complexity of the chemical composition, revealing the structural differences between lignins of various origins is a challenging task and requires the use of the most informative methods for obtaining and processing data. In the present study, a combination of two-dimensional nuclear magnetic resonance (2D NMR) spectroscopy and multivariate analysis of heteronuclear single quantum coherence (HSQC) spectra is proposed. Principal component analysis and hierarchical cluster analysis techniques demonstrated the possibility to effectively classify lignins at the level of belonging to classes and families of plants, and in some cases individual species, with an error rate for data classification of 2.3%. The reverse transformation of loading plots into the corresponding HSQC loading spectra allowed for structural information to be obtained about the latent components of lignins and their structural fragments (biomarkers) responsible for certain differences. As a result of the analysis of 34 coniferous, deciduous, and herbaceous lignins, 10 groups of key substructures were established. In addition to syringyl, guaiacyl, and p-hydroxyphenyl monomeric units, they include various terminal substructures: dihydroconiferyl alcohol, balanopholin, cinnamic acids, and tricin. It was shown that, in some cases, the substructures formed during the partial destruction of biopolymer macromolecules also have a significant effect on the classification of lignins of various origins.
