Chemometric appraisal of lignin pyrolytic assemblages from Eucalyptus woods relevant for pulping in Uruguay

Analytical pyrolysis (Py-GC/MS) at 500 °C was applied to study wood composition of Eucalyptus species (Eucalyptus grandis, Eucalyptus dunnii and Eucalyptus benthamii) which are relevant for pulping in Uruguay. Multivariate data treatments mainly principal component analysis and discriminant analysis were used to explore differences between the original wood cultivars. Multivariate analyses with automatic backwards variable selection indicated that simplified methoxyphenol patterns (up to 10 compounds) are sufficient for wood discrimination in terms of species and geographical origin but also with purposes of forecasting the ease of delignification of the resulting pulps measured as active alkali. No additional chemotaxonomical accuracy was achieved when the data sets were enlarged with carbohydrate-derived products. On the other side, discriminant or forecasting models (partial least squares and multiple linear regression) were much less significant when based on individual diagnostic compounds, groups of compounds, or the classical syringyl-to-guaiacyl (S/G) ratio. Principal component analysis indicated that the variability in lignin composition due to bioclimatic variations (spatial replications) was more significant than that due to phylogenetic differences (species and cultivars).