Linking of pyrolysis-chemical ionisation mass spectrometric and monomer compositional data of O-(2-hydroxyethyl) celluloses by canonical correlation analysis

Pyrolysis-chemical ionisation mass spectrometry (Py-CIMS) data of O-(2-hydroxyethyl) celluloses (HECs) have been correlated by canonical correlation analysis with their corresponding quantitative monomer compositional data. The monomer composition of the samples was predicted via the canonical correlation functions from the Py-CIMS data. The values denoting the degree of molar substitution derived from these predicted monomer data deviate on average by less than 10% from the values derived directly from the original monomer compositional data. The correlations between specific mass peaks in the Py-CI mass spectra and the monomer composition of the HECs are revealed by the canonical correlation loadings. The anhydrosaccharides are specifically correlated to the remaining non-derivatised cellulose fraction in the HECs, and the substituted anhydrosaccharides are correlated to 6-O-substituted glucosyl residues. The polyethylene glycol glycolaldehydes correlate well with the polyethylene glycols which are 42 lower in mass. This confirms the validity of the earlier proposed reverse aldolisation, or rather [2 + 2 + 2] cycloreversion pathway, which predicts that the glycolaldehyde ether moiety in polyethylene glycol glycolaldehydes originates from the C5 and C6 carbons in a glucosyl residue.