Evaluation of different derivatisation approaches for gas chromatographic–mass spectrometric analysis of carbohydrates in complex matrices of biological and synthetic origin

Gas chromatographic analysis of complex carbohydrate mixtures requires highly effective and reliable derivatisation strategies for successful separation, identification, and quantitation of all constituents. Different single-step (per-trimethylsilylation, isopropylidenation) and two-step approaches (ethoximation–trimethylsilylation, ethoximation–trifluoroacetylation, benzoximation–trimethylsilylation, benzoximation–trifluoroacetylation) have been comprehensively studied with regard to chromatographic characteristics, informational value of mass spectra, ease of peak assignment, robustness toward matrix effects, and quantitation using a set of reference compounds that comprise eight monosaccharides (C5–C6), glycolaldehyde, and dihydroxyacetone. It has been shown that isopropylidenation and the two oximation-trifluoroacetylation approaches are least suitable for complex carbohydrate matrices. Whereas the former is limited to compounds that contain vicinal dihydroxy moieties in cis configuration, the latter two methods are sensitive to traces of trifluoroacetic acid which strongly supports decomposition of ketohexoses. It has been demonstrated for two “real” carbohydrate-rich matrices of biological and synthetic origin, respectively, that two-step ethoximation–trimethylsilylation is superior to other approaches due to the low number of peaks obtained per carbohydrate, good peak separation performance, structural information of mass spectra, low limits of detection and quantitation, minor relative standard deviations, and low sensitivity toward matrix effects.

► Six different single- and two-step carbohydrate derivatisation approaches compared.
► Reference compounds comprised 8 monosaccharides (C5–C6), GA and DHA.
► Ethoximation–trimethylsilylation was found to be superior to other approaches.
► Analysis of carbohydrate-rich matrices of different origin is possible.
► Calculation of a novel matrix impact value is proposed to evaluate matrix effects.