Estimation of total carbohydrate amount in environmental samples by the phenol-sulphuric acid method assisted by multivariate calibration

In this paper a modification of the well known phenol-sulphuric acid method for the colorimetric determination of total carbohydrates in environmental samples is proposed. As in environmental samples different typologies of carbohydrates can be present, the solutions of glucose used as standard samples for conventional univariate calibration are replaced by solutions including neutral, uronic acid and aminosugars simultaneously to simulate better the qualitative composition of real samples. Then, principal component regression (PCR) or partial least squares (PLS) replace the univariate calibration. PCR and PLS show to be more accurate than univariate calibration because they are able to take into account the non-additive contributions to the spectral absorptions depending on the simultaneous presence in solution of neutral carbohydrates, uronic acids and aminosugars. In addition, the proposed procedure shows to have some potential applications in the speciation of neutral and uronic acid sub-fractions.