Prediction of molecular weight and density of distillation fractions from gas chromatographic—mass spectrometric detection and multivariate calibration

Abstract

Brakstad, F., Karstang, T.V., Sørensen, J. Steen, A., 1988. Prediction of molecular weight and density of distillation fractions from gas chromatographic—mass spectrometric detection and multivariate calibration. Chemometrics and Intelligent Laboratory Systems, 3: 321–328. An accurate gas chromatographic method has been developed for the prediction of molecular weight and density of distillation fractions using computerized mass spectrometric detection. Partial least squares regression, which reduces the number of fragment data, is used to quantitatively relate the significant fragmentograms of a distillation fraction to its measured mean molecular weight and density. The final regression models for molecular weight and density give relative standard errors of prediction of 1.0 and 0.3%, respectively. The advantages of this method for the simulated distillation of oils or condensates are discussed.