Size Exclusion chromatography skewing

While polydisperse model systems can nicely be resolved, the reconstruction of a broad and skewed molar mass distribution is only possible within certain limits. At this point, experimental techniques in which only a nonexponential time signal or some other integral quantity is measured and the underlying distribution is obtained from e.g. an inverse Laplace transform are inferior to fractionating techniques, like size exclusion chromatography or the field-flow fractionation techniques. The latter suffer, however, from other problems, like calibration or column-solute interaction. 

Often, size exclusion chromatograms (SEC) (compare section 11.7, Size Exclusion Chromatography) of polymers under study are expressed as differential representations of molar mass dispersity. The chromatographic retention volumes are directly transformed into the molar masses. This approach renders useful immediate information about tendencies of molar mass evolution in the course of building or decomposition polyreactions but the absolute values of molar mass can be only rarely extracted from it. As a rale, polystyrene calibrations are applied for molar mass calculation so that one deals with the polystyrene equivalent molar masses, not with the absolute values. The resulting dispersity (distribution) functions may be heavily skewed because the linear part of the calibration dependence for the polymer under study may exhibit well different slope compared with the polystyrene calibration, which was employed for the transformation of retention volumes into molar masses. 

For EMG peaks, peak skew increases with the ratio xG/oG. Figure 16-32 illustrates the characteristics of such a peak calculated for xG/aG = 1.5. In general, with xG/aG > 1 (peak skew > 0.7), a direct calculation of the moments is required to obtain a good approximation of the true value of N , since other methods give a large error (Yau et al., Modems Size-Exclusion Liquid Chromatography, Wiley, New York, 1979). Alternatively, Eq. (16-165) can be fitted to experimental peaks to determine the optimum values of fG, aG, and xG. 

---