| Abstract: | In the determination of molecular weight distributions by GPC, the traditional column arrangement is such that the fractionation process proceeds from a high-permeability limit to a low-permeability limit column. We report computer comparisons of data obtained from columns in their normal ordering (high- to low-permeability limit), reverse ordering, and random ordering. The columns had a permeability limit range from 1 × 106 Å down to 1 × 103 Å, and the polymers had a molecular weight range of 1.8 × 106 down to 2.1 × 103. The concentrations used varied from 0.05% up to 0.5%. The data show significantly different results, with the random arrangement the preferred ordering. A qualitative model for the separation mechanism is presented to account for the improvement in resolution. Additional data are presented which show that serious errors (as high as 45%, depending on concentration) will be encountered in GPC studies, unless the calibration curve is obtained at the same relative concentration as the samples, with definite overloading effects occurring at the higher concentrations. A new method of curve fitting was used in the higher molecular weight region to give meaningful calibration curves. |
