文摘
The impact of the ratio of the column diameter to the average particle size (or bed aspect ratio) on the column performance was investigated from theoretical and experimental viewpoints. The experiments were conducted for two series of 100 mm long columns, 2.1, 3.0, and 4.6 mm in diameter, packed with 2.5 ¦Ìm fully porous particles of Bridge Ethylene Hybrid (BEH) for one series and Charged Surface Hybrid (CSH) for the other. The heights equivalent to a theoretical plate (HETP) of two low molecular weight compounds, uracil (non-retained, k = 0) and naphthalene (retained, k = 2.5), were determined from the true moments of the recorded peak profiles. The results showed a systematic decrease of the column performance for uracil at high flow velocities with decreasing column inner diameter, in agreement with the theoretical predictions of the variation of the trans-column eddy dispersion HETP term with decreasing bed aspect ratio. This result is consistent with the increasing volume fraction of the wall region of the column, in which the average linear velocity of the mobile phase over a distance of 5 particle diameters from the column wall is about 10 % larger than in the bulk center of the column (infinite diameter column). For the retained compound, the discrepancies are levelled out due to the longer average residence time and larger particle diffusivities of retained compounds, which allow a more efficient relaxation of the radial concentration gradients. Further improvements of the performance of the larger I.D. columns (3.0 and 4.6 mm I.D.) may be achieved by decreasing the harmful effect of this trans-column velocity bias by injecting and/or collecting the sample molecules in a wide central zone of the column. For 2.1 mm I.D. columns, this approach would prove useful only when HPLC instruments providing a lower extra-column band broadening contribution will become available. Finally, the further minimization of the trans-column eddy dispersion HETP term and the design of new, better inlet/outlet column endfitting/frit assemblies requires newer and more accurate models of eddy dispersion in packed columns than those previously provided by Gunn and Giddings and the numerical calculation of band profiles using original functions to account for the distribution and collection of the sample molecules at the inlet and outlet of the column.