Chromatography Handbook Of Hplc - Katz Eksteen Schocnmakers Miller wiley Sons -119

Mechanisms and Importance of Zone-Spreading 109\nconcentration and, especially, the near inlet concentration, may be well in the finite concentration range. If so, the developed asymmetry near the inlet will remain in the concentration profile during further migration, only reduced to some extent by the mixing effect inherent with zone dispersion. As a result, we should not allow the inlet concentration, rather than the outlet concentration to be in excess of some limiting concentration value, yet to be specified.\n\n(c) Curvature of Partitioning Isotherms. The partition coefficient K = c/c, representing the slope of the so-called partition isotherm, c = c(c,c), is constant for any partition equilibrium at infinite dilution. In practice the linearity of the isotherm, and so the constancy of K, is limited to some restricted concentration range, above which deviations occur. The curvature of the isotherm at higher concentrations than the mentioned threshold value of the linear range, is either toward (concave, or Langmuir-type) or directed away from (convex, or anti-Langmuir) the abscissa c* . The background behind this is that, at elevated concentrations, molecules of the solute are ever more subjected to interactions with molecules of their own kind (solute-solute), rather than with solvent molecules (solute-solvent) alone, as it rare at infinite dilution. Because the interactive forces for solute-solute and solute-solvent contacts may be appreciably different, the solute molecules may show an increased preference for one of the phases, thereby influencing the ratio c/c . As a result, the partition K = c/c , the partition coefficient only for infinite dilution. If we realize now that the partition ratio determines the zone velocity of partitioning components as u = u/(1 + Kp), it is seen that once the zone velocity also becomes a function of concentration, which, depending on the curvature of the isotherm, increases (concave isotherm) or decreases (convex isotherm) with concentration. As with the migrating component the concentration is approaching zero (infinite time) and we may view the front of the zone, and shows a maximum concentration somewhere in between the rare zone center, it becomes apparent that the different regions of a component zone may migrate at different velocities through the column. For a concave isotherm, this will increase the concentration at the front and lead to a tailing peak when the more diluted parts near the bottom of the concentration profile, which produces a tailing peak on elution. For a convex isotherm, the opposite is true and the resulting peak shape is fronting. Partition isotherms cannot be expected to be linear over a solute concentration range of more than say 0.01 mol fraction, although this depends on the system at hand. Given the amount of phases present, this easily sets a limit to the maximum permissible sample size if isotherm linearity is to be maintained. In practice this limit is typically on the order of 0.01 μmol (i.e., 0.001 mg of liquid solute of assumed molecular mass 100).\n\n2. Peak Widths in Apparatus\nThe principal places that the apparatus can contribute to the zone dispersion are the injector, column frits and connectors, connecting tubes and fittings, the detector, and the data-handling system. Subreply [38] has given a comprehensive survey of these extracolumn contributions to zone-spreading. The residence times of the component zone in each of the mentioned contributing parts as well as in the column can be considered as independent and so their means and variances will be additive, as has been done with intracolumn contributions before:\n\nσpeak2 = σinject2 + σcolumn2 + σdetec2 + σconnec2 = σextracolumn2\n\nExperimentally there are two methods to obtain the extracolumn band-broadening of the chromatographic system: a linear extrapolation method using σpeak2 as a function of retention time; and a \"zero\" length column method where the column is removed from the system and replaced by a very short connecting capillary. Obviously, the latter method ignores the contribution of column frits and connectors, whereas additional broadening from the connecting capillary is introduced. Furthermore, because the pressure drop is lowered, short residence times and the resulting very narrow peaks offer experimental difficulties; this approach is very demanding and not recommended. A modification of this method, experimentally more at