Limit of the speed-resolution properties in adiabatic supercritical fluid chromatography

• The concept of Poppe plot is extended and used to compare column performance in SFC and HPLC.
• The columns were assumed to be radially homogeneous and to operate under adiabatic conditions.
• The dispersion process in adjacent column slices is assumed to be independent.
• For axial column non-uniformity, the linear addition rule applies to the time variance only.
• In RPLC and SFC, columns of core–shell particles outperform those of fully porous particles.

Kinetic Poppe plots for small retained compounds were calculated in HPLC (using pure acetonitrile) and SFC (using pure carbon dioxide) for columns having twenty one different lengths (between 3 cm and 30 m), operated under strict adiabatic conditions (no heat exchange was allowed between the column and the external environment), with a constant pressure drop of 200 bar. The outlet pressures were set at 1 and 160 bar in HPLC and SFC, respectively. The eluent inlet temperature was set at 312 K. The hold-up time t0 and the apparent column efficiency 〈N〉 were calculated by taking into account the longitudinal variations of the eluent pressure, its temperature, density, viscosity, heat capacity, thermal expansion coefficient, equilibrium constant, and diffusion coefficient along the column length. Three different classes of stationary phase were considered: fully porous particles and core–shell particles of different diameter and structure, and silica monolithic columns of the second generation. The reduced plate heights of these stationary phases were taken from experimental data obtained with liquid eluents (acetonitrile/water mixtures). The columns were assumed to be radially homogeneous. The Henry's constant of the compound was fixed at Ka = 2.0 at the column inlet. The results demonstrate the potential advantage of using sub-3 μm core–shell particles for fast analysis in both LC and SFC, regardless of the intra-particle diffusivity through the stationary phase. In RPLC conditions, the contribution of surface diffusion to intra-particle diffusivity is important and the 4.6 μm core–shell particles can perform as well as sub-2 μm fully porous particles and silica monolithic columns of the second generation. In the absence of surface diffusion or for localized adsorption onto the stationary phase, sub-2 μm particles and silica monolithic column of the second generation outperform the 4.6 μm core–shell particles because the solid–liquid mass transfer resistance controls the column efficiency at high speeds. Eventually, for the same stationary phase and speed of analysis, SFC methods using pure CO2 may provide at least a twice column efficiency than LC methods using pure acetonitrile. For a constant pressure drop and resolution power, SFC methods may generate four times faster analyses than LC methods. Ultimately, a standard commercial 4.6 mm × 50 mm long column packed with 2.6 μm core–shell particles, operated with an inlet flow rate of 25 mL/min in fast SFC (200 bar back pressure, 40 °C) may provide a hold-up time of about 1 s requiring data acquisition at a frequency of 400 Hz, with a variance of 0.35 μL2. This performance will require the use of new, ultra-low dispersion SFC system.