Ultra high pressure liquid chromatography. Column permeability and changes of the eluent properties

The behavior of four similar liquid chromatography columns (2.1mm i.d. x 30, 50, 100, and 150 mm, all packed with fine particles, average d(p) approximately 1.7 microm, of bridged ethylsiloxane/silica hybrid-C(18), named BEH-C(18)) was studied in wide ranges of temperature and pressure. The pressure and the temperature dependencies of the viscosity and the density of the eluent (pure acetonitrile) along the columns were also derived, using the column permeabilities and applying the Kozeny-Carman and the heat balance equations. The heat lost through the external surface area of the chromatographic column was directly derived from the wall temperature of the stainless steel tube measured with a precision of +/-0.2 degrees C in still air and +/-0.1 degrees C in the oven compartment. The variations of the density and viscosity of pure acetonitrile as a function of the temperature and pressure was derived from empirical correlations based on precise experimental data acquired between 298 and 373 K and at pressures up to 1.5 kbar. The measurements were made with the Acquity UPLC chromatograph that can deliver a maximum flow rate of 2 mL/min and apply a maximum column inlet pressure of 1038 bar. The average Kozeny-Carman permeability constant of the columns was 144+/-3.5%. The temperature hence the viscosity and the density profiles of the eluent along the column deviate significantly from linear behavior under high-pressure gradients. For a 1000 bar pressure drop, we measured DeltaT=25-30 K, (Deltaeta/eta) approximately 100%, and (Deltarho/rho) approximately 10%. These results show that the radial temperature profiles are never fully developed within 1% for any of the columns, even under still-air conditions. This represents a practical advantage regarding the apparent column efficiency at high flow rates, since the impact of the differential analyte velocity between the column center and the column wall is not maximum. The interpretation of the peak profiles recorded in UPLC is discussed.