Half-fraction and full factorial designs versus central composite design for retention modelling in reversed-phase...

In a previous paper (J. O. De Beer, C. V. Vandenbroucke and D. L. Massart, J. Pharm. Biomed. Anal., 12, (1994) 1379-1396) liquid chromatographic (LC) retention modelling of the cough-syrup compounds methyl para-hydroxybenzoate (MPHB) and propyl para-hydroxybenzoate (PPHB), phenylephrine hydrochloride (PE) and chlorphenamine maleate (CPM) was studied using a face-centred central composite design. It is examined whether smaller half-fractional and full factorial designs with fewer experiments tend to reliably predict retention times of the latter compounds as well. Simplified regression modelling, however, neglecting more first-order and interactive effects and disregarding pure second-order effects, has to be set up. These smaller designs finally satisfy the prediction of the retention of MPHB, PPHB and PE also. Retention prediction of CPM is much less accurate. CPM has a pKa value of 4.0, which is encompassed by the examined mobile phase pH limits 3.0 and 5.0. Since the largest retention shifts occur near the pKa value, retention prediction in this area becomes more complex. CPM retention modelling from a full factorial design is useful if the mobile phase pH is fixed at 5.0 for methanol as well as for acetonitrile as organic modifers. The full factorial design, applied with acetonitrile as organic modifer, enables the selection of suitable LC parameter combinations for fast and complete separation of the four compounds in cough-syrup analysis.