Premium
Optimization of the separation of β‐blockers by ion‐pair capillary electrophoresis in non‐aqueous media using univariate and multivariate approaches
Author(s) -
Servais AnneCatherine,
Fillet Marianne,
Chiap Patrice,
Abushoffa Adel M.,
Hubert Philippe,
Crommen Jacques
Publication year - 2002
Publication title -
journal of separation science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.72
H-Index - 102
eISSN - 1615-9314
pISSN - 1615-9306
DOI - 10.1002/1615-9314(20021101)25:15/17<1087::aid-jssc1087>3.0.co;2-a
Subject(s) - chemistry , univariate , chromatography , fractional factorial design , atenolol , factorial experiment , capillary electrophoresis , chemometrics , acetonitrile , electrolyte , analyte , multivariate statistics , central composite design , analytical chemistry (journal) , response surface methodology , mathematics , statistics , medicine , electrode , blood pressure , radiology
The separation of four β‐Blocking drug substances (atenolol, sotalol, betaxolol, and metoprolol) selected as model basic analytes has been investigated in non‐aqueous capillary electrophoresis (NACE) using the principle of ion‐pair formation. Camphorsulphonate was selected as the counter‐ion in a background electrolyte (BGE) made up of formate buffer in a mixture of acetonitrile/methanol or acetonitrile/ethanol. The influence on resolution of the concentration of the counter‐ion, the nature and proportion of the organic solvents, as well as the concentration of the ionic components of the BGE was first studied by a univariate approach. An experimental design was then applied to estimate possible quadratic effects and first‐order interactions. To identify the most important factors affecting the separation of the four β‐blockers, a two‐level fractional factorial design with 16 experimental points was applied as a screening test. A three factor Box‐Behnken design with 12 experimental points was used to predict and optimize the selectivity. Finally the optimal conditions obtained by the univariate and the multivariate approaches were compared.