The response surface methodology (RSM) was used to optimize the separation conditions in a supercritical fluid chromatography-mass spectrometry (SFC/MS) analysis. This optimization was applied on data obtained from a central composite design (CCD). The target compounds were two fatty acids – arachidonic acid (AA) and docosahexaenoic acid (DHA) – along with various lipid classes, including phosphatidylcholines (PC), phosphatidylglycerols (PG), and phosphatidylethanolamines (PE). Three operational variables – gradient time, pressure, and modifier percentage – were investigated to evaluate their influence on the resolution between AA and DHA and the symmetry factor of their respective peaks.

The CCD included 3 levels for each variable within a range of gradient time between 1 and 2 min, the pressure between 125 and 225 bar, and the modifier percentage from 2 to 5 %.

The data of RSM for the critical pair of acids indicated that pressure and modifier percentage were the parameters that most influenced both the resolution and symmetry factor, whereas gradient time was not a significant variable.

The importance of this study lies in the capability of the response surface to show not only a linear interaction (relationship) but also a non-linear interaction between the significant variables – pressure and modifier percentage. Consequently, small changes in one of the significant variables affected the resolution AA/DHA and symmetry factors of their respective peaks significantly. This type of interaction between variables can only be elucidated by the design of experiments (DOE) approach employed in this work, which is not demonstrated in similar work of lipid anbalysis.