In this study, a separability parameter is introduced to determine the selection of optimum operating parameters for DEP separation of a cell pair. The separability parameter is defined as a function of cells' ClausiusMossotti (CM) factors. T-cell leukemia Jurkat and mouse melanoma B16 cells are tested to validate the separability parameter. CM factors of cells are measured using a recently developed microfluidic impedance spectroscopy device. Separability maps are generated for varying values of field frequency and buffer conductivity. Cell separation is tested using a planar interdigitated electrode array at different buffer conductivities. Impedance measurements of the DEP device are performed at various buffer conductivities. Electrode polarization effects and energy allocation for dielectrophoretic manipulation of cells are computed from the impedance data utilizing an equivalent circuit model. Cell separation results are explained in the light of the impedance measurements.