In the present study, the internal interactions of an analyte/modifier/matrix complex were investigated systematically. Target compounds (ethylphenol, quinoline, and nitronaphthalene), test matrices (silica gel, silica modified with aniline, and XAD-2), and modifying solvents (acetic acid, tetrahydrofuran, and toluene) were chosen based on their alpha, beta, or pi* characteristics. The SFE recoveries of target solutes were variable depending on the type of matrix, suggesting that the most critical factor in the SFE process is to eliminate the matrix-analyte interaction. The action of the modifier appears to be competition with the analyte for sorption sites. From alpha-matrix (silica gel), ethylphenol (alpha) was quantitatively extracted by alpha-modifier (acetic acid) and tetrahydrofuran (beta) was the most effective modifier for quinoline (beta) extraction (73% recovery). From beta-matrix (silica modified with aniline), 104% recovery of nitronaphthalene (pi) was achieved with pi*-modifier (toluene). The modifier effect of tetrahydrofuran (beta) for extracting quinoline (beta) was superior to acetic acid (alpha) and toluene (pi*), although quantitative recovery was not obtained even with tetrahydrofuran (73% recovery). From XAD-2 substrate (pi*), all three target compounds were readily extracted with unmodified supercritical fluid (SCF) CO2, showing there were no significant interactions between the target analytes and the XAD-2 matrix. The preliminary study was applied to alumina and it was found that combined solvatochromic parameters of modifiers work more effectively and thus methanol, having relatively high solvatochromic values (alpha, beta, and pi*), remains the best in most cases.