Currently available studies on supercritical fluid phase separations are limited to chemically nonreactive systems at low pressure. The present study is concerned with a possible influence of chemical reactions on a supercritical phase change above 1 GPa. We will first give a brief review of statistical mechanical theory, which is designed to handle chemically reactive systems, as used in this work. We next apply the theoretical formulation to chemically reactive systems containing species composed of C, H, N, O, F atoms. These systems produce mixtures such as CO, CO2, H20, N2, HF, etc. Our earlier calculations without F atoms predicted that these molecular systems separate into an N2-rich and an N2-p00r fluid phases at high pressure and high temperature. This prediction has been experimentally confirmed in part for a N2+H20 mixture [2]. Addition of F atoms complicates the chemical equilibrium, as the chemical species can react with H or C atoms to produce HF and CF4. The chemical equilibrium calculations described below predict that fluorine occurs mostly as HF in the N2-poor phase up to a certain pressure, beyond which it appears mostly as a constituent in CF4 in the N2-rich phase. But the shift in fluorine chemistry is sensitive to intermolecular potentials involving HF and can be abrupt in thermodynamic sense, thereby enhancing the character of the N2-fluid phase change. Relevance of the present prediction to detonation properties of high explosives containing fluorine binders is discussed.