ESR experiments with 2,2,6,6-tetramethyl-4-piperidone (4-oxo-TEMP) and the spin-trap 5,5-dimethyl pyrroline-N-oxide (DMPO) have been performed on a series of new phthalocyanines: the bis(tri-n-hexylsiloxy) silicon phthalocyanine ([(nhex)3SiO]2SiPc), the hexadecachloro zinc phthalocyanine (ZnPcCl16), the hexadecachloro aluminum phthalocyanine (AlPcCl16), the hexadecachloro aluminum phthalocyanine sulfate (HSO4AlPcCl16), whose photocytotoxicity has been studied against various leukemic and melanotic cell lines. Type I and Type II pathways occur simultaneously in DMF although the Type II seems to be prevalent. These results are not changed when the bis(tri-n-hexylsiloxy) silicon phthalocyanine is entrapped into liposomes. By contrast, the Type I process is favored in membrane models for all the perchlorinated phthalocyanines. This modified behavior may be accounted on a possible stacking of phthalocyanines in membranes and a preventing effect of axial ligands against aggregation in the case of the bis(tri-n-hexylsiloxy) silicon phthalocyanine. The photodynamic action of zinc perchlorinated phthalocyanine is not dependent on singlet oxygen, phototoxicity of this molecule being essentially mediated by oxygen free radicals. Quantitation of the superoxide radical was accomplished, with good agreement, by two techniques: the cytochrome c reduction and the ESR quantitation based on the double integration of the first derivative of the ESR signal. The disproportionation of the superoxide radical or degradation of the spin-trap seem to be avoided in aprotic solvents such as DMF.