Environmentally persistent free radicals (EPFRs) have been recognized as one of the important emerging contaminants with biological toxicity, environmental persistence, and global mobility. Previous studies have identified the catalytic role of surface metal oxides in EPFRs formation and illustrated the metal-dependence of EPFRs by studying on various metal oxide nanoparticles and single crystals. However, there is still lack of an understanding on the formation of EPFRs from the point of view of metal sites. Various factors (e.g., crystalline phases and surface species) of metal oxides are regarded to contribute to the generation of EPFRs, which present profound difficulties for scientists to tease apart the impact of metal type. Herein, a laboratory investigation, in terms of the acidity and oxidation strength of metal cations, was conducted by selecting metal-variable isostructural metal-organic frameworks as material platforms. Specifically, we evaluated EPFRs generation on MIL-100(M) (M = Al, Cr, Fe) from chlorine-substituted phenol vapor and catechol under thermal conditions. It is found that high Lewis acidity of metal sites is crucial for capturing the above two phenolic precursors, activating the O-H bond and promoting EPFRs formation. Radical species with half-life as long as 70 days were generated on MIL-100 rich in 5-fold coordinated Al3+ sites. The unpaired electron spin density donation was further confirmed by using 27Al solid-state nuclear magnetic resonance spectroscopy. Despite their higher oxidation power than Al3+, the exposed Cr3+ and Fe3+ sites show undetectable catalytic activity for the formation of EPFRs, because of their insufficient Lewis acidity. Our results suggest that the surface species rather than Lewis acid sites may be a major contributor to the formation of EPFRs on metal oxides like Fe2O3.