The 680-nm-absorbing "peroxide state" of the Escherichia coli cytochrome d terminal oxidase complex, obtained by addition of excess hydrogen peroxide to the enzyme, is shown to be a ferryl intermediate in the catalytic cycle of the enzyme. This ferryl intermediate is also created by aerobic oxidation of the fully reduced enzyme. Resonance Raman spectra with 647.1-nm excitation show an Fe(IV) = O stretching band at 815 cm-1, a higher frequency than noted in any other ferryl-containing enzyme to date. The band shows an O-16/O-18 frequency shift of -46 cm-1, larger than that observed for any porphyrin ferryl species. The Fe(IV) = O formulation was unambiguously established by oxidations of the reduced enzyme with (O2)-O-16, (O2)-O-18, and O-16 O-18. Only the use of a mixed-isotope gas permitted discrimination between a ferryl and a peroxo structure. A catalytic cycle for the cytochrome d terminal oxidase complex is proposed, and possible reasons for the high nu(Fe = O) frequency are discussed.