The creation of efficient catalysts for splitting water into H2 and O2 is one of the greatest challenges for chemists working on the production of renewable fuel. The water oxidizing center (WOC) within photosynthetic organisms is the only natural system able to efficiently photooxidize water using visible light, and is thus a blueprint for catalyst design. One of the atomic structural models of the WOC derived from X-ray diffraction involves a “cubelike” core comprised of a {CaMn3O4} unit tethered to a fourth manganese atom through one or two bridging oxo units. A few nonbiological tetramanganese complex mimics of this site have been prepared that contain an incomplete or distorted cubic {Mn4Ox} core [4–7] or are part of a larger Mnx–oxo lattice. [4] However, none of these have shown activity towards water oxidation. We have previously synthesized a prototypical molecular manganese–oxo cube [Mn4O4] n+ in a family of “cubane” complexes [Mn4O4L6], where L is a diarylphosphinate ligand (p-R-C6H4)2PO2 (R=H, alkyl, OMe). The diphenylphosphinate complex (1, R=H, Figure 1) assembles spontaneously from manganese(II) and permanganate salts in high yield in non-aqueous solvents. The release of O2 by the {Mn4O4} 6+ core in 1 was shown to be possible on thermodynamic grounds, but cannot take place because of the rigidity of the core arising from the six diarylphosphinate ligands, which bridge pairs of manganese atoms on the six cube faces. The assembly of 1 is also driven by intramolecular van der Waals forces that attract three aryl rings from adjacent phosphinate ligands. The cubic core in 1 is a much stronger oxidant than any known dimanganese complex with {Mn2O2} 3+ cores. Cubane 1 abstracts hydrogen atoms from various organic substrates by breaking O H and N H bonds with dissociation energies greater than 390 kJmol . Titrations of 1 against compounds containing either amine or phenol groups reach an end point after the abstraction of four successive hydrogen atoms, yielding two water molecules (from corner oxo groups) plus [L6Mn4O2], the so-called “pinned butterfly” complex 2 (Scheme 1). {Mn4O4} cubane complexes are unique in releasing an O2 molecule upon photoexcitation of the Mn !O charge transfer band, which reaches a maximum at 350 nm. This process, which occurs with high quantum efficiency only in the gas phase, involves the core oxygen atoms and is triggered by ejection of one phosphinate ligand, thereby generating the [L5Mn4O2] + “butterfly” complex 3 (Scheme 1). In contrast, noncuboidal manganese molecular complexes possessing {Mn2O}, {Mn2O2}, and {Mn3O6} cores in the Mn or Mn oxidation states fail to release O2, but instead photodecompose into multiple fragments. Thus, O2 release is favored by complexes with a {Mn4O4} cubane core. The composition of the butterfly complexes 2 and 3 differs only by one phosphinate ligand (Scheme 1). This finding suggests the possibility of creating a catalytic cycle that could oxidize two water molecules bound to 2 along the reverse pathway in Scheme 1 (1-3H!1-2H!1-H!1), eventually forming 3 by photochemical release of O2 and a phosphinate ligand. Thus far it has proved impossible to realize a catalytic cycle, as in Scheme 1, because O2 is not photodissociated from 1 or 1 (the one-electron oxidized cubane) in condensed phases. This was attributed to a large activation barrier for O2 release when all the phosphinate ligands remain ligated or re-ligate by fast geminate recombination. Figure 1. X-ray crystal structure of 1.