Structural isomers of [UO2(oxalate)(3)](4-), [UO2(oxalate)F-3](3-), [UO2(oxalate)(2)F](3-), and [UO2(oxalate)(2)(H2O)](2-) have been studied by using EXAFS and quantum chemical ab initio methods. Theoretical structures and their relative energies were determined in the gas phase and in water using the CPCM model. The most stable isomers according to the quantum chemical calculations have geometries consistent with the EXAFS data, and the difference between measured and calculated bond distances is generally less than 0.05 Angstrom. The complex [UO2(oxalate)(3)](4-) contains two oxalate ligands forming five-membered chelate rings, while the third is bonded end-on to a single carboxylate oxygen. The most stable isomer of the other two complexes also contains the same type of chelate-bonded oxalate ligands. The activation energy for ring opening in [UO2(oxalate)F-3](3-), DeltaU(double dagger) = 63 kJ/mol, is in fair agreement with the experimental activation enthalpy, DeltaH(double dagger) = 45 +/- 5 kJ/mol, for different [UO2(PiCOlinate)F-3](2-) complexes, indicating similar ring-opening mechanisms. No direct experimental information is available on intramolecular exchange in [UO3(oxalate)(3)](4-). The theoretical results indicate that it takes place via the tris-chelated intermediate with an activation energy of AV = 38 kJ/mol; the other pathways involve multiple steps and have much higher activation energies. The geometries and energies of dioxouranium(VI) complexes in the gas phase and solvent models differ slightly, with differences in bond distance and energy of typically less than 0.06 Angstrom and 10 kJ/mol, respectively. However, there might be a significant difference in the distance between uranium and the leaving/entering group in the transition state, resulting in a systematic error when the gas-phase geometry is used to estimate the activation energy in solution. This systematic error is about 10 kJ/mol and tends to cancel when comparing different pathways.