Sorption of U(VI) to goethite is a fundamental control on the mobility of uranium in soil and groundwater. Here, we investigated the sorption of U on goethite using EXAFS spectroscopy, batch sorption experiments and DFT calculations of the energetics and structures of possible surface complexes. Based on EXAFS spectra, it has previously been proposed that U(VI), as the uranyl cation UO 2 2þ , sorbs to Fe oxide hydroxide phases by forming a bidentate edge-sharing (E2) surface complex, >Fe(OH) 2 UO 2 (H 2 O) n . Here, we argue that this complex alone cannot account for the sorption capacity of goethite (a-FeOOH). Moreover, we show that all of the EXAFS signal attributed to the E2 complex can be accounted for by multiple scattering. We propose that the dominant surface complex in CO 2 -free systems is a bidentate corner-sharing (C2) complex, (>FeOH) 2 UO 2 (H 2 O) 3 which can form on the dominant {101} surface. However, in the presence of CO 2 , we find an enhancement of UO 2 sorption at low pH and attribute this to a (>FeO)CO 2 UO 2 ternary complex. With increasing pH, U(VI) desorbs by the formation of aqueous carbonate and hydroxyl complexes. However, this desorption is preceded by the formation of a second ternary surface complex (>FeOH) 2 UO 2 CO 3 . The three proposed surface complexes, (>FeOH) 2 UO 2 (H 2 O) 3 , >FeOCO 2 UO 2 , and (>FeOH) 2 UO 2 CO 3 are consistent with EXAFS spectra. Using these complexes, we developed a surface complexation model for U on goethite with a 1-pK model for surface protonation, an extended Stern model for surface electrostatics and inclusion of all known UO 2 -OH-CO 3 aqueous complexes in the current thermodynamic database. The model gives an excellent fit to our sorption experiments done in both ambient and reduced CO 2 environments at surface loadings of 0.02-2.0 wt% U.