Ceria-based metal oxides are promising redox materials for solar H 2 O/CO 2 -splitting thermochemical cycles. Density functional theory (DFT) computations are applied to elucidate the underlying mechanism of the role of dopants in facilitating ceria based redox cycles; specifically, we explain why some dopants increase performance, while others do not. Firstly, we find that Zr and Hf dopants increase the oxygen exchange capacity of ceria because they store energy in tensilely strained Zr-or Hf-O bonds which is released upon O-vacancy formation. This finding corrects a long held assumption that Zr and Hf decrease the O-vacancy formation energy by compensating for ceria expansion upon reduction.Although the released strain energy decreases the O-vacancy formation energy, O-vacancy formation remains sufficiently endothermic to split H 2 O and CO 2 . Secondly, we show that two electrons must be promoted into the high energy Ce f-band during reduction if the O-vacancies are to store sufficient energy to drive the oxidative gas splitting step. This means that di-and trivalent dopants are not suitable for this process. Lastly, we show that dopants which break O bonds due to their small size or strongly covalent character, such as Ti and the pentavalent dopants, substantially decrease the O-vacancy formation energy because only three O bonds must break during reduction. These vacancies, therefore, are too low in energy to drive gas splitting. Based on these findings, we develop guidelines for new ceria doping strategies to facilitate solar thermochemical gas splitting cycles.