“…In terms of lattice mismatch, F occupancy of equatorial X 1 sites might be expected (giving perovskite blocks of a similar width to the [Bi 2 O 2 ] 2+ layers) but tilting of the Ti X 6 octahedra, about in-plane and out-of-plane axes, can also reduce this interfacial mismatch while maintaining satisfactory Ti– X bond lengths. The bond valence sum analysis and our understanding of strain suggest that F occupancy of equatorial X 1 sites is certainly possible in Bi 2 TiO 4 F 2 and is consistent with the conclusions of Needs et al 39 However, this contrasts with several other n = 1 Aurivillius and Ruddlesden-Popper oxyfluorides including Bi 2 NbO 5 F, 54 Sr 2 ScO 3 F, 67 and Sr 2 MnO 3 F, 68 which are reported to have F occupancy of apical anion sites despite similar arguments involving stacking strain between [Bi 2 O 2 ] 2+ and Nb(O,F) 2 layers ( a p = 3.95 Å and a p = 4.08 Å for F – in equatorial X 1 or apical X 2 sites, respectively) or between SrO and Mn(O,F) 2 layers (ideal a SrO ≈ 3.65 Å; for Sr 2 ScO 3 F a p = 4.12 Å and a p = 4.19 Å for F – in equatorial or apical sites, respectively; for Sr 2 MnO 3 F a p = 3.86 Å and a p = 3.99 Å for F – in equatorial or apical sites, respectively). It is striking that Bi 2 TiO 4 F 2 has relatively small stacking strain compared to these examples, suggesting the possibility that anion distribution might be tuned by strain engineering, for example.…”