Density-functional Theory (DFT) approaches have recently been used to judge the topological order of various materials despite its well-known band gap underestimation. Use of the more accurate quasi-particle GW approach reveals here few cases where DFT identifications are false-positive, possibly misguiding experimental searches of materials that are topological insulators (TI) in DFT but not expected to be TI in reality. We also present the case of false-positive due to the incorrect choice of crystal structures and adress the relevancy of such choice of crystal structure with respect to the ground state one and thermodynamical instability with respect to binary competing phases. We conclude that it is then necessary to consider both the correct ground state crystal structure and the correct Hamiltonian in order to predict new TI. 72.25.Hg, 3D bulk topological insulators (TI) have a band inversion, e.g. the s-like conduction band below the (p, d)-like valence band at the time reversal invariant momentum such as Γ. This order of bands is defined through the inversion energy ∆ i = ε s − ε p,d being negative (see Figure 1). Specifically, for normal insulators such as CdTe, ∆ i is positive while for topological insulators such as HgTe, ∆ i is negative. The excitation gap defined as the energy difference between the highest occupied state and the lowest unoccupied state can be either zero (Fig 1b) or positive (Fig 1c) even in TI. When one tuncates a 3D bulk TI to create a 2D surface or interface, new states will appear inside the 2D excitation band gap: these will have linear dispersion ('massless fermions') crossing each other ('gapless') within the 2D excitation gap and maintain spin-polarization without net charge current.Interest in their exotic properties has motivated recent search and discovery efforts of materials that would be topological insulators [1][2][3][4][5][6][7][8]. Theory predicts interesting cases expected to be TI's and pleas are then made to experimentalists to make and probe such materials in the laboratory. As exciting as such a close level of theoryexperiment interaction is, there are two potential issues that can impede progress. First, most of these calculations rely on density functional theory (DFT) or variants of it [1-8], yet these methods are known to systematically underestimate band gaps specifically by placing the conduction band too low and the valence band too high, thereby creating the potential of 'false positive' prediction of a material being TI in DFT but not in reality, to the detriment of experiment-theory interaction. Second, sometimes [1-8] theoretical predictions of negative inversion energy ∆ i are conducted on assumed crystal structures of given candidate materials; in some cases those assumed structures are energetically quite far from the ground state structure. This results in predicting an * Electronic address: alex.zunger@gmail.com exciting property (e.g., TI) in unrealizable structures: we present here a case where the assumed crystal structure makes the comp...

No abstract

No abstract

No abstract

No abstract