Gas Doping on the Topological Insulator $\mathbf{B}{\mathbf{i}}_2\mathbf{S}{\mathbf{e}}_3$ Surface

Gas molecule doping on the topological insulator ${\mathrm{Bi}}_2{\mathrm{Se}}_3$ surface with existing Se vacancies is investigated using first-principles calculations. Consistent with experiments, ${\mathrm{NO}}_2$ and ${\mathrm{O}}_2$ are found to occupy the Se vacancy sites, remove vacancy-doped electrons, and restore the band structure of a perfect surface. In contrast, NO and ${\mathrm{H}}_2$ do not favor passivation of such vacancies. Interestingly we have revealed a ${\mathrm{NO}}_2$ dissociation process that can well explain the speculative introduced “photon-doping” effect reported by recent experiments. Experimental strategies to validate this mechanism are presented. The choice and the effect of different passivators are discussed. This step paves the way for the usage of such materials in device applications utilizing robust topological surface states.

Figure 1

Atomic structures and band structures for (a) a perfect ${\mathrm{Bi}}_2{\mathrm{Se}}_3$ surface, (b) a surface with a Se vacancy, (c) ${\mathrm{NO}}_2$ adsorption, (d) O adsorption from the dissociated ${\mathrm{NO}}_2$, (e) ${\mathrm{O}}_2$ adsorption, and (f) ${\mathrm{H}}_2$ adsorption. Bands are aligned to their valence band maximum at $\overline{\Gamma }$. The Fermi energy is shifted to zero. (a)–(d) show structures from the side view, whereas (e),(f) show structures from the top view. Red (thick) lines indicate the surface states. The transparent ball indicates the missing Se at the vacancy in (b). Band structures are calculated from a $3\times 3$ supercell of 7 QLs thick. Colored circles denote Dirac points.

Figure 2

(a) Schematic of ${\mathrm{NO}}_2$ molecule adsorption and dissociation processes and (b) corresponding reaction energetics on the ${\mathrm{Bi}}_2{\mathrm{Se}}_3$ surface with a Se vacancy. (SOC is not included in these energies) The Se vacancy is indicated by the white dot (step I) and donates two electrons that can dope the surface. During ${\mathrm{NO}}_2$ exposure, the molecule occupies the vacancy site quickly with one O binding to three Bi atoms from the second atomic layer (step II). The corresponding N-O bond becomes weaker compared to before adsorption (step I). The stable adsorption structure is shown in (b) together with the charge density difference before and after adsorption. The donor electrons transfer partially from the vacancy to ${\mathrm{NO}}_2$. Under external photon exposure (step III), the weakened N-O bond breaks, resulting in ${\mathrm{NO}}_2$ dissociation into $\mathrm{NO}+\mathrm{O}$ (step IV). The NO molecule leaves the surface, while O passivates the vacancy. In this case, two donor electrons transfer to the O atom and the vacancy is compensated. White balls stand for Bi atoms, red for O, and blue for N. In the isovalue surface plots of the charge difference, solid blue (wired gray) color denotes charge depletion (gain).

Figure 3

Density of states (DOS) for (a) the perfect pristine surface and the one with a Se vacancy, (b) ${\mathrm{NO}}_2$, (c) a single O atom, (d) ${\mathrm{O}}_2$, and (e) ${\mathrm{H}}_2$ passivated. To be more clear, the adsorbate’s partial DOS has been scaled by $N_X/N_{\mathrm{Tot}}$, where $N_X$ is the number of atoms in the adsorbate and $N_{\mathrm{Tot}}$ is the total number of atoms in the supercell.