Motivated by recent nuclear magnetic resonance (NMR) experiments, we present a microscopic sp 3 tight-binding model calculation of the NMR shifts in bulk Bi2Se3 and Bi2Te3. We compute the contact, dipolar, orbital and core polarization contributions to the carrier-density-dependent part of the NMR shifts in 209 Bi, 125 Te and 77 Se. The spin-orbit coupling and the layered crystal structure result in a contact Knight shift with strong uniaxial anisotropy. Likewise, because of spin-orbit coupling, dipolar interactions make a significant contribution to the isotropic part of the NMR shift. The contact interaction dominates the isotropic Knight shift in 209 Bi NMR, even though the electronic states at the Fermi level have a rather weak s-orbital character. In contrast, the contribution from the contact hyperfine interaction to the NMR shift of 77 Se and 125 Te is weak compared to the dipolar and orbital shifts therein. In all cases, the orbital shift is at least comparable to the contact and dipolar shifts, while the shift due to core polarization is subdominant (except for Te nuclei located at the inversion centers). By artificially varying the strength of spin-orbit coupling, we evaluate the evolution of the NMR shift across a band inversion but find no clear signature of the topological transition. arXiv:1602.02649v2 [cond-mat.mes-hall]

Recent theoretical work has established the presence of hidden spin and orbital textures in nonmagnetic materials with inversion symmetry. Here, we propose that these textures can be detected by nuclear magnetic resonance (NMR) measurements carried out in the presence of an electric field. In crystals with hidden polarizations, a uniform electric field produces a staggered magnetic field that points to opposite directions at atomic sites related by spatial inversion. As a result, the NMR resonance peak corresponding to inversion partner nuclei is split into two peaks. The magnitude of the splitting is proportional to the electric field and depends on the orientation of the electric field with respect to the crystallographic axes and the external magnetic field. As a case study, we present a theory of electric-field-induced splitting of NMR peaks for 77 Se, 125 Te and 209 Bi in Bi2Se3 and Bi2Te3. In conducting samples with current densities of 10 6 A/cm 2 , the splitting for Bi can reach 100 kHz, which is comparable to or larger than the intrinsic width of the NMR lines. In order to observe the effect experimentally, the peak splitting must also exceed the linewidth produced by the Oersted field. In Bi2Se3, this requires narrow wires of radius 1 µm. We also discuss other potentially more promising candidate materials, such as SrRuO3 and BaIr2Ge2, whose crystal symmetry enables strategies to suppress the linewidth produced by the Oersted field. arXiv:1709.02376v2 [cond-mat.mtrl-sci]