Figure 1

(a) The crystal structure of Bi${}_2$Se${}_3$ can be described with space group $R\overline{3}m$, where the quintuple layer is the building block with Se2 (blue) in the middle and Se1 (green) near the van der Waals gap. The inset shows the BiSe${}_6$ octahedron and the Se1 (or Bi${}_{\text{Se}1}$ antisite in dashed circle) environment. (b) Crystals grown using $\text{Bi}:\text{Se}=2:3$ and $2:(3+x)$ of stoichiometric and Se-rich fluxes, with a (001) plane index mostly perpendicular to the growth direction. (c) Crystals grown using $\text{Bi}:\text{Se}=(2+x):3$ of Bi-rich flux show a (001) plane index along the growth direction.Reuse & Permissions

Figure 2

STEM-HAADF images of Bi${}_2$Se${}_3$. (a) A high resolution image shows that the quintuple of Bi${}_2$Se${}_3$ consists of two bright spots (Bi) and three fainter spots (Se). (b) For the Bi-rich flux growth, a high density of Bi can be found in the grain boundary, as shown by the thick and bright lines at low magnification. (c) A low density of Bi${}_2$-layer intercalated patches can always be found in the van der Waals gap for the as-grown crystals, whether using Bi- or Se-rich flux. The intercalated Bi${}_2$-layer patches would deform the quintuple layer locally by opening up the van der Waals gap. (d) A large area of perfectly ordered quintuple layers without intercalated Bi${}_2$-layer patch induced deformation can be found in samples after 300 ${}^{\circ }$C annealing.Reuse & Permissions

Figure 3

Lattice parameters vs lattice volume of all crystals studied, where c axes are shown as empty circles and a axes are shown as empty squares. All studied samples are clustered in groups of Bi-rich flux growth before annealing (Bi-rich), Se-rich flux growth before annealing (Se-rich), and Se-rich plus post annealing at 300–600 ${}^{\circ }$C (Se-rich+300–600 ${}^{\circ }$C), in an increasing trend, as indicated by the connected experimental data points. A list of Bi${}_2$Se${}_3$ lattice parameters reported in the literature since 1960 are shown as various solid symbols (Ref. 8).Reuse & Permissions

Figure 4

Drude mobility ($\mu$) vs carrier concentration ($n$) for all samples studied are shown; all are $n$ type. Crystals grown from Bi-rich flux (Bi-rich) have the highest carrier concentration but the lowest mobility, and 300 ${}^{\circ }$C postannealing (Bi-rich+300 ${}^{\circ }$C) reduces the carrier concentrations to the same level of $\sim$$-15\times {10}^{18}$. Crystals from stoichiometric and Se-rich flux growth (Se-rich) have the lowest carrier concentration and highest mobility, and the 300 ${}^{\circ }$C postannealing (Se${}_{3.10}$+300 ${}^{\circ }$C) shows no significant impact on modifying the carrier concentration and mobility.Reuse & Permissions

Figure 5

(a) STEM-HAADF image of a bulk region with thickness $\sim$25 nm along the incident beam direction. The Se1 column pointed out by the white arrow shows an anomalously enhanced imaging contrast. Further STEM-EDX chemical mappings of the corresponding boxed region are enlarged in the right panels (cyan, Bi map; yellow, Se map). The gray (white) circles in the mapping represent the Bi (Se) columns. The enhanced contrasts of the Se1 columns (signified by the red Se1 text) are shown to be correlated with the presence of Bi on the sites, i.e., the formation of Bi${}_{\text{Se}}$ antisite defects. (b) The enhanced Se1 contrasts are more regularly observed at the edge of the same specimen, as indicated by the white arrows, with a corresponding enhanced blowup of the white boxed region in the inset in the lower left corner.Reuse & Permissions