Polar discontinuity governs surface segregation and interface termination: A case study of $\mathrm{LaIn}{\mathrm{O}}_3/\mathrm{BaSn}{\mathrm{O}}_3$

We combine (scanning) transmission electron microscopy, electron energy loss spectroscopy (EELS), and x-ray photoelectron spectroscopy (XPS) with density functional theory (DFT) to investigate the complex interplay of surface energetics and polar discontinuity compensation in the interface formation of (001) ${\mathrm{LaInO}}_3/{\mathrm{BaSnO}}_3$-based heterostructures. We present evidence from both experiment and theory that the $\mathrm{BaSn}{\mathrm{O}}_3$ surface with BaO termination is energetically favored over a wide range of chemical potentials. However, overgrowth of the nonpolar BaO-terminated surface with $\mathrm{LaIn}{\mathrm{O}}_3$ results in a $\mathrm{Sn}{\mathrm{O}}_2$-terminated interface. EELS and XPS show that this interfacial termination exchange is mediated by Ba segregation to the growth surface. Our DFT calculations show that the efficient compensation of the polar discontinuity, attributed to polar distortions within $\mathrm{BaSn}{\mathrm{O}}_3$ at the $\mathrm{Sn}{\mathrm{O}}_2$-terminated interface, serves as the driving force behind the observed Ba segregation. This intricate interplay underscores the importance of polar discontinuity compensation as a pivotal factor influencing interface formation in perovskite systems.

Figure 1

(a) $\mathrm{BaSn}{\mathrm{O}}_3$ (001) single crystal and (b) thin film surfaces viewed along the [110] direction under TEM high-vacuum conditions. HRTEM simulations are shown as white-framed insets, while white arrows shown in (b) are pointing to the 1-unit-cell-high surface steps.

Figure 2

$\mathrm{BaSn}{\mathrm{O}}_3$ single crystal (001) surface seen along [110] projection under high-vacuum conditions and electron beam (a) before and (b) after 15 s. HRTEM simulations are shown as white-framed insets, while yellow arrow in (b) shows the direction of the atomic reconstruction of the surface exposed to the electron beam.

Figure 3

(a) High-resolution dark-field STEM image of the region where the EELS intensities are extracted. (b) Magnified dark-field STEM image of the region inside the yellow box shown in (a) with overlaid supercell built by DFT. (c) Laterally averaged EELS intensities of the Ba, Sn, La, and In $M_{4,5}$ edges showing the LaO/$\mathrm{Sn}{\mathrm{O}}_2$ plane termination at the $\mathrm{LaIn}{\mathrm{O}}_3/\mathrm{BaSn}{\mathrm{O}}_3$ interface shown in (a).

Figure 4

(a) High-resolution dark-field STEM image of the $\mathrm{LaIn}{\mathrm{O}}_3/\mathrm{BaSn}{\mathrm{O}}_3$ heterostructure grown by MBE showing the area used for the EELS analysis. (b) EELS spectra showing Ba $M_{4,5}$ and La $M_{4,5}$ edges extracted from the regions shown in (a).

Figure 5

(a) HAADF STEM image of the $\mathrm{LaIn}{\mathrm{O}}_3/\mathrm{BaSn}{\mathrm{O}}_3$ heterostructure grown by PLD showing the area used for the EELS line scan. The green arrow highlights the interface profile of 5 nm length where the EELS signal was collected, while the yellow line indicates the area of lateral integration of the EELS signal. (b) The high-loss EELS line spectrum of the region marked with the green arrow shown in (a). (c) Corresponding EELS spectra of the regions marked with blue (1), red (2), and green (3) shown in (a) and (b).

Figure 6

Ba $3d_{5/2}$ core level spectra of a $\mathrm{LaIn}{\mathrm{O}}_3$ (LIO)/$\mathrm{BaSn}{\mathrm{O}}_3$ (BSO) heterostructures at varying photon energies. SXPS data of (a) 4-nm-thick and (b) 8-nm-thick $\mathrm{LaIn}{\mathrm{O}}_3$ layer at photon energy of 1.5 keV. HAXPES data of 8-nm-thick $\mathrm{LaIn}{\mathrm{O}}_3$ layer at photon energies of (c) 3.4 keV and (d) 6.0 keV. The black lines represent the measured data, while the green and red lines result from peak fitting using the CasaXPS software, assuming two contributions, one from Ba which comes from the $\mathrm{BaSn}{\mathrm{O}}_3$ layer (S2) and one from a Ba signal that either results from a Ba compound inside the $\mathrm{LaIn}{\mathrm{O}}_3$ layer or from a Ba surface layer (S1). The blue line represents the Shirley background (BG).

Figure 7

(a) Ba from $\mathrm{BaSn}{\mathrm{O}}_3$–to–Ba from BaO surface layer ratio derived from XPS analysis (solid icons) and from the DDF model (empty icons) assuming the structure shown in the lower right corner. (b) Ba from $\mathrm{BaSn}{\mathrm{O}}_3$–to–Ba from $\mathrm{LaIn}{\mathrm{O}}_3$ contamination ratio derived from XPS analysis (solid icons) and from DDF model (empty icons) assuming the structure in the lower right corner.

Figure 8

Surface energy diagram for the $\mathrm{BaSn}{\mathrm{O}}_3$ (001) surface as a function of the chemical potentials of Sn and O, in eV. The $\mathrm{BaSn}{\mathrm{O}}_3$ phase can be formed within the green area, while the favorable termination will be BaO. The boundaries with other precipitates are indicted by the black solid lines.

Figure 9

(a) Top and side views of the symmetric superlattice that is considered to simulate the $p$- and $n$-type $\mathrm{LaIn}{\mathrm{O}}_3/\mathrm{BaSn}{\mathrm{O}}_3$ interfaces. The in-plane lattice parameter is fixed to that of bulk $\mathrm{BaSn}{\mathrm{O}}_3$. (b) Variation of the formation energy per atom as a function of the thickness of $\mathrm{BaSn}{\mathrm{O}}_3$ and $\mathrm{LaIn}{\mathrm{O}}_3$ layers. PC denotes pseudocubic unit cell as highlighted in (a).

Figure 10

Side view of the nonperiodic $\mathrm{LaIn}{\mathrm{O}}_3/\mathrm{BaSn}{\mathrm{O}}_3$ interface with (a) BaO-terminated and (b) $\mathrm{Sn}{\mathrm{O}}_2$-terminated $\mathrm{BaSn}{\mathrm{O}}_3$. The orientations of the formal polarization $P_0$ and polarization induced by structural distortions Δ$P$ are highlighted by black and magenta arrows, respectively, in both panels. The top view of (b) is shown in (c).