Suppression of Three-Dimensional Charge Density Wave Ordering via Thickness Control

Barium bismuth oxide (${\mathrm{BaBiO}}_3$) is the end member of two families of high-$T_c$ superconductors, i.e., ${\mathrm{BaPb}}_{1-x}{\mathrm{Bi}}_x{\mathrm{O}}_3$ and ${\mathrm{Ba}}_{1-x}{\mathrm{K}}_x{\mathrm{BiO}}_3$. The undoped parent compound is an insulator, exhibiting a charge density wave that is strongly linked to a static breathing distortion in the oxygen sublattice of the perovskite structure. We report a comprehensive spectroscopic and x-ray diffraction study of ${\mathrm{BaBiO}}_3$ thin films, showing that the minimum film thickness required to stabilize the breathing distortion and charge density wave is $\approx 11$ unit cells, and that both phenomena are suppressed in thinner films. Our results constitute the first experimental observation of charge density wave suppression in bismuthate compounds without intentionally introducing dopants.

Figure 1

Structural characterization of the BBO films. (a) A schematic diagram of the perovskite lattice structure. The blue arrows show the atomic motion associated with the oxygen breathing phonon mode. (b) A schematic diagram of the static breathing distortion, showing the three-dimensional checkerboard ordering of nonequivalent ${\mathrm{BiO}}_6$ octahedra. (c) XRD $\theta \text{−}2\theta$ patterns, showing the (002) reflections of BBO and the STO substrates for film thicknesses in the range 6–170 u.c. The measurements employed the Cu $K_{\alpha }$ line. (d) XRD $\theta \text{−}2\theta$ patterns of thick 120-BBO, showing the (222) series of reflections. (e) The presence of a static breathing distortion results in a doubling of the unit cell, as well as the appearance of a half-integer ($3/2$, $3/2$, $3/2$) reflection. The data in (d),(e) were acquired using a synchrotron source with $E=11.03\mathrm{keV}$. All reflections are indexed to pseudocubic notation.

Figure 2

XRD reciprocal space maps of the BBO films, acquired using (103) reflections. The reciprocal coordinates are provided in terms of the substrate lattice parameter $a_{\mathrm{STO}}=3.905\AA$. The dashed lines indicate the positions of (103) reflections expected for an undistorted cubic lattice. (a) A thick 120-BBO film. The reflections from the substrate and from BBO are well separated, demonstrating that the BBO film is relaxed. (b)–(e) Evolution of the BBO (103) reflections with the film thickness. (b) 120, (c) 17, (d) 13, and (e) 9 u.c. The color scale was adjusted separately for each sample.

Figure 3

Spectroscopic characterization of the BBO films. (a) The Raman response at $\lambda =633\mathrm{nm}$. The spectral features comprise a mode at $\sim 565{\mathrm{cm}}^{-1}$, which results from the BBO oxygen breathing phonon (see the inset), as well as first- and second-order modes of the substrate (STO, 2STO). For $d\le 9\mathrm{u}.\mathrm{c}.$, the oxygen breathing mode is not Raman active. (All Raman spectral data are shown as measured without normalization or vertical shifts.) (b) The amplitude of the Raman response for the oxygen breathing mode relative to the STO response baseline. (c) The evolution of optical conductivity at the absorption maximum, ${\sigma }_1(2\mathrm{eV})$, with sample thicknesses. The dashed line highlights discontinuous behavior at $d_c\simeq 11\mathrm{u}.\mathrm{c}.$ The solid lines are guides to the eye. (d) The optical conductivity ${\sigma }_1$ measured via spectroscopic ellipsometry. The absorption maximum at $\sim 2\mathrm{eV}$ marked by the dashed line reflects the CDW gap in BBO.