BaBiO3 has attracted a lot of research attention since it was discovered as the parent compound for the high-Tc superconducting BaPbxBi1–xO3 and Ba1–xKxBiO3. In its pure state, BaBiO3 is an insulator due to the presence of a breathing distortion of the BiO6 octahedra. The distortion is attributed to the valency of Bi in the compound being charge-ordered in the form of Bi3+ and Bi5+ along the lattice, resulting in alternating expanded or contracted BiO6 octahedra. The interaction between the electronic properties and the thickness of the thin film is crucial to study. We conducted a thorough study to investigate the effect of the thickness reduction of BaBiO3-δ grown on SrTiO3-buffered Si substrates on the optical properties as well as the Bi electronic structure of the thin films. We conclude that modifications in the valency of Bi in the ultra-thin film regime result in an optically conducting layer.

The perovskite structure, with the chemical formula ABO3, has allowed for exploring a variety of interesting physics such as ferroelectricity,1 magnetism,2 and superconductivity.3 In this representation, the B cation is octahedrally coordinated by oxygen ions, forming a network of BO6 across the crystal, which contributes to shaping the material’s overall band structure. Structural distortions of the octahedra, therefore, play a crucial role in defining the electronic and magnetic properties of the perovskites. BaBiO3 (BBO3) has been studied for its use in applications such as photovoltaics,4 photocatalysis for water splitting,5 and superconducting devices,6 and it is well represented by this phenomenon. BBO3 has a unique phase diagram where it undergoes a phase transition at 430 K from monoclinic in space group C2/m to rhombohedral in space group R3̄ and finally to a cubic structure in space group Pm3m above 820 K.7 At room temperature, according to neutron profile measurements, the monoclinicity of BBO3 is crystallographically characterized by having two distinguishable Bi–O bond lengths of 2.12 to 2.28 Å, forming a charge density wave (CDW).8 The insulating or metallic behavior of BBO3 has been widely attributed to the structural properties of the perovskite and the presence or absence of CDW in the BBO3 lattice.

In the ionic model, the charge disproportionation over the Bi ions has been considered for the material to have a chemical formula of Ba2Bi+3Bi+5O6. The charge ordering is associated with a static breathing lattice distortion where Bi resides within expanded and contracted octahedra in an alternating fashion. According to band theory, BBO3 is expected to be metallic with its half-filled valence band; however, the above-mentioned distortions are responsible for opening an energy gap in the conduction band, rendering the material a Peierl’s insulator with an optical gap of 1.85–2.05 eV.9–13 However, as a result of the experimental observation of no distinction among the valency of the two crystallographically inequivalent Bi ions, another bond disproportionation image came into the picture.14–17 In this regime, the strong hybridization between Bi 6s and O 2p orbitals excludes the charge ordering in favor of having only Bi3+ in the structure; therefore, one oxidation state but different local environments.18–20 The neutron diffraction data and charge neutrality are then justified by an unequal transfer of charge from O 2p to the two distinct Bi states.15 

Research interest has been devoted to the parent material since the discovery of its superconductivity upon p-type doping. At x = 0.3 and 0.4, respectively, BaPbxBi1–xO3 and Ba1–xKxBiO3 become superconducting, where a phase transition to a cubic structure occurs, which recovers the expected metallicity of the oxide.21,22 This phase transition is characterized by the suppression of the characteristic symmetric A1g Raman-active mode observed for the pristine material around 570 cm−1, which is attributed to the breathing phonon mode .23 Confinement of the material in thin films has been alternatively investigated to understand the link between the structural and electronic properties of BBO3 without chemically altering the composition. Thickness induced metal-to-insulator transition for SrVO3,24 LaNiO3,25 and SrIrO326 is a relevant example for the interplay between the electronic properties and the structural transition in perovskite materials. Kim et al. studied the effect of reducing the film thickness on both the structural and optical properties of BBO3.27 A phase transition to a cubic structure driven by the suppression of breathing distortion, which is characterized by the absence of the characteristic Raman-active mode for BBO3 upon thickness reduction, was observed. However, the electrical transition to a metallic system for an ultra-thin film has not yet been demonstrated. More recent studies followed, where modifications to the electronic and optical properties are obtained by reducing the thickness of the BBO3 thin films grown on SrTiO3 (STO) substrates by pulsed laser deposition (PLD),28,29 but without a conclusion on the driving force of this effect.

In this work, we performed spectroscopic characterization on BaBiO3-δ (BBO3-δ) thin films epitaxially grown by molecular beam epitaxy (MBE) on STO-buffered Si substrates. This is to understand the origin of the lattice distortion and how confinement alters the physical and optical properties of BBO3-δ thin films.

BBO3-δ thin films of different thicknesses are grown in an ultra-high vacuum oxide chamber of a Riber 49 MBE reactor on Si(001) substrates with 15 nm STO buffer layers. A detailed description of the MBE adsorption-controlled growth process can be found in our previous study,30 where a Bi-rich environment along with plasma oxygen are used to ensure the epitaxy of the volatile BiO species. The sample is then cooled down in oxygen for a short time before being transferred to the loadlock. Reflection high-energy electron diffraction (RHEED) with a 20 kV operated e-gun is utilized to monitor the thin films’ growth mode and morphology. For the crystal perfection assessment, a Phillips X’Pert Panalytical high resolution x-ray diffraction (HR-XRD) setup is used. Symmetric ω–2θ scans based on the Cu Kα1 radiation line (λ 15.4 Å) are collected from an 18° scattering angle up to 35.2°. The crystalline quality of the films is evaluated by the full width at half-maximum (FWHM) of each layer’s rocking curve (RC) peak. To investigate the effect of thickness scaling on the properties of the layers, MBE is used to grow films that are 3, 4, 25, and 47 nm thick, as measured by the x-ray reflectivity (XRR) technique.

Raman spectra are gathered upon exciting the heterostructures with 633 nm laser light on the Horiba Jobin-Yvon LabRAM HR800 Raman tool in backscattering configuration. Photoluminescence (PL) response at room temperature is also collected on the same tool with an excitation laser of 532 nm. X-ray photoemission spectroscopy (XPS) for the thin films is performed with monochromatized Al Kα1 radiation (beam energy of 1486.6 eV) in the angle integrated mode using a Physical Electronics QUANTES instrument. The optical constants of the different thin films are calculated based on spectroscopic ellipsometry (SE) measurements performed on a J. A. Woollam RC2. CompleteEASE software is used for fitting the obtained data. For transmission electron microscopy (TEM) inspection, a Titan tool is used with an operating voltage of 200 kV. Lamella is prepared by using a focused ion beam (FIB) with reduced ion current, and no scanning electron microscopy (SEM) is carried out while preparing the sample. A capping layer of 3 nm of Al2O3 is deposited in situ to protect the sample from degradation while being measured.

Figure 1(a) illustrates a high-angle annular dark-field (HAADF) scanning-TEM (STEM) image of the STO buffer layer as well as the 25 nm BBO3-δ layer with atomic resolution (ZBa = 56, ZBi = 83). Due to the 11.7% lateral lattice mismatch between the active layer and buffer layer, BBO3-δ growth is achieved via domain matching epitaxy, where each 8 unit cell is accommodated on 9 unit cells of STO.31,32 This is clearly visible in Fig. S1 based on the fast Fourier transform (FFT) filtered image and the RHEED pattern evolution at the interface as compared to that of the surface after cooling down. The epitaxial relationship of STO(001)||BBO(001), as seen in Figs. 1(c) and 1(d), is achieved without the use of a buffer layer of BaO, as was previously reported.33 Energy dispersive x-ray spectroscopy (EDS) is used to study the composition of the epilayers; a quantified profile averaged along the growth direction is shown in Fig. 1(b). A drop in the oxygen signal within BBO3-δ with respect to the STO layer is observed. This observation is attributed to a lower oxygen content in the BBO3-δ layer. Besides, a periodic modulation within the BBO3-δ epilayer can be noticed in every three perovskite unit cells along the growth direction, as the intensity is varied, in the HAADF-STEM image. Ordered oxygen vacancies are common among other oxygen deficient perovskite oxides such as SrFeO334 and SrCoO3.35 Therefore, oxygen deficiency within BBO3-δ is assumed to form an ordered oxygen vacancy superlattice. The superstructure, resulting from the ordered oxygen vacancies, is also manifested by the doubly spotty pattern in the reciprocal space image of BBO3-δ as compared to the single diffraction spot of STO, as highlighted by the blue and green ovals in Figs. 1(c) and 1(d), respectively.

FIG. 1.

(a) HAADF-STEM image showing the heterostructure of the STO buffer layer and 25 nm BBO3-δ thin film, (b) EDS profile quantified throughout STO, BBO3-δ, and Al2O3 capping layer, and (c) and (d) reciprocal space images based on the FFT analysis for STO and BBO3-δ with zoomed-in images, respectively.

FIG. 1.

(a) HAADF-STEM image showing the heterostructure of the STO buffer layer and 25 nm BBO3-δ thin film, (b) EDS profile quantified throughout STO, BBO3-δ, and Al2O3 capping layer, and (c) and (d) reciprocal space images based on the FFT analysis for STO and BBO3-δ with zoomed-in images, respectively.

Close modal

Figure 2(a) depicts the out-of-plane symmetric scans for the different samples of BBO3-δ with thicknesses of 3, 4, 25, and 47 nm. Scans are calibrated based on the Si(004) Bragg reflection. The STO(002) Bragg reflection from the buffer is observed for the four different samples at ω = 23.13°, and the BBO3-δ(002) reflection is probed at ω = 20.66° with a strong intensity-to-thickness correlation. Indeed, the diffraction peak is more pronounced for the thicker layers of 25 and 47 nm, while for the 3 and 4 nm films, the diffraction is barely captured in terms of its strength due to the limited thickness and more room for mosaic spread. The lattice constant of the virtual substrate STO is, on average, 3.92 Å. For BBO3-δ, the average out-of-plane pseudocubic lattice constant is 4.370 and 4.379 Å for the 25 and 47 nm films, respectively. On the other hand, both layers have almost the same crystalline quality, as determined by the FWHM around 0.6° of the diffraction peaks at the rocking curve configuration. Reflectivity spectra of the combined oscillations from the STO and BBO3-δ layers are shown in the XRR scans, which are used to accurately extract the thickness of the thin films based on the Fourier transform of the signals and then calculate the growth rate of the layers to be 1.2 nm/min. Figure 2(c) schematically illustrates the heterostructure grown on a Si(001) substrate alongside the RHEED patterns recorded after the full growth of the layers is accomplished. The streak-to-streak distance denotes the formation of the perovskite, as referred to in our previous study.30 The streaky patterns ensure a layer-by-layer growth mechanism with smooth surfaces for all films. In addition, the 25 and 47 nm RHEED patterns show that a ×6 surface reconstruction along ⟨110⟩ is taking place.

FIG. 2.

(a) XRD’s out-of-plane ω–2θ scans showing Si substrate, STO buffer layer, and BBO3-δ diffraction peaks for the different BBO3-δ thin films, (b) XRR scans showing thickness oscillations due to the STO buffer and BBO3-δ active layer, and (c) schematic diagram showing the heterostructure grown on Si(001) substrate and the streaky RHEED patterns for the four thin films.

FIG. 2.

(a) XRD’s out-of-plane ω–2θ scans showing Si substrate, STO buffer layer, and BBO3-δ diffraction peaks for the different BBO3-δ thin films, (b) XRR scans showing thickness oscillations due to the STO buffer and BBO3-δ active layer, and (c) schematic diagram showing the heterostructure grown on Si(001) substrate and the streaky RHEED patterns for the four thin films.

Close modal

Tilting of the BiO6 octahedra around the pseudocubic [110] axis is taking place to stabilize the structure in the presence of the breathing lattice distortion.36 This enables the lowering of the symmetry of the structure from a cubic to a monoclinic structure at room temperature and the doubling of the unit cell. Even though the diffraction peak of BBO3-δ is observed at a position close to that expected for a cubic structure, it is not straightforward to assume that the structure is cubic. This is due to the fact that the lowering of symmetry is mainly occurring because of oxygen movements, and XRD diffraction is governed by planes containing heavier atoms like those of Ba and Bi in the structure, which makes it more difficult to differentiate between the two structures by simple out-of-plane symmetric scans. However, for convenience, pseudocubic Miller indices are assigned to the diffraction peaks.

As illustrated in Fig. 3(a), the Raman spectra for the different samples are plotted and compared with a Si reference spectrum. An intense peak can be observed at a phonon frequency of 521 cm−1, corresponding to the Raman-active mode of the Si substrate. An additional peak at a phonon frequency of 565 cm−1 corresponds to the phonon breathing mode existing for BBO3 with long range A1g symmetry.23 Normalizing the spectra relative to the substrate Raman mode allows for a more qualitative view of the Raman response of the BBO3-δ layers. This dominant Raman peak for BBO3-δ is observed for film thickness between 4 and 47 nm; however, it is completely vanishing for the 3 nm film. If the 3 nm film is expected to show lower intensity due to the low Raman cross-section, the complete absence of the Raman peak indicates a different response than that of the 4 nm thin films. While an ideal cubic perovskite does not have Raman-active modes according to the polarization selection rules, this suggests that our films have a lowered symmetry due to the breathing distortion of the lattice, which is not persistent at a thickness of 3 nm. Similarly, in K-doped BBO3 sputtered film, at the composition of the cubic phase, the Raman peak is not observed and is accompanied by an insulating-to-metal transition.23 

FIG. 3.

(a) Raman spectra for the different samples showing the characteristic Raman-active modes for Si and BBO3-δ thin films at 521 and 565 cm−1, respectively, and (b) XPS measurements of the Bi 4f core-level with different film thicknesses.

FIG. 3.

(a) Raman spectra for the different samples showing the characteristic Raman-active modes for Si and BBO3-δ thin films at 521 and 565 cm−1, respectively, and (b) XPS measurements of the Bi 4f core-level with different film thicknesses.

Close modal

Since this breathing distortion of BBO3 has been linked to the presence of two oxidation states for bismuth (being a valence skipping element), the Bi 4f XPS core-level spectra for the set of samples are collected and shown in Fig. 3(b). As BBO3 oxide is known to be an insulating material, charge neutralization is needed for carrying out the XPS measurements. To compare the oxidation states, all collected spectra are binding energy calibrated according to a fixed C 1s binding energy. For 4, 25, and 47 nm, thin films have the same line shape, which is fitted by two doublets. Each doublet represents an electronic angular momentum (j = 5/2 and 7/2) of Bi 4f core-level with an energy splitting of 5.3 eV. The peak fitting is then carried out using Casa XPS software with a mixed Gaussian–Lorentzian peak shape. Utilizing two doublets refers to the obvious presence of two different oxidation states with an energy separation of 0.75 eV. However, for the thinnest 3 nm film, there is only one oxidation state according to the fitting, as shown in Table I.

TABLE I.

Bi 4f′s core-level illustrating the binding energies of the different oxidation states for the four BBO3-δ  thin films.

Binding energy (eV)
4f5/24f7/2
Bi core-levelBi5+Bi4+Bi3+Bi5+Bi4+Bi3+
Thickness (nm) 47 164.1  163.3 158.78  158.0 
25 164.0  163.3 158.7  158.0 
164.0  163.3 158.7  158.0 
 163.78   158.5 
Binding energy (eV)
4f5/24f7/2
Bi core-levelBi5+Bi4+Bi3+Bi5+Bi4+Bi3+
Thickness (nm) 47 164.1  163.3 158.78  158.0 
25 164.0  163.3 158.7  158.0 
164.0  163.3 158.7  158.0 
 163.78   158.5 

Bismuth can exist in different oxidation states for the various compounds: Bi3+ for BiFeO3 and Bi2O3, Bi5+ for NaBiO3 and Bi2O5, and sometimes even a mixture of Bi3+ and Bi5+ like for Bi2O3–Ga2O3 and CaBiO3.37–41 In the literature, the average values for the binding energies for different Bi-containing compounds and BBO3 are Bi3+4f7/2 = 158.1 eV and Bi5+4f7/2 = 158.8 eV.17,42,43 With a separation of 0.75 eV between the two peaks, the values according to the deconvolution shown in Table I suggest two distinct oxidation states, which are Bi5+ and Bi3+, agreeing with the widely known ordered perovskite formula of Ba2 Bi5+Bi3+O6. This is, however, not observed for the 3 nm film, where one oxidation state, Bi4+, exists with a binding energy between those of Bi3+ and Bi5+, as shown in Table I. Only one oxidation state for Bi 4f′s core-level is also evident by the symmetric line shape of the peak measured as well as the shorter FWHM of the peak, as illustrated in Table S1. The origin of the breathing distortion in some cases is related to the charge ordering; however, in the literature, there is limited evidence of mixed valency for BBO3.44,45 Zapf et al. related the distinction of the Raman-active mode to the presence of a dead layer that does not host the breathing distortion below 3 nm thickness.29 Yet, our XPS data show a strong correlation between the breathing distortion, hence the Raman response, and the electronic structure of the thin film.

TABLE II.

PL spectra peak position and FWHM of each of the peaks of 25 and 47 nm films.

PL fitted peak
(nm)P1P2P3P4P5
47 nm Position 640.16 619.16 602.96 583.93 565.79 
FWHM 95.71 19.50 13.21 21.87 6.71 
25 nm Position 640.15 619.79 602.81 583.86 565.70 
FWHM 99.27 18.72 12.03 16.19 5.17 
PL fitted peak
(nm)P1P2P3P4P5
47 nm Position 640.16 619.16 602.96 583.93 565.79 
FWHM 95.71 19.50 13.21 21.87 6.71 
25 nm Position 640.15 619.79 602.81 583.86 565.70 
FWHM 99.27 18.72 12.03 16.19 5.17 

To further investigate the effect of the different electronic structures on the optical conductivity of the thin films, SE measurements were carried out. Optical conductivity of the different thin films is calculated using the refractive index (n) and absorption coefficient (k) as extracted from the SE measurements with a range of wavelengths from 500 up to 1240 nm. Figure 4(a) shows how the layers are absorbing the light excitation with a peak centered around 2.1 eV for 4, 25, and 47 nm, which does not exist for the 3 nm film. This response has been linked in the literature to the CDW, which is attributed to the breathing distortion of the BiO6 octahedra and the strong hybridization of the Bi 6s and O 2p orbitals.46 The CDW generates a density of states above and below the material’s Fermi level, resulting in this absorption peak in optical conductivity. This justifies the noticeable increase in the spectral intensity of the Raman response of the 25 nm layer when excited with a laser of 633 nm wavelength compared to other lasers with wavelengths of 785, 532, and 405 nm, as shown in Fig S2. This, as pointed out previously,23 is because the laser radiation is in resonance with the optical bandgap of BBO3-δ, which is around 2.1 eV. The intensities of BBO3-δ’s Raman peak as well as optical conductivity at a photon energy of 2.1 eV for the different samples are plotted against the thickness of the layer, as shown in Fig. 4(b). The abrupt decrease in the signals going from a thickness of 4 nm down to 3 nm indicates that the effect is beyond the spectroscopic information due to the limited measured volume.

FIG. 4.

(a) Optical conductivity of the thin films as extracted from SE optical constants data, (b) spectral intensity for both Raman spectra at 565 cm−1 and optical conductivity at 2.1 eV showing the vanished intensity for the 3 nm film, as indicated by the same color codes from (a) and (c) PL spectra for the different films, and (d) schematic diagram showing the Pm3m cubic structure of BaBi4+O3 as well as the charge ordered monoclinic C2m structure of Ba2Bi3+Bi5+O6.

FIG. 4.

(a) Optical conductivity of the thin films as extracted from SE optical constants data, (b) spectral intensity for both Raman spectra at 565 cm−1 and optical conductivity at 2.1 eV showing the vanished intensity for the 3 nm film, as indicated by the same color codes from (a) and (c) PL spectra for the different films, and (d) schematic diagram showing the Pm3m cubic structure of BaBi4+O3 as well as the charge ordered monoclinic C2m structure of Ba2Bi3+Bi5+O6.

Close modal

In Fig. 4(c), the PL responses of the BBO3-δ layers are shown and compared to a Si reference spectrum at room temperature with a range from 560 to 830 nm. PL spectra of 25 and 47 nm thin films show a sharp peak (P5) with a FWHM of 5.17 and 6.71 nm and positions of 565.70 and 565.79 nm (2.19 eV), respectively. In addition, according to the fitting of the data after subtracting the background shown in Fig. S3, there is a broad emission peak (P1) around 640 nm (1.94 eV) with a FWHM of 99.27 and 95.71 for the 25 and 47 nm thin films, respectively. Other peaks are also observed (P4, P3, P2), corresponding to energies of 2.12, 2.06, 2, and 1.94 eV, respectively. Extracted information from fitting the PL data can be found in Table II.

The sharpest peak in the spectra around 2.19 eV is assigned to the band-to-band radiative transition, hence corresponding to the optical bandgap of the compound, agreeing with previously reported data for the direct bandgap of BBO3.4,47,48 As observed before for BBO3 nanoparticles,49 wide emission spectra can be related to the presence of defects within the samples, specifically doubly charged oxygen vacancies, V0++, which are responsible for the yellow emission with a range of 550–610 nm.50 The presence of the broad emission peak around 640 nm is attributed to the presence of Bi3+ in the thin films, which causes this redshift.51 According to a density functional theory (DFT) study, the transition across the bands involves the transfer of photogenerated carriers from the 6s orbitals of Bi3+ to the 6s orbitals of Bi5+.52 So not observing these transitions for the 4 nm thin film is puzzling, and further understanding can be gained by collecting PL spectra at lowered temperatures.53 

These results along with the different oxidation states observed for the different films indicate that by decreasing the thickness of the film below 4 nm, the breathing distortion is altered, resulting in the original cubic structure with one oxidation state, as shown in Fig. 4(d). Admittedly, we do not yet have the evidence that this film is electrically conductive, and this will be our next research question. Results do not exclude any of the two images but ensure the occurrence of both charge orderings, which results in Bi 6s O 2p hybridization and CDW that are responsible for opening an optical bandgap. With the reduction of the thickness, the charge ordering and the associated breathing distortion are suppressed, and the cubic structure is retrieved within the perovskite portion of the thin film; therefore resulting in the original band structure of BBO3 with no optical bandgap.

In this work, we shed light on the optical properties of the compound BaBiO3-δ and their link to its structural properties and electronic structure. Even though there are ordered oxygen vacancies within the structure as observed by STEM, the perovskite structural and spectroscopic properties are still observed for the grown thin films. Unlike other interpretations where only Bi3+ is present and the disproportionation exists with a hole density on the oxygen site (O2− and O) ,54 here it is shown, according to XPS results, that charge disproportionation happens on the Bi site. It is demonstrated that the suppression of the charge ordering with a thickness lower than 4 nm results in the suppression of the Raman-active mode around 565 cm−1 and, hence, the breathing distortion in the BBO3 lattice. As a result, this closes the optical bandgap, as observed from the spectroscopic ellipsometry data, and retrieves the original band structure of the BaBiO3 perovskite domains in the thin film with only one oxidation state for Bi ions.

supplementary material includes four subsections: A. Analysis of the STEM image and RHEED pattern, B. Raman intensity as a function of the used excitation wavelength, C. FWHM of XPS peaks for different samples, and D. PL data fitting. In Fig. S1, the FFT filtered STEM image of the 25 nm BBO3-δ thin film and RHEED patterns collected at the start of BBO3-δ growth and after cooling down the layer illustrate that the growth follows domain matching epitaxy. In Fig. S2, the graph displays how the breathing distortion’s Raman peak is changing in intensity as the laser excitation’s wavelength is varied. The maximum Raman intensity is shown at a wavelength of 633 nm; the excitation is in resonance with the optical bandgap of the material (2.1 eV) in this case. Table S1 shows the FWHM of the Bi 4f′s 5/2 and 7/2 core-level peaks, which clarifies the motivation behind the fitting used in the graph of Fig. 3(b). In Fig. S3, the PL data is shown with fitted peaks, as described in the Fig. 4 discussion.

This work was part of the IMEC Industrial Affiliation Program. The authors would like to acknowledge process and hardware engineers Hans Costermans and Kevin Dubois for their great support of the MBE cluster tool and Moloud Kaviani, who created Fig. 4(d) using DFT. This research has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (Grant Agreement No 864483).

The authors have no conflicts to disclose.

I. Ahmed: Conceptualization (lead); Data curation (equal); Formal analysis (equal); Investigation (lead); Methodology (lead); Visualization (lead); Writing – original draft (lead); Writing – review & editing (equal). M. Korytov: Data curation (equal); Formal analysis (equal); Investigation (equal); Methodology (equal); Software (equal); Writing – review & editing (equal). S. Sergeant: Data curation (equal); Formal analysis (equal); Methodology (equal). T. Nuytten: Formal analysis (equal); Investigation (equal); Methodology (equal); Writing – review & editing (equal). T. Conard: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Investigation (equal); Methodology (equal); Writing – review & editing (equal). S. De Gendt: Investigation (equal); Resources (equal); Supervision (equal); Validation (equal); Writing – review & editing (equal). C. Merckling: Funding acquisition (lead); Project administration (lead); Resources (lead); Supervision (equal); Validation (equal); Writing – review & editing (equal).

The data that support the findings of this study are available from the corresponding author upon reasonable request.

1.
K. J.
Choi
et al, “
Enhancement of ferroelectricity in strained BaTiO3 thin films
,”
Science
306
(
5698
),
1005
1009
(
2004
).
2.
U.
Lüders
et al, “
Room-temperature magnetism in LaVO3/SrVO3 superlattices by geometrically confined doping
,”
Phys. Rev. B
80
(
24
),
241102
(
2009
).
3.
J. G.
Bednorz
and
K.
Alex Müller
, “
Possible high Tc superconductivity in the Ba–La–Cu–O system
,”
Z. Phys. B: Condens. Matter
64
(
2
),
189
193
(
1986
).
4.
A. S.
Chouhan
et al, “
BaBiO3: A potential absorber for all-oxide photovoltaics
,”
Mater. Lett.
210
,
218
222
(
2018
).
5.
J.
Tang
,
Z.
Zou
, and
J.
Ye
, “
Efficient photocatalysis on BaBiO3 driven by visible light
,”
J. Phys. Chem. C
111
(
34
),
12779
12785
(
2007
).
6.
A.
Kussmaul
et al, “
Superconductor-insulator-superconductor tunneling in Ba1−xKxBiO3 grain boundaries
,”
Appl. Phys. Lett.
63
(
20
),
2824
2826
(
1993
).
7.
D. E.
Cox
and
A. W.
Sleight
, “
Mixed-valent Ba2Bi3+Bi5+O6: Structure and properties vs temperature
,”
Acta Crystallogr., Sect. B: Struct. Crystallogr. Cryst. Chem.
35
(
1
),
1
10
(
1979
).
8.
G.
Grüner
, “
The dynamics of charge-density waves
,”
Rev. Mod. Phys.
60
(
4
),
1129
(
1988
).
9.
S.
Tajima
et al, “
Optical study of the metal-semiconductor transition in BaPb1−xBixO3
,”
Phys. Rev. B
32
(
10
),
6302
(
1985
).
10.
R. P. S. M.
Lobo
and
F.
Gervais
, “
Bismuth disproportionation in BaBiO3 studied by infrared and visible reflectance spectra
,”
Phys. Rev. B
52
(
18
),
13294
(
1995
).
11.
H.
Namatame
et al, “
Electronic structure and the metal-semiconductor transition in BaPb1−xBixO3 studied by photoemission and x-ray-absorption spectroscopy
,”
Phys. Rev. B
48
(
23
),
16917
(
1993
).
12.
M.
Khraisheh
,
A.
Khazndar
, and
M. A.
Al‐Ghouti
, “
Visible light-driven metal-oxide photocatalytic CO2 conversion
,”
Int. J. Energy Res.
39
(
8
),
1142
1152
(
2015
).
13.
R.
Vesto
,
H.
Choi
, and
K.
Kim
, “
Observation of bandgap closing in SrxBa1−xBiO3 films: Evidence toward topological order in BaBiO3
,”
J. Appl. Phys.
132
(
1
),
015102
(
2022
).
14.
G. K.
Wertheim
,
J. P.
Remeika
, and
D. N. E.
Buchanan
, “
Electronic structure of BaPb1−xBixO3
,”
Phys. Rev. B
26
(
4
),
2120
(
1982
).
15.
M.
Bharath
et al, “
Interplay of lattice distortion and electronic structure in BaBiO3
,”
J. Phys.: Condens. Matter
32
(
5
),
055504
(
2019
).
16.
J. T. W.
De Hair
and
G.
Blasse
, “
Determination of the valency state of bismuth in BaBiO3 by infrared spectroscopy
,”
Solid State Commun.
12
(
7
),
727
729
(
1973
).
17.
N. C.
Plumb
et al, “
Momentum-resolved electronic structure of the high-Tc superconductor parent compound BaBiO3
,”
Phys. Rev. Lett.
117
(
3
),
037002
(
2016
).
18.
K.
Foyevtsova
et al, “
Hybridization effects and bond disproportionation in the bismuth perovskites
,”
Phys. Rev. B
91
(
12
),
121114
(
2015
).
19.
A.
Khazraie
et al, “
Oxygen holes and hybridization in the bismuthates
,”
Phys. Rev. B
97
(
7
),
075103
(
2018
).
20.
G. M.
Dalpian
et al, “
Bond disproportionation, charge self-regulation, and ligand holes in s–p and in d-electron ABX3 perovskites by density functional theory
,”
Phys. Rev. B
98
(
7
),
075135
(
2018
).
21.
A. W.
Sleight
,
J. L.
Gillson
, and
P. E.
Bierstedt
, “
High-temperature superconductivity in the BaPb1−xBixO3 system
,”
Solid State Commun.
88
(
11–12
),
841
842
(
1993
).
22.
L. F.
Mattheiss
,
E. M.
Gyorgy
, and
D. W.
Johnson
Jr.
, “
Superconductivity above 20 K in the Ba-K-Bi-O system
,”
Phys. Rev. B
37
(
7
),
3745
(
1988
).
23.
S.
Tajima
et al, “
Raman-scattering study of the metal-insulator transition in Ba1−xKxBiO3
,”
Phys. Rev. B
46
(
2
),
1232
(
1992
).
24.
K.
Yoshimatsu
et al, “
Dimensional-crossover-driven metal-insulator transition in SrVO3 ultrathin films
,”
Phys. Rev. Lett.
104
(
14
),
147601
(
2010
).
25.
R.
Scherwitzl
et al, “
Metal-insulator transition in ultrathin LaNiO3 films
,”
Phys. Rev. Lett.
106
(
24
),
246403
(
2011
).
26.
P.
Schütz
et al, “
Dimensionality-driven metal-insulator transition in spin-orbit-coupled SrIrO3
,”
Phys. Rev. Lett.
119
(
25
),
256404
(
2017
).
27.
G.
Kim
et al, “
Suppression of three-dimensional charge density wave ordering via thickness control
,”
Phys. Rev. Lett.
115
(
22
),
226402
(
2015
).
28.
R. L.
Bouwmeester
,
A.
Brinkman
, and
K.
Sotthewes
, “
Thickness-dependent band gap modification in BaBiO3
,”
Nanomaterials
11
(
4
),
882
(
2021
).
29.
M.
Zapf
et al, “
Structural and stoichiometric modifications in ultrathin epitaxial BaBiO3 films
,”
Phys. Rev. B
99
(
24
),
245308
(
2019
).
30.
I.
Ahmed
,
S.
De Gendt
, and
C.
Merckling
, “
Self-regulating plasma-assisted growth of epitaxial BaBiO3 thin-film on SrTiO3-buffered Si (001) substrate
,”
J. Appl. Phys.
132
(
22
),
225304
(
2022
).
31.
M.
Zapf
et al, “
Domain matching epitaxy of BaBiO3 on SrTiO3 with structurally modified interface
,”
Appl. Phys. Lett.
112
(
14
),
141601
(
2018
).
32.
J.
Narayan
and
B. C.
Larson
, “
Domain epitaxy: A unified paradigm for thin film growth
,”
J. Appl. Phys.
93
(
1
),
278
285
(
2003
).
33.
T. M.
Takehiko Makita
and
H. A.
Hitoshi Abe
, “
Control of crystal orientation for BaBiO3 thin film on SrTiO3 (100) substrate using BaO buffer layer
,”
Jpn. J. Appl. Phys.
36
(
2A
),
L96
(
1997
).
34.
L.
Wang
et al, “
Brownmillerite phase formation and evolution in epitaxial strontium ferrite heterostructures
,”
Appl. Phys. Lett.
114
(
23
),
231602
(
2019
).
35.
Y.
Gu
et al, “
Oxygen-valve formed in cobaltite-based heterostructures by ionic liquid and ferroelectric dual-gating
,”
ACS Appl. Mater. Interfaces
11
(
21
),
19584
19595
(
2019
).
36.
B. J.
Kennedy
et al, “
Structures and phase transitions in the ordered double perovskites Ba2BiIIIBiVO6 and Ba2BiIIISbVO6
,”
Acta Crystallogr., Sect. B: Struct. Sci.
62
(
4
),
537
546
(
2006
).
37.
O.
García-Zaldívar
et al, “
BiFeO3 codoping with Ba, La and Ti: Magnetic and structural studies
,”
J. Adv. Dielectr.
05
(
04
),
1550034
(
2015
).
38.
V.
Vijayakrishnan
,
G. U.
Kulkarni
, and
C. N. R.
Rao
, “
Valence states of bismuth and thallium in superconducting cuprates: A Bi–L3 and Tl–L3 x-ray absorption spectroscopic study
,”
Mod. Phys. Lett. B
04
(
07
),
451
461
(
1990
).
39.
Y.
Shimizugawa
,
N.
Sugimoto
, and
K.
Hirao
, “
X-ray absorption fine structure glasses containing Bi2O3 with third-order non-linearities
,”
J. Non-Cryst. Solids
221
(
2–3
),
208
212
(
1997
).
40.
H.
Fan
,
G.
Wang
, and
L.
Hu
, “
Infrared, Raman and XPS spectroscopic studies of Bi2O3–B2O3–Ga2O3 glasses
,”
Solid State Sci.
11
(
12
),
2065
2070
(
2009
).
41.
K.
Rokesh
,
M.
Sakar
, and
T.-O.
Do
, “
Calcium bismuthate (CaBiO3): A potential sunlight-driven perovskite photocatalyst for the degradation of emerging pharmaceutical contaminants
,”
ChemPhotoChem
4
(
5
),
373
380
(
2020
).
42.
M. S.
Hegde
et al, “
Electronic structure of high-Tc Ba0.6K0.4BiO3 by x-ray photoelectron spectroscopy
,”
Phys. Rev. B
39
(
7
),
4752
(
1989
).
43.
R.
Itti
et al, “
Photoelectron spectroscopic study of Ba1−xRbxBiO3
,”
Phys. Rev. B
43
(
1
),
435
(
1991
).
44.
L. F.
Mattheiss
, “
Electronic structure of BaPb1-xBixO3
,”
Jpn. J. Appl. Phys.
24
(
S2
),
6
(
1985
).
45.
H.
Namatame
et al, “
Effects of hole doping and electron-phonon interaction on the electronic structure of Ba1−xKxBiO3 studied by photoemission spectroscopy
,”
Phys. Rev. B
50
(
18
),
13674
(
1994
).
46.
H. G.
Lee
,
R.
Kim
et al, “
Anisotropic suppression of octahedral breathing distortion with the fully strained BaBiO3/BaCeO3 heterointerface
,”
APL Mater.
6
(
1
),
016107
(
2018
).
47.
N.
Kumar
,
S. L.
Golledge
, and
D. P.
Cann
, “
Synthesis and electrical properties of BaBiO3 and high resistivity BaTiO3–BaBiO3 ceramics
,”
J. Adv. Dielectr.
06
(
04
),
1650032
(
2016
).
48.
B.
Men
et al, “
Photoluminescence, surface photovoltage and photocatalytic properties of BaBiO3 powders
,”
J. Mater. Sci.: Mater. Electron.
29
,
12729
12734
(
2018
).
49.
K. V.
Shilna
,
S. C.
Sahoo
, and
K. J.
Thomas
, “
Novel ferromagnetism and negative magnetoresistance in BaBiO3 nanoparticles
,”
Appl. Mater. Today
27
,
101427
(
2022
).
50.
A.
Van Dijken
et al, “
The luminescence of nanocrystalline ZnO particles: The mechanism of the ultraviolet and visible emission
,”
J. Lumin.
87–89
,
454
456
(
2000
).
51.
L.
Kumari
,
J.-H.
Lin
, and
Y.-R.
Ma
, “
Laser oxidation and wide-band photoluminescence of thermal evaporated bismuth thin films
,”
J. Phys. D: Appl. Phys.
41
(
2
),
025405
(
2008
).
52.
P. B.
Allen
and
I. B.
Bischofs
, “
Self-trapped exciton defects in a charge density wave: Electronic excitations of BaBiO3
,”
Phys. Rev. B
65
(
11
),
115113
(
2002
).
53.
C. J.
Ellison
et al, “
Confinement and processing effects on glass transition temperature and physical aging in ultrathin polymer films: Novel fluorescence measurements
,”
Eur. Phys. J. E
8
,
155
166
(
2002
).
54.
S.
Balandeh
et al, “
Experimental and theoretical study of the electronic structure of single-crystal BaBiO3
,”
Phys. Rev. B
96
(
16
),
165127
(
2017
).

Supplementary Material