3.1. Crystal Structure of BST Ceramic Thin Film
As an example, X-ray diffraction patterns of the BST ceramic thin films deposited by the sol-gel method on stainless-steel substrates and fired (crystallized) at
T = 700 °C for 1 h are shown in
Figure 1 (dots). The visual inspection of the diffraction data has shown that three diffraction lines, namely, 2
Θ = 51.02°, 2
Θ = 59.65°, and 2
Θ = 89.41°, are due to the stainless-steel substrate (
Figure 1a,c, dotted line).
The line profile analysis was performed after raw data processing and the results of calculations are also presented in
Figure 1a,c (solid line). The Williamson-Hall plots are shown in
Figure 1b,d for so-called “upgraded” (
x = 5-4-3, the first significant digits of the Sr mole fraction were used for abbreviation) and “downgraded” (
x = 3-4-5) BST thin films, respectively. One can see from the Williamson-Hall plots that the average crystallite size is <
D> = 215 nm for “upgraded” BST film, whereas for “downgraded” BST thin film, <
D> = 220 nm. The average strain (<
ε>), which is a measure of micro-deformations (Δ
dhkl/
dhkl, where
dhkl is an interplanar distance) equals to <
ε> = 0.6% for both thin-film-layered structures.
The unit cell search calculations were performed on the basis of the X-ray diffraction patterns. Once possible unit cells were found, the next step was to refine a selected unit cell and to decide on a lattice and possible space group. The results are presented in
Figure 2, and
Table 1 and
Table 2.
It was found that the “upgraded” (
x = 5-4-3) BST thin film exhibits an orthorhombic structure with a space group
Imma (74) (
Figure 2a,
Table 1), whereas the crystal structure of the “downgraded” (
x = 3-4-5) BST thin film is well-described by a cubic symmetry with a space group
P23 (195) (
Figure 2b,
Table 2).
One can see from
Table 1 and
Table 2 that the global parameters of the fitting process confirm the good quality of the obtained results, with the value of the GOF (goodness-of-fit) parameter ranging from 1.71 to 1.88. It is worth noting that the bottom panels in
Figure 2 show the difference between the intensities of the observed and calculated patterns. One can see that the substantial value of the difference appeared only at the diffraction angles where the stainless-steel substrate manifested itself.
Detailed structural analysis of the X-ray diffraction patterns was performed with the Rietveld refinement method. The Rietveld method was used to determine precise lattice constants from a measurement. Resulting diffraction patterns are presented in
Figure 3 and results of the calculations are provided in
Table 3 and
Table 4.
As a model structure of BST films with an “upward” (Sr molar fraction: x = 0.5, x = 0.4, and x = 0.3) and “downward” gradient of the chemical composition, the following powder diffraction patterns were taken as initial structures for structural parameters’ refinement:
Barium Strontium Titanate, chemical formula: Ba0.67O3Sr0.33Ti1-tetragonal, space group: P4mm(99), ICSD collection code: 54150, PDF number (calculated powder diffraction data) 01-089-0274.
Barium Strontium Titanium Oxide, chemical formula: Ba0.592O3Sr0.408Ti1-cubic, space group: Pm-3m(221), ICSD collection code: 90006, PDF number (calculated powder diffraction data): 01-070-3628.
Barium Strontium Titanate, chemical formula: Ba0.45O3Sr0.55Ti1-cubic, space group: Pm-3m(221), ICSD collection code: 154403, PDF code: 00-039-1395.
Results of the calculations were performed for the XRD profile modeled with a Pseudo-Voigt function. Global parameters of the Rietveld fitting are provided in
Table 3.
One can see from
Table 3 that the global parameters of the fitting process performed for XRD diffraction patterns of the thin films with both configurations of the layers, i.e., structures with upward and downward chemical composition gradient, exhibited low values of
R-parameters, typical for thin films deposited on stainless-steel [
40].
The relevant parameters of BST phases used as initial structures for structural parameters’ refinement of “upgraded” (
x = 5-4-3) and “downgraded” (
x = 3-4-5) BST thin films are presented in
Table 4.
One can see from
Table 4 that the weight fraction of the phases constituting the BST thin films and retrieved on the basis of the Rietveld refinement of the crystal structure consists mainly of the tetragonal phase, namely, 49.4% for “upgraded” (
x = 5-4-3) and 43.3% for “downgraded” (
x = 3-4-5) BST thin film. The weight fraction of the “middle layer” (i.e., for
x = 0.4) of the three-layer-type structure of (0-2) connectivity differs slightly—26.4% for “upgraded” and 25.4% for the “downgraded” structure. Calculated density for all phases is higher in the case of the “downgraded” (
x = 3-4-5) BST thin film.
3.2. Morphology Studies of the Thin Film
The microstructure of the BST thin films was studied by Atomic Force Microscopy (AFM). The morphology of the “upgraded” (
x = 5-4-3) BST thin-film surface of 30 × 30 μm is shown in
Figure 4, whereas the morphology of the “downgraded” (
x = 3-4-5) BST thin film is shown in
Figure 5.
Statistical analysis of the surface roughness of the “upgraded” (x = 5-4-3) BST thin film deposited on the stainless-steel substrate has shown that the average roughness of the thin-film’s surface is Sa = 185.31 nm and root mean square, Sq = 232.024 nm.
On the other hand, the surface roughness of the “downgraded” (x = 3-4-5) BST thin film deposited on the stainless-steel substrate exhibited the average roughness of the thin-film’s surface equal to Sa = 166.997 nm and root mean square, Sq = 207.321 nm.
It is worth noting that the layer-type structure (i.e., “upgraded” (
x = 5-4-3) and “downgraded” (
x = 3-4-5)) BST thin films were deposited on polished stainless-steel of AISI-304–type, which was characterized with an arithmetic mean height of the surface, Sa = 0.05 μm [
41]. Therefore, one can conclude that the BST thin-film surface roughness was caused by conditions of the thin-film growth rather than the roughness of the substrate.
To perform the grain size analysis, the source data AFM images shown in
Figure 4 and
Figure 5 were processed with the help of the Image Analysis P9 program by NT-MDT [
42]. First, the sharpening filter was utilized, and then the procedure of grain identification with a watershed transformation was performed. Thus, the obtained picture with clearly visible grains and grain boundaries was statistically analyzed. It was found that statistical distribution of the thin-film grain sizes depended on the synthesis conditions [
43]. It can be seen from the obtained data (
Table 5 and
Table 6) that the “downgraded” (
x = 3-4-5) BST thin films deposited on stainless-steel exhibited smaller values of calculated average geometric parameters by about 24–28%, namely, area, average size (26%), length (25%), mean width (28%), volume, perimeter (24%), and diameter (26%).
As an example, histograms of the distribution of average size are shown in
Figure 6.
3.3. X-ray Photoelectron Studies of Graded BST Thin Films
X-ray photoelectron survey spectra for sol-gel-derived BST thin films are shown in
Figure 7.
One can see from
Figure 7 that intensities of the X-ray photoelectron survey spectra, collected from the surface layer comprising several atomic layers, depend on the deposition sequence of the barium strontium titanate layers constituting the BST thin film. The intensity of the spectrum for the layer-type structures with the Ba
0.
7Sr
0.
3TiO
3 layer on the top of the layered structure (i.e., BST with
x = 5-4-3-type structure, blue line in
Figure 7) is higher than for the case when the Ba
0.
5Sr
0.
5TiO
3 layer is on the top of the structure (i.e., BST with
x = 3-4-5-type structure, red line in
Figure 6) for the binding energy value within the range
EB = 1.4–0.7 keV. On the contrary, for the lower values of the binding energy (i.e., smaller than 0.7 keV), the difference in intensities is less visible.
In principle, the peak positions in terms of binding energy provide information about the chemical state for a material. The data in
Figure 7 provide evidence for the following series of peaks corresponding to photoemissions from the different core shells:
s,
p, and
d for barium (Ba 3p1, Ba 3p3, Ba 3d3, Ba 3d5, Ba 4s, Ba 4p, Ba 4d), strontium (Sr 3s, Sr 3p1, Sr3 p3, Sr3 d), and titanium (Ti 3p).
It is commonly known, e.g., see [
44,
45], that the best way to compare XPS intensities is via percentage atomic concentrations. The atomic concentrations of the three main metals, namely titanium, strontium, and barium, calculated from the survey spectra (
Figure 7) of the BST thin films, are presented in
Table 7.
Taking into consideration the chemical composition of the upper layer of the BST thin film with the sequence of layers x = 3-4-5 (i.e., “downgraded” BST), where the Ba0.5Sr0.5TiO3 layer is on the top of the structure, the atomic concentration (C) calculated on the basis of the number of moles of an element in relation to the total moles of the elements in the compound (oxygen excluded), is: CBa = 25 at.%, CSr = 25 at.%, and CTi = 50 at.%. For the BST thin film exhibiting the sequence of layers x = 5-4-3 (“upgraded” BST), where the Ba0.7Sr0.3TiO3 layer is on the top of the layered structure, the calculated atom percent (at.%) is: CBa = 35 at.%, CSr = 15 at.%, and CTi = 50 at.%.
High-resolution X-ray photoelectron spectra for BST thin films differing in the sequence of the deposition of the barium strontium titanate layers constituting the thin film are shown in
Figure 8.
From a visual inspection of the spectra shown in
Figure 8, one can see smooth and monotonic curves for C 1s, O 1s, Ba 3d, Sr 3d, Ti 2p, and Fe 2p photoelectron lines. However, it can be seen that intensities of the relevant photo-peaks recorded under the same measuring conditions are higher for BST (
x = 3-4-5) thin films as compared to BST (
x = 5-4-3). The most intense photoelectron lines are usually symmetrical in shape and exhibit the smallest width (they are usually the narrowest lines) in the observed spectrum [
44].
The binding energy of the core electrons basically depends on the elements to which it is bonded. The observed “chemical shift” is due to the charge transfer. Therefore, the bonding environment (chemical state) can be established. Chemical shifts can arise due to several reasons, e.g., variation in electronegativity, molecular environment, positions in the lattice, oxidation states, etc.
Figure 8a illustrates the chemical shift effect of C 1s in the BST thin film. The main line at binding energy c.a.
EB = 285 eV can be ascribed to C-C and C-H bonds. The weaker line at
EB = 288.5 eV can be ascribed to C-O and C=O bonds. The core level binding energy of the carbon atom increases as the electronegativity of the neighboring atoms, in this case oxygen, increases, resulting in the chemical shift. There is also a difference in the peak positions of the main line at the C 1s binding energy spectrum between the BST (
x = 5-4-3) and BST (
x = 3-4-5) (blue and red line, respectively,
Figure 8a). The Gauss approximation yields
EB = 285.19 eV and
EB = 284.98 eV, for BST (
x = 5-4-3) and BST (
x = 3-4-5) thin-film structures, respectively.
Figure 8b illustrates the O 1s photoelectron lines’ spectrum. The main line at c.a
EB = 530 eV can be ascribed to possible oxides, whereas the weaker line at
EB = 531.7 eV provides information on possible surface impurities, likely H
2O and OH [
46].
One can see from
Figure 8c that the Ba 3d photoelectron lines are split. The line at
EB = 778.8 eV may be ascribed to BaCrO
4, whereas the line at
EB = 780 eV can be recognized as BaCO
3. The photoelectron lines of Sr 3d are shown in
Figure 8d. The titanium peak corresponding to photoemission from core shell
p (Ti 2p line, shown in
Figure 8e) exhibits a binding energy similar to BaTiO
3, i.e.,
EB = 458.4 eV.
Peaks corresponding to photoemissions from Fe 2p, Cr 2p, Mn 2p, and Mo 3d are shown in
Figure 8f–i, respectively. It is worth noting that the stainless-steel substrate used for the thin-film growth may contain the above-mentioned elements. Therefore, the photoemission line that peaks at
EB = 711.1 eV can be assigned to oxidized iron, Fe
2O
3 (
Figure 8f). Chromium Cr 2p photoemission lines are found to be split (
Figure 8g). The line at
EB = 577 eV can be assigned to Cr
2O
4, whereas the line at
EB = 580 eV is likely to be Cr
2O
3. Wide asymmetrical lines are visible in
Figure 8h. The photoemission line at
EB = 641.7 eV may be ascribed to Mn
2O
3, whereas the shoulder that appeared at approximately
EB = 643.5 eV could be assigned to MnO
2. Peaks corresponding to photoemission from molybdenum are shown in
Figure 8i. It can be seen that both BST (
x = 5-4-3) and BST (
x = 3-4-5) show the photoemission peak at
EB = 233 eV (Mo 3d3) that can be ascribed to MoO
2. On the other hand, the Mo 3d5 photoemission line that peaks at
EB = 230 eV was visible for BST (
x = 3-4-5) thin film.
Figure 8j illustrates the N 1s photoelectron lines’ spectrum that is caused by NH
3-containing compounds, probably.
On the basis of the XPS high-resolution spectra, the atomic concentrations were calculated and the results relevant to the three main metals, namely, Ti, Sr, and Ba, are presented in
Table 8.
One can see from
Table 8 that the atomic concentration of titanium is less than the theoretical one, namely,
CTi = 50 at.% for Δ
CTi = (2.2–1.94) at.%. Additionally, the atom percentage for barium was found to be lower than the theoretical one calculated for the upper-layer composition of the three-layered BST structure for c.a. Δ
CBa = (3.53–1.74) at.%. On the contrary, the strontium atomic concentration was found to be higher that the calculated theoretical one for Δ
CSr = (5.73–3.68) at.%. The obtained results are within the limits of the quantitative accuracy of the atomic percent values calculated from the major XPS peaks (90–95% for each peak). One can also conclude that the differences in the atomic concentrations indirectly proved the preservation of spatially heterogeneous chemical composition of the layers comprising the three-layer BST structure after the heat treatment of the films.