Real time in situ x-ray diffraction study of the crystalline structure modification of Ba0.5Sr0.5TiO3 during the post-annealing

We report about an in situ study of crystalline structural changes during thermal treatment of a Ba0.5Sr0.5TiO3 (BSTO) film grown on MgO. The study covers the complete cycle of heating, annealing and cooling and reveals simultaneous phenomena of phase transitions and strain evolution, which have been characterized by in situ 2D reciprocal space mapping (2D-RSM) using high-resolution synchrotron x-ray diffraction in coplanar and grazing incidence geometries. In this way, temperature induced phase transformation from the BSTO2 to the BSTO1 phase has been monitored and the appearance of a further crystalline phase was detected. Moreover, for both BSTO phases, transitions between in-plane compressive and tensile states have been determined during thermal treatment. Furthermore, a contraction of the out-of-plane lattice components has been observed during the annealing phase while the in-plane lattice components remain leading to the change of the residual in-plane strain towards tensile state. The in situ 2D-RSM findings provide valuable and versatile insights into strain engineering and structure modification upon thermal treatment.


Supplementary Note 1
X-ray Photoelectron Spectroscopy (XPS) measurements were performed in the ANKA UHV analysis Laboratory using a Phoibos 150 Analyzer and an unmonochromated XR-50 Mg Kα X-ray source from SPECS. The angle between the analyzer and the X-ray source was fixed to 45°. The base pressure inside of analysis chamber was about 3 x 10 -10 mbar. All measurements were done at room temperature with pass energy set at 20 eV and an energy step at 0.05 eV. The deconvolution of the peaks was performed by the Fityk software using Voigt profiles and a linear background [1].The binding energy was calibrated using the Au 4f 7/2 (84.00 eV) photoemission line of a corresponding reference material [2]. Due to the charging effect, the adventitious C 1s (284.5 eV) was required as an additional calibration of the binding energy [3]. The spectra of the Ba 3d recorded on the surface of BSTO film in the as-deposited state is shown in the supplementary figure (1a). Each of the Ba 3d 3/2 and Ba 3d 5/2 can be fitted by two peaks. The lower binding energy of the spin-orbit-split (SOS) pair at 3d 3/2 BE Ba1 = 793.28 eV and 3d 5/2 BE Ba1 = 777.95 eV, denoted as Ba1 is assigned to Ba1 atoms in perovskite lattice. The higher binding energy SOS pair at 3d 3/2 BE Ba2 = 794.85 eV and 3d 5/2 BE Ba2 = 779.52 eV, denoted Ba2, assigned to Ba2 in different chemical environment. Similar binding energy values were recorded for Ba1 and Ba2 by Baniecki et al [4]. The energy separation between Ba1 and Ba2 was 1.5 ± 0.1 eV.
Supplementary figures (1a, 1b) compare the XPS spectra of Ba 3d in the as-deposited and in the postannealed states. Upon annealing, the binding energy of Ba atoms was shifted and the shape of the profiles was changed. A new phase has appeared leading to the detection of Ba peak, denoted as Ba3 with binding energy BE=781.75 eV due to regions with BaTiO3-like behavior as it is explained by Rodrigues et al [5] and Ba4 atoms, which probably corresponds to a Ba-O bond in the relaxed surface phase.
Supplementary figures (1c, 1d) compare the XPS spectra of the O 1s in the as-deposited and in the postannealed states (1d). The lower binding energies (BE O1 = 528.65 eV, BE O1 = 528.88 eV) and (BE O2 = 530.55 eV, BE O2= 530.60 eV), labelled as O1 and O2 respectively, are assigned to O 2ions in the BSTO phases. While the O3 atoms with (BE O3 = 531.60 eV, BE O3 = 531.85 eV) and O detected at (BE O = 533 eV BE O = 533.30 eV), are attributed to an oxidation state O x-(0<x<2), corresponding to chemisorbed species as well as to oxygen vacancies and to the non-perovskite structure of BSTO. The O4 atoms with a binding energy BE O4 = 527 eV corresponds to the Sr-O bond in the SrO oxide component.

Supplementary Note 2
Supplementary Figure (2a) illustrates the 2D-RSMs of the symmetric XRD002 reflection over the three post-thermal treatment steps: heating, annealing and cooling. The first row starts with the 2D-RSM at room temperature and ends with 2D-RSM at the annealing temperature 900 °C. The first three 2D-RSMs of the second row correspond to the annealing step with duration of 180 min. The cooling step starts with the last 2D-RSM of the second row at 750 °C until the next-to-last 2D-RSM at 300 °C.
During the heating phase (see supplementary figure (3a)), the diffraction profile of the XRD002 reflection is asymmetric and it could be fitted by three Voigt profiles (blue for BSTOres, magenta for BSTO1 peak and green for BSTO2 peak) indicating the presence of three different crystalline phases. As the temperature increases, the intensity of the BSTO1 peak increases on expenses of the intensity of BSTOres and BSTO2 peak, leading to the disappearance of the BSTOres phase at T= 750 °C.
The annealing phase (see supplementary figure (3b)) was characterised by a slight increase of the intensity BSTO2 peak on the expenses of the intensity of BSTO1 peak.
During the cooling step, the diffraction radial profiles of BSTO, shown in see supplementary figure (3c), were fitted by two peaks because of the profile asymmetry, originating from the existence of two crystalline phases.

Supplementary Note 4
All the corresponding BSTO (1,2) radial diffraction profiles of the GID200 reflections (see supplementary figure (4)) are found to be symmetric. This indicates that the two BSTO (1, 2) phases of the film have different out-of-plane lattice parameters, but identical in-plane lattice parameters. Moreover, we find the GID curve profiles shown in the supplementary figure (4) to stay very stable during thermal treatment where the asymmetric shape of the XRD002 reflections develops with thermal treatment.

Supplementary Note 5
Diffraction radial profiles corresponding to three different temperatures RT, T=600 °C and 750 °C recorded during heating (black) and during cooling (red line) of a) GID200 and b) XRD002.
No reversibility recorded for GID200 reflections while the XRD200 profiles corresponding during heating and cooling to a certain temperature are shifted for the same chosen heating and cooling rate of 150 °C/min. This would lead to a hysteresis in the variation of the out-of-plane lattice parameter with the temperature as it could be demonstrated in figure (5b).

Supplementary figures
Supplementary figure (1) The XPS spectra of the Ba 3d and O1s recorded on the surface of BSTO film in the as-deposited state (a, c) and in the post-annealed state (b, d).
2D-RSMs of the XRD002 reflection. The first row corresponds to the heating step. The first three 2D-RSMs of the second row correspond to the annealing step and the third row represents the 2D-RSMs recorded during the cooling step.