Evolution of the structural phase and orientation in ultrathin SrRuO₃ film with SrTiO₃ capping layer

Abstract

In a perovskite oxide of , the octahedral rotation and tilting are known to result in a structural distortion and transform the crystalline structure to a lower symmetry phase, such as the tetragonal phase (T-phase) and distorted orthorhombic phase (O-phase). Here, we present the observation of such structural transformation in 6 unit cell (uc) ultrathin films of (SRO) driven by an additional (STO) capping layer. A series of nominal STO (x uc)/SRO (6 uc) films with x ranging from 0 to 8 were grown on STO (001) substrates using an oxide molecular beam system with adsorption-controlled growth technique for SRO and layer-by-layer growth technique for STO. Systematic high-resolution X-ray diffraction measurements revealed a clear transition from the T-phase for x = 0 to the O-phase for x = 2 and 8, which was accompanied by a change of the structural orientation from (001)-oriented SRO in T-phase to (110)-oriented SRO in O-phase. Our findings highlight the possibility of intriguing interface-driven structural phase and orientation transformations in a perovskite oxide system.

Introduction

In perovskites, the rotation of octahedra plays an important role for the resulting symmetry and physical properties^1–4. The magnitudes and directions of these rotations about the three principal pseudocubic axes can be described using Glazer notation^5–7. For SRO, the Glazer notation for O-phase and T-phase was determined to be and ^2,8–11, respectively, where the octahedral rotations happen for all three principal axes in O-phase. A number of reports^{1,3,4,8,9,12–50} on SRO-based heterostructures have demonstrated the structural evolutions and their influences on the physical properties by varying structural defects^9,40–44, strain ^{1,3,4,8,45–48,51}, and dimensionality^49,50. For example, SRO films grown on STO (001) substrate undergo an O-to-T phase transition below a critical thickness of 10–20 uc, depending on growth conditions^49,50. However, SRO films grown in low-oxygen environments remain in the T-phase up to 80 nm⁹. It was also reported previously that the O-phase of SRO transforms to the T-phase with either capping layer on SRO/STO (001) film¹⁴ or with the buffer layer in between SRO film and STO (001) substrate³³. On the other hand, in the SRO-based superlattices, such as [/^15,28 and [/(⁵², it was found that the structural phase of ultrathin SRO can be regulated by varying the . These results revealed that the structural phase of the SRO film can be modulated with the film thickness, capping, and buffer layers.

In this work, we systematically investigated the structural phase and orientation of SRO in the MBE-grown high-crystalline thin films of STO (x uc)/SRO (6 uc) on STO (001) substrates, where x = 0, 2, and 8, as schematically shown in Fig. 1a. In order to identify the exact structural phase and orientation of SRO, various possible structural orientations in either O-phase^49,53–56 or T-phase of SRO were considered for the structural determinations. There are two possible orientations of SRO film (see Supplementary Table S1) on STO (001) substrate, as illustrated in Fig. 1b–e. The first orientation is SRO STO (-oriented), and the second orientation is SRO STO (-oriented), where the subscripts o and c refer to the orthorhombic and cubic lattices, respectively. In the O-phase, four -oriented domains A, B, C, D, and two -oriented domains E and F can be present^49,53–57, as shown in Fig. 1b,c. On the other hand, the T-phase of SRO leads to two possible -oriented domains and and a -oriented domain , as shown in Fig. 1d,e. A detailed study to identify these structural domains and orientations in either O-phase or T-phase is crucial to explore the intrinsic and novel properties of SRO ultrathin film.

Fig. 1.

SRO orientation. Schematic of STO (x uc)/SRO (6 uc)/STO (001) film and possible orientation and domains in the O-phase (b-c) and T-phase (d-e) of SRO film grown on the STO (001) substrate. (a) Schematic of nominal 6 uc SRO film grown on STO substrate with no STO capping x = 0, and with the 2 uc (x = 2) and 8 uc (x = 8) STO capping layers. (b) Four -oriented and 90° rotated (see arrows at each domain) domains A, B, C, and D with the orientation of domain A is SRO STO and SRO STO ⁴⁹. The rotation pattern of the leads to angle 90° in the O-SRO, as illustrated in domain B. (c) Two -oriented and 90° rotated domains E and F. The orientation of domain E is SRO STO and SRO STO . (d) Two -oriented and 90° rotated domains and with the orientation of domain being SRO STO and SRO STO . (e) -oriented domain with the orientation of SRO STO and SRO STO . The figure was constructed using Microsoft PowerPoint 2013.

Through rigorous structural studies using high-resolution X-ray diffraction technique, we determined the structural phase and orientation of SRO ultrathin film. The x = 0 film exhibits the T-phase, which transforms into an O-phase for x = 2 and 8. In addition, a change in structural orientation was found during such a structural phase evolution. The T-phase SRO (x = 0) ultrathin film on STO (001) substrate turns out to be -oriented, whereas the -orientated SRO ultrathin film is the favorable orientation for O-phase (x= 2 and 8).

Results

The growth dynamic of STO(x uc)/SRO (6 uc) thin films was monitored in-situ by reflective high energy electron diffraction (RHEED). First, the spot-like RHEED pattern (Fig. 2a) of the bare terminated STO substrate transforms into a streak-line feature with secondary streak lines appearing between the principal spots (Fig. 2b) after the growth of an optimum initial SrO layer⁴⁹. The excess ruthenium (Ru) flux was then supplied for the growth of SRO films in the adsorption controlled growth regime. Figure 2c illustrates the time evolution of the RHEED intensity of the region shown by a red line across the (00)-spot as shown in Fig. 2a. The secondary streak lines gradually disappeared after supplying excess Ru flux. The RHEED intensity of the specular (00) spot remained nearly constant during the SRO growth, indicating a step-flow growth process⁵⁸. On the contrary, an oscillating (00) line intensity was observed during the STO growth, suggesting a different growth dynamics of atomic layer-by-layer growth^58,59. Fig. 2d displays the RHEED pattern after the growth of a nominal 6 uc SRO, and Fig. 2e shows the RHEED pattern after subsequent growth of the nominal 8 uc of STO on SRO. In both cases, the presence of streak lines alongside the principal spots indicates the uniform growth of both SRO and STO layers. To characterize the thickness of both SRO and STO layers and the uniformity of the interfaces, cross-sectional scanning transmission electron microscope (STEM) imagings with atomic resolution were carried out (see Supplementary Figure S1), and the thickness of the SRO and STO layers are close to the nominal thicknesses with the thickness uncertainty for each layer is less than 1 uc ( 0.4 nm), justifying the excellent uniformity and sharp interfaces in our ultrathin films of STO(x uc)/SRO(6 uc) on STO (001) substrate.

Fig. 2.

Structural evolution during the growth of STO (x uc)/SRO (6 uc)/STO (001) film. (a) RHEED pattern of a terminated STO substrate. (b) RHEED pattern after the growth of the optimum initial SrO layer, showing the secondary streak lines for superstructure⁴⁹. (c) Time-dependent RHEED intensity along the red line shown in (a). The RHEED intensity of the (00)-spot remains nearly unchanged during the growth of SRO, and it shows an oscillating nature during the growth of STO. (d) RHEED pattern after the growth of 6 uc of SRO, and (e) RHEED pattern after the growth of STO (x uc)/SRO (6 uc)/STO (001) film with x = 8, revealing the smooth and uniform growth.

For further investigation of the structural evolution in STO (x uc)/SRO (6 uc)/STO (001), synchrotron-based X-ray diffraction was employed. Figure 3a shows the specular reflection around the STO of STO(x uc)/SRO (6 uc)/STO (001) samples with different x value. SRO reflection appeared on the slightly lower L side of the STO reflection^49,60. The presence of Laue oscillations along the crystal truncated rods (CTR) further reveals the high-quality thin film and excellent crystallinity in STO (x uc)/SRO (6 uc)/STO (001). The SRO thin-film exhibits two possible structure phases: the O-symmetry with space group Pbnm (No. 62)^1,3,4 (see Supplementary Table S3 and Supplementary Figure S4 for the space group determination), and the tetragonal one with space group I4/mcm (No.140)^1,3,4. To distinguish the structure phase of SRO in STO(x uc)/SRO/STO system with different x values, two off-normal reflections SRO and SRO have been measured. According to the space group symmetry⁶¹, the reflection and in Pbnm space group are diffractions allowed, where the same reflection and in I4/mcm are forbidden. Figure 3b and c show the L-scan of the SRO and reflections, respectively, for different x values. Pronounced peaks of SRO and only appeared for x = 2 and 8, confirming a structural phase transition from the T-phase in the x = 0 film to the O-phase in the x = 2 and 8 films.

Fig. 3.

Crystalline quality and structural phase in STO (x uc)/SRO (6 uc)/STO (001) films. (a) STO CTRs of the STO (x uc)/SRO (6 uc)/STO (001) films with different x values. Pronounced fringes appeared along the CTR, suggesting the high crystallinity in the SRO films along z-direction. (b) and (c) show the L-scan across the SRO and SRO reflections, respectively, for different x values. SRO and SRO reflections are allowed for the O-phase and forbidden in the T-phase. Pronounced peaks of SRO and SRO reflections appeared for x = 2 and 8, but no such peak appeared for x = 0 films. This result suggests the T-phase in the x = 0 film, and O-phase in the x = 2 and 8 films. (a–c) share the same color code.

To further identify the structural orientation of ultrathin SRO on STO (001) substrate, X-ray measurements on a series of reflections that provide clear distinctions between possible orientations were carried out (see Supplementary Table S1 and S2). Figure 4a and b show the azimuthal -scan for the SRO (02±1) Bragg diffraction⁴⁹ of the -oriented SRO (Fig. 1b). Four peaks of SRO (02±1) pairs, with nearly equal peak intensity, appeared. The peak locations agreed with the peak positions for the -oriented SRO (Fig. 1b), revealing the presence of four structural domains A, B, C, and D with nearly equal volume fractions in the O-SRO STO (x uc)/SRO (6 uc)/STO (001) films with x = 2 and 8. We also checked for the possible domains E and F of -oriented SRO (Fig. 1c) in the O-phase by seeking for the SRO reflection, and no observable peak for the -oriented SRO reflection appeared within the detection limits, revealing the negligible amount of domains E and F of -oriented SRO in the O-SRO film (see Supplementary Figure S2)^54,57. These results show that the -oriented SRO is the preferred orientation in the x = 2 and 8 films.

Fig. 4.

Structural domains and orientation in O-SRO films. (a) Azimuthal -scan for the SRO reflection for x = 2 and (b) for x = 8 films. A set of four peaks, with nearly equal peak intensity, for both SRO and SRO reflections appeared. Peaks location agreed with the -oriented SRO⁴⁹, suggesting the four -oriented structural domains in the O-SRO (x = 2 and 8) films.

We now turn to discuss the procedure for the determination of structural orientation in x = 0 film with T-phase. The reflection for the -oriented SRO overlaps with the SRO reflection for -oriented SRO. Since reflection is forbidden due to the violation of the symmetry⁶¹, the absence of the reflection for -oriented SRO can thus confirm the absence of -oriented SRO on STO (001) substrate in the T-phase of x = 0, and this is exactly the case shown in Fig. 5a, where no noticeable peaks were found in the -scan for the reflection from -oriented SRO. For consistency check, we performed similar -scan of the reflection for -oriented SRO that overlaps with reflection for -oriented SRO, which are both allowed by the reflection conditions⁶¹, and the results are shown in Fig. 5b, showing pronounced peaks as expected. These results reveal that the -oriented SRO is the preferred orientation in the T-phase of SRO on STO (001) substrate, and the peaks in Fig. 5b derive from the reflection for -oriented SRO. We note that a similar scheme was applied to and reflections for -oriented SRO, which gives a consistent conclusion (see Supplementary Table S2 and Supplementary Figure S3).

Fig. 5.

Structural domains and orientation in the tetragonal (x = 0) SRO film. (a) Azimuthal -scan for the Bragg reflection of -oriented SRO, and the Bragg reflection of the -oriented SRO. The -oriented SRO plane overlaps with the -oriented SRO plane. (b) Azimuthal -scan for the -oriented SRO and -oriented SRO reflections. The -oriented SRO plane overlaps with the -oriented SRO plane. All possible reflections at each peak location are labeled in both the figures. The SRO reflection is forbidden in the tetragonal SRO (space group I4/mcm). No noticeable peaks appeared when only -oriented SRO reflection is present. These results suggest that the -oriented SRO is the dominant orientation in the tetragonal SRO film, and the peaks in figure (b) correspond to -oriented SRO reflection.

We remark that the transformation from T-phase to O-phase is associated with a subtle change in the octahedral rotations pattern^{1–4,8–11,62,63}. Based on earlier work, the orientation and rotation pattern of the octahedra for the tetragonal SRO with space group I4/mcm (No. 140)^1,3,4 and O-SRO with space group Pbnm (No. 62)^{2–4,9–11,63} are illustrated in Fig. 6a and b, respectively. For T-phase SRO, the adjacent octahedra exhibit out-of-phase rotation only along the pseudocubic direction. On the other hand, for O-phase SRO, the octahedral shows out-of-phase rotation along the , in-phase rotation along the , and out-of-phase rotation along the directions of STO. Our experimental results revealed that the additional STO capping layer favors the O-phase SRO with octahedral rotations along all three principal axes.

Fig. 6.

Orientation and rotation of octahedra in T-SRO and O-SRO. (a) crystal structure of the T-SRO (space group I4/mcm), where the O-Ru-O bond angle () is . The dotted black lines show the unit cell of T-SRO and the solid blue lines show the pseudocubic unit cell. The octahedra show out-of-phase rotation about the axis only. (b) crystal structure of the O-SRO (space group Pbnm) with , dotted black lines show the unit cell of O-SRO and the solid blue lines show the pseudocubic unit cell. The octahedra rotate out-of-phase about the and axes, and in-phase rotation about . Sr atoms are omitted for clarity. (c) schematic of distorted orthorhombic SRO unit cell under compressive strain. The color bar shows the strain (u) as a function of z, where the strain is maximum near the interfaces. The figure was constructed using Crystal Maker software version 9.2.8 and Microsoft PowerPoint 2013.

Discussion

In order to better resolve the intrinsic structures of these ultrathin SRO films at a macroscopic scale, the structural phase and orientation determinations were all carried out by non-destructive and high resolution X-ray diffraction measurements^25,54–56. In some SRO-based superlattice systems, such as [/(⁵² and [/^15,28, an increase in the or STO layer thickness ( 6 uc) can drive a structural phase transition in SRO layer from O-phase to T-phase. However, such a transition was not observed in our STO (x uc)/SRO (6 uc)/STO (001) samples, where SRO remains to be in O-phase with x up to 8 uc. As pointed out earlier, the difference between O-phase and T-phase is rooted in the subtle difference in the octahedral rotations along three principal axes, which is highly sensitive to films’ vacancy defects^9,40–44,60, stoichiometry^1,50,64 and strain^2,4,46,48,51. For example, it was reported previously that the oxygen vacancies at the apexes of the octahedra may suppress the octahedral rotations along certain principal axes and thus favor the T-phase SRO^{1,9,27,44,63–68}.

In light of the vacancy defects and stoichiometry problem, we used the adsorption controlled growth technique with an oxide MBE for the growth of STO (x uc)/SRO (6 uc) on STO (001) substrate. The stoichiometry in SRO was thermodynamically self-regulated under excess Ru flux condition, which was demonstrated by several groups^49,60 to achieve a low defect level with a high residual resistivity ratio in SRO thin films.

Due to the difference in the ionic sizes⁶⁹, STO (001) substrate is known to give rise to a compressive strain on SRO with a constraint of = 2, where and are the O-phase SRO lattice parameters, and is the cubic STO (001) lattice parameter. Such a constraint results in the distorted O-phase in SRO and a small tilting of away from with as shown in Fig. 6c. Therefore, the atomic strain gradient across the oxide interface inevitably imposes another influence on the octahedral rotations pattern and thus the resulting structural symmetry, which has been demonstrated in several oxide heterostructures^{2–4,15,34,46,49–51,70}. In our x = 0 film, the atomic strain gradient along the direction introduces the global inversion symmetry breaking mechanism in the SRO layer, which may drive the change of the octahedral rotations pattern in SRO layer. Conversely, as illustrated in Fig. 6c, in the x = 2 and 8 films, the SRO layer experiences atomic strain gradients (du/dz) in opposite directions due to the STO substrate and capping layer. This opposing strain distribution suppresses inversion symmetry breaking mechanism, which may stabilize the O-phase in the SRO layer.

Conclusion

Using an oxide-MBE and adsorption-controlled growth technique with an optimized initial SrO growth condition, a series of STO(x uc)/SRO (6 uc) thin films with x = 0, 2, and 8 were grown on STO (001) substrates. A systematic study using high-resolution synchrotron X-ray diffraction revealed the T-phase in the x = 0 film, and it transforms into the O-phase for x = 2 and 8. The change from T-phase to O-phase in SRO (6 uc) by simply additional STO capping layer is intimately related to a subtle change in the rotational pattern.

The structural orientation was further identified by systematic XRD measurements on several SRO reflections, giving -oriented and -oriented SRO on STO (001) substrate for T-phase and O-phase, respectively. Our results demonstrated an effective approach for modulating the structural phase and orientation in an ultrathin oxide perovskite thin film by an additional ultrathin STO capping layer.

Methods

Using an oxide-MBE technique, a series of nominal STO (x uc)/SRO (6 uc)/STO (001) films, where x = 0, 2, and 8 were grown on -terminated STO (001) substrates⁴⁹, as shown schematically in Fig. 1a. The films were grown at a substrate temperature of about 680 °C, which was measured with a pyrometer. Distilled ozone was introduced into the growth chamber whenever the substrate temperature was above 150 °C. The partial pressure of the ozone was maintained at about 5 torr throughout the growth process. Effusion cells were used for the evaporation of Sr and Ti, while Ru was evaporated using the e-beam. The source-to-substrate distance is approximately 50 cm for effusion cells and around 75 cm for the e-beam source. The atomic fluxes of all the elements were pre-calibrated with a quartz crystal microbalance. The fluxes of Ti and Sr were about 9.4 and 1.7 , respectively, and the flux of Ru was about twice the flux of Sr. To achieve the high crystallinity in the SRO film, an adsorption-controlled growth technique with the optimized initial SrO layer was utilized for the SRO growth^49,60, and a layer-by-layer growth technique^58,71 was used for the STO growth on the SRO layer. Further details on the adsorption-controlled growth of SRO can be found in our earlier work⁴⁹. For STO growth, Sr and Ti doses were optimized using the in-situ RHEED patterns and intensity oscillation profile⁵⁹. To achieve better temperature and flux uniformity on the substrate larger than 55 , we used an angular rotation of the substrate during growth. After growth, the sample was cooled to about 150 °C in the ozone environment. STEM was used to verify the layer thicknesses of the SRO and STO films. A detailed structural evolution was studied using high-resolution X-ray diffraction (wavelength 1.0332 Å) at the beamline TPS 09A of the NSRRC, Taiwan.

Author contributions

Akhilesh Kr Singh grew the samples and performed X-ray measurements and analyses with Song Yang, Uddipta Kar, Guan-Ruei Chen and Shih-Chang Weng. Akhilesh Kr Singh, Chia-Hung Hsu and Wei-Li Lee designed the experiment. Akhilesh Kr Singh, Song Yang, Uddipta Kar, Shih-Chang Weng, Chia-Hung Hsu and Wei-Li Lee wrote the manuscript.

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-00408-5.