|
|
September 21, 2005 Perovskite Thin Films Under StrainF. He1,*, B.O. Wells1, and S.M. Shapiro2 Strontium titanate (SrTiO3) nanoscale films were studied as a model oxide film system. Highly strained films were grown on different substrates, resulting in either compressive or tensile strain. The measured strain-temperature phase diagram is qualitatively consistent with theory; however, the change in transition temperature is much larger than predicted. Further, the film is constrained because it is epitaxially clamped to the substrate, which causes the SrTiO3 to be tetragonal at all temperatures. Therefore, the phase transition involves only a lowering of symmetry. This leads to the unique situation in which the low temperature phase under tensile strain has an orthorhombic Cmcm space group but a tetragonal lattice, a situation not possible for bulk materials.
In epitaxial films only several nanometers thick, the film atoms tend to align with the underlying substrate atoms. If the lattice parameters of the film and substrate material are different, the film will be strained from its natural bulk atomic spacing. An important area of research in recent years has focused on understanding exactly why the properties of strained films differ from related bulk materials. An ultimate goal is to learn how to use these differences to enhance the performance of devices, or as a tool to probe the fundamental physics of materials with strong electron interactions. The many fascinating properties of perovskite-related materials, such as ferroelectricity and high-temperature superconductivity, are particularly sensitive to strain. Our experiments used synchrotron x-ray scattering to investigate the structural phase transitions in very highly strained films of various perovskite-based film materials. Our main model system has been SrTiO3, which has a well-known structural phase transition involving rotations of the internal TiO6 octahedral, as shown in Figure 1(a). In this case, the transition is from a high symmetry cubic phase at a high temperature to a lower symmetry tetragonal phase below the transition temperature. The structural phase transition in SrTiO3 is monitored through the superlattice diffraction peaks associated with the low-temperature phase.
We produced SrTiO3 films with different amounts and types (compressive or tensile) of strain by carefully growing them on different underlying crystals. Under compressive strain, the phase transitions in SrTiO3 films go from a high-symmetry tetragonal structure to a low-symmetry tetragonal phase. Under tensile strain, the transition is from high-symmetry tetragonal to orthorhombic. The structural phase transition temperature Ts is enhanced under both compressive strain and tensile strain. The main difference between the two cases is that the orientation of the rotation axis is different for compressive strain and tensile strain, as illustrated in Figure 1(b). A very similar effect happens to the spontaneous polarization in related ferroelectric films. An interesting phenomenon in epitaxial films is that, although internally there are structural changes corresponding to the phase transitions, externally the dimensions of the unit cells do not change because of the strong bonding, or clamping, from the substrates. Therefore, the tetragonality of the unit cell, which is the secondary order parameter for bulk SrTiO3, is no longer a good order parameter for phase transitions in films. The internal transitions without the external changes in the shapes of the unit cells create unique morphologies that are not possible in bulk materials. For example, the low-temperature phase in SrTiO3 films under tensile strain has an orthorhombic space group (Cmcm) but a tetragonal lattice as a consequence of strain and substrate clamping. The strain phase diagram based on our experimental data is compared with the theoretical calculations by Pertsev et al, shown in Figure 2. Many aspects of the strain-temperature phase diagram, such as the domain structures and the general trend of the phase boundary, are well described by current theory, though improvements are necessary to describe the magnitude of the increase in transition temperature. BEAMLINE FUNDING PUBLICATIONS FOR MORE INFORMATION |