From Zirconia to Yttria
The Quest for the YSZ Thin Film Phase Diagram
- Fig. 1: Schematics of our self-built sputter source,  and an illustration of the thin film preparation process: deposition on NaCl single crystals and subsequent floating off in water, after which the film can be collected via gold grids for transmission electron microscopy.
- Fig. 2: Selected area electron diffraction (top) and the unit cell heights (bottom), which allowed us to distinguish between the tetragonal and cubic polymorphs due to the recession of the lattice parameter, c. Based on data from .
- Fig. 3: AFM surface topology images of the unsupported thin films. Adapted from .
By means of ion beam sputtering, unsupported yttria-stabilized zirconia thin films of various compositions are prepared on NaCl single crystal substrates, and are subsequently investigated using transmission electron microscopy and atomic force microscopy. Studies of the unit cell dimensions (by means of selected area electron diffraction) reveal the phase transition from the tetragonal to the cubic polymorph to lie between 8 and 20 mol% Y2O3.
Towards a Greener Energy Future
Yttria-stabilized zirconia (YSZ), a solid solution of yttrium oxide, Y2O3, in zirconium oxide, ZrO2, is a widely known compound that features a large ionic conductivity at elevated temperatures, and is hence used as an electrolyte in, for instance, oxygen sensors or solid oxide fuel cells. The latter application is especially promising for providing a more efficient and, thus, cleaner way of obtaining energy from hydrocarbons. There are, however, still some challenges like decreasing the operating temperature of these cells. One of the proposals to achieve this is to reduce the electrolyte thickness to the nanometer range. However, there are discrepancies in literature regarding the phase diagram of YSZ – i.e. when it comes to answering the question of which composition yields which crystal structure. In order to elucidate the crystallographic and morphologic properties of such thin films, which are detrimental to the final properties of the fuel cells, we decided to employ our model thin film approach to investigate YSZ samples with varying yttria contents (3, 8, 20 and 40 mol% Y2O3).
No Contradiction: Unsupported, Yet Epitaxially Grown Thin Films
The main challenge in creating an appropriate thin film is maintaining the composition while keeping the thickness to such a low level as to still be able to conduct high-resolution TEM and EELS studies. Additionally, it being a model system, a well-ordered, epitaxially grown thin film is favorable. While evaporation by resistive heating would yield the necessary low growth rates, it is not suitable to deposit films of complex oxide mixtures due to different vapor pressures of the constituents.
Magnetron sputtering, as used in commercial setups, is able to retain the composition, but has a high deposition rate, not ideal to prepare very thin films (in the range of 25 nm). To overcome these problems, we developed our own sputtering source that works at much lower pressures (10-5 mbar instead of the 10 mbar range) by emitting electrons from a heated filament and consequently ionizing argon atoms that are then accelerated towards a target that is kept at a high negative voltage [1,2]. The setup is shown in figure 1. Due to the high temperatures reached in the vicinity of the target, it is possible to sputter oxidic films, although the voltage used is only DC, which is usually restricted to conductive targets. This source allows for very low deposition rates of about 0.1 nm min‑1. Furthermore, in order to be able to investigate the thin films using TEM without having to resort to FIB techniques to remove it from the deposition substrate, we deposit them on NaCl(001) single crystals at 573 K. This has the advantage that the crystals can be dissolved in water, leaving the unsupported, yet epitaxially grown thin film to float on the surface, ready to be collected using TEM gold grids (fig. 1).
It’s Impossible to Distinguish Tetragonal and Cubic Phases using Diffraction – Is It Really?
The problem with determining the crystallographic phases for zirconia and YSZ polymorphs is that the different crystal structures feature lattice planes with exactly the same spacings, meaning that a set of spacings obtained from a selected area diffraction pattern that is assigned to tetragonal spacings could also fully be attributed to cubic lattice planes. This can be seen qualitatively and quantitatively in the diffraction patterns shown in figure 2 that look the same and also feature more or less the same lattice spacings. Moreover, in the case of a textured film due to an ordered growth on a cubic substrate, no distinction can be made from the epitaxy either. This, however, is not a problem inherent to electron diffraction inside the microscope – it also applies to X-ray diffractometry, where distinguishing them is nearly impossible. For this, usually, more structure-sensitive methods such as Raman spectroscopy are required.
However, we can exploit the fact that the heights of the tetragonal and cubic unit cells of ZrO2 are equivalent: they are both approximately 0.51 nm. Thus, in both cases, the (002) planes can be attributed to the same diffraction spot. Hence, this unit cell height, which is the lattice parameter c, can be calculated for each sample – and not only from the (002) spot, but from every diffraction ring visible in the pattern to get a statistically more valid result. The crystal structure that is assumed during this calculation does not influence the result: the same parameters are obtained for cubic and tetragonal values. For didactic purposes, we will assume a tetragonal crystal structure for all the samples. To calculate the unit cell height in this case, we need to know the a/c ratio that can be obtained from the crystal structure of pure ZrO2 (under the assumption that the incorporation of Y2O3 into the lattice does not distort the lattice too much and just isotropically increases the unit cell volume). The results from this calculation are shown in figure 2 as a function of the Y2O3 content.
There, it can be seen that there would be a recession of the unit cell height between 8 mol% and 20 mol% Y2O3. However, the latter specimen contains a drastically larger amount of Y3+, which is a larger ion than Zr4+. This requires the unit cell volume to be larger than for 8 mol%. This stagnation can thus only be explained if the unit cell instead expands in the lateral directions (a and b) and so still increases its volume – this is exactly what happens if there is a phase transition occurring from the tetragonal to the cubic polymorph. Hence, the calculation of the lattice parameter can, in this case, be used as a tool to determine the point of the phase transition from tetragonal to cubic YSZ.
Single Crystals, Flakes and Nanocrystals: The Surface Morphology
The AFM topology images shown in figure 3 have been recorded using the intermittent contact mode. For this, the TEM grids were used; that is, these topologies of the substrate-less specimens reflect the true nature of the thin films, with no influence from the underlying substrate at all, which is otherwise pronounced for films of such low thicknesses. It is, however, not straightforward to measure these free-standing samples as they are prone to tearing or ‘folding’ due to the scanning process. The images show a trend from 3 mol% to 20 mol% yttria: while for the lowest yttria content, the surface features rectangular platelets, and even small tetragonal single crystals, they become more irregular and ragged for 8 mol% and 20 mol% Y2O3, looking more like flakes. The thin film for 40 mol% Y2O3 on the other hand looks distinctly different, however: it has a rather smooth surface covered in pits. The smoothness likely stems from the nanocrystallinity that can also be observed in the respective diffraction pattern in figure 2: the crystallites are too small to be imaged using the tips employed. The recessions, or pits, are also interesting because in the phase image [1,3], they appear bright, indicating that they are crystallites that are recessed into the surface. This has been observed for YSZ and CaF2 single crystals and is attributed to defects in the structure that cause an increase in the diffusivity in their vicinity, creating these holes in the crystal structure .
 Thomas Götsch, Wolfgang Wallisch, Michael Stöger-Pollach, Bernhard Klötzer and Simon Penner: From zirconia to yttria: Sampling the YSZ phase diagram using sputter-deposited thin films, AIP Adv. 6, 025119 (2016) DOI 10.1063/1.4942818
 Lukas Mayr, Norbert Köpfle, Andrea Auer, Bernhard Klötzer and Simon Penner: An (ultra) high-vacuum compatible sputter source for oxide thin film growth, Rev. Sci. Instrum. 84, 094103 (2013) DOI 10.1063/1.4821148
 Thomas Götsch, Lukas Mayr, Michael Stöger-Pollach, Bernhard Klötzer and Simon Penner: Preparation and characterization of epitaxially grown unsupported yttria-stabilized zirconia (YSZ) thin films, Appl. Surf. Sci. 331, 427–436 (2015) DOI 10.1016/j.apsusc.2015.01.068
 Richard G. Green, Luke Barré and Javier B. Giorgi: Nano-structures in YSZ(1 0 0) surfaces: Implications for metal deposition experiments, Surf. Sci. 601, 792–802 (2007) http://dx.doi.org/10.1016/j.susc.2006.11.007
Thomas Götsch1 and Simon Penner1
1 University of Innsbruck, Institute of Physical Chemistry, Innsbruck, Austria