Thin film functional oxides for catalytic water splitting-characterization

Katarzyna A. Janik
External collaborators: DTU CINF, iNano Århus University
Project coordinator: C. D. Damsgaard
Staff at DTU Cen: J. B. Wagner, A. Bastos da Silva Fanta
Funding: DTU Cen

Water splitting, in photoelectrochemical cells, utilizing solar light for the production of hydrogen is an obvious way of converting non-storable solar energy directly into a storable fuel (hydrogen). The design of such devices, operating on a two photon principle, requires materials synthesized in a thin film form. Materials found by theoretical screening, i.e. promising photocathode, photoanode, transparent shielding, as much as conventional materials are grown by physical vapor deposition (PVD) techniques for this purpose. However, the PVD-made films often do not replicate the corresponding bulk materials. This means that before the thin film materials reach the theoretically promised electrochemical functionality, the structural parameters of such films have to be fine-tuned. This tuning is accomplished by a feedback process comprising a suite of analytical techniques used to disclose the composition, morphology, and crystallography of the thin films. These include: Rutherford backscattering spectrometry (RBS), energy dispersive X-ray spectrometry (EDX), scanning electron microscopy (SEM) utilizing secondary and backscattered electrons, focused ion beam (FIB) combined with scanning transmission electron microscopy (STEM) in SEM and transmission electron microscopy (TEM), electron backscatter diffraction (EBSD), grazing incidence X-ray diffraction (GIXRD), and X-ray reflectometry (XRR).


The PVD growth of perovskite oxynitrides is realized in a couple of steps.  The first step consists of an oxide deposition. In this context, an example Ta-O material deposited on a single crystal silicon substrate is presented. The thin film Ta-O material proves to be tetragonal Ta2O5 phase (established by GIXRD) contaminated by a residum of the Ar processing gas (established by EDX, RBS). The polycrystallinity of the film proven by GIXRD  on a macroscopic film area is also indicated by the SEM technique (a)-(b) on a microscale. The SE (a)  and BSE  (b) images that depict the same film area show striking resemblance. Areas of various contrast in both types of images are associated with different film grains. Moreover, the SEM combined with the EBSD technique (c) performed on various film grains reveal that the crystallographic structure of the film can be more complex than the one inferred from GIXRD. Namely, the EBSD measurements specify film areas (indicated by arrows in (a)-(b)) that are likely amorphous, as they do not bring about any EBSD patterns. Althougth these crystallographic findings have to be further explored by TEM, this hypothesis seems to be very likely especially when the  thermal history of the Ta-O thin film is taken into account.

Christian Danvad Damsgaard
Associate Professor
DTU Nanolab
+45 45 25 64 87
Jakob Birkedal Wagner
DTU Nanolab
+45 45 25 64 71
Alice Bastos da Silva Fanta
Senior Researcher
DTU Nanolab
26 FEBRUARY 2020