LSF has been proposed as a possible anode material and a number of investigations have analyzed its electrocatalytic activities. However, it has been shown that iron exsolution occurs in reductive environments at higher temperatures, leading to the formation of Fe rods on the surface.[1] Other experiments revealed that this segregated metallic Fe also appears when applying cathodic biases, causing a pronounced enhancement of the electrochemical water-splitting kinetics.[2] However, this structural instability would render this perovskite unsuitable for the use in SOFCs. It is therefore necessary to investigate the details of the exsolution in order to find ways to eventually suppress it.

Utilizing an array of complementary in situ and ex situ techniques such as XPS, in situ XRD, scanning EXAFS, in situ TGA/DSC, and in situ TEM, we characterize this phenomenon regarding its thermodynamic and kinetic properties. According to in situ XRD, there are two distinct processes taking place: in the lower temperature regime, the perovskite exhibits a non-linear shift of the diffraction peaks to larger unit cell volumes, while, at temperatures above 773 K, Fe starts to crystallize. The shift allows the hypothesis to be made that the Fe atoms are relocated from their lattice positions to interstitial sites, thereby possibly forming diffusion channels, before they segregate to the surface and crystallize as metallic iron agglomerates.

With XPS, the iron oxidation state in the surface-near region can be observed as a function of the reduction treatment in dry, flowing hydrogen. Starting from a state with only Fe(III) present, the perovskite is continuously reduced at temperatures above 673 K, resulting in the presence of only Fe(II) after reduction at 773 K. However, upon raising the temperature further, the observed Fe(II) concentration drops again: at 873 K, only 57% of the sampled iron is in the form of Fe(II), whereas the amount is zero after reduction at 973 K. This increase in the Fe(III) content is readily explained by TEM investigations: due to the transport of the samples in air, the exsolved iron particles are oxidized at their surface, forming a 2 nm thick oxide overlayer.[1] This shell prevents the surface-sensitive XPS measurements from sampling the metallic core. However, since the surface-near region of the perovskite itself only consisted of Fe(II) at 773 K, any Fe(III) in the spectrum originates from the oxide hull of segregated metallic Fe particles. Since the observed Fe(III) concentration, and thus the amount of segregated iron, increases above 773 K, the XPS results agree well with the in situ XRD measurements. Scanning EXAFS measurements of single LSF particles show that there is a difference in the electronic structure between the surface and bulk regions, corroborating the findings from the XPS analyses.

[1] Thalinger, R.; Gocyla, M.; Heggen, M.; Klötzer, B.; Penner, S. J. Phys Chem. C 2015, 119, 22050–22056

[2] Opitz, A. K.; Nenning, A.; Rameshan, C.; Rameshan, R.; Blume, R.; Hävecker, M. Knop-Gericke, A.; Rupprechter, G.; Fleig, J.; Klötzer, B. Angew. Chem. Int. Ed. 2015, 54, 2628­–2632

Figure 1: Temperature-programmed in situ XRD measurements in hydrogen/argon. Arrows mark the peaks resulting from segregated metallic Fe.