Structural Degradation of Freeze-Cast Iron-Based Foams during Redox Cycling

Wilke, Stephen

doi:https://doi.org/10.21985/n2-5tnv-8987

Work

Structural Degradation of Freeze-Cast Iron-Based Foams during Redox Cycling

Public

Download PDF

The high-temperature oxidation/reduction behaviors of iron and its oxides are important to a variety of energy conversion and storage technologies, such as the solid-oxide iron-air battery and chemical looping combustion. The useful lifetime of iron redox materials is limited, however, by structural degradation arising from (i) sintering, accelerated by phase transformation volume changes and stresses, and (ii) irreversible microstructural dynamics inherent to the oxidation and reduction processes. I aim to better understand the mechanisms underlying this degradation, with a specific focus on highly porous, iron-based foams created by freeze casting. New mechanistic insights motivate investigations on foam structure variations (lamellar vs. dendritic pore architectures) and incorporation of sintering inhibitors or alloying elements to the foams. To begin, lamellar iron foams (48-65% porous) are prepared using directional freeze casting of aqueous Fe2O3 suspensions, followed by reduction with H2 and sintering. These foams comprise iron lamellae separated by lamellar pores, or channels, which provide an open network for gas flow. For redox cycling, this foam architecture offers sintering resistance and provides sufficient porosity for each lamella to expand and contract, without contacting and exerting stress on neighboring lamellae. These foams are cycled under conditions simulating operation of a solid-oxide iron-air battery: repeated oxidation by H2O and reduction by H2 at 800 °C. After 5 or 10 cycles, however, the foams exhibit densification (9-22% remaining porosity) and formation of a dense, gas-blocking shell surrounding the foam exterior. These degradation effects slow gas transport and redox kinetics, effectively rendering much of the foam unusable for cycling. To investigate the mechanisms underlying foam degradation, redox cycling is performed using in operando X-ray microtomography, from which the evolution of foam structure, porosity, and extent of reaction are temporally resolved. Combined with post-cycling microanalysis, these data reveal the mechanism leading to shell formation: the Kirkendall effect, arising from cation vacancy diffusion in FeO, causes microporosity formation at the Fe/FeO interface during oxidation. The micropores irreversibly widen lamellae, lead to delamination at the Fe/FeO interface, and cause the outer tips of neighboring lamellae to contact, sinter, and form the shell. To complement these experiments, a finite element model is developed for the coupled diffusion and elastoplastic deformation of FeO during oxidation, which predicts stress concentration in the rounded tips of lamellae, consistent with the observed fracture and delamination. Addition of Y2O3-stabilized ZrO2, CeO2, or ZrO2 sintering inhibitor (SI) particles in the freeze casting suspension creates Fe-composite foams (5-15 vol. % SI) that better retain the desirable channel porosity after 5 redox cycles. After 20 cycles, however, these foams still exhibit densification, which correlates with spatial segregation of Fe and SI particles within individual lamellae. Again, during oxidation, Fe cation diffusive flux from the cores to the surfaces of lamellae is the root of the problem. Finally, Fe-Ni foams are prepared (7-25 at. %) by incorporating NiO to the freeze casting suspension. During oxidation, the lamellae in these alloyed foams transform into composite structures with Ni-rich metallic cores and Fe3O4 surface layers. The ductile core provides structural support and prevents delamination at the metal/oxide interfaces. Upon reduction, newly formed Fe at the metal/oxide interface diffuses back into the Ni-rich core, reversing the outward flux from oxidation. The resulting redox-reversible, self-healing microstructure enables Fe-25Ni foams to maintain stable channel porosity and minimal microporosity for at least 20 redox cycles. Future work is proposed for employing the alloying approach with different metals (Co and Cu), and for using in situ X-ray nanotomography to further reveal the effects of size (i.e., lamella thickness) and interfacial curvature (pore morphology) on Kirkendall microporosity formation. Supplementary Material accompanies this thesis: four videos of the structural evolution in iron-based foams during redox cycling with in operando X-ray tomography.

Creator