Peggy Y. Hou

Characterization of iron-based alloy interconnects for reduced temperature solid oxide fuel cells

Abstract The oxidation kinetics and electrical properties of oxide scales thermally grown on the surface of a commercial ferritic alloy have been investigated on the un-oxidized and pre-oxidized alloys as functions of temperature and time under oxidizing atmospheres with four different electrodes. Oxidation kinetic studies with the un-oxidized alloys show a nearly parabolic dependence on time of oxide-scale growth rate, but a significantly increased growth rate with a coating of LSCo (La 0.6 Sr 0.4 CoO 3− δ ) compared to those without and with the coatings of LSM (La 0.85 Sr 0.15 MnO 3 )+LSGM (La 0.8 Sr 0.2 Ga 0.83 Mg 0.17 O 2.815 ) and platinum. Short-term resistance measurements in stagnant air as a function of temperature with pre-oxidized alloys indicate that the oxide scale has a semiconducting transport property. The overall activation energy includes a term from small-polaron hopping inside the oxide scale Δ H m and terms Δ H i and Δ H j from charge transfers at the electrode/oxide-scale and alloy/oxide-scale interfaces, respectively. For the LSCo electrode, long-term resistance measurements as a function of time with un-oxidized alloys reveal a secondary oxidation mechanism related to the formation of an insulating spinel phase in addition to a primary oxidation mechanism associated with the formation of Cr 2 O 3 . SEM observations show that oxidation of the un-oxidized alloy in the presence of an oxide electrode results in considerable interdiffusion of Cr and the electrode cations, especially Co, across the interfaces. Since the ASR values of the oxide scale measured with oxide electrodes quickly approach the permitted limit of a practical SOFC, highly recommended for prevention of a secondary electrochemical oxidation of iron-based alloy interconnects is (1) the use of an oxide coating having purely electronic conductivity and/or (2) prior-to-use conditioning of the alloys via pre-oxidation.

The effect of reactive element additions on the selective oxidation, growth and adhesion of chromia scales

Abstract The addition of minor amounts of reactive elements to Cr 2 O 3 or Al 2 O 3 forming alloys has been known to produce a number of beneficial effects in improving their oxidation resistance. The mechanism by which these elements influence oxidation is, however unclear. In this paper, the effect of reactive element addition, whether as surface or alloying additives, on the development, growth and adhesion of Cr 2 O 3 scales is discussed. It is pointed out that the development of a continuous external Cr 2 O 3 layer and the elimination of base metal oxidation are not coupled processes; reactive element additions can eliminate the latter, but not always affect the former. The incorporation of a reactive element in oxide scales is essential for modifying scale growth. However, a slower oxidation rate or a change in the overall oxide growth direction are not critical factors in improving scale adhesion. In light of these and other experimental results, current theories concerning the reactive element effect are critically evaluated.

Oxide scale adhesion and impurity segregation at the scale/metal interface

The chemistry at scale/metal interfaces was studied using scanning Auger microscopy after removal of the scale in ultra-high vacuum using an in situ scratching technique. Al2O3 and Cr2O3 scales formed between 900°C and 1100°C on Fe-18 wt.% Cr-5 wt.% Al and on Ni-25 wt.% Cr alloys, respectively, were investigated. The adhesion of these scales was determined qualitatively by way of micro-indentation and scratching on the surface oxide. All of the alumina scales fractured to the same degree to expose the metal surface, regardless of the oxidation temperature. The chromia-forming alloy on the other hand, developed more adherent scales at lower oxidation temperatures. About 20 at.% sulfur was found at the metal surface in all cases, and its presence was not only detected on interfacial voids, but also on areas where the scale was in contact with the alloy at temperature. Results from this study clearly demonstrated that sulfur as an alloying impurity does segregate to the scale/alloy interface. However, for alumina scales and chromia scales, the effect of this segregation on oxide adhesion is noticeably different.

The influence of ion-implanted yttrium on the selective oxidation of chromium in Co-25 wt.% Cr

Specimens of Co-25 wt.% Cr, Co-25 wt.% Cr-1 wt.% Y, and yttriumimplanted Co-25 wt.% Cr alloy were oxidized at 1000°C in 1 atm O2. The implantation dosage ranged between 1016 to 1018 ions/cm2. The unimplanted binary alloy oxidized to a duplex Co-rich scale, but the Y-containing ternary alloy formed a continuous Cr2O3 layer. When the implantation dosages were lower than a nominal 1018 ions/cm2, the alloy failed to develop a similar continuous Cr2O3 layer as that observed with the Y-containing alloy. A temporarily stable external Cr2O3 scale was formed on the most heavily implanted specimen (1×1018 Y+/cm2). This Cr2O3 scale consisted of very fine-grained oxide, which is permeable to the outward transport of Cr and Co. Internal oxidation pretreatment of the ion-implanted specimens converting the Y metal to its oxide prior to the oxidation experiment, can enhance the development of an external Cr2O3 scale, but this scale is also unstable. Results suggest that the selective oxidation of chromium in an ordinarily non-Cr2O3 -forming alloy can be due to the reactive element oxides acting as preferential nucleation sites on the alloy surfaces, but the subsequent growth of these scales may require a continuous supply of reactive elements in the alloy.

Reduced Area Specific Resistance for Iron-Based Metallic Interconnects by Surface Oxide Coatings

The effects of various reactive-element coatings on the oxidation kinetics of two commercial iron-based alloys, Ebrite and ANSI 446, at elevated temperatures were investigated. Although NiO is not regarded as a reactive element oxide, it was chosen as a coating due to the enhanced conductivity of Cr2O3 by doping with NiO. Hot-dipping was used to make the desired coatings. The results consistently show a pronounced reduction in oxidation rate by either a Y2O3 coating or an equally mixed Y2O3/NiO double coating, particularly at low temperature, but only a slight influence by NiO coating. Examination of the scale microstructures revealed a continuous denser scale strongly bonded to the underlying metal for a Y2O3 coating, but visible micro-cracks for NiO coating. Electrical resistance measurements after a Y2O3 coating indicate a corresponding reduction in area specific resistance (ASR) with platinum and oxide electrodes, but an increase in ASR with oxide electrodes for a NiO coating. With a Y2O3/NiO double coating, a careful control of the NiO level in the scale has led to a denser, thinner and conducting Cr2O3 scale.

Strain determination in thermally-grown alumina scales using fluorescence spectroscopy

By exploiting the strain dependence of the ruby luminescence line, we have measured room-temperature residual strains in thermally-grown alumina scales. Measurements were made on two alloys Fe-5Cr-28Al and Fe-18Cr 10Al (at.% bal. Fe), oxidized between 300–1300°C. Significantly different levels of strain buildup were observed in scales on these alloys. Results on similar alloys containing a dilute reactive element (RE) are also presented. Scales formed on RE-containing alloys (Zr or Hf) could support significantly higher strains at T ≥ 1000°C. Strain relief associated with spallation thresholds is readily observed. In early-stage oxidation, the evolution of transition phases is monitored using Raman and fluorescence spectroscopies. The fluorescence technique also provides a sensitive probe of early-stage formation of α-Al2O3. It appears that, in the presence of Cr2O3 or Fe2O3, the α-phase of Al2O3 can form at anomalously low temperatures.

Residual stresses in convoluted oxide scales

Alumina scales that grow during oxidation of FeCrAl alloys can develop a convoluted morphology. Although convolution relieves the overall growth stress, high thermal stresses develop locally and can be detrimental to the scale or interface integrity. Ruby fluorescence measurements and finite element simulations are used to examine residual thermal stresses and strains that result when the convoluted scales are cooled to room temperature. Unlike a flat scale that is in biaxial compression, a convoluted scale contains significant gradients, with tensile stress components along the outside and near the interface of the convoluted peaks. The experimental results are in good agreement with model calculations and provide much needed verification of the model assumptions. Because the ruby fluorescence technique provides only the hydrostatic stress averaged over an excited volume that includes the entire alumina scale thickness, modeling provides detail and insight to the experimental measurements.

Analysis of Pore Formation at Oxide-Alloy Interfaces—I: Experimental Results on FeAl

Pores, or voids, at oxide–alloy interfaces are commonly observed after high temperature oxidation when the alloy does not contain a reactive element. In order to understand the pore-nucleation and growth processes, the density, size and depth of interfacial pores on Fe–40 at.%Al as a function of oxidation time at 1000°C were examined. Scanning-electron microscopy (SEM) and atomic force microscopy (AFM) were used to characterize the pores after removal of the surface Al2O3 scale. The nucleation of pores was most rapid during the initial stage of oxidation where cation-transported-alumina growth dominates. Pore growth involves widening as well as deepening, where the deepening rate is slower for larger pores. Growth is accomplished by aluminum evaporation after ∼20 min or by surface diffusion before that time. Pore shape within an alloy grain stays constant and is dictated by the balance of surface and interface energies.

The effect of aluminum as an alloying addition or as an implant on the high-temperature oxidation of Ni-25Cr

The oxidation behavior of aluminum-implanted Ni-25Cr and Ni-25Cr containing 1 wt.% Al has been studied at 1000°C and 1100°C in oxygen. As did Y alloying addition or Y-implantation, 1 wt.% Al added to Ni-25Cr prevented nodular formation of Ni-containing oxides, improved spalling resistance of the scale upon cooling to a similar degree, and eliminated the formation of large voids between the alloy and the scale at the oxidation temperature. However, the Al addition did not alter the rate of growth of the Cr2O3 scale, nor did it change the growth direction. Al-implantation produced no effect even when the maximum concentration and depth of penetration were adjusted to be identical with those of the yttrium in the Y-implanted alloy. The implications of these results concerning the reactive element effect are discussed.

Sulfur Segregation to Al2O3–FeAl Interfaces Studied by Field Emission-Auger Electron Spectroscopy

Using a 30-nm field-emission Auger spectroscopy probe, the segregation of sulfur to a growing oxide–metal interface was studied. The interfaces were formed by the oxidation of a Fe–40at.% Al alloy at 1000°C for various times. Both the oxide and the alloy sides of the interface were examined after spalling the surface Al2O3 layer in ultra-high vacuum. Results were compared with similar studies performed using conventional AES and related to scale development and the interface microstructure. Sulfur started to segregate to the interface only after a complete layer of α-Al2O3 developed there, its concentration then increased slowly with further oxidation until reaching a level close to half a monolayer. Higher amounts were observed on interfacial-void surfaces, where Al and S cosegregated. The study showed that sulfur segregation to oxide–alloy interfaces depended on the type of interface, indicating possible relationships between segregation energies and interface microstructure.