Search

Ni/YSZ cermet is a widely used SOFC anode material, particularly due to its combination of good electric conductivity, high catalytic activity and low cost. Nonetheless, Ni/YSZ cermet anodes are facing unique challenges such as nickel agglomeration and reoxidation, eventually resulting in anode degradation. A solution to these problems might be obtained via a better comprehension of the underlying manufacturing processes, particularly the final H2-induced reduction reaction of NiO to Ni. Experimental approaches have contributed immensely toward a better understanding of these phenomena, yet are generally limited in terms of spatial and temporal resolution, thus motivating us to investigate the problem from a multiscale simulation viewpoint. In this study, we developed a reactive force field (ReaxFF) for nickel/oxygen interactions, which was specifically tailored toward molecular dynamics (MD) simulations of NiO reduction in H2 environments. The force field parameters were derived by fitting the ReaxFF energies and forces of various nickel/oxygen structures to the respective values of first-principles calculations. The structures considered here include bulk NiO, Ni2O3 and Ni3O4 at various densities, NiO under shear, the NiO(100) surface, oxygen vacancies in bulk and surface NiO as well as oxygen adsorbed onto the NiO(100) surface. The developed force field was subsequently used to predict the reduction kinetics of nano-scale NiO crystals in a H2 atmosphere at different temperatures and hydrogen partial pressures. Here, we observed a significantly improved structural stability of bulk NiO at temperatures over 700 K, compared to our previous ReaxFF nickel/oxygen description [1] and similar ReaxFF parameter sets published by other authors [2]. Besides quantifying the reduction rate, further core phenomena such as the induction period before the onset of a steady state reduction process, the pore formation in NiO at lower temperatures, as well as the importance of oxygen vacancies during the reduction are discussed and compared with experimental results. Fig. 1 NiO nanocrystal after 0.2 ns of MD reduction simulation at a global temperature of 850 K and hydrogen partial pressure of 10 bars, using our previous ReaxFF parameter set (left) and the parameter set introduced in this publication (right). References [1] Oberhoffer, S., Iskandarov, A., Umeno, Y., "Reduction of NiO in a H2 Environment: A ReaxFF Molecular Dynamics Study",2016 Asian SOFC Symposium (2016.09.04-07), The University of Tokyo, Japan. [2] Zou, C., Shin, Y. K., van Duin, A. C., Fang, H., Liu, Z. K. "Molecular dynamics simulations of the effects of vacancies on nickel self-diffusion, oxygen diffusion and oxidation initiation in nickel, using the ReaxFF reactive force field.", Acta Materialia 83 (2015): 102-112. Figure 1

The electrode performance in solid oxide fuel cell depends both on electrode materials and microstructure. Therefore, it is important to breakdown the complexity of multi-physics in the porous electrode into processes in different scales. To clarify the multi-physics toward better electrode design, the use of both the advanced measurement and the simulation techniques is inevitable. For this purpose, we have taken the combined top-down and bottom-up approaches. In this manuscript, we discuss the practical issues in multi-scale, multi-physics simulation, focusing on the local activity of Ni-YSZ anode.

Porous nickel/yttria-stabilized zirconia (YSZ) composite is a common choice as an anode material for solid oxide fuel cells (SOFCs). Because of its complicated structure, there are several processes occurring in the anode that affect total performance of SOFCs. For instance, oxygen ion diffusion provides oxygen transport to triple phase boundaries (TPBs) where fuel is oxidized. Another important process is nickel sintering, which causes decreasing of the active TPB length and thus leads to degradation of SOFCs. Understanding the mechanism of each individual phenomenon in Ni/YSZ anodes can contribute to improving total performance and stability of SOFCs. Atomistic simulations can provide a deep insight into the processes taking place in Ni/YSZ anodes. While the density functional theory method is one of the most rigorous atomistic approaches used in materials science, its applicability is limited to small systems of hundreds of atoms. Larger systems can be studied by molecular dynamics (MD) and Monte-Carlo simulations, which are based on empirical potentials. However, most of the potentials for YSZ are based on simple pairwise interactions and designed in the way to reproduce specific bulk properties. Unfortunately, their validity for studying surface and interface properties of Ni/YSZ systems is usually only presumed [1], which can lead to questionable results for such processes as nickel sintering and oxygen surface diffusion. Thus, in the present paper we develop a DFT-based model that purposefully aims for studying a wide range of processes in Ni/YSZ systems, including oxygen diffusion in bulk and surface regions of YSZ, yttrium surface segregation, and nickel sintering. Our model combines three types of interatomic interactions. The first one is based on the dipole model [2] and was parameterized for bulk and surface YSZ regions in our previous work [3]. It reproduces with a good accuracy not only bulk but also surface properties of YSZ. The second describes nickel by the embedded atom method [4]. The third is for interactions between Ni and YSZ and is described by a pairwise Buckingham potential, which was parameterized individually for different Ni/YSZ interfaces. Having confirmed that the model well reproduces the bulk properties of YSZ, we employed the developed model to demonstrate that it reproduces with good accuracy the surface structures and bulk conductivity of YSZ. This allowed us to use the model to further study oxygen ion diffusion near (111) and (110) YSZ surfaces where YSZ structure can be very much distorted [5], and thus oxygen diffusion can be affected. Obtained results suggest suppressed oxygen diffusion along (111) because of the lower oxygen vacancy concentration near the surface. In contrast, oxygen vacancies tend to migrate to the (110) surface, which, together with activation of jumps along directions, facilitates diffusion along the surface. This indicates that it can be easier for oxygen ions to migrate to TPBs near the (111) surfaces from bulk regions than along the surface, while near the (110) surface ions prefers to migrate along the surface. We also demonstrate how the model performs for the nickel sintering problem, based on the structures consisting of two Ni particles being 30 Å in diameter and placed on a YSZ slab. We studied the effect of binding energy between the particles and the slab by considering different interfaces at the contact areas and using corresponding interactions. Simulations at 1250 K indicate that a relatively low binding energy of 0.5 J/m2, which corresponds to the Ni(111)/YSZ(111) interface, allows the nanoparticles to slide over the YSZ support and coalesce. When the binding energy is above 1.0 J/m2, the sliding is restricted and particle migration may occur only by means of nickel atom diffusion on nickel surface. [1] H. Lee et al., Acta Materialia 58 (2010) 2197. [2] P. Tangney et al., J. Chem. Phys. 117 (2002) 8898. [3] A. Iskandarov et al., J. Phys.: Condens. Matter 25 (2015) 015005. [4] Y. Mishin et al., Phys. Rev. B 59 (1999) 3393. [5] G. Ballabio et al., Phys. Rev. B. 70 (2004) 075417.

Though a number of atomistic-model studies of yttria-stabilized zirconia (YSZ) have been reported to elucidate its properties, most of them have employed simple pairwise potential functions to express interactions between atoms, which limits the transferability of the models. We have developed a Tangney-Scandolo dipole model potential for YSZ. Energy, stress and forces on atoms calculated by the ab initio (first-principles) density functional theory are provided as reference data for potential fitting. The developed potential successfully reproduces cubic-tetragonal phase transition at a range of yttria concentration relevant with SOFC application. The potential can well reproduce the barrier energy of oxygen vacancy migration. Molecular dynamics simulations of oxygen diffusion in bulk and at grain boundaries are demonstrated.

Molecular dynamics (MD) simulations of processes occurring in SOFC anodes require accurate description of atomic interactions. Typical SOFC anode represents nanoporous compound of Ni and yttria-stabilized zirconia (YSZ), and the main process, fuel oxidation, takes place in triple phase boundaries (TPB). Thus it is of importance, in particular, to reproduce bulk and surface properties of YSZ in order to describe interactions in TPBs with reasonable accuracy. In present paper we incorporate many-body interatomic potential based on dipole model for cubic YSZ. Parameters of the model have been determined by fitting energies, stresses and forces to their values derived from ab initio calculations. Classical MD simulations with the developed potential revealed its ability to describe YSZ surface energies and structures. The effect of temperature and yttria concentration on oxygen ions is studied and discussed.

In typical porous Ni/yttria-stabilized zirconia (YSZ) anode of solid oxide fuel cells (SOFCs), hydrogen oxidation reaction (HOR) occurs at the pore/nickel/zirconia triple-phase boundaries (TPBs). This anode compound i) allows gaseous hydrogen molecules to move in pores to the reaction sites, ii) makes current flow possible from the TPBs to current collector via nickel matrix, iii) allows oxide ions to transfer to the reaction sites by means of oxygen ion diffusion in zirconia. The last point is possible only when zirconia is doped by cations with oxidation states different from that of Zr4+. For instance, doping by Y3+implies creation of an oxygen vacancy for a pair of dopant cations. So, it is actually presence of the oxygen vacancies that eventually makes the oxygen diffusion possible in SOFC anode. Moreover, oxygen diffusion provides an oxygen-ion flow from cathode to anode through the solid electrolyte made of doped-zirconia. Considering high importance of the dopants in zirconia for SOFC performance, there have been intensive studies of their effect on various phenomena: oxygen diffusion, nickel sintering, phase stability, etc. However, as for details of HOR on atomic scale, all up-to-date theoretical studies considered only Ni/YSZ systems [1-3]. In the present work we perform atomistic modeling study based on density-functional theory (DFT) to see how qualitative and quantitative characteristics of HOR can be affected by selection of dopant species. The TPB structure for our study is based on the (111) zirconia and (111) Ni slabs that are attached to each other as shown in Figure 1 [4]. Two dopants are introduced instead of two zirconium cations near the TPB (sites A and B). We also introduce oxygen vacancies into zirconia slab, whose number is appropriate to formal oxidation state of the dopants and to the fact that we consider one-oxygen rich zirconia slab in reactant structures (product structures are stoichiometric). In this work we consider three trivalent (Y, Sc, Al) and two divalent dopants (Ba, Ca). Three reaction mechanisms are considered: interface reaction, O spillover, and H spillover (terms are used according to notation by Shishkin et al. [5]). All calculations were performed with Vienna Ab initio Simulation Package (VASP) [6], which can perform electronic structure calculations based on the projector-augmented wave method and the generalized gradient approximation. To evaluate transition states, we performed nudge elastic band calculations as implemented in VASP by Henkelman et al. [7]. Obtained DFT results for the total energy of the structures were processed with open-circuit voltage (OCV) correction in a similar concept given by Nørskov et al. [8]. Energy profiles built for the HOR mechanisms allow to determine the most energetically favorable mechanism for each dopant. This judgement is based on energy barriers of the rate-limiting steps (activation energy). Obtained results for Y-doped TPB agree well with literature data that predicts interface reaction mechanism to be dominant with the rate-limiting step being H transfer from the nickel surface to oxygen ion at the TPB (as indicated in Figure 1 by blue line) [2-3,5]. Qualitatively similar results were obtained for Sc and Ba. In contrast, Al and Ca dopants make O spillover mechanism favorable with the rate-limiting step being O transfer from TPB region onto the nickel surface (as indicated in Figure 1 by red line). Considering dopants promoting the same HOR mechanism, activation energies for trivalent dopants are lower, which makes them better dopant candidates. Specifically, low activation energy of O spillover mechanism in Al-doped system makes Al best candidate to promote this mechanism. Background behind switching of the reaction mechanism with respect to dopant species is analyzed and discussed in this paper. References [1] Shishkin, M.; Ziegler, T. Phys. Chem. Chem. Phys. 2014, 16, 1798 [2] Cucinotta, C. S.; Bernasconi, M.; Parrinello, M. Phys. Rev. B 2011, 107, 206103 [3] Liu, S.; Ishimoto, T.; Monder, D. S.; Koyama, M. J. Phys. Chem. C 2015, 119, 27603 [4] Tada T. ECS Transactions 2015, 68, 2875 [5] Shishkin, M.; Ziegler, T. J. Phys. Chem. C 2010, 114, 11209 [6] Kresse, G.; Hafner, J. Phys. Rev. B. 1993, 47, 558 [7] Henkelman, G.; Jónsson, H. J. Chem. Phys. 2000, 113, 9901 [8] Nørskov, J.K; Rossmeisl, J.; Logadottir, A.; Lindqvist, L. J. Phys. Chem. B 2004, 108, 17886 Caption: Figure 1. TPB model consisting of nickel and zirconia slabs. Nickel atoms are shown in black, oxygen - in red, zirconium - in white. Positions of cation dopants A and B are shown in green. Red line indicates O transfer step of O spillover mechanism, blue line - H transfer step of interface reaction mechanism. Figure 1

Electrode processes in solid oxide fuel cells are intensively studied to increase the system efficiency to compete with that of the state-of-the-art LNG-fueled combined cycle thermal power plant. Because the electrode performance depends both on electrode materials and microstructure, we are obliged to breakdown the complexity of phenomena in the porous electrode into elementary processes by best utilizing both the advanced measurement and the simulation techniques. Closely examining the efforts in literature [1-7], we have taken the combined top-down and bottom-up approach. The most difficult issue has lied on the meso-scale modeling, which typically is based on microkinetic modeling [5-8]. To guarantee the predictivity of the modeling, it is inevitable to realize the simulation without arbitrary fitting parameters that are not necessarily based on realistic physics. For this purpose, we first exhaustively summarized literature data followed by the careful examinations. Next, we have made exhaustive simulations assuming different combinations of parameters in literature. Finally, we obtained a set of parameters that matches with experimental observations within reasonable differences. The details together with the results of top-down and bottom up simulations will be introduced together with future perspectives. Acknowledgements The activities of INAMORI Frontier Research Center is supported by Kyocera Co. Ltd. The activities of Advanced Automotive Research Collaborative Laboratory in Hiroshima University are supported by Mazda Corporation. The part of the research is supported by CREST, Japan Science and Technology Agency. References [1] J. Mizusaki, H. Tagawa, T. Saito, T. Yamamura, K. Kamitani, K. Hirano, S. Ehara, T. Takagi, T. Hikita, M. Ippommatsu, S. Nakagawa, K. Hashimoto, Solid State Ionics, 70/71, 52 (1994) [2] W. Yao and E. Croiset, Can. J. Chem. Eng., 93, 2157 (2015) [3] A. Bieberle, L. P. Meier, L. J. Gauckler, J. Electrochem. Soc., 148, A646 (2001) [4] A. Utz, H. Störmer, A. Leonide, A. Weber, E. Ivers-Tiffée, J. Electrochem. Soc., 157 B920 (2010) [5] W. G. Bessler, S. Gewies, M. Vogler, Electrochim. Acta, 53, 1782 (2007) [6] W. G. Bessler, Solid State Ionics, 176, 997 (2005) [7] M. Vogler, A. Bieberle-Hütter, L. Gauckler, J. Warnatz, and W. G. Bessler, J. Electrochem. Soc., 156, B663 (2009) [8] H. Kohno, S. Liu, T. Ogura, T. Ishimoto, D. S. Monder, K. Karan, M. Koyama, ECS Trans., 57(1), 2821 (2013)

A solid oxide fuel cell (SOFC) system anode material is usually composed of a mixture of Ni and other oxide ceramics, such as doped-CeO2, and Sc stabilized ZrO2, due to their high activity and structural stability. Among them the nickel/yttria-stabilized zirconia (Ni/YSZ) cermet is widely used today as an anode material of SOFC due to its good electrocatalytic property against low price. For SOFC performance, it is essential to improve the stability of the Ni/YSZ cermet anode. Combination of Ni and YSZ has been studied experimentally, but the interaction is not well understood in the boundary between Ni and YSZ. However, there is no evidence about the atomic-scale Ni/YSZ interface dynamics and chemical changes in the structure during the operation of SOFC. Unfortunately, most classical force fields are incapable of describing chemical reactions. To overcome this difficulty, reactive force field (ReaxFF) potential was developed by van Duin and co-workers. Recently, Ni/YSZ/H2 interactions are reported by ReaxFF. But in this interaction Ni/YSZ interface stability is not taken into account. As a first step we developed our own ReaxFF potential to investigate Ni/ZrO2 interface and for determining stable Ni/YSZ interface for modeling of the triple-phase-boundary in SOFCs. Big difference is found between our developed and reported ReaxFF and that is work of separation. Using our ReaxFF potential, work of separation is in excellent agreement with literature. It is found that O around Zr4+cation is reduced in case of bulk and interface regions. Although interface region exhibited less reduction compared to bulk region. More details and Ni/YSZ interface stability through O vacancy may present at conference.