0
Research Papers

Simulation Technology on SOFC Durability With an Emphasis on Conductivity Degradation of ZrO2-Base Electrolyte

[+] Author and Article Information
Harumi Yokokawa

Institute of Industrial Science,
The University of Tokyo,
Tokyo 153-8505, Japan
e-mail: yokokawa@iis.u-tokyo.ac.jp

Haruo Kishimoto, Taro Shimonosono, Katsuhiko Yamaji

National Institute of Advanced Industrial
Science and Technology (AIST),
Ibaraki 305-8565, Japan

Mayu Muramatsu, Keiji Yashiro, Tatsuya Kawada

Graduate School of Environmental Studies,
Tohoku University,
Sendai 980-8579, Japan

Kenjiro Terada

International Research Institute
of Disaster Science,
Tohoku University,
Sendai 980-0845, Japan

1Present address: Department of Chemistry, Biotechnology, and Chemical Engineering, Kagoshima University, 1-21-40 Korimoto, Kagoshima 890-0065, Japan.

Manuscript received June 20, 2016; final manuscript received February 15, 2017; published online March 28, 2017. Assoc. Editor: Jan Van herle.

J. Electrochem. En. Conv. Stor. 14(1), 011004 (Mar 28, 2017) (19 pages) Paper No: JEECS-16-1083; doi: 10.1115/1.4036038 History: Received June 20, 2016; Revised February 15, 2017

Attempts have been made to simulate numerically the conductivity degradation of solid oxide fuel cell (SOFC) YSZ electrolyte; physicochemical model has been constructed on the basis of experimental conductivities of Pt/1%NiO-doped YSZ/Pt cells under OCV condition. The temperature effect was extracted from the time constant for degradation caused by one thermal activation process (namely Y-diffusion), whereas the oxygen potential effect was determined by those Raman peak ratios between the tetragonal and the cubic phases which linearly change in relation to the conductivity. The electrical properties of the YSZ electrolyte before and after the transformation are taken into account. The time constant is directly correlated with Y-diffusion with proper critical diffusion length (∼10 nm), while the Y-diffusion can be enhanced on the reduction of NiO; this gives rise to the oxygen potential dependence. The most important objective of simulating the conductivity degradation is to reproduce the oxygen potential profile shift on transformation. Detailed comparison between experimental and simulation results reveal that the shift of oxygen potential profile, therefore, the conductivity profile change inside the YSZ electrolyte can well account for the Raman spectra profile. This also reveals that with decreasing temperature, there appear other kinetic factors of weakening or diminishing enhancing effects by NiO reduction. This may be important in interpreting the ohmic losses in real stacks, because there are differences in time constant or in magnitude of degradation between the pellets and those industrial stacks in which transformation was confirmed by Raman spectroscopy.

Copyright © 2017 by ASME
Your Session has timed out. Please sign back in to continue.

References

Figures

Grahic Jump Location
Fig. 1

Current status of performance analyses made by CRIEPI (as of July 2015)

Grahic Jump Location
Fig. 2

(a) Schematic distributions of oxygen potential, electrochemical potentials of oxide ions, and electrons inside YSZ electrolyte: comparison with the individual contributions from cathode/anode overpotential, cathode/anode Nernst term, ohmic loss in electrolyte, cathode and anode to be determined as average values in stack performance test in the NEDO durability projects. Although only the total ohmic loss is detected, other contributions will determine oxygen potential values at two ends of electrolyte. Available electrical data inside electrolyte enable division into the oxidative and the reductive regionin terms of the NiO reduction. (b) The electron conductivity minimum line which provides a borderline between the electron- and hole-dominant conductive regions. It compares with oxygen potentials in air and in fuel (H2/H2O) and also with that for the NiO/Ni redox equilibrium. (c) Factors of shifting the border position between the reductive and oxidative regions inside YSZ electrolytes. Those factors are originated from electrical characteristics of electrolyte and their degradation, change in operation temperature, electrode performance characterized as overpotential, and its change with operation time.

Grahic Jump Location
Fig. 3

Phase diagram for the ZrO2-YO1.5 system based on Yashima et al. [73] phase data for transformation of t′ phase into coherent t and c phase mixtures are plotted as open (tetragonal) and solid (cubic) symbols. From conductivity measurements, stable cubic phase region is obtained and shown as solid circle, whereas region in which the transformation takes place is plotted as open symbols. Composition in the coherent mixture is also plotted as diamond symbols.

Grahic Jump Location
Fig. 5

Experimentally determined area for transformed tetragonal phase detected by Raman spectra. Intensity scale is derived from the peak area ratio between cubic and tetragonal signals in Raman spectra (see Fig. 4(a)) on 8YSZ electrolyte in the flatten tubular cells fabricated as anode-support cells: (a) sample tested for 5000 h at 800 °C in CRIEPI site, (b) sample tested for 5000 h at 775 °C in CRIEPI site, and (c) cells used in demonstration program which was operated for 4070 h under real service environments (operation temperature is thought to be less than 800 °C).

Grahic Jump Location
Fig. 6

Degradation of electrical conductivity of cells consisting of Pt/1 mol % NiO-doped 8YSZ/Pt placed between air and 1.2% H2O/H2 with a parameter of operation temperature. The vertical axis in the right-hand side is the terminal voltage for cells operated under the OCV condition.

Grahic Jump Location
Fig. 9

The electron conductivity of 1 mol % NiO added YSZ [58] compared with no NiO-doped YSZ [83]. The electron conductivity is lowered, whereas the hole conductivity is unchanged, leading to shift of the borderline between the oxidative and reductive regions to the fuel side.

Grahic Jump Location
Fig. 10

Comparison in position of redox borderlines between inflection point of Raman data and those two NiO reduction fronts calculated from different sets of electrical conductivities for the initial and the final stages of transformation

Grahic Jump Location
Fig. 11

Two logarithmic time constants for conductivity degradation in 1 mol % NiO-doped YSZ as a function of inverse temperature; solid circles are for the anode side, small open squares being for the cathode side. Large square and triangle solid symbols are for 1 mol % NiO-doped 8YSZ in fuel [50,51], and large square and triangle open circles being for nondoped YSZ in air [37,4144]. Solid line is modeled time constant for NiO-doped 8YSZ in anode, the dashed line is obtained for YSZ in air under the assumption that the activation energy should be the same as that in the anode side. This means that at any temperature, about two orders of magnitude difference exists between two cases.

Grahic Jump Location
Fig. 12

Simulation results at 1173 K for LSCF/1%NiO YSZ/Ni-YSZ cells: (a) time-dependent oxygen potential distribution in the cells. After the development of oxygen potential distribution under dual atmospheres, oxygen potential profile shifts due to changes in conductivities by transformation, (b) conductivity of oxide ions, indicating decrease in the reducing side, and (c) conductivity of electrons, showing that the minimum point, corresponding to the border, is shifting to the reducing side after 1,000,000 s. The dashed lines correspond to the initial and final fronts of NiO reduction given in Fig. 10.

Grahic Jump Location
Fig. 13

Normalized Raman peak area ratio of tetragonal-related peaks to cubic-related peak as a function of temperature. Normalization is made using the peak ratio for 3YSZ in which the tetragonal peak becomes significantly large; solid circle is for 1 mol % NiO-doped 8YSZ (8 mol % stabilized ZrO2) and solid square being for 1 mol % NiO-doped 10YSZ (10 mol % Y2O3 stabilized ZrO2).

Grahic Jump Location
Fig. 14

Summary of phase evolution of 8 at % Y-doped zirconia as a function of Hollomon–Jaffe aging parameter derived from temperature and holding time of annealing test for thermal barrier coating materials. HF parameter is selected at 45,000 and 50,000 as the starting and ending stages of development of mixture of Y-lean tetragonal and Y-rich cubic phases (Reproduced with permission from Lipkin et al. [79]. Copyright 2013 by Wiley).

Grahic Jump Location
Fig. 15

Comparison in cation diffusivity between experimentally determined values (solid lines [8588] and open symbols [89] and model-fitting values based on Hollomon–Josef parameters for phase evolution in thermal barrier coating materials (dashed lines) and the present degradation behavior of conductivity of YSZ with and without NiO doping (dotted lines) given in Table 2

Grahic Jump Location
Fig. 16

Logarithmic time constants for conductivity degradation in YSZ as a function of logarithmic NiO concentration. Solid circle [53] and other symbols [48,50] are for NiO-doped 8YSZ; open triangles are from literature for 8YSZ without NiO doping [42,43] (see Table 1).

Grahic Jump Location
Fig. 17

Schematic drawing for growth of coherent t″ + c mixture in the t′ phase accompanied with high defect concentrated growing front at the interfaces between the t′ phase and the mixtures. NiO in 8YSZ is also accumulated in such growing fronts and increases the cation vacancies on reduction of NiO: (a) pure 8YSZ and (b) 1 mol % NiO-doped 8YSZ.

Grahic Jump Location
Fig. 19

Effective area of the two time-constant model in the oxygen potential versus temperature plot; this represents kinetic effects of transformation of 8YSZ into the t″ + c mixtures. The instability temperature of the cubic phase is given around 1373–1423 K; below that temperature, the cubic phase undergoes the diffusionless transformation to the t′ phase which has the same cation configuration but the different oxygen configuration. Above the nucleation temperature of the t″ + c mixtures, the coherence t″ + c mixtures grow their size with time due to the Y-diffusion from the Y-poor t″ phase to the Y-rich c phase. Between those two temperatures, the transformation-induced conductivity degradation can be represented in terms of the two time constants except for the area where the NiO reduction is delayed due to possible kinetic barriers for nucleation of Ni metals or surface reactions of forming Ni along grain boundaries, both of which are represented by the increasing Gibbs energy drop at the Ni precipitation sites (indicated by the dashed line) with decreasing temperature.

Tables

Errata

Discussions

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In