THERMASOFC

a innovative method to predict the thermal radiative properties of the ceramics used in the design of SOFCs from the knowledge of their chemical composition and their microstructure

       In the next years, a deep change of the generation and usage of energy is expected. The decrease of classical fossil sources and the ageing of the actual nuclear power plants influence this evolution. In this sense, emerging energy conversion technologies are encouraged by governments around the world (United Sate, Japan, and Germany, among others). In particular, Fuel Cell devices can supply electricity and heat by consuming oxygen and hydrogen that are continually replenished. From an energetic viewpoint, the most interesting Fuel Cells are the Solid Oxide Fuel Cell (SOFC) since their energetic efficiency can reach 70%. Most common cells are based on the association of yttria stabilized zirconia (YSZ) dense ceramic as solid electrolyte, doped lanthanum manganite (LSM) porous ceramic (porosity, p ~ 40 %) as cathode and a mixed composite nickel-(YSZ) porous cermet (p ~ 40 %) as anode. This SOFC operates generally around 1000°C to ensure a good ionic transport trough the electrolyte [1]

 However, these benefits are counteracted by the thermo-mechanical failure of the delicate ceramic sandwich cell that occurs with elevated temperature. In particular, the effects of the thermal mechanical stresses seem hardly to be modelled by a simple mathematical equation. Furthermore, undesirable chemical reactions governed by the thermal gradient can occur at the interfaces, reducing then the electrochemical performances. The accurate knowledge of the field of temperature appears here vital.

Now few research teams around the world deal with this subject. One can notice the works carried out by Fedorov and co-workers at the Georgia Institute of Technology (Atlanta, US) in the framework of the SECA program [2] supported by the US Department Of Energy. They have performed approximations to simplify the resolution of the radiative transfer equation (RTE) in order to calculate the radiative heat flux. Thus the Schuster-Schwaterzchild two flux approximation and the Rosseland approximation have been respectively used to deal with the electrolyte and the electrodes [3]. Even if it appears questionable, the authors used furthermore room temperature radiative input data to justify their assumptions. Such rough calculations are also reported elsewhere [4]. It leads to a controversial debate on the real role of the radiation on the total energy balance.

On other drawback reported in the literature lies on the complexity to estimate finely the part played by the radiative contribution on the total thermal conductivity. In consequence the lack of accurate thermal radiative data is here striking. It prevents fine calculations of the temperature for the transient or the stationary regim.

¨ To fill in this need of thermal radiative properties (TRP), we plan to measure and to model them within the THERMASOFC project. To reach this objective, we base firstly our strategy on an experimental approach. We plan to perform emittance measurement with a set-up realized by CEMHTI (Orléans, France) and based on infrared spectrometry. The bench has been developed for a temperature range going from 300 K to 2 500 K and for a wavelength range that fully covers the spectral range of thermal radiation (0.8 to 1 000 µm) [5]. It is completely adapted to the thermo-optical characterization of the SOFCs at their operating temperature. The comparison of the spectra aquired on the porous ceramics can be used as a guideline to develop a numerical approach.

We recall that the heterogeneities (pores, grains) embedded inside a ceramic must greatly influence its thermal radiative behaviours [6]. The chemical composition, the texture and the porosity (~ 40%) of the ceramic play here a key role. The term texture stands for the spatial arrangement of heterogeneities within the host matrix as their size distributions. Experimental works have shown that the spectral emissivity of a rough layer made with micronic grains of a mixed ionic and electronic conducting oxides (Pr₂NiO_4+d) works like black body (T = 1000°C). The enhancement of the intrinsic spectral emissivity is around 30 % [7,8]. In the other hand, when the materials become semi-transparent, volumic contributions can act on the thermo optical data.

¨ To understand more clearly the connection between the microstructure and the TRP, CEMHTI has developed since 4 four years a fundamental research in this scientific field. Numerical works have shown that the volumic surface of a fused silica glass containing gaseous bubbles (porosity of 12.9%), influences clearly the volumic backscattering of the infrared light. The computing code consists to use a Monte Carlo Ray Tracing procedure (MCRT). It consists to follow the propagation at the microscopic scale of a very high number of photons that are launched trough a sample. Such a statistical process allows to reproduce the macroscopic TRP. It integrates both the 3D internal structure of the glass and the intrinsic absorption mechanisms of the corresponding homogenous phase of the glass [9]. Particularly, the spatial arrangement of the bubbles is given by X-Ray µ tomography. The MCRT program is very efficient to reproduce the TRP of materials with sub millimetric pores. The key parameters of the code are the chemical composition and the 3D image of the microstructure. In the other hand, it allows to select specific morphological parameters (porosity, specific surface, pore size distribution) to act on these properties.

More recently, the CEMHTI has developed a new methodology based on the use of the Effective Medium Approximation (EMA) and of the MCRT program. It allows to model the optical behaviour of the opaque ceramic of Pr₂NiO_4+d with heterogeneities spanning on two lengths scale i.e. 0.5 µm and 20 µm. The MCRT approach is well adapted to treat the case of the biggest scatterers whereas the EMA approach is well adapted for dealing with the smallest scatterers. This hybrid approach was presented at a French meeting organized by the “GDR ITSOFC-PACTE” in October 2006. To reinforce this work, the CEMHTI has recruited since 1^st October 2005, a ph-D student [10], Hector Gomart in order to develop and promote new methodologies adapted for the case of opaque materials.

The CEMHTI focuses also its efforts to identify the radiative equivalent parameters (extinction coefficient, albedo, phase function) of porous samples. In the framework of the French Federation of Heat and Mass Transfer, a partnership between the Centre for Thermal Science of Lyon (Dominique Baillis) has been established to interpret the numerical results calculated with the MCRT code [11].

¨ To predict efficiently the TRP of SOFCs, one highlights the growing interest for developing a method based on (i) the perfect characterization of the material‘s microstructure and on (ii) the computing of new methods to propagate photons at the micronic scale. The know how of CEMHTI in the development of numerical approach will be use as an anchor.

Recent high-resolution tomographic data (best resolution around 0.28 µm) obtained at the beamline ID19 of the European Radiation Synchrotron Facilities indicate that submicronic pores can be easily evidenced within a porous ceramic of Pr₂NiO_4+d. It strongly motivates us to adapt the MCRT code to the case of materials used in the design of SOFCs. The modelling planned within this project can also draw on the recent development in the textural analysis of the ceramics (X-Ray and Neutron Small-Angle scattering) [12]. In the other hand, theoretical development concerning the interaction of light with small scatterrers (Mie Scattering [13], Resonant Effect [14]) are planned.