Effect of trivalent and tetravalent co- doping on phase stability and ionic conductivity of scandia stabilized zirconia. Shyam Kumar C N, Ranjit Bauri

Transcription

1 Effect of trivalent and tetravalent co- doping on phase stability and ionic conductivity of scandia stabilized zirconia * Shyam Kumar C N, Ranjit Bauri Department of Metallurgical and Materials Engineering, Indian Institute of Technology Madras, Chennai , India * shyamkumarcn@gmail.com ABSTRACT Scandia stabilized zirconia (SSZ) shows highest conductivity amongst all zirconia based electrolytes and hence, is a suitable candidate for intermediate temperature solid oxide fuel cells (IT-SOFCs). However, SSZ exhibits other low conducting phases on exposure to high temperatures. In the present study SSZ was co-doped with Yb (trivalent) and Ce (tetravalent) and its effect on the phase stability, high temperature aging behaviour and ionic conductivity was studied. Both binary (10 mol% SSZ) and the ternary (9 mol % SSZ co-doped with 1 mol % of Yb 2 O 3 and CeO 2 ) compositions were found to be in single cubic phase in the as-processed condition. However, post sintering the binary composition exhibited the rhombohedral β phase whereas the ternary compositions remained in the single cubic phase. The sintered pellets were aged at a temperature of 900 C for 500 hours in air. TEM analysis showed that small amount of tetragonal phase formed in the Yb co-doped sample while the Ce co-doped sample was in single cubic phase after aging. Impedance spectroscopy was used to measure the electrical conductivity. The co-doped samples showed higher conductivity before and after aging compared to the binary composition. The Yb co-doped sample displayed considerable conductivity degradation after high temperature aging while the conductivity reduction in the Ce co-doped sample was minimal. 705

2 1. INTRODUCTION Solid oxide fuel cells (SOFC) are potential source of clean, efficient and reliable energy which can meet the future energy needs. Degradation of properties of fuel cell components, because of its high operating temperature and prolonged operating time, is a major concern for the realization of the SOFC technology (Vlasov 1987, Badwal 2001, Tu 2001). Electrolyte material should show phase stability and stable ionic conductivity for the long term use of SOFC systems. Stabilized zirconia systems which show high conductivity and stability at higher temperatures are widely used as the electrolyte in SOFC (Badwal 1192). Because of less size mismatch of Sc 3+ with the host Zr atoms, scandia stabilized zirconia (SSZ) shows the highest conductivity among the zirconia based electrolytes and hence, is a promising candidate as electrolyte to lower the operating temperature of SOFCs. However, scandia stabilized zirconia can exhibit other low conducting phases as the cubic lattice can be easily distorted due to the low elastic interactions. This leads to considerable reduction in conductivity during prolonged operation at high temperatures (Strickler 1965, Lakshmi 2011, Spiridonov 1970, Haeringa 2005, Badwal 1992) Co doping stabilized zirconia is a widely used strategy to improve the conductivity. Several co-dopants like, Y, Yb, Gd, Ce, Ca, Ni, Al etc have been explored by researchers (Badwal 1998, Abbas 2011, Spirin 2012, Ota 2010, Lakshmi 2011, Joo 2009). In the present study, the co-doping approach has been adapted to enhance the phase stability of SSZ. 9 mol % scandia stabilized zirconia was co-doped with 1 mo % of Yb 2 O 3 and CeO 2 to improve the stability of the cubic phase and reduce the high temperature aging of SSZ. The choice of the co-dopants is based on the fact that Yb 2 O 3 -ZrO 2 binary system does not exhibit the rhombohedral (β) phase (Wang 2007) and large atoms like Ce will prefer 8-fold coordination thereby stabilizing the cubic phase (Bechepeche 1999). Combustion synthesis is a simple yet useful technique to produce very fine and highly crystalline ceramic oxides (Venkatachari 1995, Patila 2002) and is very effective in synthesizing multi-component systems including SSZ. Therefore, combustion synthesis was used in the present study to process the binary (SSZ) and ternary (codoped) compositions. 2. EXPERIMENTAL PROCEDURE 2.1 Powder Synthesis The chosen binary and ternary compositions were processed by combustion synthesis. The nitrate salts, zirconium oxynitrate ZrO(NO 3 ) 2.6H 2 O (Sigma aldich), scandium nitrate Sc(NO 3 ) 3.6H 2 O (Alfa aesar), ytterbium nitrate Yb(NO 3 ) 3.5H 2 O (Alfa aesar) and cerium nitrate Ce(NO 3 ) 3. 5H 2 O (Alfa aesar) were used as the precursors. The fuel used for the combustion reaction was glycine (NH 2 CH 2 COOH). Stoichiometric amounts of these starting materials were calculated based on the thermo-chemical concepts used in propellant chemistry. The solution is made in deionised water with required amount of salts in a silica basin. Polyethylene glycol was used as a dispersant. The solution was heated up to 250 C until a thick viscous solution is formed on a hot plate. It was then transferred to a pre heated furnace at 706

3 600 C. Completion of combustion is confirmed by the appearance of fumes which lasted for only a few minutes. As-processed powders were dry, fragile, and foamy. The powders were converted to fine powders by manual crushing. The as-processed powder was calcined at 800 C for 1 hr to remove the carbonaceous residues. The calcined powders were ball milled in a Fritsch planetary ball mill for 4 hours using zirconia balls at 300 rpm in terpineol medium in order to break the agglomerates. The ball to powder ratio was taken as 20:1. The milled powder was dried in an oven overnight at 150 C and pelletized in a 12 mm diameter die by cold isostatic pressing. The green compacts were sintered at 1300 C for 4 hours in air inside a tubular furnace. Archimedes principle was used to find out the density of the sintered pellets. High temperature aging of the sintered pellets was carried out at 900 ºC for 500 hours in a tubular furnace in air atmosphere to evaluate the high temperature phase stability. 2.2 Characterization Phase analysis was done by X-ray diffraction using a Philips PANalytical diffractometer with CuK α (λ=0.154 nm) radiation for all as-processed powders, calcined powders, sintered pellets and the aged pellets. After aging, polished and thermally etched (at1250 C for five minutes) pellets were observed under HR SEM (FEI Inspect F). Thin slices were cut from the pellets by diamond saw wheel for TEM analysis. These slices were polished down to 100 m using emery paper and then thinned down to 60 m using a disc grinder (Gatan disc grinder model 623) followed by ion milling (Gatan precision ion polishing system model 691) for more than 4 hours. Observations were made in a Philips CM12 TEM at 120 kv. 2.3 Electrical conductivity measurement The ionic conductivity was measured from 400 C to 1000 C at an interval of 50 C, in the frequency range of 0.1 Hz to 1 MHz using a Solatron 1260 impedance analyzer. Samples were coated with platinum paste on both sides for electrical contact and two platinum wires were used as current collector. The platinum paste was cured at 1000 C for 1 hr. The Nyquist plots were obtained and the conductivity was calculated from the derived resistances using a proper equivalent circuit. 3. RESULT AND DISCUSSION 3.1 Phase analysis and microstructure Fig. 1(a) shows the X-ray diffraction (XRD) pattern of the as-sintered samples. The SSZ sample showed the fluorite cubic phase with some low intensity peaks around the major cubic peaks. These low-intense peaks correspond to the rhombohedral β phase which was confirmed using the JCPDS database. The presence of these low conducting β phase in scandia stabilized zirconia has been found before, for compositions greater than 9 mol% of Sc 2 O 3. This phase transforms back to cubic phase at temperature above 600 C (Spirindonov 1970). Thus it is clear that a fraction of the cubic phase transforms to the rhombohedral phase during sintering giving rise to the extra peaks. XRD patterns of the co-doped samples (Fig. 1 a) showed only the 707

4 cubic phase and no other undesirable phases were found in these ternary compositions. XRD analysis of the samples after high temperature aging (Fig. 1b) revealed that no additional phases appear even after aging in the co-doped samples. This indicates that the co doping approach with a trivalent (Yb) and tetravalent (Ce) dopant is effective in preventing the transformation from cubic to rhombohedral β phase and improving the stability of the cubic phase even during high temperature exposure. This is in line with the fact that the β phase is absent inn the binary phase diagram of ZrO 2 -YbO 1.5 and hence, the addition of the trivalent dopant (YbO 1..5) takes the composition away from the stability domain of the β phase (Wang 2007). In the case off larger tetravalent atoms like Ce, which prefers eight-fold coordination, the stabilization of the cubic phase field is by inducing a local cubic symmetry in the oxygen sub lattice (Bechepeche 1999). (a) (b) Figure 1. XRD patterns of (a) sintered samples and (b) agedd samples. Microstructural characterisation was carried out on the aged samples to get more insight into the aging mechanism. SEM micrographs of the thermally etched samples after aging are shown in Fig. 2. Even after highh temperature aging for 500 hours the avarage grain size of the samples as determinedd by line intercept method are found to be in the range of 500 to 700 nm. The binary SSZ sample andd the Yb co-doped sample (Yb-SSZ) showed a bimodal grain size distribution with a number of small grains with size less than 100 nm (Fig.. 2, a and b respectively) whilee the Ce co- doped sample showed a uniform grain structure (Fig. 2 c). 708

5 (a) (b) (c) Figure 2. SEM micrographs of the aged sampless (a) SSZ, (b) Yb-SSZ and (c) Ce-SSZ TEM analysis of the samples after aging was carried out to find out the presence of any other low conducting phase in small amounts which cannot c be detected by XRD. The presence of the rhombohedral phase in the SSZ sample whichh is also shown by the XRD analysis is shown in Figuree 3 (a). The characteristic herringbone structure confirmed it to be the rhombohedral β phase (Moghadam 1983, Sakuma 1985, Du 2008). For the cubic phase to be stabilized at room temperature, the host Zr atoms should have the preferablee seven fold coordination. This can happen if the 8-fold coordinated atom binds to an oxygen vacancy. The primary dopant, Sc 3+, is of similar size as Zr 4+ and hence, will compete for the oxygen vacancies. This will leave lesser 709

6 vacancies to the Zr 4+ ions reducing the stability of the cubic phase leading to its disintegration into other low-conducting However, thee aged Yb co-doped sample exhibited lath kind of phases. The rhombohedral phase was absent in the co doped samples. precipitates growing from the grain boundaries towards grainn interior. These precipitates were confirmed to be tetragonal zirconia phase in a cubic matrix by the corresponding diffraction patterns (Fig. 3b). No such precipitates were observed in case of the tetragonal co-dopant or thee Ce co-doped (Ce-SSZ) sample and it exhibited clean grain and grain boundaries (Fig. 3 c). This clearly shows that tetravalent co doping, i.e. co doping with Ce is moree effectivee in improving the stability of the t cubic phase than trivalent dopants. (a) (b) (c) Figure 3. TEM micrographs of the aged sampless (a) SSZ, (b) Yb-SSZ and (c) Ce-SSZ The effect of trivalent and tetravalent dopant is different in the stabilization of cubic phase. Addition of trivalent dopants creates oxygen vacancies to maintain the charge neutrality. Over sized trivalent dopantss like Yb are oriented as next nearest neighbour (NNN) to the oxygen vacancies and hence, the vacancies are positioned as nearest neighbour (NN) to the host Zr atoms. Thus, the host Zr 4+ will remain in its favourable seven fold coordination configuration (Li 1994) by bondingg to an oxygen vacancy and 710

7 hence, will stabilize the cubic phase. Tetragonality is also decreased by creation of oxygen vacancies (Li 1994). Considering the tetragonal structure as a layered structure with two different Zr-O bond lengths (Michel 1983), decrease in tetragonality is due to the introduction of oxygen vacancies around Zr 4+ that reduces the oxygen overcrowding between the layers (Li 1994). The reduction in tetragonality is more dependent on the amount of oxygen vacancies and not on the size of the over sized trivalent dopants. In the present study, the concentration of the larger size trivalent co-dopant may be insufficient to reduce the tetragonality, especially during long term thermal exposure and hence, the presence of the tetragonal precipitates. Oversized tetravalent dopants like Ce do not create any vacancies and the reduction in the tetragonality is by the adjustment of Zr-O I and Zr-O II bond length in the tetragonal layer structure. This results in more symmetric eight fold coordination and destroys the layered structure (Beechepeche 1999, Li 1994, Michel 1983). Thus addition of Ce that has large size difference with the host atom is effective in stabilizing cubic field. It should be noted here that the stability of the cubic phase in SSZ strongly depends on the processing route employed and there are some contradictory results on the effectiveness of co-doping Ce in SSZ in stabilizing the cubic phase (Du 2008 Lee 2005, Arachi 2001). In the present study, addition of Ce was found to be effective in improving the stability of the cubic phase especially during high temperature exposure. 3.2 Electrical Conductivity Conductivity of the pellets was measured using an impidance analyzer as a function of temperature from 400ºC to 1000 ºC at an interval of 50 ºC. Grain and grain boundary contributions were identified based on the frequency dispersion in the Nyquist plots. The conductivity was calculated from the real intercept and the Arrhenius plots were obtained. The activation energies were estimated from the slope of the Arrhenius plot. Fig. 4 shows the Arrhenius plots of grain (Fig. 4 a and b) and grain boundary conductivity (Fig. 4 c and d) of the compositions studied before and after aging. The activation energy (E a ) values are summarized in Table 1. Figure 5 shows the grain and grain boundary conductivity of samples at 1000 ºC before and after aging. Trivalent doping (Yb co-doped sample) which creates vacancies shows highest bulk conductivity before aging (as sintered) followed by Ce-SSZ. The tetravalent co-dopant Ce (Ce 4+ ) does not introduce any oxygen vacancies and hence the conductivity of the Ce-SSZ sample is lower than Yb-SSZ in as-sintered condition. However, the Yb-SSZ showed significant degradation in conductivity after aging. The low-conducting tetragonal phase formed during the high temperature aging can be attributed to this. The Ce-SSZ sample, on the other hand, did not exhibit any lowconducting phase and hence, showed the highest bulk conductivity and lowest degradation in the conductivity after aging. Activation energy for bulk conductivity of co doped samples is less compared to the binary SSZ sample, before and after aging. All samples showed an increase in the activation energy after aging. After aging, Ce-SSZ showed the lowest activation energy for grain conductivity and that can be attributed to the improved phase stability imparted by Ce. 711

8 (a) (b) (c) (d) Fig.4. Arrhenius plots of (a) and (b) grain conductivity before and after aging respectively and (c) and (d) corresponding plots for grain boundary conductivity. Fig. 5. Grain and grain boundary conductivity at 1000 C before and after aging. 712

9 Grain boundary conductivity showed considerable reduction in all samples after aging (Fig. 5). In the case of SSZ and Ce-SSZ, the reduction in grain boundary conductivity is low compared to Yb-SSZ. It is well known that grain size, impurities, presence of second phase, density, inter granular porosity etc affect the grain boundary conductivity. The considerable reduction in the grain boundary conductivity in Yb-SSZ samples can be attributed to the tetragonal low-conducting phase that formed along the grain boundaries and grown towards the grain interiors during the high temperature aging. It is also important to note that the samples which show the bimodal distribution in the grain size (SSZ and Yb-SSZ) showed substantial grain boundary conductivity reduction. Ce-SSZ having uniform grain structure and single cubic phase without any low conducting phase showed the lowest degradation in the grain boundary conductivity as well. Table 1: Activation energy (E a ) of samples before and after aging Composition E a (ev) Before aging Grain Grain boundary E a (ev) After aging Grain Grain boundary 10SSZ Yb-SSZ Ce-SSZ CONCLUSION Trivalent (Yb) and tetravalent (Ce) dopants were co-doped in scandia stabilized zirconia in order to improve the phase stability. The conduction behaviour of the binary (SSZ) and the ternary (Yb-SSZ, Ce-SSZ) compositions were characterized and correlated with the phase stability and high temperature aging. Following conclusions can be drawn from the study. 1. Co-doping with Yb and Ce was found to be effective in preventing the formation of the rhombohedral β phase after sintering and even after aging. 2. Some tetragonal phase was found in the Yb-SSZ composition after high temperature aging. 3. Ce co-doped composition (Ce-SSZ) remained in single cubic phase even after high temperature aging. 4. The conductivity of the ternary co-doped samples was higher than the binary SSZ sample. 5. The Ce-SSZ exhibited least degradation in conductivity after high temperature aging. 713

10 REFERENCES Abbas H. A., Argirusis C., Kilo M., Wiemhöfer H.D., FHammad. F., Hanafi Z. M. (2011), Preparation and conductivity of ternary scandia-stabilised zirconia, Solid State Ionics, 184, 6 9. Arachi Y.,Asai T.,Yamamoto O.,Takeda Y.,Imanishi N.,Kawate K.,Tamakoshib C. (2001), Electrical conductivity of ZrO 2 -Sc 2 O 3 doped with HfO 2, CeO 2,and Ga 2 O 3, J. Electrochem. Soc., 148, Badwal S.P.S. (1992), Zirconia-based solid electrolytes: microstructure, stability and ionic conductivity, Solid State Ionics, Badwal S.P.S. (2001), Stability of solid oxide fuel cell components, Solid State Ionics, Badwal S.P.S.,Ciacchi F.T.,Rajendran S.,Drennan J. (1998), An investigation of conductivity, microstructure and stability of electrolyte compositions in the system 9 mol% (Sc 2 O 3 ZrO 2 ) ZrO 2 (Al 2 O 3 ) Systems,Solid State Ionics, 109, Badwal S.P.S., Drennan J. (1992), Microstructure/conductivity relationship in the scandia-zirconia syste, Solid State Ionics, 53-56, Bechepeche P.,Treu Jr O.,Longo E.,Paiva-Santos C. O.,Varela J. A. (1999), Experimental and theoretical aspects of the stabilization of zirconia, J. Mater. Sci., 34, Costa G.C.C., Muccillo R. (2010), Comparative studies on properties of scandiastabilized zirconia synthesized by the polymeric precursor and the polyacrylamide technique, J. Alloys Compd., 503, Du K., Kim C.-H., Heuer A. H., Goettler R., Liu Z. (2008), Structural evolution and electrical properties of Sc 2 O 3 -stabilized ZrO 2 aged at 850ºC in air and wet-forming gas ambient, J. Am. Ceram. Soc., 91,[5], Haeringa C., Roosena A., Schichlb H., Schnfller M. (2005), Degradation of the electrical conductivity in stabilised zirconia system part II: Scandia-stabilised zirconia, Solid State Ionics, 176, Inaba H., Tagawa H. (1996), Ceria-based solid electrolytes Solid State Ionics, 83, Joo J. H., Choi G. M. (2009), Effect of Ni doping on the phase stability and conductivity of scandia-stabilized zirconia, Solid State Ionics, 180, Lakshmi V. V., Bauri R., Gandhi A. S., Paul S. (2011). Synthesis and characterization of nanocrystalline scsz electrolyte for sofcs, Int. J. Hydrogen Energy, Lakshmi V. V., Bauri R. (2011), Phase formation and ionic conductivity studies on ytterbia co-doped Scandia stabilized zirconia (0.9ZrO Sc 2 O Yb 2 O 3 ) electrolyte for sofcs,solid State Sciences, 13, Lee D.-S., Kim W.S., Choi S.H., Kim J., Lee H.-W., Lee J.-H. (2005), Characterization of ZrO 2 co-doped with Sc 2 O 3 and CeO 2 electrolyte for the application of intermediate temperature sofcs Solid State Ionics, 176, Li P., Chen I.-W, Penner-Hahn J. E. (1994), Effect of dopants on zirconia stabilizationan x-ray absorption study: i, trivalent dopants, J. Am. Ceram. Soc., 77,[1], Michel D., Mazerolles L., Perez M. Jorba Y. (1983), Fracture of metastable tetragonal Zirconia crystals, J. Mater. Sci., 18,

11 Moghadam F. K., Yamashita T., Sinclair R., Stevenson D.A. (1983), Transmission electron microscopy of annealed ZrO 2 +8 mol% Sc 2 O 3,J. Am. Ceram. Soc., 66,[3], Ota Y., Ikeda M., Sakuragi S., Iwama Y., Sonoyama N., Ikeda S., Hirano A., Imanishi N.,Takeda Y., Yamamotoc O. (2010), Crystal structure and oxygen ion conductivity of Ga 3+ co-doped scandia-stabilized zirconia, J. Electrochem. Soc., 157,(11), b1707-b1712. Patila K. C., Arunab S.T., Mimania T. (2002), Combustion synthesis: an update, Curr. Opin. Solid State Mater. Sci., 6, Sakuma T., Yoshizawa Y.-I., Suto H. (1985), The rhombohedral phase produced in partially-stabilized Zirconia,J. Mater. Sci. Lett., 4, Spiridonov F.M. (1970), On the phase relation and electrical conductivity in the system ZrO 2 -Sc 2 O 3,J SOLID STATE CHEM., Spirin A., Ivanov V., Nikonov A., Lipilin A., Paranin S., Khrustov V., Spirina A. (2012), Scandia-stabilized zirconia doped with yttria: synthesis, properties, and ageing behaviour, Solid State Ionics, 225, Strickler D. W., Carlson W. G. (1965), Electrical conductivity in the ZrO 2 -rich region of several M 2 O 3 -ZrO 2 systems, J. Am. Ceram. Soc., 48, Tu H., Liu X., Yu Q. (2011), Synthesis and characterization of scandia ceria stabilized zirconia powders prepared by polymeric precursor method for integration into anode-supported solid oxide fuel cells, J. Power Sources, 196, Tu H., Stimming U. (2004), Advances, aging mechanisms and lifetime in solid-oxide fuel cells, J. Power Sources, Venkatachari K., Huang R. D., Ostrander S. P., Schulze W. A., Stangle G. C. (1995), A combustion synthesis process for synthesizing nanocrystalline zirconia powders, J. Mater. Res., 10, 3, Vlasov A. N., Perfiliev M.V., (1987), Ageing of ZrO 2 -based solid electrolytes, Solid State Ionics, Wang C., Zinkevich M., Aldinger F., (2007), Phase diagrams and thermodynamics of rare-earth-doped zirconia ceramics, Pure Appl. Chem., 79,