Room-Temperature Synthesis and Electrical ...

8 downloads 0 Views 141KB Size Report
We show in this paper the possibility of using mechanical milling to prepare apatite- type La, Nd and Gd silicates starting from stoichiometric mixtures of the ...
Mater. Res. Soc. Symp. Proc. Vol. 972 © 2007 Materials Research Society

0972-AA09-05

Room-Temperature Synthesis and Electrical Properties of La, Nd and Gd Apatite-Type Silicates Antonio F. Fuentes1, Luis G. Martinez-Gonzalez1, Karla J. Moreno1, Evelyn RodriguezReyna1, and Ulises Amador2 1 Unidad Saltillo, Cinvestav, Carretera Saltillo-Monterrey Km. 13, Ramos Arizpe, Coahuila, 25900, Mexico 2 Departamento de Quimica, Facultad de Farmacia, Universidad San Pablo-CEU, Boadilla del Monte, Madrid, 28668, Spain ABSTRACT We show in this paper the possibility of using mechanical milling to prepare apatitetype La, Nd and Gd silicates starting from stoichiometric mixtures of the constituent oxides. XRD patterns collected after grinding the starting mixtures for 9 hours contain only the characteristic reflections of the target materials with no other phase apparently present. Electrical conductivity data were successfully fitted to a Jonscher-type empirical expression with fractional exponent n included in the 0.35-0.75 range. Activation energies for oxygen migration were found to decrease as the size of the rare-earth cation increases. Therefore, the highest conductivity values were found for the apatite-type lanthanum silicate.

INTRODUCTION Apatite-type rare earth silicates, RE10-x(SiO4)6O3-1.5x, have recently gained considerable attention in the field of solid oxide ion conductors because of their high ionic conductivity at low temperatures (e.g. 0.01 S·cm-1 at 700ºC), higher even than that obtained with yttria stabilized zirconia (YSZ) (for a recent review on apatite-type ion conductors see ref 1). The crystal structure of these compounds is built up of isolated SiO4 tetrahedra with the extra oxide ions responsible for the ionic conduction, occupying the center of one dimensional channel running through the structure along the c-axis at [0,0,z]. In addition, there are two RE3+ sites, one fully occupied on the periphery of the same tunnel containing the mobile oxygen atoms and another one, partially occupied, in the centre of a smaller tunnel running also parallel to the c-axis at [1/3,2/3,z]. While rare earth atoms located in the first site are coordinated to 7 oxygen atoms, those on the second one are 9-coordinated The presence of cation vacancies on the smaller channel allows the SiO4 tetrahedra substructure to relax towards the empty site enhancing oxide ion conduction along the larger channel. Therefore, ionic conductivity in single crystal apatite-type silicates have been found to be highly anisotropic with that parallel to the c-axis in the hexagonal lattice about one order of magnitude higher than that perpendicular to the same axis. Studied initially by Nakayama and co-workers (2,3) and despite their promising performance, their application has been hindered by the high temperatures and long firing cycles needed for their synthesis and sintering (>1400ºC). In this work, we will show that is possible to prepare apatite-type La, Nd and Gd silicates at room-temperature, by mechanically milling mixtures of the corresponding oxides. We will also analyze the electrical properties of as-prepared silicates by using impedance spectroscopy.

EXPERIMENTAL Stoichiometric mixtures of high purity (>99.9%) RE2O3 (RE = La, Nd, Gd) and amorphous silica (Aldrich Chem. Inc., 99.8%; particle size = 0.014 µm; surface area = 200 ± 25 m2/g) in a 5:4 molar ratio (SiO2: RE2O3) corresponding to a reaction product with a chemical formula of RE9.60(SiO4)6O2.4, were weighed out and placed in 125 ml zirconia containers together with six 20 mm diameter zirconia balls (mass ≈ 24 g; balls-to-powder mass ratio equal to 10:1). Rare earth oxides were dried overnight at 900°C in air before using in order to decompose hydroxides and/or oxycarbonates present. In a typical experiment, a 15 g batch of reactants were dry milled in air in a Retsch PM400 planetary ball mill by using a rotating disc speed of 350 rpm with reversed rotation every 20 min. Phase evolution with milling time was followed by using X-ray powder diffraction in a Philips X'Pert diffractometer using Ni-filtered CuKα radiation (λ = 1.5418 Å). Impedance studies were carried out from 300 to 900ºC on pellets (12mm diameter and ~1mm thickness) obtained by uniaxially pressing the fine-milled powders obtained by milling. To increase their mechanical strength and obtain dense samples, pellets were sintered at 1500ºC for 12 hours (heating and cooling rates 2ºCmin-1). AC impedance measurements were carried out in air using a Solartron 1260 Frequency Response Analyzer over the 100 Hz to 1MHz frequency range. Electrodes were made by coating opposite faces of the pellets with SPI-ChemTM conductive platinum paint.

RESULTS AND DISCUSSION Synthesis Figure 1 shows an X-ray diffraction study of the evolution of the Nd2O3-SiO2 starting mixture with milling time and the temperature of post-milling thermal treatment. As this Figure shows, new reflections emerge in the XRD pattern after grinding for 6 hours, mainly in the 20-35° region (2θ). After comparing this pattern with those included in the ICDD database, it became evident that they belong to the hexagonal apatite-type neodymium silicate. Thus, the presence of reflections appearing at ~21.5, 22.5, 25.4, 27.6 and 28.5° (2θ), identified as characteristic lines of the target phase, confirms that milling the starting mixture at room-temperature for only 6 hours at a moderate rotating disc speed, is enough to induce a chemical reaction and to obtain a product consisting primarily of Nd9.6(SiO4)6O2.4. The XRD pattern collected after milling for 9 hours is very similar to that obtained after 6 hours with no additional reflections belonging to the starting reagents or to any other mixed oxide, present suggesting a complete reaction after this milling time. Low intensity and broad reflections are typical of materials prepared by mechanical milling and indicate small crystallite size and low crystallinity. Similar results were observed in the starting mixtures incorporating amorphous silica and Gd2O3 or La2O3 after milling for 9 hours confirming the feasibility of using mechanical milling as a powder processing method for preparing rare-earth silicates. As Figure 1 shows, firing the as-prepared powders at temperatures of up to 600ºC does not produce noticeable changes in the XRD patterns. However, treating the milled powders at temperatures ≥ 800ºC produces peaks with increasing intensity and decreasing peak broadening corresponding to better ordered samples with increasing crystallite size. Figure 2 shows a comparison between the XRD patterns obtained for the three rare-earth silicates prepared by mechanical milling and fired 12 hours at 1000ºC evidencing a shift towards lower 2θ values as the size of the rare-earth cation decreases (RGd = 1.107 Å < RNd = 1.163 Å

< RLa = 1.216 Å, all 9-coordinated). Interestingly, no signs of the presence of the very stable RE2SiO5 and/or RE2Si2O7 impurities frequently present when these materials are prepared by traditional solid state reaction, were observed in any of the samples under study.

Fig.1: XRD study of the evolution of the Nd2O3 and SiO2 starting mixture with milling time and after milling for 9 hours and firing 12 hours at different temperatures

Fig.2: XRD patterns of the as-prepared apatitetype rare-earth silicates, RE9.6(SiO4)6O2.4 (RE = Gd, Nd and La) after firing 12 hours at 1000ºC.

Electrical properties Figure 3 shows the frequency dependence of the real part of the conductivity at several temperatures for as-prepared La9.6(SiO4)6O2.4 powders; plots obtained for the remaining rareearth silicates studied in this work were similar to the one presented. The ac conductivity is well described by the so-called Jonscher empirical expression [4]: σ ' (ω ) = σ o [(1 + (ω / ω p ) n ] (0 ≤ n < 1) which is the sum of the dc conductivity, σo, and a power-law term with fractional exponent n. The characteristic frequency ωp is the cross-over frequency and marks the onset of the powerlaw regime (σ’(ω) ∝ ω n) and is found to be thermally activated with about the same activation energy, Edc, of the dc conductivity. This fact indicates that the frequency dependent electrical conductivity, σ’(ω), originates from migration of ions and that the conductivity observed at low frequencies (ω < ωp) and the power-law dependence observed at high frequencies (ω > ωp) are actually due to the same ionic transport mechanism. There is in addition another contribution to the electrical conductivity which has to be considered in the in the above mentioned equation consisting of an almost linear frequency dependent term of the form σ’(ω) ≈ Aω). At sufficiently low temperatures or high frequencies, this term dominates over the power law dependence with the fractional exponent n. This nearly frequency independent dielectric loss (nearly constant loss, NCL) contribution might be added to the Jonscher type contribution and the total ac conductivity can be described by the extended Jonscher’s form: σ ' (ω ) = σ o [(1 + (ω / ω p ) n ] + Aω

Experimentally, A is not thermally activated and has temperature dependence much milder than σo, or ωp. There are two basic theories which describe the Jonscher type behaviour in ionic conductors: (i) independent separate ion hopping events with a broad distribution of relaxation times (parallel conduction mechanism) and (ii) collective motion with each hopping ion having strong Coulomb-type interactions with the surrounding charge carriers (series conduction mechanism). However, experimental evidences mostly support the latter and the frequency dispersion of the ac conductivity is believed to be related with the degree of ion-ion interactions existing in the ionic hopping process which on turn, will determine the value of the fractional exponent n; i.e. in the case of independent random ion hopping, the exponent n would be 0 while n would tend to 1 for a completely correlated ion motion (5, 6). The dc conductivity values for the title compounds can be directly obtained from the high frequency plateau at each temperature in Figure 3 (dc plateau) as well as from the traditional complex impedance plane analysis. By fitting the frequency dependence of the isothermal conductivity data to the above mentioned extended Jonscher-type expression, we calculated the value of n for the three compounds studied in this work which are plotted in Figure 4 as a function of the rare-earth cation radii. Interestingly, the value found for the lanthanum silicate 0.35, is lower than expected for ionic conductors, typically between 0.6 and 1. Although we do not have an explanation at the moment, further data analysis is currently underway to find a possible reason for this value. The conductivity decrease observed in figure 3 at high temperatures (above 500ºC) and low frequencies, is due to blocking effects at grain boundaries and such data points were omitted from the fitting process. The temperature dependence of the dc conductivity in the title compounds was analyzed by using an Arrhenius law of the form: σdcT = σoexp(-Edc/kBT), where σo is the pre-exponential factor (which is related amongst others to the effective number of mobile species), kB is the Boltzmann constant, T is the absolute temperature and Edc denotes the activation energy for the conduction process. Figure 4 shows such representation for the three compounds studied in this work where the dashed lines are fits to an Arrhenius law with regression coefficients in each case better than 0.998, confirming that the ionic diffusion process is thermally activated. Activation energies calculated from the slope of such linear fits were found to decrease with increasing ionic radii of the rare-earth element, with values of 0.85, 0.89 and 1.02 eV for the La, Nd and Gd silicates respectively. Decreasing activation energy implies increasing conductivity and the maximum value at each temperature is found for the La-containing silicate. In general, dc conductivity in the lanthanum compound increases almost by two orders of magnitude when compared to those obtained for the Nd and Gd similars. Figure 6 shows a graphic representation of the Edc and dc conductivity values (at 550ºC) found for each silicate versus the rare earth elements radii. Increasing conductivity has been related with the larger size of the hexagonal channel where mobile oxygens are located as the size of the RE3+ cation increases. Although the high oxide ion conduction mechanism in apatite-type silicates has not yet been fully understood, conductivity is apparently more affected by cation vacancies mostly at the 9-coordinated 4f site (that located at the centre of the smaller channel), rather than oxygen ion vacancies. Thus, as the number of cation vacancies decreases the total conductivity also decreases (7). By using computer modeling studies (8), two possible migration paths have been recently proposed: a direct linear path through [0,0,z] direction and a sinusoidal-like migration through a new interstitial oxygen site also inside the hexagonal channel. The presence of such interstitial oxygen ions is still the subject of some controversy because of the difficulties in including them in the analysis of X-ray and neutron powder diffraction data since they would lie very close to other atoms in the normal apatitetype crystal structure (9).

Fig.3: Frequency dependence of the real part of the conductivity for La9.6(SiO4)6O2.4 at several temperatures. Solid lines are examples of best fits to a Jonscher-type empirical expression.

Fig.5: Arrhenius plots of dc conductivity for the La, Nd and Gd apatite-type silicates studied in this work.

Fig.4: The value of n as a function of the ionic radii of the rare-earth cation, RE3+, in apatite-type RE9.6(SiO4)6O2.4.

Fig.6: Activation energy ( ) and dc conductivity at 550ºC ( ) as a function of the rare-earth cation radii in apatite-type silicates

Thus, some authors found no evidence of them in X-ray diffraction analysis of apatite-type lanthanum silicate single crystals (10). In any case, it seems that depending on chemical composition and temperature the actual mechanism is a mixture of interstitial and vacancy-based oxygen migration with the transition from one to another bringing about an ionic conductivity decrease.

CONCLUSIONS We have shown that is possible to easily prepare apatite-type rare-earth silicates by milling for a short period of time stoichiometric mixtures of the constituent oxides in a planetary ball mill using a moderate rotating disc speed. The proposed method avoids using high temperature and long firing cycles otherwise needed for preparing this type of compounds by traditional solid state reaction and as a solvent-free technique, improves existing and time consuming sol-gel methods. Electrical properties are comparable to those described for their analogues prepared by the above mentioned routes with increasing ionic conductivity as the size of the rare-earth element increases.

ACKNOWLEDGEMENTS Work supported by Mexican Conacyt (SEP-2003-C02-44075).

REFERENCES 1. 2. 3. 4.

P.R. Slater, J.E.H. Sansom, J.R. Tolchard, Chem. Rec. 4 (2004) 373. S. Nakayama, T. Kageyama, H. Aono, Y. Sadaoka, J. Mater. Chem. 5 (1995) 1801. S. Nakayama, H. Aono, Y. Sadaoka, Chem. Lett. 24 (1995) 431. A.K. Jonscher, Dielectric Relaxation in Solids, Chelsea Dielectric Press, London 1983. 5. K. Funke, J. Non-Cryst. Solids 172-174 (1994) 1215. 6. K.L. Ngai, K.Y. Tsang, Phys. Rev. E 60 (1999) 4511. 7. V.V. Kharton, A.L. Shaula, M.V. Patrakeev, J.C. Waerenborgh, D.P. Rojas, N.P. Vyshatko, E.V. Tsipis, A.A. Yaremchenko, F.M.B. Marques, J. Electrochem. Soc. 151 (2004) A1236. 8. J.R. Tolchard, M.S. Islam, P.R. Slater, J. Mater. Chem. 13 (2003) 1956. 9. L. León-Reina, J. Porras-Vázquez, E.R. Losilla, M.A.G. Aranda, Solid State Ion. 177 (2006) 1307. 10. H. Okudera, Y. Masubuchi, S. Kikkawa, A. Yoshiasa, Solid State Ion. 176 (2005) 1473.