Synthesis, structural characterization, electrical properties and chemical stability of a (ZrO₂)_0.97(Y₂O₃)_0.03−x(MgO)_2x solid solution

Abstract

This paper is concerned with ternary solid state ZrO₂–Y₂O₃–MgO where the zirconia tetragonal phase is stabilized by incorporation of yttrium and/or magnesium in the ZrO₂ lattice. This subject is especially important due to the broad application of yttria stabilized zirconia (YSZ) in (among others) electroceramic systems in particular as potential component of the composite anode material for SOFC technology. A series of samples with a starting composition corresponding to 3 mol% yttria-stabilized zirconia (3YSZ) and increasing substitution of yttrium by magnesium were synthesized. The resulting samples were examined in terms of structural properties, phase composition, electrical properties and chemical stability. The obtained results show close correlation between the amount of incorporated magnesium and structural parameters of the tetragonal phase as well as electrical properties of all samples. The substitution of yttrium with magnesium in the 3YSZ system leads to a significant decrease of conductivity and the appearance of a monoclinic phase in the system for x > 0.015 (3 mol% of Mg²⁺). Moreover, it was confirmed that exposing the ZrO₂–Y₂O₃–MgO ternary system to a CO₂ and H₂O atmosphere can significantly decrease its chemical stability. From the point of view of both, basic and application research, these studies make an important contribution to the current knowledge of the properties of the ZrO₂–Y₂O₃–MgO ternary solid solution system.

---

1. Introduction

Tetragonal zirconia (t-ZrO₂), due to its potential (and practical) applicability, is one of the most intensively studied ceramic materials. The problem with this particular ZrO₂ phase is, that it is thermodynamically stable above 1170 °C, while below this temperature martensitic phase transition to monoclinic phase occurs, resulting in irreversible structural damage (due to approx. 4.5% volume increase during process). In tetragonal phase each zirconium cation is surrounded by eight oxygen anions which form distorted cube while in monoclinic phase the number of surrounding anions is reduced to 7. At ambient conditions t-ZrO₂ phase is thermodynamically metastable and only an increase of energy barrier (total free energy change essential for phase transition) can prevent phase transition to low symmetry phase. Tetragonal phase stabilization can be achieved in one of the two ways: either by optimization of preparation conditions in order to obtain the proper particle sizes (below critical one) which leads to an increase in surface free and internal energy changes, or by the incorporation of suitable dopants which may cause a decrease of chemical free energy change. According to Garvie¹ the surface free energy of nanocrystalline ZrO₂ powders drastically increases for crystallites with diameter below 30 nm and when diameter drops below 17 nm, only the tetragonal phase should be observed. The later investigations^2,3 confirmed Gravies' results. In practice, the latter of mentioned methods (the incorporation of aliovalent cations such as Y³⁺, Gd³⁺, La³⁺, Ca²⁺ or Mg²⁺ in Zr lattice) is more frequently used. Stabilization of t-ZrO₂ results in structural changes that impact directly other properties, e.g. increase of conductivity due to the formation of oxygen vacancies. Such defects are created as a charge-compensating lattice deficiencies accompanying ZrO₂ doping with lower valence cation according to equation:

The best examined doped ZrO₂ system is 3 mol% yttria-stabilized zirconia (3YSZ) which is abundantly used in various applications due to its thermal and chemical stability, mechanical durability and high ionic conductivity. By doping zirconia with MgO one can obtain material with thermal expansion coefficient (TEC) significantly higher than in the case of pure ZrO₂ or Y₂O₃–ZrO₂ solid solution (since TEC for pure 3YSZ and MgO is equal to 10.5 × 10⁻⁶ K⁻¹ and 13.9 × 10⁻⁶ K⁻¹ (ref. 4) respectively, the incorporation of certain amount of magnesia to 3YSZ should improve its thermal properties without significant change of other important ones). Therefore thorough examination of ternary solid solution MgO–Y₂O₃–ZrO₂ is very important, especially that this system is considered as a good material for electrochemical devices, in particular as potential component of the composite material of anode for SOFC technology. Recent studies also indicate that the introduction of MgO to the system will result in improvement of catalytic activity of the anode reforming reaction of hydrocarbons.^5,6 In contrast, double systems of solid solutions of yttria-stabilized zirconia (YSZ)^7–10 and magnesia – stabilized zirconia^11,12 as well as composite system YSZ – MgO^13,14 – mainly for magnesia amount well over 20 mol% – have already been examined intensively and used in practical applications, there is still quite difficult to find information about this ternary system for composition range related to full solubility of all these three components. As long as you can occasionally find such information. Similar studies has been previously conducted for such ternary system, but for regular system (8YSZ – 12YSZ), thus since it is well known that mechanical strength of 3YSZ system is superior comparing to 8YSZ one (while unfortunately electrical conductivity is smaller in the former one¹⁵), it is worth to study ternary system MgO – 3YSZ from the point of view of possible improvement of electrical properties of pure 3YSZ system.

2. Experimental

2.1. Preparation

A series of samples of zirconia doped with Y₂O₃ or/and MgO (in order to study the influence of gradual substitution of yttrium by magnesium in the 3 mol% yttria stabilized zirconia – respective compositions are presented in Table 1) were synthesized by modified citric method¹⁶ as follows: saturated solution of zirconyl nitrate (prepared by dissolving zirconia in nitrate acid) was mixed with yttrium nitrate solution and magnesium oxide. Next, the stoichiometric amount of citric acid (with around 10 mol% excess) was added and the final solution after initial heating in dryer (ca. 240 °C) was calcined in furnace at 800 °C (in air). Obtained samples were crushed in mortar, milled in atritor in isopropyl alcohol, dried at 240 °C in air and calcined in furnace at 400 °C for 3 h. Resulting powders were formed as cylindrical pallets and sintered in the air atmosphere in alundum crucible on the platinum wires in order to prevent the reaction between magnesium and Al₂O₃. The sintering temperatures were set to 1200 °C for one series of samples (pallets for electrical measurements) and 1300 °C for the second (for XRD analysis and chemical stability test).

Table 1 The composition of obtained samples

Sample label	Number of Y^3+ mole substituted by Mg^2+ (in 3YSZ)/mol%	Number of Y2O3 remaining in ZrO2 lattice/mol%	x value in the formula of (ZrO2)0.97(Y2O3)0.03−x(MgO)2x	Mol% of metal atoms
				Zr	Y	Mg
Mg0.5_YSZ	0.5	2.75	0.0025	94.175	5.340	0.485
Mg1.0_YSZ	1	2.5	0.005	94.175	4.854	0.971
Mg1.5_YSZ	1.5	2.25	0.0075	94.175	4.369	1.456
Mg2.0_YSZ	2	2.0	0.01	94.175	3.883	1.942
Mg2.5_YSZ	2.5	1.75	0.0125	94.175	3.398	2.427
Mg3.0_YSZ	3	1.5	0.015	94.175	2.913	2.913
Mg3.5_YSZ	3.5	1.25	0.0175	94.175	2.427	3.398
Mg4.0_YSZ	4	1.0	0.02	94.175	1.942	3.883
Mg4.5_YSZ	4.5	0.75	0.0225	94.175	1.456	4.369
Mg5.0_YSZ	5	0.5	0.025	94.175	0.485	5.340
Mg6.0_YSZ	6	0.0	0.03	94.175	0.000	5.825

---

After sintering the samples were left in furnace for cooling to room temperature. The citric method was chosen for materials preparation as the most suitable one, due to required properties of potential SOFC anode materials (high porosity, small grains). The resulting samples, after sintering at 1200 °C consist of nanocrystallites and have the total porosity (as determined by relative geometrical density measurements method) equal to 12.5 ± 2%.

2.2. Apparatus and methods

XRD analysis was performed using Philips X'Pert Pro diffractometer (CuK_α = 1.5406 Å, 2Θ = 20–90°). The amounts of particular phases were determined using Rietveld method by means of High Score Plus program.

The energy dispersive X-ray spectrometer (EDS-Oxford Instruments) coupled with scanning electron microscopy (SEM) was used to determine the presence of magnesium and yttrium in the samples. Bulk and grain boundaries conductivity were determined using impedance spectroscopy method (Solartron SI 1260 Impedance/Gain-Phase Analyzer). The measurements were carried out at 400, 500 and 600 °C in the flow of mixture 5% H₂/95% Ar and Pt-paste applied as the electrode, for frequency range 0.1 Hz to 10⁶ Hz and amplitude of the sinusoidal voltage set to 10 mV. The impedance spectra were analysed using ZPLOT software package delivered by Solartron.

Termogravimetry (TG) with simultaneous recording of mass spectra (MS) of released gaseous products was used for evaluation of chemical stability of obtained samples in CO₂ and H₂O atmosphere. TG measurements were carried out by means of SDT 2960 TA Instruments and mass spectrometry analysis using QMD 300 ThermoStar (Balzers). The samples with two chosen compositions (Mg2.0_YSZ and Mg6.0_YSZ) were exposed to the atmosphere enriched in CO₂ and H₂O (7% of CO₂ in air, 100% RH) at around 20 °C for about 700 h (four weeks). The fragments of pallets (of mass about 45 mg) after exposition test were heated in synthetic air atmosphere in platinum crucible with 10 deg min⁻¹ heating rate. SDT apparatus was coupled with quadruple mass spectrometer by quartz capillary which was heated (up to 200 °C) during all measurements and used for determination of mass loss and identification of released gases for samples before and after exposition test.

3. Results and discussion

3.1. Structural characterization

The structure analysis (phase composition, lattice parameters and cell volume) was carried out for a series of (ZrO₂)_0.97(Y₂O₃)_0.03−x(MgO)_2x samples sintered at 1300 °C. X-ray diffraction patterns (Fig. 1) confirmed that in samples with magnesium amount up to x = 0.015 only one, tetragonal phase of ZrO₂ (P4₂/nmc space group) is present, while for x ≥ 0.0175 second (monoclinic) phase appeared. The amount of tetragonal phase of ZrO₂ (estimated using Rietveld analysis with structural data from crystallography open database http://www.crystallography.net/, COD no. 1526427 and 9007485 for tetragonal¹⁷ and monoclinic¹⁸ ZrO₂ respectively, used as a starting point for refinement procedure) gradually decreases from 31.3 wt% for (ZrO₂)_0.97(Y₂O₃)_0.0125(MgO)_0.035 sample to 4.2 wt% for (ZrO₂)_0.97(MgO)_0.06. The detailed results obtained for all compositions are shown on Fig. 2. Besides mentioned tetragonal and monoclinic ZrO₂ there were no other phases in the tested samples and the presence of yttrium and magnesium was confirmed based on EDS analysis.

Fig. 1 XRD analysis of samples after sintering at 1300 °C. m – peaks corresponding to monoclinic phase.

Fig. 2 t-ZrO₂ phase participation for all samples.

The problem of tetragonal phase stabilization is very complicated – due to the insufficient data and degree of complex processes and factors involved in stabilization process it is difficult to give fully convincing, unambiguous explanation of behaviour as shown in Fig. 2 but the following reasoning, based on available information provides well justified explanation.

Taking into account the results reported by Fabris et al.,¹⁹ which show that dominant factor in the stabilization of the tetragonal phase are oxygen vacancies (with an increase in the number of vacancies, the degree of stabilization of the tetragonal phase increases) the addition of yttrium to ZrO₂ results mainly in an increase in the number of oxygen vacancies (created spontaneously in order to neutralize total charge of the system) and does not lead to significant changes in local dopant's environment in cationic sublattice in comparison with the host lattice – this is due to the same electronic configuration of Zr⁴⁺ and Y³⁺ cations, and hence the tendency to form the same type of bonds (directional covalent-ionic) in the same coordination polyhedral (distorted cube).

As a result, the larger the addition of yttrium, the stronger the effect of stabilization of higher symmetry phase. The results of magnesium introduction into cation sublattice are significantly different due to two mutually oppositely operating factors: the introduction of magnesium results in oxygen vacancies formation, analogously as in the case of yttrium doping (but the amount of vacancies is twice per mole bigger than in the case of yttrium) acting as stabilizing factor and, on the other hand, due to the completely different nature of chemical bonds between magnesium and surrounding oxygen atoms (spherically symmetric and strong non-directional ionic bonds) and magnesium preference for 6-fold coordination (according to Pauling's rules r_Mg/r_O = 0.464 – hence octahedron as a coordination polyhedron) magnesium introduction acts in destabilizing way, strongly disturbing its local (as well as neighbouring host cations') environment.

For substantially larger amounts of magnesium first factor probably becomes dominant so that it is possible to stabilize the higher symmetry phases of ZrO₂ by doping with sufficiently large amount of magnesium or – by analogy – calcium. In the latter case destabilizing effect of alkali earth metal introduction is significantly reduced due to specific properties of Ca²⁺ cations – larger ionic radius and hence weaker ionic interactions with surrounding oxygen ions and a tendency to adopt 6- or 8-fold coordination (the higher coordination for partial charge transfer, due to covalent bonding of oxygen ions with zirconium cations and thus also net charges smaller than formal ones – exactly the case here).

Based on the above analysis, one can expect that when we remove yttrium from the structure by substituting it with magnesium, the destabilizing effect coming from magnesium cation starts to dominate and at x > 0.015 this effect is already strong enough for monoclinic phase to appear (since such amount of magnesium is far too small for stabilizing effect of oxygen vacancies began to dominate the destabilizing influence of magnesium cations).

Further substitution of magnesium for yttrium leads to creation of larger quantities of monoclinic phase and the likely migration of part of magnesium cations into this phase resulting in an increase of the unit cell size of tetragonal phase. The latter hypothesis can be justified by the comparison of cation coordination polyhedra in both phases: in the case of undoped tetragonal phase¹⁷ such polyhedron (distorted cube) is created by 8 bonds, 4 longer (approx. 2.10 Å) and 4 shorter (approx. 2.38 Å) with the average bond length of 2.23 Å, whereas in the case of monoclinic phase¹⁸ polyhedron is formed by 7 bonds, each one of different length (from 2.05 Å to 2.28 Å) with the average bond length of 2.15 Å. Taking into account that in MgO ionic crystal²⁰ the length of Mg–O bonds is equal to 2.11 Å and thus very similar to the average Mg–O bond length in monoclinic phase, one can infer that the transfer of magnesium from tetragonal to monoclinic phase will reduce the local stresses in respective magnesium coordination polyhedra (due to 7-fold instead of 8-fold coordination, closer to preferred by magnesium octahedral one and smaller variance of bond lengths and more uniform distribution of neighboring oxygen ions) and thus will be energetically more favourable.

As expected, incorporation of Mg²⁺ ions in the ZrO₂ tetragonal lattice leads to decrease of a cell parameter as a result of magnesium ion radius being smaller than zirconium one (Fig. 3).

Fig. 3 Unit cell a and c parameters and volume for all samples.

Taking into account ionic radii of 8-coordinate Zr⁴⁺, Y³⁺ and Mg²⁺ (as in t-ZrO₂) – 0.98 Å, 1.16 Å and 1.03 Å respectively,²¹ one could expect some decrease of cell parameters and volume after Y³⁺ substitution by Mg²⁺. However, as it can be observed in Fig. 3, up to x = 0.015 cell volume decreases with increasing amount of magnesium substitution for yttrium and then starts to increase. Careful analysis of cell parameters in tetragonal structure (Fig. 3) shows that such behaviour of cell volume change is connected with an increase of c parameter accompanying the changes of a value and closely connected with structural changes in MgO–Y₂O₃–ZrO₂ system.

Comparing the changes of crystallite sizes in tetragonal phase with composition (Fig. 4), one can observe that crystallite sizes are similar in the case of samples with x from 0.0025 to 0.0125 (for which only tetragonal phase exists in the system). For x between 0.0175 and 0.03 the crystallite sizes drastically decrease and this is correlated with an increase of cell volume.

Fig. 4 The changes of crystallite sizes in tetragonal phase and cell volume for synthesized samples (dashed line separates single phase tetragonal samples from two-phase, tetragonal and monoclinic ones).

At the same time additional, monoclinic phase appears in samples, with crystallite sizes (estimated by means of Scherrer formula based on FWHM of most intensive peak) equal to 26 nm, 29 nm, 32 nm and 48 nm for Mg3.5_YSZ, Mg4.0_YSZ, Mg4.5_YSZ and Mg6.0_YSZ samples, respectively. Taking into account the above results we can propose the hypothesis that when magnesium content exceeds x = 0.015, due to insufficient stabilization in the bulk of tetragonal crystallites, statistically distributed nanodomains of monoclinic phase begin to form, resulting in an effective decrease of the size of crystallites of tetragonal phase, “separated by” monoclinic phase crystallites (which additionally stabilizes tetragonal phase, since as it is well known, in smaller crystallites tetragonal phase is preferred over monoclinic one).

To sum up, one can observe that for x ≥ 0.0175 to x = 0.03 the ratio of Mg/Y dopants in the system is not sufficient for effective stabilization of tetragonal form. The weaker stabilization effect of magnesium in comparison with yttrium can be explained on the basis of electronic structure of these metals – the presence of d-orbitals in yttrium valence shell allows formation of directional, covalent bonds with oxygen atoms, which is impossible in the case of s-type valence orbitals in magnesium (spherically symmetric ionic interactions). Therefore, coordination polyhedra formed around yttrium cations are more stable than magnesium ones. Moreover, the effect of tetragonal phase stabilization by ZrO₂ particle sizes (below critical diameter) seems to be – at least – of equal importance. The amount of t-ZrO₂ phase decreases, in favour of m-ZrO₂, with increasing amount of Mg²⁺ (and thus depletion of Y³⁺) which is at the same time related to an increase of crystallite sizes of tetragonal phase.

3.2. Electrical properties

The results of electrical properties measurements show an interesting behaviour of studied system: it follows from Fig. 5 (where dependence of total conductivity (σ) on temperature for samples with selected compositions is shown) that while for all tested samples σ increases with temperature, the respective conductivity values can be divided into two separate groups. The first one consists of samples with tetragonal structure only (Mg0.5_YSZ and Mg2.5_YSZ) with electrical conductivity significantly higher (more than an order of magnitude) than in pure 3YSZ system⁹ and the samples from the second group consisting of samples containing also monoclinic phase (Mg3.5_YSZ and Mg4.5_YSZ). Additionally, for the samples with highest possible magnesium amount (with complete yttrium substitution by magnesium; Mg6.0_YSZ) the further dip in σ values can be observed. It is well known that in case of ceramic materials the bulk and grain boundary conductivities contribution to total conductivity of sample is very important. The employment of impedance spectroscopy method allows the analysis of particular types of conductivity.

Fig. 5 The total conductivities σ of selected samples with temperature.

The obtained impedance spectra (Fig. 6 as an example of Nyquist plots for Mg0.5_YSZ and Mg3.5_YSZ samples) consist of three semicircles related to bulk conductivity (σ_b), grain boundary conductivity (σ_gb) and electrode conductivity. In the case of analysed samples, irregular complex semicircle composed of two (or three) semicircles corresponding to appropriate conductivity can be observed. The bulk conductivity (σ_b) and the grain boundary conductivity (σ_gb) were calculated concerning the geometry of the sample and bulk resistance (R_b) or grain boundary resistance (R_gb) obtained based on the EIS measurements: σ_b = L/(AR_b) and σ_gb = L/(AR_gb), where L and A are the sample thickness and the electrode area respectively. The results of analysis of these spectra for the measurements carried out at 500 °C are shown in Fig. 7. There are two series of points here – one related to σ_b the other to σ_gb. One can observe that the values of both, σ_b and σ_gb decrease at first, with addition of magnesium up to x = 0.0175 (Mg3.5_YSZ sample), then conductivities slightly increase and for x = 0.03 (Mg6.0_YSZ) strongly decline. The local minima at the curves (corresponding to Mg3.5_YSZ sample) can be related to unpredicted appearance of monoclinic phase (resulting in some disorder in previously ordered tetragonal phase). Generally, incorporation of further amounts of magnesium in the place of yttrium results in conductivity decrease and complete substitution of Y³⁺ by Mg²⁺ leads to a drop of both, σ_b and σ_gb by approx. two orders of magnitude.

Fig. 6 The examples of Nyquist plots for Mg0.5_YSZ and Mg3.5_YSZ samples.

Fig. 7 The comparison of values σ_b and σ_gb for the all samples (dashed line separates single phase tetragonal samples from two-phase, tetragonal and monoclinic ones).

3.3. Chemical stability

The possibility of MgO formation is disadvantageous phenomena which can lead to a loss of good mechanical and electrical properties of materials designed for potential component in electrochemical devices (for example anode material for SOFC technology). Moreover, as a result of secondary reaction of magnesia with surrounding atmosphere, some other compounds can be created. Since in the air the most chemically active component is carbon dioxide and water vapour, the behaviour of two of obtained compositions were examined in environment abundant in CO₂ and saturated with H₂O. The first composition contained slight amount of introduced magnesium (Mg2.0_YSZ), while the second one the maximum amount possible (Mg6.0_YSZ; thus providing the opportunity to obtain maximum amount of secondary reaction products). The other two probable oxides in studied ternary solid solution system (i.e. Y₂O₃ and ZrO₂) cannot react with H₂O and/or CO₂ (neither in ambient nor in work conditions of SOFC). In case of magnesium oxide several compounds can be formed, e.g. magnesium carbonate, magnesium carbonate hydrates, hydroxide or mixed-carbonate-hydroxide hydrates. The amount of products created is strictly connected with chemical stability of examined system. For the analysis of secondary reaction products the thermal decomposition method was used and measured decomposition curves are presented in Fig. 8. Comparing decomposition curves obtained for original sample and sample after exposition test one can observe significant differences between them in both, Mg2.0_YSZ (Fig. 8a) and Mg6.0_YSZ (Fig. 8b) systems. The mass losses in original samples are significantly lower than in ones after exposition. In the case of Mg2.0_YSZ original sample total mass loss up to 1000 °C is ca. 0.063 ± 0.002 wt% while for the sample after exposition to CO₂ and H₂O the mass loss in the same temperature range is 0.076 ± 0.02 wt%. For Mg6.0_YSZ sample the mass losses are equal to 0.10 ± 0.01 wt% and 0.20 ± 0.01 wt% respectively. Additionally, mass spectra registered simultaneously with TG curves show that larger amounts of CO₂ and H₂O are released in samples subjected to exposition. The shape (character) of ionic current lines (registered for H₂O and CO₂) is strictly associated with thermal decomposition processes visible on TG curves. There are almost no distinct peaks on mass spectra in the case of Mg2.0_YSZ sample, only for material after exposition one can observe extended peak above 200 °C. Below this temperature only the liberation of adsorbed water can be noticed. The behaviour of Mg6.0_YSZ is entirely different – TG curves and mass spectra have more complex character what corresponds to three stage decomposition of tested materials. The first step below 200 °C matches the release of water attributed to dehydration and liberation of adsorbed H₂O. The second step from 200 °C up to around 450 °C is connected with evolving of CO₂ and small amount of H₂O and the last one (above 450° to around 700 °C) again with CO₂ release. In case of original samples these steps are less visible, due to a smaller amount of secondary products. The potential products of MgO reaction with CO₂ and H₂O can be described by general formula xMgCO₃·yMg(OH)₂·zH₂O. According to some authors, 3MgCO₃·Mg(OH)₂·3H₂O²² or 4MgCO₃·Mg(OH)₂·4H₂O²³ can be formed in this system. Taking into account TG/MS data presented above, as well as thermal analysis results reported by Khan,²³ we can conclude that in the samples exposed on CO₂ and H₂O mixed carbonate-hydroxide hydrates are formed. The creation of this type of compound would definitely reduce chemical stability of magnesium doped ZrO₂.

Fig. 8 TG curves and selected ionic currents (MS spectra) registered during TG measurements for original and after exposition samples (a) Mg2.0_YSZ; (b) Mg6.0_YSZ sample.

4. Conclusions

The studies of structural properties of 3YSZ system doped with magnesium demonstrate that systematic substitution of yttrium by magnesium in the structure allows maintaining the stable tetragonal phase up to certain amount of magnesium admixture (x = 0.015) and the stabilization is realized by both dopants, while for higher x there is no sufficient stabilization effect due to the weaker stabilization ability of magnesium cations, which results in appearance of additional, monoclinic phase. Furthermore, electrical properties test indicate that for small amount of magnesium (x < 0.015), when the resulting samples remain tetragonal, the electrical conductivity is significantly higher than in the case of pure 3YSZ samples. The higher amount of magnesium results, however, in appearance of second, monoclinic phase and significant decrease of electrical conductivity. Additionally, incorporation of magnesium into 3YSZ leads to worsening of chemical stability (but still 3YSZ has higher mechanical strength than 8YSZ). Based on presented results, we can state, that the samples with small (up to x < 0.015) amount of magnesium can still be considered as promising material composite for SOFC anode.

Acknowledgements

I would like to thank Dr Jan Wyrwa for performing the electrical measurements. This work was financially supported by the National Science Centre of the Republic of Poland, Grant No. 2014/14/EST5/00763.

References

1. R. C. Garvie, J. Phys. Chem., 1965, 69, 1238.
2. T. Chraska, A. H. King and C. C. Berndt, Mater. Sci. Eng., A, 2000, 286, 169.
3. S. Shukla and S. Seal, Int. Mater. Rev., 2005, 50(1), 45.
4. Y. S. Toloukian, R. K. Kirby, R. E. Taylor and T. Y. L. Lee, Thermophysical Properties of Matter, IFI/Plenum, New York, USA, 1977, vol. 13, p. 288.
5. Y. J. O. Asencios, J. D. A. Bellido and E. M. Assfat, Appl. Catal., A, 2011, 397, 138.
6. A. Djaidja, H. Messaoudi, D. Kaddeche and A. Barama, Int. J. Hydrogen Energy, 2015, 40, 4989.
7. T. Sakuma, Y.-I. Yoshizawa and H. Suto, J. Mater. Sci., 1985, 20, 2399.
8. H. Hayashi, A. Ueda, A. Suino, K. Hiro and Y. Hakuta, J. Solid State Chem., 2009, 182, 2985.
9. E. Drozdz, M. Jelonek, J. Wyrwa and M. Rekas, Arch. Metall. Mater., 2015, 60(2), 998.
10. T. Sakuma, Trans. Jpn. Inst. Met., 1988, 29(11), 879.
11. Y. Yin and B. B. Argent, J. Phase Equilib., 1993, 14(4), 439.
12. H. Gocmez and H. Fujimori, Mater. Sci. Eng., B, 2008, 148, 226.
13. F. C. Fonseca and R. Muccillo, Solid State Ionics, 2000, 131, 301.
14. Y. Shiratori, F. Tietz, H. P. Buchkremer and D. Stöver, Solid State Ionics, 2003, 164(1–2), 27.
15. A. Pawłowski, M. M. Bućko and Z. Pędzich, Mater. Res. Bull., 2002, 37, 425.
16. E. Drożdż and J. Therm, Anal. Calorim., 2014, 118, 1345.
17. N. Igawa and Y. Ishii, J. Am. Ceram. Soc., 2001, 84, 1169–1171.
18. D. K. Smith and H. W. Newkirk, Acta Crystallogr., 1965, 18, 983.
19. S. Fabris, A. T. Paxton and M. W. Finnis, Acta Mater., 2002, 50(20), 5171.
20. S. Sasaki, K. Fujino and Y. Takéuchi, Proc. Jpn. Acad., Ser. B, 1979, 55(2), 43.
21. R. D. Shannon, Acta Crystallogr., Sect. A: Cryst. Phys., Diffr., Theor. Gen. Crystallogr., 1976, 32, 751.
22. N. Khan, D. Dollimore, K. Alexander and F. W. Wilburn, Thermochim. Acta, 2001, 367–368, 321.
23. K. J. Fricker and A.-H. A. Park, Chem. Eng. Sci., 2013, 100, 332.

---