Thermodynamic Understanding of Chromium Poisoning

Diss. ETH No.
18139
Thermodynamic Database of the La-Sr-Mn-Cr-O Oxide System and Applications to Solid Oxide Fuel Cells
DISSERTATION for the degree of DOCTOR OF SCIENCES of the SWISS FEDERAL INSTITUTE OF TECHNOLOGY ZURICH
presented by ERWIN POVODEN-KARADENIZ Mag. rer. nat. born on March 18, 1973 Citizen of Austria
accepted on the recommendation of Prof. Dr. Ludwig J. Gauckler, examiner Prof. John T.S. Irvine, co-examiner Dr. Ming Chen, co-examiner
Zurich, 2008
Dedicated to my parents
Whatever creates or increases happiness or some part of happiness, we ought to do; whatever destroys or hampers happiness, or gives rise to its opposite, we ought not to do.
Aristoteles
Acknowledgements
I am deeply grateful to my supervisor Professor Gauckler. He gave me a great chance by taking me into the boat: a boat that is not only sailed to scientific success. It took me away from an insecure float wobbling in the surf and approaches a promising future. Endless gratitude belongs to my wife who stays by my side throughout highs, downs, and distances, preventing me from losing the way; she is my firm anchor. I am greatly indebted to Nicholas Grundy for open doors, his patience of a saint, great teaching, and picky reviewing. He catalyzed my way into the field of thermodynamic modeling, airing the modeling is fun approach at any time. I would further like to thank Ming Chen for continuing scientific support and advising; he was often motivating me to spin the wheel of accurate and fast modeling and publishing. I owe thanks to Franc and Flavia Dugal-Borsari who saved me from an unintentional outdoor adventure in Zurich during a time when it was extremely difficult to find a new accommodation. It was a very pleasant time in Zollikon. I thank Brandon Brgler and Jennifer Rupp for their pleasing office companionship at the beginning of my work: they facilitated my jump into the ETH waters. I would also like to thank Toni Ivas for abiding collegiality, cooperation, and friendship. I thank the rest of the office crew, Thomas Ryll and Rene Tlke, for always enjoyable working hours.
Zurich, December 2008
Table of Contents
Table of Contents
Summary Zusammenfassung 7 9
Degradation of planar solid oxide fuel cells (SOFC) with LSM cathodes and Cralloy interconnects 1.1 Introduction 1.1.1 1.1.2 1.2 1.3 Principles of SOFC The problem of chromium poisoning 11 11 11 13 14 17 17 22 24 24 27 28 36 36 37 45 46 46 47 47 48 49
Volatilization of Cr2O3 Literature survey 1.3.1 1.3.2 1.3.3 1.3.4 1.3.5 1.3.6 Degradation of SOFC caused by chromium from the interconnect The role of current load on electrical losses of degraded SOFC Impedance spectroscopy measurements and implications on the degradation process Microstructures in degraded SOFC Amounts of chromium in SOFC tested with and without current load Critical assessment of proposed mechanisms of chromium poisoning
1.4
Proposed strategies against chromium poisoning and their effectiveness 1.4.1 1.4.2 Increasing the Cr-tolerance of conventional SOFC with Cr-interconnects and LSM cathodes New ways alternative materials
2 3
Aim of study Method 3.1 3.2 Benefits of the thermodynamic La-Sr-Mn-Cr-O oxide database for the understanding of Cr-poisoning of SOFC Thermodynamic modeling 3.2.1 3.2.2 3.2.3 Stoichiometric solid oxides Solid solution phases the Compound Energy Formalism (CEF) Vacancies and the concept of reciprocal reactions 4
Table of Contents
3.2.4 3.3 4 4.1
Calculation of defect chemistry using the Calphad approach
51 52 53 53 53 54 58 59 66 67 73 77 77 78 86 89 93 96 97 101 102 103 103 109 117 128 134 134 135 137 140 143
Optimization of model parameters Thermodynamic reassessment of the Cr-O system in the framework of SOFC research 4.1.1 Technology 4.1.2 4.1.3 4.1.4 4.1.5 4.1.6 Experimental data Previous assessments of the Cr-O System Thermodynamic modeling Optimization of parameters Results and discussion
Thermodynamic assessments
4.1.7 Conclusions 4.2 Thermodynamic assessment of the Mn-Cr-O system for SOFC materials 4.2.1 Introduction 4.2.2 4.2.3 4.2.4 4.2.5 4.2.6 4.3 Experimental Thermodynamic modeling Optimization of parameters Results Discussion
4.2.7 Applications on SOFC Thermodynamic assessment of the La-Cr-O system 4.3.1 Introduction 4.3.2 4.3.3 4.3.4 4.3.5 4.4 Literature review of the La-Cr system Literature review of the La-Cr-O system Thermodynamic modeling and optimization Results and discussion
4.3.6 Conclusions Thermodynamic La-Sr-Mn-Cr-O oxide database for SOFC applications 4.4.1 Introduction 4.4.2 4.4.3 4.4.4 Assessment of data from the literature Modeling and optimization Results and discussion
4.4.5 Conclusions 5 Thermodynamic calculations of impacts of chromium on Sr-doped 5
Table of Contents
manganite (LSM) cathodes for SOFC 5.1 5.2 5.3 Introduction Method Results 5.3.1 5.3.2 5.3.3 5.3.4 5.4 5.5 Thermodynamic calculations of La0.9Sr0.1MnO3- contaminated by chromium Thermodynamic calculations of (La0.8Sr0.2)0.9MnO3- contaminated by chromium Thermodynamic testing of LSM with Mn-deficiency Formation of Cr2O3
148 149 150 152 152 157 160 162 163 165
Discussion Conclusions
Appendix Thermodynamic La-Cr database Thermodynamic La-Sr-Mn-Cr-O-(H) oxide database
170 169 172
Curriculum Vitae
190
Summary
Summary
The thermodynamic La-Sr-Mn-Cr-O oxide database is established by assessing oxide subsystems using the CALPHAD (Calculation of phase diagrams) approach. The new database is applied to the problem of chromium poisoning of Sr-doped lanthanum manganite cathodes ((La1xSrx)1-yMnO3- or LSM) for Solid Oxide Fuel Cells (SOFC) stemming from gaseous Cr species from the high-Cr containing alloy of the interconnect. The chromium is known to deteriorate the electrical performance of the cathodes. In chapter 1 the basics of planar SOFC are briefly explained, and previous findings of chromium poisoning of SOFC are critically reviewed. Based on the findings from the literature it gets clear that several questions about the key mechanisms of the chromium poisoning have not been answered yet, and the aim of this study (chapter 2) is to gather a deeper understanding of these unsolved problems by using thermodynamics. In the third chapter the reader learns, how thermodynamic calculations can lead to a better understanding of a system, even if the system is in a thermodynamic non-equilibrium state, and the modeling approach used in this study is presented. Chapter 4 deals with the construction of the La-SrMn-Cr-O oxide database based on the assessments of subsystems. The new database is applied to the problem of chromium poisoning of SOFC with Cr-interconnects and LSM cathodes in chapter 5: a consistent phenomenological description of the process of chromium poisoning of SOFC cathodes is established that is in line with both experimental findings reported in the literature and thermodynamic calculations using the presented database. It is shown that chromium poisoning of SOFC cathodes is a rather complex process consisting of several steps, some of them probably occurring simultaneously. Some of these processes are governed by thermodynamics, and some are kinetically controlled. A key role is played by the adsorption of gaseous CrO3(g) (g = gaseous) and chromiumoxyhydroxides stemming from the interconnect on LSM and reaction of chromium with LSM. The associated chemical changes of the LSM phase, as well as the formation of a new spinel phase occur under thermodynamic control: decreasing concentrations of vacancies in LSM that contains chromium are calculated at decreased oxygen partial pressure reflecting SOFC operation at high current load. This has calamitous consequences for the electrochemical properties of the cathode. Furthermore spinel blocks pores and thus impedes the oxygen
Summary
reduction required for the function of the cell. Cr2O3(s) (s = solid) that hampers the diffusion of oxygen into the electrolyte is a metastable phase in LSM contaminated by chromium. With this contribution the prevailing question is resolved, which of the phenomena in a chromium-poisoned LSM cathode are governed by thermodynamics. Appropriate measures can be foreseen preventing the long-term degradation of SOFC cathodes in combination with high-chromium containing interconnects.
Zusammenfassung
Zusammenfassung
Die thermodynamische La-Sr-Mn-Cr-O Oxid-datenbank wird basierend auf dem Assessment oxidischer Subsysteme mit dem CALPHAD-ansatz (Berechnung von Phasendiagrammen) aufgebaut. Die neue Datenbank wird auf das Problem der Chromvergiftung von Srdotierten Lanthan-Manganit-kathoden ((La1xSrx)1yMnO3- oder LSM) fr FestoxidBrennstoffzellen (SOFC) angewandt, welches von gasfrmigen Cr spezies der hochgradig Crfhrenden Interkonnektor-Legierung herrhrt. Es ist bekannt, dass das Crom die elektrische Leistung der Kathoden verschlechtert. In Kapitel 1 werden die Grundlagen von planaren SOFC kurz erklrt, und es wird ein kritischer berblick ber bisherige Erkenntnisse der Chromvergiftung von SOFC gegeben. Basierend auf den Erkenntnissen aus der Literatur wird klar, dass einige Fragen, welche die Schlsselmechanismen der Chromvergiftung betreffen, noch nicht beantwortet wurden. Das Ziel dieser Studie (Kapitel 2) ist es, unter Verwendung der Thermodynamik ein tieferes Verstndnis dieser ungelsten Probleme zu erlangen. Im dritten Kapitel lernt der Leser, wie thermodynamische Berechnungen zu einem besseren Verstndnis eines Systems fhren knnen, selbst wenn dieses System sich in einem thermodynamischen Ungleichgewichtszustand befindet, und der in dieser Studie verwendete Modellansatz wird vorgestellt. Kapitel 4 beschftigt sich mit der Konstruktion der La-Sr-Mn-Cr-O OxidDatenbank, basierend auf den Assessments der Subsysteme. In Kapitel 5 wird die neue Datenbank auf das Problem der Chromvergiftung von SOFC mit Cr-interkonnektoren und LSM-kathoden angewandt: Eine konsistente phenomenologische Beschreibung des Prozesses der Chromvergiftung von SOFC-kathoden wird gegeben, welche sowohl im Einklang mit experimentellen Erkenntnissen in der Literatur als auch mit thermodynamischen Berechnungen unter Verwendung der prsentierten Datenbank steht. Es wird gezeigt, dass Chromvergiftung von SOFC-kathoden ein ziemlich komplexer Vorgang mit mehreren, teilweise gleichzeitig in der Zelle ablaufenden Schritten ist. Manche dieser Prozesse sind thermodynamisch kontrolliert, und manche laufen unter kinetischer Kontrolle ab. Eine Schlsselrolle spielt die Adsorbtion von gasfrmigem CrO3(g) (g = gasfrmig) und Chromium-oxyhydroxiden, welche vom Interkonnektor stammen, an LSM und die Reaktion von Chrom mit LSM. Die damit verbundenen chemischen nderungen der LSM-phase und die Bildung einer neuen Spinellphase finden unter thermodynamischer Kontrolle statt. Die 9
Zusammenfassung
Berechnungen ergeben abnehmende Konzentrationen der Leerstellen in Cr-hltigem LSM unter erniedrigtem Sauerstoffpartialdruck, und somit bei Betrieb von SOFC unter hohem Laststrom . Das hat katastrophale Konsequenzen fr die elektrochemischen Eigenschaften der Kathode. Weiters blockiert Spinell Poren und behindert so die fr die Funktion der Zelle notwendige Sauerstoffreduktion. Cr2O3(s) (s = fest), welches die Diffusion von Sauerstoff in den Elektrolyt erschwert, ist eine metastabile Phase in Cr-kontaminiertem LSM. Mit diesem Beitrag werden einige der vorherrschenden Fragen ber Chromvergiftung von SOFC geklrt, und geeignete Manahmen zur Verhinderung der Langzeitdegradation von SOFC-kathoden in Kombination mit hochgradig Chrom-fhrenden Interkonnektoren knnen vorhergesagt werden.
10
Degradation of planar solid oxide fuel cells (SOFC) with LSM cathodes and Cr-alloy interconnects
E. Povoden and L.J. Gauckler, to be submitted to Int. J. Mater. Rev. For the use of LSM cathodes in planar SOFC a comprehensive understanding of the mechanisms of the cell degradation caused by chromium diffusing from the interconnects into the cell is needed. This poisoning has been intensively investigated over the last decade. In this paper the affects of Cr on the degradation of SOFC with LSM cathodes and Cr-alloy interconnects are reviewed: the suggested models of chromium poisoning of planar SOFC with chromium-alloy interconnects and (La1-xSrx)1-yMnO3- (LSM) cathodes from the literature are critically assessed. Taking into account all available experimental findings on the affects of chromium on Sr-doped lanthanum manganite cathodes in planar solid oxide fuel cells, it can be concluded that several poisoning processes contribute to the deterioration of the cell performance. The review of all available experimental findings on the degradation of SOFC caused by chromium allows predictions, as to how the extent of degradation caused by chromium depends on the current load, operation temperature, operation time, as well as the amount of chromium diffusing from the interconnect.
1.1
Introduction
1.1.1 Principles of SOFC A fuel cell directly converts chemical energy into electrical energy. A solid oxide fuel cell consists of two porous electrodes that are separated by a dense, oxygen ion-conducting electrolyte. A simple schematic of the electrochemical process is shown in Fig. 1.1.1 (next page).
11
Fig. 1.1.1 Scheme of the electrochemical processes in a fuel cell with O2 oxidant and H2 fuel. White circles symbolize pores. a = anode, c = cathode, e = electrolyte. The oxygen, supplied at the cathode reacts with electrons from the external electric circuit to form oxygen ions. These ions migrate through the electrolyte to the anode. At the anode the oxygen ions react with hydrogen of the fuel to form water and release electrons. The electrons flow from the anode through the external electric circuit to the cathode. The direct-current electricity is produced by the electron flow through the external electric circuit. In an SOFC, cathode and electrolyte consist of refractory solid oxide ceramics, and ceramicmetal composites are used for the anode. The materials for the cell components need to have a sufficient chemical and structural stability at rather high temperatures up to 1273 K that occur during cell production as well as during cell operation. The electrodes are required to have high reactivity and the electrolyte must allow high oxygen ion diffusion. All the components of the cell need to be matched in their thermal expansion in order to minimize mechanical stresses. A single cell produces a voltage of 0.7 to 1 V and power around 0.5 to 1 W cm-2. Normally many cells are electrically connected in series by an interconnect, as shown in Fig. 1.1.2 (next page) for the widely used planar-design SOFC, forming a cell stack to obtain higher voltage and power.
12
Fig. 1.1.2 Planar design of SOFC The interconnect separates the fuels and oxidants in adjacent cells. It is the most demanding component in a planar SOFC as it should have a high electronic conductivity, a low ionic conductivity; stability in both oxidizing and reducing atmospheres at the high cell operating temperature (from about T = 973 K to 1273 K); low permeability for oxygen and hydrogen to minimize direct combination of oxidant and fuel during cell operation; a thermal expansion coefficient close to that of the cathode and the anode; and chemical compatibility (no reactions) with other cell materials. 1.1.2 The problem of chromium poisoning
In the 1990is LaCrO3-based ceramics were intensively investigated for interconnect applications in SOFC, as the thermal expansions of LaCrO3-based interconnect and conventional perovskite cathode materials are similar, the electronic conductivity of several LaCrO3-based ceramics under SOFC operating conditions is high, and their thermal and redox-stability is satisfying[1]. However high costs of these materials, difficulties in sintering and manufacturing and low mechanical strength[2] required the development of alternative interconnect materials. Nowadays the state-of-the-art interconnect is commonly a chromiumcontaining metal plate[3-5], as chromium alloys come close to all desired properties. However high-valent gaseous Cr-oxide and Cr-oxyhydroxides diffuse from the Cr2O3(s) scale covering the interconnect into the cathode up to the cathode-electrolyte interface and cause the degradation that results in the strong deterioration of the electrical performance of SOFC.
13
In the last decade a lot of effort was made to elucidate the degradation mechanisms, and it was suggested that high-valent gaseous Cr-oxide and Cr-oxyhydroxides detrimentally affect the O2 -adsorbtion, -reduction, and -diffusion processes.
1.2
Volatilization of Cr2O3
Early investigations[6-9] revealed that oxidation of Cr-containing alloys at high temperatures leads to the redeposition of Cr2O3(s) crystals at cooler parts of the experimental apparatus from the gas phase. This was a surprising result, as the film of Cr2O3(s) covering the alloy specimen would have been expected to act as a diffusion barrier preventing the migration of Cr that has a high vapour pressure from the alloy through the Cr2O3 layer. Furthermore, neither the vapour pressure of Cr2O3(s) nor its dissociation pressure is high enough to account for the quantities of deposits observed[10]. Because of the high vapour pressure of Cr it was thus first considered that the metal itself would diffuse along oxide grain boundaries of the barrier film, or at discontinuities such as fractures in the film and would then evaporate. But when it was learned that Cr2O3(s), in the absence of metal, lost weight when heated in oxygen, it became evident that a volatile Cr-oxide was being formed[10]. Caplan and Cohen[10] investigated the evaporation of Cr2O3(s) by measuring the weight loss when Cr2O3(s) pellets with 1.2 cm in diameter and height were heated at T = 1273 K in stagnant air, and at T = 1373 and 1473 K in flowing dry and wet oxygen as well as in dry and wet argon. Appreciable volatilization occurred in dry oxygen, the weight loss being 0.6 mg at T = 1373 K and 2.3-2.6 mg at T = 1473 K at a gas flow rate of 200 ml min-1 after 20 h. The volatilization in wet oxygen was significantly higher after 20 h at the same gas flow rate: 2.1 mg at T = 1373 K and 5.6 mg at T = 1473 K. In stagnant air the volatilization was 0.3 mg at T = 1273 K after 72 h. Volatilization of Cr2O3(s) was neither observed in dry nor in wet argon. The observation that no loss of Cr2O3(s) occurs in argon confirms that volatilization does neither occur by dissociation of the oxide nor as Cr2O3(g) vapor. Since weight loss takes place under oxidizing conditions, the volatile species must be a higher oxide of chromium. A known volatile oxide of Cr is CrO3, but its formation by the reaction
Cr2 O3(s) + 3 2O2(g) 2CrO3(g)
(1.2.1)
14
is thermodynamically unfavourable at high temperatures[10], G of Eq. 1.1 is calculated to be +321 kJ at T = 1273 K using assessed thermodynamic data for Cr2O3(s) from Povoden et al.[11], data for O2(g) from Dinsdale[12], and for CrO3(g) from Ebbinghaus[13]. As gaseous CrO3(g) was detected experimentally by mass spectrometry when Cr2O3(s) was heated under oxidizing conditions, the formation of CrO3(g) occurs under kinetic control[10]. The existence of gaseous Cr-oxyhydroxides as oxidation products (Eqs. 1.2.2 and 1.2.3) in wet atmosphere was experimentally proven in several studies[14-17]. Their formation by oxidation of Cr2O3 in wet air reads:
Cr2 O3(s) + 1 2O2(g) + 2H 2 O(g) 2CrO2 (OH)2(g)
(1.2.2)
Cr2 O3(s) + 1 2O2(g) + H 2 O(g) 2CrO2 (OH)(g)
(1.2.3)
G of reaction 1.2 is calculated to be 158 kJ at T = 1273 K using combined data from Opila et al.[18] and Ebbinghaus[13], and G of reaction 1.3 is calculated to be +134 kJ at T = 1273 K using combined data from Kim and Belton[14] and Ebbinghaus[13]. Ebbinghaus[13] estimated a significantly higher partial pressure of CrO2(OH)2(g) compared to CrO3(g) in wet atmosphere up to T = 1600 K based on available thermodynamic data of gaseous Cr-species, and this was affirmed by thermodynamic modeling[19]. However, in a recent combined experimental and modeling study[18] these earlier findings are rejected for high temperatures: in wet atmosphere CrO2(OH)2(g) is predominant in the gas from T 1173 K, whereas at higher temperatures the gas phase mainly contains CrO3(g) and CrO2OH(g). This tendency is shown in Fig. 1.2.1 (next page): the calculated amounts of main Cr-species in the gas phase as a function of temperature in humid air of pH2O = 2000 Pa at constant chemical potential of oxygen, (O) being 300 J mol-1 referred to pure oxygen gas result from combined thermodynamic data[11-14,18] cited above. These findings are supported by the higher volatilization of Cr2O3(s) in wet air.
15
Fig. 1.2.1 Calculated amounts of gas molecules in Cr-gas as a function of temperature for constant pH2O = 2000 Pa at (O) = 300 J mol-1 referred to 100000 Pa O2(g) Transpiration experiments of Cr2O3(s) from T = 673 K to 1223 K resulted in the following partial pressures of Cr at a flow rate of 150 m min-1: pCr = 2.12x10-5 Pa at T = 673 K and increases as a function of increasing temperature, reaching pCr = 4.57x10-3 Pa at T = 1223 K[20]. Mass loss of Cr2O3(s) at T = 973 K and 1073 K was measured in air with different amounts of water, the mass loss being higher at higher water content and higher temperature: the constant rate of mass loss was 0.6 g h-1 for 3 mol% H2O in air at T = 973 K, 3.2 g h-1 for 3 mol% H2O in air at T = 1073 K, and 18.3 g h-1 for 25 mol% H2O in air at T = 1073 K. Cr-vaporization in SOFC: Konysheva et al.[21] used a transpiration method proposed by Gindorf et al.[20] to measure the vaporization rate of Cr from Cr5Fe1Y2O3 (Ducrolloy) and Crofer22APU (high-Cr ferritic steel), two high-chromium alloy interconnects widely used in SOFC, at T = 1073 K for a time period of about 500 h. The Cr-vaporization rate of Cr5Fe1Y2O3 exceeds that of Crofer22APU by about a factor of 3 in the temperature range from T = 1023 K to 1173 K, and the
16
vaporization rate increases with increasing temperatures for both alloys. With increasing humidity the difference in the vaporization rates between the two alloys increases. However, for an SOFC setup with a Cr5Fe1Y2O3 interconnect plate, an LSM cathode, an yttriumstabilized zirconia (YSZ) electrolyte and a Ni-zirconia cermet (ceramic-metallic composite) anode operated at T = 1173 K and 1273 K, Badwal et al.[22] mentioned significant amounts of deposited Cr2O3(s) in the air exhaust of the cell; thus quantitative chromium poisoning rates affecting the cathode are difficult to determine. This is in line with the experimental observation[21] that only a fraction of the chromium deposited at the cathode side contributes to the strong degradation of SOFC with LSM cathodes and Cr-alloy interconnects that were tested under a current load of 200 mA cm-2 for 393 h. The amount of Cr in these degraded cells was 140 g cm-2 with Cr5Fe1Y2O3, this value being by about a factor of 2.5 higher than with Crofer22APU.
1.3
Literature survey
1.3.1 Degradation of SOFC caused by chromium from the interconnect Considering the experimental data from Caplan and Cohen[10], the volatilities of gaseous CrO3(g) and gaseous Cr-oxyhydroxides are negligible under the low oxygen partial pressure at the fuel side of the cell, and the chromium problem is restricted to the interconnect-cathodeelectrolyte region of SOFC. Experimental results[21-29,31,32,34,35] of the degradation of SOFC with LSM cathodes caused by chromium are listed in Table 1.3.1, pp. 18-21.
17
Table 1.3.1 Results of chromium poisoning of SOFC with and without Cr-containing interconnects with LSM cathodes collected from the literature
18
19
20
The degrading effect of gaseous chromium species that form at the Cr2O3 scale under oxidizing conditions and diffuse into the cathode on the cell performance was first reported in 1995 by Taniguchi et al.[23]. These authors measured an increase of cathode polarization and decrease of cell voltage in an SOFC consisting of an LSM cathode with the compositions La0.9Sr0.1MnO3-, a YSZ electrolyte and a NiO/YSZ anode with a piece of a Ni-Cr-Fe-alloy (Inconel 600) attached on top of a Pt mesh used as current collector. The cell was electrochemically tested at T = 1273 K under a current load of 300 mA cm-2 for 400 h. The cell voltage decreased over this time from initially about 0.7 V to about 0.1 V, and intensity measurements using electron probe microanalysis showed that Cr was concentrated at the cathode-electrolyte interface. In a test of the same setup under open circuit conditions for
21
300 h no deterioration of the cell performance was observed, and Cr was randomly distributed across the cathode. Taniguchi et al.[23] thus linked the cell degradation to the time that the discharge current was applied, and to chromium deposited at the LSM-YSZ interface filling pores, hindering the supply of oxygen gas and decreasing the number of reaction sites for the oxygen reduction. Later these results were confirmed by Badwal et al.[22]: these authors tested the cell performance of an SOFC with a Cr5Fe1Y2O3 interconnect plate, an LSM cathode, a YSZ electrolyte and a Ni-zirconia cermet anode at T = 1273 K and a current density of 250 mA cm-2. Experiments without a Cr-based interconnect plate with Pt mesh serving as current collectors were conducted at T = 1205 K and 188 mA cm-2 current density using the same electrodes and electrolyte. The cell performance without interconnect plate deteriorated only little by less than 0.1 V during an operation time of 2500 h. On the other hand the voltage of the cell with Cr-Fe-alloy interconnect decreased rapidly as a function of operation time, the voltage drop being 0.4 V after only 16 h. A comparison of measurements of the overpotentials of SOFC with LSM cathode and high-Cr alloy interconnect with measurements without interconnect[24-27] or LaCrO3-based interconnect[28] led to the following results: the overpotentials without interconnects or with LaCrO3-base interconnects consistently became less negative with time, whereas the opposite was observed for SOFC with Cr-alloy interconnect. Simner et al.[29] presented cell performance data of an SOFC with LSM cathode, a Sm2O3-CeO2 interlayer between cathode and electrolyte, a YSZ electrolyte and a Pt counter electrode (in the following SOFC with Pt counter electrode are denoted as half-cell) with and without Crofer22APU interconnect at T = 1073 K, holding the cells at 0.7 V: without the interconnect steel, the cell performance was stable for 110 h, the power density being 0.48 W cm-2. But using a Cr-Fe-alloy interconnect, the cell started to degrade severely after 20 h of testing, its power density decreasing from a maximum of 0.2 W cm-2 to 0.05 W cm-2 after 110 h. All these experimental studies[22-29] unambiguously proved that chromium stemming from the alloy interconnect causes the degradation of SOFC.
1.3.2
The role of current load on electrical losses of degraded SOFC
Badwal et al.[22] reported that the degradation rate of SOFC with a Cr5Fe1Y2O3 interconnect plate, an LSM cathode, a YSZ electrolyte and a Ni-zirconia cermet anode at T = 1173 K and T = 1273 K was more related to the period of current passage and was less dependent on the time when no current was flowing through the cell: Badwal et al.[22] ascribed the voltage decrease to increasing losses of cathodic overpotential, whereas ohmic losses (resistance to 22
flow of electrons through the cathode) increased only insignificantly during the cell tests. The overpotential losses were higher and the cell deterioration was faster at higher current density. At the beginning of the cell tests, losses increased sharply and reached their maximum values after only 15 h of cell operation. Konysheva et al.[21], using a half-cell with a Cr5Fe1Y2O3 interconnect, a La0.65Sr0.3MnO3-/LSM+YSZ double layer cathode and a YSZ electrolyte confirmed the strong dependence of the voltage drop on the current density during 450 h cell tests at T = 1073 K: at 70 mA cm-1 the voltage decrease was 0.07 V, whereas at 280 mA cm-1 the voltage dropped by 0.6 V. A plateau of degradation was reached after about 400 h of testing. Konysheva et al.[30] tested the reversibility of degradation in a half-cell setup with CrFe-alloy interconnect, LSM cathode and YSZ at a current density of 100 mA cm-2 at T = 1073 K and found that the rapid degradation was reversible and disappeared after switching off the current load, in agreement with earlier findings[22,23,28,31]. However the cell degraded rapidly again when the current was switched on again. Matsuzaki and Yasuda[28] measured an overpotential loss from initially 500 mV to 2000 mV after 14 h in a half-cell setup with an Inconel 600 interconnect, La0.6Sr0.4MnO3- cathode and a YSZ electrolyte at 300 mA cm-2 current density. Zhen et al.[27] measured a rapid decrease of cell polarization from initially 350 to 750 mV after only 10 minutes in a half-cell with Cr-Fe-alloy (RA446) interconnect, LSM cathode and a YSZ electrolyte tested at T = 1173 K and a current density of 200 mA cm-2. In earlier studies using the same setup a rapid decrease from 360 to 560 mV after 10 minutes[25,26] was observed. In reference tests without Cr-Fe-alloy interconnect the results were opposite to the tests with Cr-Fe-alloy interconnects: the polarization increased from 550 to 300 mV[27] or 420 to 170 mV [24,25] at 1173 K and a current density of 200 mA cm-2. The polarization behavior of SOFC with Crcontaining interconnect was explained by the strong inhibiting effect of gaseous Cr-species on the oxygen reduction in LSM[27], in general agreement with Jiang et al.[24,25,32,33]. Paulson and Birss[34] reported rapid deterioration of the performance of a half-cell setup with a stainless steel disk containing 15.21 % Cr on top of a (La0.8Sr0.2)0.98MnO3- cathode and a YSZ electrolyte over 5 to 10 h at T = 1073 K applying a 0.5 V potential, with a tendency of stabilization of the cell performance after this testing period at much lower magnitude of output current. The total polarization resistance (Rpol) of a half-cell setup using a Cr-Fe-alloy interconnect, an LSM/LSM-YSZ cathode double layer and a YSZ electrolyte tested for 400 h was markedly dependent on the thickness of the LSM-YSZ layer, Rpol being 0.5 Ohm cm2 for a thickness of
23
50 m, 0.75 Ohm cm2 for 13 m, and 2 Ohm cm2 for 7 m[30]. The total polarization resistance was also higher at higher current load. Influence of temperature on the degradation: SOFC with LSM cathodes and high-chromium 430 stainless steel were tested at T = 973 K and 1073 K for 300 h[35]. The degradation was higher at higher temperatures at 0.7 V; it deteriorated from 500 mA to 350 mA at T = 1073 K and from 300 mA to 150 mA at T = 973 K over the testing time of 200 h and then remained constant. Jiang et al.[32] observed less overpotential losses at lower temperatures in a half-cell with CrFe-alloy interconnect, LSM cathode and a YSZ electrolyte: they measured an overpotential change from initially 300 mV to 650 mV after 10 minutes at T = 1173 K, from 900 to 1200 mV at T = 1073 K, and from 800 to 1120 mV at T = 973 K after 10 minutes.
1.3.3 Impedance spectroscopy measurements and implications on the degradation process Badwal et al.[22] observed an increased size of the high frequency arc during the current passage in half-cell tests using a Cr5Fe1Y2O3 interconnect plate, an LSM cathode and a YSZ electrolyte as a function of operation time, reflecting increasing cathode resistance. Mazusaki and Yasuda[28] concluded from the interpretation of impedance spectra of a half-cell with an Inconel 600 interconnect, a La0.6Sr0.4MnO3- cathode and a YSZ electrolyte operated at T=1073 K and 300 mA cm-2 current load that the degradation in the electrode caused by chromium was due to the increase in both charge-transfer resistance and surface diffusion resistance, but not due to the increase in ohmic resistance. Zhen et al.[27] reported the existence of a high frequency and a low frequency arc in impedance spectra of a half-cell with Cr-Fe-alloy interconnect, LSM cathode and YSZ electrolyte tested at T = 1173 K and a current density of 200 mA cm-2. The increase of both arcs over the testing time was ascribed to the affect of Cr on the oxygen diffusion processes in the LSM cathode and across the LSMelectrolyte interface and is in line with the interpretations from Jiang[25,33] and Jiang et al.[24,26,32].
1.3.4
Microstructures in degraded SOFC
Cathodic polarization: Taniguchi et al.[23] were the first who reported the occurrence of Cr-Mn-spinel in Crpoisoned SOFC with an LSM cathode by using XRD analysis. This finding was confirmed 24
by Badwal et al.[22] who detected small amounts of Cr-Mn spinel after 100 h of operation of SOFC with a Cr5Fe1Y2O3 interconnect plate, an LSM cathode, a YSZ electrolyte and a Nizirconia cermet anode at T = 1173 and 1273 K under current load. For the same setup large quantities of spinel had formed after 2000 to 2500 h of cell operation, in some of the experiments forming a dense layer of several microns at the cathode-electrolyte interface. In some cases these authors also observed Cr2O3(s) at the cathode-electrolyte interface; unfortunately the specific conditions for its formation were not given in more detail. The amount of spinel at the cathode-electrolyte interface was much larger than within the LSM cathode particularly after a period of current load. Badwal et al.[22] further observed spinel in the contact region between interconnect and cathode. Using a half-cell setup with Inconel 600, LSM cathode and YSZ electrolyte Matsuzaki and Yasuda[31] reported the formation of a dense layer of Cr-deposit at the LSM-YSZ interface after a cell test at T = 1073 K and a current density of 300 mA cm-2 over 100 h of polarization. Zhen et al.[27] observed dense Cr-Mn spinel-deposits exclusively at the LSM-YSZ interface after a half-cell test of an SOFC with Cr-Fe-alloy interconnect operated for 20 h at T = 1173 K and a current density of 200 mA cm-2. Using the same setup, Jiang et al.[25,36] documented spinel formation at the LSM-YSZ interface already after 4 h. The grain size of spinel was about 0.17 m after 4 h of cell testing and increased to about 0.7 m after 50 h. The zone of these large faceted crystals was followed by a zone of very fine grains (about 0.05 m) of Cr2O3 towards the cathode-electrolyte interface. The deposition zone broadened as the polarization time increased from about 60 m after 50 h to 89 m after 129 h. Under the same testing conditions as above, but under anodic polarization very fine grains of Cr2O3 were forming exclusively at the LSM-YSZ interface. Using the same setup as above, no Cr-deposits formed after 50 h of testing under open circuit conditions[25,36]. Very small grains of Cr-deposits formed at T = 1373 K under open circuit conditions, but further details on their spatial distribution and composition were not given. In a half-cell setup with a Cr-Fe-alloy interconnect, an LSM/LSM-YSZ cathode double layer and a YSZ electrolyte chromium-deposits were only found in the LSM layer under open circuit conditions, and the cell-degradation was weak[21]. On the other hand chromiumdeposits were also found in the LSM-YSZ layer and on the surface of the YSZ electrolyte on increasing the current density up to 280 mA cm-2, and the cell degradation was strong[21]. In the same setup without LSM-YSZ functional layer no chromium-deposits were detected without current at T = 1073 K over 393 h. On the other hand Cr-Mn spinel formed already at
25
very low current load of 2 mA cm-2 on the surface of the electrolyte and near the LSMelectrolyte contact. After 160 h of testing at 100 mA cm-2 current load, large gaps between YSZ grains formed, and spinel crystals were found on the surface of a thin Cr2O3-layer adjacent to the electrolyte. Cr2O3 was also found inside YSZ, up to 10 m away from the LSM-YSZ interface after 160 h of testing, and 30 m from the LSM-YSZ contact after 940 h. Cr2O3 completely filled gaps between YSZ grains. Transmission electron microscopy analyses revealed the layered structure of the composites: a 0.5 m thick layer directly adjacent to the YSZ containing mainly Cr2O3 is covered by a spinel layer. Paulson and Birss[34] investigated the microstructures in a half-cell setup with a stainless steel disk containing 15,21 % Cr attached on top of a 4 mm2 square (La0.8Sr0.2)0.98MnO3- cathode that rested on a 144 mm2 square YSZ electrolyte, after the half-cell was tested by a sequence of 8 chronoamperometry experiments at 0.5 V and T = 1073 K. These authors observed the formation of a 500 m broad zone of 8 individual, dense Cr2O3-layers at the edge of LSM on the YSZ surface. Cr-deposits consisting of Cr2O3 and Cr-Mn-spinel were concentrated in an about 2 m broad region at the LSM-YSZ interface. A reference test without polarization did not lead to these features. Anodic polarization: After a half-cell test of an SOFC with Cr-Fe-alloy interconnect operated for 6 h at T = 1173 K and a current density of 200 mA cm-2, Jiang et al.[36] reported the formation of very fine particles of Cr2O3, exclusively covering the YSZ surface, and the deposition was less pronounced at T = 1073 K and 973 K. The morphology of the particles was different than the morphology of the deposits under cathodic polarization. No spinel formation was observed in these experiments. Influence of temperature: Microstructural changes during half-cell tests of a setup consisting of a Cr-Fe-alloy interconnect, a La0.72Sr0.18MnO3- cathode and a YSZ electrolyte were systematically investigated as a function of time and temperature at a current load of 200 mA cm-2 by Jiang et al.[32]. After 5 minutes of testing at T = 1173 K very fine Cr-deposits (< 100 nm) already formed on the YSZ-surface, and the density and size of deposits increased by time. After 20 h spinel formation was observed forming a 40 to 50 m wide band at the LSM-YSZ interface. In direct contact with YSZ about 500 nm large Cr2O3-grains were detected, almost completely
26
covering the YSZ surface, associated with large grains (about 1 m) of spinel. The amount of Cr-deposits was significantly smaller after 20 h of testing and 200 mA cm-2 current load at T = 973 K: isolated fine particles (about 100 to 200 nm) were detected on the YSZ surface. The decreasing degradation at lower temperatures was ascribed to slower diffusion and lower partial pressure of gaseous Cr-species.
1.3.5
Amounts of chromium in SOFC tested with and without current load
Konysheva et al.[21] compared the total amounts of chromium in half-cells with Cr-Fe-alloy interconnects, a La0.65Sr0.3MnO3-/LSM+YSZ double layer cathode and a YSZ electrolyte after tests without current showing very small degradation, and under 200 mA cm-2 current density showing strong degradation as a function of testing time at T = 1073 K. The amounts of chromium in the half-cell operated under a current was 100 g cm-2 after 150 h and 150 g cm-2 after 500 h; this is only 15 to 20 % higher than in the cell operated without current, although the degradation of the polarized cell was remarkably higher. This was explained by the following: only under current load chromium deposits are concentrated in the functional region of LSM close to the contact to YSZ where they inhibit oxygen reduction and diffusion processes. In SOFC with LSM cathodes and a Cr-Fe-steel interconnect that were tested at T=1073 K for 300 h at 0.7 V, Krumpelt et al.[35] measured about 2.8 wt.% of Cr at 10 m distance from the cathode-electrolyte interface. The Cr-content dropped to about 0.5 wt.% at 14 m distance from the contact to the electrolyte. It further decreased slightly towards the interconnectcathode interface. In a half-cell with LSM cathode, a Sm2O3-CeO2 interlayer between cathode and electrolyte and a YSZ electrolyte with a Crofer22APU interconnect operated at T = 1073 K, holding the cell at 0.7 V for 120 h, Simner et al.[29] measured 5 at.% Cr in the Sm2O3-CeO2 interlayer, but no Cr in LSM using energy dispersive scanning electron microscopy. By using molecular dynamics simulation techniques it was stated recently that only 890 ppm Cr3+ in LSM significantly increase the formation energy of O2- vacancy within about 10 around Cr3+. This may lead to a dramatic drop of the oxygen diffusion coefficient in LSM by about 60%[36] and pers. comm.
27
1.3.6
Critical assessment of proposed mechanisms of chromium poisoning
For the mechanisms of chromium poisoning two models have been proposed: 1) Reduction of gaseous Cr-species under polarization[21-23,30-31,33], and 2) Chemical dissociation of Crspecies on the LSM surface[24-26,32,33,37]. Both reduction and chemical dissociation processes reflect non-equilibrium conditions. 1) Several authors[21-23,230,31,33] ascribe the mechanism of chromium poisoning to the reduction of gaseous CrO3(g) and Cr-oxyhydroxides at the cathode-electrolyte interface, described by inverse Eq. 1.2.1. In an LSM cathode the reduction of CrO3(g) is expected to be localized at the triple phase boundary (TPB) between LSM, YSZ, and gas, where the reaction partners for the reduction, electron-donating LSM and oxygen-accepting YSZ are available[38]. This reduction reaction would compete with the oxygen reduction and would lead to blocking of the active sites at the TPB and subsequent formation of Cr-Mn spinel by the reaction of Cr2O3(s) with LSM. Badwal et al.[22] emphasized that chromium poisoning would consist of several processes, including changes of the chemical composition of the LSM surface, reduction of CrO3(g) at the cathode-electrode interface competing with the normal oxygen reduction reaction, and blocking of pores by Cr-Mn spinel and Cr2O3(s). In particular they suggest the tight intercalation between changes of the chemical composition at the surface of LSM particles and the oxygen adsorption and surface diffusion kinetics in the early stage of chromium poisoning. Badwal et al.[22] ascribe a key role for the late stages of cell degradation to the formation of Cr-Mn spinel that would block pores and lead to substantial decrease of the TPB area. Cr-Mn spinel is interpreted by these authors to form in a solid-solid reaction between Cr2O3(s) and LSM that may have the simplified form of Eq. 1.3.1 when solid solubility of Cr in LSM[39] is considered and spinel contains the molar fraction of Cr, X(Cr) = 2X(Mn):
La1 xSrx MnO3 + 3 2 y Cr2 O3(s) La1 xSrx Mn1 y Cry O3 + yMnCr2 O4 + 1 4 y O2(g)
(1.3.1)
Eq. 1.3.2 is a possible reaction for the formation of spinel with a higher amount of Mn, X(Mn)= 2X(Cr):
La1 xSrx MnO3 + 1 2 y Cr2 O3(s) La1 xSrx Mn1 2 y Cry O3 + yMn 2 CrO4 + 5 4 y O 2(g)
(1.3.2)
28
For equal y in Eqs. 1.3.1 and 1.3.2 less Cr2O3 is needed for the formation of Mn2CrO4, more oxygen is produced, and LSM gets more deficient in Mn. Oxygen production stems from the reduction of Mn3+ in perovskite to Mn2+ in spinel. However the formation of spinel can also be interpreted as a direct solid-gas reaction. Fig. 1.3.1 is a simplified illustration of possible reaction paths that lead to the end product Cr-Mn spinel.
Fig. 1.3.1 Possible reaction paths for the spinel formation as a function of Gibbs energy. The true shape of the curves depends on the activation energy Ea and is thus not known. red = reduction, sp-form = spinel formation. Which of the possible reaction paths is realized, depends on the activation energy, Ea of the concerning reaction, and this is not known. The shape of the curves in Fig. 1.3.1 was chosen based on the consideration that the diffusionless reduction of Cr-gas may have a lower activation energy than the solid-solid reaction between LSM and Cr2O3(s), and the mobility of the gas phase is high, thus assuming a lower activation energy for the LSM-Cr-gas reaction. These assumptions would mean that fast reduction of Cr-gas to Cr2O3(s) occurs as one process, and the LSM-Cr-gas reaction occurs as a parallel process leading to the formation of spinel. On the other hand it may last a long time for the Cr2O3(s) that was formed by the reduction reaction to transform into spinel in the solid-solid reaction with LSM. As it is not assured if spinel in fact forms in a solid-solid reaction, reactions of direct formation of spinel by the interaction between Cr-gas and LSM can be formulated (Eqs. 1.3.3 to 1.3.5) considering the main chromium molecules that interact with LSM for spinel with X(Cr) = 2X(Mn). Such gas-solid reaction can be split into two reaction steps: formation of
29
Cr2O3(s) from the gas and subsequent spinel formation from Cr2O3 + LSM. The differences of oxygen contributions to respective reactions stem from the reaction step of Cr2O3(s) formation:
La1 xSrx MnO3 + 3 yCrO3(g) La1 xSrx Mn1 y Cry O3 + yMnCr2 O4 + 5 2 y O2(g)
(1.3.3)
La1 xSrx MnO3 + 3 yCrO2 (OH)2(g) La1 xSrx Mn1 y Cry O3 + yMnCr2O4 + 3 yH 2O(g) + 5 2 y O2(g)
(1.3.4)
La1 xSrx MnO3 + 3 yCrO2 (OH)(g) La1 xSrx Mn1 y Cry O3 + yMnCr2 O4 + 3 2 y H 2 O(g) + 3 4 y O2(g)
(1.3.5)
2) The chemical dissociation of gaseous Cr-species on the LSM surface for the cell degradation was proposed as the key process for the degradation of SOFC caused by chromium by another research group[24-27,32,33,37]: Mn2+ on the surface of LSM at reduced oxygen partial pressure close to the cathode-electrolyte interface would react with gaseous Crspecies to Cr-Mn-O nuclei, and consequently to Cr-Mn spinel and Cr2O3(s). As Mn2+ is associated to vacancy formation in LSM that is necessary for the oxygen diffusion,[24-27,32,33,37] oxygen diffusion is inhibited by the nuclei-formation. It can be seen from Eqs. 1.3.1 to 1.3.5 that oxygen is produced during the formation of spinel. Thus the pO2 at the locations of the spinel formation is expected to increase. This in turn will also lead to less Mn2+ in LSM[40] and consequently lower oxygen diffusion in LSM. The role of the oxygen vacancy diffusion mechanism in an LSM cathode has been considered controversially: Mogensen and Skaarup[41] concluded from the low oxygen self-diffusion coefficients of the order of 410-14 cm2 s-1 at T = 1173 K[42] that long range bulk migration of oxygen ions cannot play a significant role for the cathode performance. However they did not discuss the dependence of oxygen diffusion upon pO2 . Huang et al.[43] confirmed these early suggestions by evaluating the ionic conductivity of LSM from pure oxygen to pO2 = 300 Pa at temperatures from 953 K to 1153 K using YSZ as blocking electrode. The ionic conductivity was lower at lower oxygen partial pressures, opposite to the trend that would be expected under the control of the vacancy diffusion mechanism. On the other hand the measured oxygen tracer diffusion coefficient in LSM strongly increases when the oxygen partial pressure is decreased from pure oxygen to pO2 = 200 Pa[44]. Yasuda et al.[44] concluded that 30
oxygen ions in the bulk of LSM diffuse by the vacancy diffusion mechanism. The activation energy for the diffusion of oxygen for LSM is in the range of 250 to 300 kJ mol-1. This is close to 270 kJ mol-1 for La0.9Sr0.1CoO3- in which oxygen ions are transported by the vacancy mechanism. This indicates that a vacancy diffusion mechanism also applies to LSM[45]. In an investigation of active sites for the oxygen reduction at the O2/LSM/YSZ interface[46] for three different overvoltages of cathode polarization ( = 0.336 V, = 0.185 V, and = 0.090 V) using isotopic oxygen exchange and secondary ion mass spectrometry it was found that oxygen ions can only diffuse through dense LSM at the high overvoltage of = 0.336 V corresponding to pO2 = 10-4 Pa[46]. The calculated amount of oxygen vacancies () in La0.8Sr0.2MnO3- at 973 K and pO2 = 10-4 Pa is = 2.4x10-6, compared to = 3.94x10-9 in air[23]. This confirms the suggestion that the formation of oxygen vacancies in LSM contributes to the oxygen diffusion at high current loads[22]. Based on the findings from the literature it can be summarized that in LSM oxygen diffuses through grain boundaries at high
pO2 , as oxygen vacancies are simply not available under these conditions. We believe that the
oxygen vacancy diffusion mechanism contributes to the oxygen diffusion under high current loads, when the oxygen partial pressure at the cathode-electrolyte interface is decreased significantly, as it was directly proven by isotopic and tracer diffusion[44] experiments. Contradictory interpretations[43] from the dependence of the ionic conductivity on pO2 need to be judged with care due to the difficulty of controlling the numerous factors that can influence the results of the blocking electrode method used. The electrochemical reduction of CrO3(g) was rejected by the authors favoring the chemical dissociation approach[24-27,32,33,37]. It is necessary to test the arguments for this claim of exclusive validity: a strong tendency exists for CrO3(g) to get reduced to Cr2O3(s) at the TPB, as G of the reduction being the inversion of reaction Eq. 1.2.1 has a large negative value. It was also mentioned in the early paper of Caplan and Cohen[10] that substantial precipitation of Cr2O3(s) from CrO3(g) occurred in the cooler part of the experimental setup. Reduction of CrO3(g) to Cr2O3(s) was such predominant as to make sampling of gaseous CrO3(g) difficult. This strong tendency for the precipitation of Cr2O3(s) makes a rejection of the reduction of CrO3(g) as a possible process contributing to the cell degradation doubtful. Paulson and Birss[34], as well as Konysheva et al.[21] observed the extension of dense Cr2O3layers into YSZ. This phenomenon was well explained by continuous feeding of an initial
31
Cr2O3-layer with CrO3(g), the latter becoming reduced at the new TPB consisting of YSZ and electron-donating Cr2O3(s)[21,34], whereas an explanation by the chemical dissociation approach is not satisfying. Thus it is obvious that the explanation of the poisoning process by the chemical dissociation approach alone is not without doubt. Some indications for two independent chromium poisoning mechanisms can be found in the work from Jiang et al.[24,26,37]: the two phases formed in the scope of a polarized LSM cathode exhibit distinctive microstructures: spinel forms large grains, whereas the phase that was most likely identified as Cr2O3 occurs in fine-grained, partly layered structures. The region of spinel formation extends several microns from the TPB into the cathode, whereas Cr2O3 is always located directly at the cathode-electrolyte interface. From the occurrence of finegrained Cr2O3 the existence of a large number of nuclei for its formation is concluded, which does not seem to be the case for the spinel phase. Furthermore, from impedance spectra analyses it was in fact possible to distinguish two distinctive depositions of Cr-species, one with a lower rate on the LSM surface, and the second with a higher rate on the YSZ electrolyte surface. Also two different diffusion processes were distinguished, which both seemed to be inhibited by chromium poisoning. If CrO3(g) is electrochemically reduced to Cr2O3(s) in a cell, Cr2O3(s) deposition should also occur under open-circuit conditions, which was definitively not observed. In this case the contribution of reduction to the Cr-poisoning has to be rejected, and the degradation can be associated to the dissociation process[24-27,32,33,37] and subsequent formation of spinel. However the situation changes if the reduction of CrO3(g) is under the main control of the oxygen partial pressure gradient towards the cathode-electrolyte interface, which is increasing as a function of increasing polarization. In this case no chromium will be deposited at the cathode-electrolyte interface under open-circuit conditions, whereas in a polarized cathode the reduction of CrO3(g) takes place and competes with the oxygen reduction leading to Cr2O3(s) deposition. This explanation is in line with the microstructural features of tested cells both under open circuit voltage and under current load, and it is also in line with the observed temporary reversibility of the cell deterioration[22,23,28,30,31]: by switching off the polarization the competing reduction of CrO3(g) no longer occurs, and the normal charge transfer can take place by switching it on again. But how can one explain the strictly localized deposition of Cr2O3 that also occurs under anodic polarization[37]? Under oxidizing conditions little Mn2+ is expected to be present in LSM[40], thus the formation of nuclei by the proposed LSM-Cr interaction wont occur. This is in contrast with the complicated mechanism for the formation of Cr2O3(s) under anodic 32
polarization established by Jiang et al.[37] that includes diffusion of Mn3+/Mn2+ driven by the oxygen evolution reaction at the cathode/electrolyte interface, again with Mn2+ acting as agent for the formation of Cr-Mn-O nuclei, the number of the latter being less than under cathodic polarization, and thus lack of spinel formation. Alternatively, the following simple explanation for strictly localized Cr2O3(s) formation under anodic polarization can be given: in an LSM cathode the reduction of gaseous Cr-species is expected to be localized at the triple phase boundary, where the reaction partners for the reduction, electron-donating LSM and oxygen-accepting yttrium-stabilized zirconia (YSZ) are available[38]. Oxygen deficiency is negligible in LSM under high pO2 [40], and thus under these conditions LSM has no tendency at all to accept oxygen, contrary to the situation of a strong pO2 gradient under cathodic polarization. This is a simple and consistent explanation for a strict localization of Cr2O3(s) formed by reduction of gaseous Cr-species even under anodic conditions. In cell tests of a polarized platinum electrode using a Cr-containing interconnect no Cr was observed, contrary to an LSM electrode. This different behavior of Pt and LSM electrodes under Cr-poisoning was used as an evidence for the exclusive validity of the dissociation approach, based on an early finding that LSM behaves like a metallic electrode at low polarization potentials[47] that was not quantified. However this conclusion was not tested in the light of the oxygen partial pressure gradient towards the electrode-electrolyte interface: contrary to platinum, vacancies are expected to form in LSM under increasing polarization, and in LSM a pO2 gradient is expected under polarization, which is indeed not the case in a platinum cathode. This once again may favor the reduction of CrO3(g) and gaseous Croxyhydroxide resulting in Cr2O3(s) deposition at the cathode-electrolyte interface in LSM, opposite to the situation with a platinum cathode. It was further mentioned that the existence of Cr-containing products away from the TPB would be in disagreement with the reduction approach. This is indeed true for the case of CrO3(g) and Cr-oxyhydroxide reduction being the only Cr-poisoning mechanism. However, if both the chemical dissociation as well as the reduction of gaseous Cr-species is occurring with different proportions, this apparent antagonism is abolished. In this context experimental results of a half-cell test with Cr-Fe-alloy (RA446) interconnect, LSM cathode and a YSZ electrolyte at T = 1173 K and a current density of 200 mA cm-2 from Zhen et al.[27] are particularly interesting: the slope of the cathode polarization curve (Fig. 4 b[27]) as a function of time reveals an inflection point after about 6 1/2 h. This is an indication against one unique poisoning mechanism, but several processes may lead to the deterioration of cell
33
performance, and their respective influence on the cell deterioration may vary as a function of time. From the considerations in this chapter we conclude that no sustainable arguments exist for the rejection of the reduction of gaseous Cr-species as one of the controlling mechanisms of Cr-poisoning of SOFC. The important role of decreased oxygen activity at the LSM-YSZ interface under current load for the cell degradation was already suggested by Taniguchi et al.[23]. Konysheva et al.[30] give the following explanation, why the strong oxygen partial pressure gradient in the LSM cathode under high current densities plays a key role for the degradation: the LSM cathode has a low electrochemically active area (TPB) near the interface with the electrolyte only. Under polarization, the oxygen ions formed at this interface are transported from the cathodeelectrolyte interface through the electrolyte. This results in a lower oxygen partial pressure at the interface as compared to that in air. The higher the current density under SOFC operation, the lower is the oxygen partial pressure at the contact between LSM and YSZ. The deposition of chromium followed by its reduction near this interface blocks direct oxygen access to the electrochemically active sites, thereby still more decreasing the oxygen partial pressure at a newly formed Cr2O3(s)/electrolyte interface[30]: the TPB between LSM and YSZ diminishes more and more by the blocking of Cr2O3(s), and oxygen cannot access the TPB. As Cr2O3(s) has a small electronic conductivity of 0.8 S m-1 at T = 1282 K[48] and pO2 = 1 Pa, the oxygen ions from this new, weak catalytic reaction diffuse into YSZ, and the chemical activity of the cell is furthermore deteriorating due to the lack of oxygen supply through the rather dense Cr2O3 layer to the new TPB. The temporary reversibility of the deterioration by switching the cell off and on again can also be explained: in contrast to current load operation, under open circuit the LSM-Cr interactions occur randomly throughout the cathode, thus the remaining TPB/YSZ active sites are almost unaffected under open current circuit. The small area close to the new TPB that was strongly depleted of oxygen under current load is filled with air leaking through remaining pores between LSM and Cr2O3. Applying a current load, the LSMCr interaction is again favored in the region close to YSZ as pO2 decreases at the TPB, even though the decrease is expected to be less due to less LSM/YSZ active sites caused by the first degradation. Oxygen is mainly reduced at the new TPB between Cr2O3(s) and YSZ. Reduction already takes place at higher pO2 at the beginning of the current load operation, as electronic conductivity of Cr2O3 is significantly higher at higher pO2 (1.8 S m-1 in air[48]). 34
Oxygen ions diffuse into YSZ, but new oxygen is not supplied to the new TPB. Active LSM/YSZ sites further diminish by ongoing formation of spinel and Cr2O3 deposits, and the degradation increases as a function of time. Fig. 1.3.2 is a visualisation of the microstructural consequences of chromium in an LSM cathode. The reported dependence of structural features of the degraded cell on the operation temperature, current load, and chromium content is schematized in the picture.
Fig. 1.3.2 Model of chromium poisoning of an SOFC with Cr-interconnect and LSM cathode based on the findings in the literature. Numbers refer to locations of processes that are decisive for the degradation Number 1 in Fig. 1.3.2 denotes the interconnect-cathode interface region where oxidation of Cr2O3(s) to gaseous Cr-oxides and Cr-oxyhydroxides by Eqs. 1.2.1 to 1.2.3 occurs, followed by diffusion of the gaseous products into the cathode. Number 2 denotes the region of interactions between LSM and chromium leading to spinel formation by solid-solid reaction, Eqs. 1.3.1 to 1.3.2, or gas-solid reaction, Eqs. 1.3.3 to 1.3.5, and number 3 denotes the reduction of gaseous Cr-species by the reverse of Eq. 1.2.1 leading to the redeposition of Cr2O3(s) at the cathode-electrolyte interface.
35
1.4
Proposed strategies against chromium poisoning and their effectiveness
1.4.1
Increasing the Cr-tolerance of conventional SOFC with Cr-interconnects and LSM cathodes
More than ten years ago Badwal et al.[22] proposed that coating of the Cr-interconnect with a protective electrically conductive dense layer would be an effective strategy against the diffusion of Cr-species into the cathode. Several promising materials for coating applications were developed in the following years that act as chromium diffusion barrier and hinder growth of chromia scale at the alloy surface, thereby improving the electrical conductivity of the interconnect-cathode interface[49-69]. However, so far volatilization could not be suppressed completely. Application of the following coatings upon the interconnect has been shown to considerably reduce the diffusion of chromium into the cathode thus decreasing the cell degradation: Electroplated metallic Co[49,50], Co-Mn, or Cu-Mn[51], sputtered Co, Ni, or Cu[52], Mn, La, or Mn2CrO4[52], Co3O4[53], MnCo2O4[54-61], Cu1.4Mn1.6O4[53,54], La0.65Sr0.3MnO3[58,65], Ce0.05Mn1.475Co1.475O4[62], La0.85Sr0.15MnO3-[50], (La,Sr)CoO3[63], La0.67Sr0.33MnO3[64],
La0.6Sr0.4Co0.8Fe0.2O3[58], La0.8Sr0.2Mn0.5Co0.5O3-[66], La0.8Sr0.2Mn0.5Co0.5O3[65], La0.8Sr0.2FeO3-[67], two-segment Cr-Al-Y-O nanocomposite and (Mn,Co)3O4[68], as well as (Ti,Al)N[69]. However, as Cr in the ppm range significantly influences the oxygen diffusion in the LSM cathode[36], coating alone does not solve the problems associated to chromium poisoning completely, but a combination of the quoted strategies is advisable to further improve the long-time stability of SOFC performance. The formation of a dense electrically isolating Cr2O3 layer is probably preventable by using electrolyte materials or a functional layer between LSM cathode and YSZ electrolyte that can incorporate Cr in solid solution without affecting the electrical conductivity. Furthermore such a buffer layer may act as a sink for CrO3(g) thus diminishing nuclei formation on LSM. If the buffer layer contains an ionic conductor, more active sites for the oxygen reduction will result in a higher Cr-tolerance. This was recently shown for a cell with a YSZ-LSM functional layer: a functional LSM-YSZ layer adjacent to the YSZ electrolyte led to a lower cell degradation[30]: increasing the ionic conductivity of the LSM cathode that is predominantly electronically conducting down to pO2 = 10-7 Pa[70] by admixture of YSZ results in an 36
expanded area of active sites for the oxygen reduction away from the TPB. Thus the number of active sites is increased, and the cell is more tolerant against chromium[30]. Besides, the reduction of gaseous chromium will not be restricted to the small area at the TPB due to a smaller oxygen partial pressure gradient, leading to the formation of more scattered reduction products instead of a dense layer: thus the block of oxygen diffusion into the electrolyte can be avoided. The ionic conductivity can be increased by doping the B-site of ABO3 perovskite with reducible cations. Ideally the selected dopants decrease the mobility of Mn2+ and thus prevent the formation of nuclei for the adsorption of CrO3(g) without influencing the formation of vacancies.
1.4.2
New ways alternative materials
Badwal et al.[22] already considered alternative cathode materials to reduce or stop the formation of the spinel phase. Matsuzaki and Yasuda[31] concluded from insignificant Crdeposits in tested SOFC with Cr-Fe-alloy interconnect, La0.6Sr0.4Co0.2Fe0.8O3- (LSCF) cathode and Ce0.8Sm0.2O1.9 electrolyte that the ratio of the reduction of gaseous CrO2(OH)2(g) to that of O2(g) at the electrode/electrolyte interface is controlled by the electrochemical properties of the interface. Based on these findings they predicted that highly Cr-tolerant cathodes can be developed. In recent time it was found that new cathode materials such as La1-xSrxCo1-yFeyO3- (LSCF), La(Ni,Fe)O3- (LNF), and (La,Ba)(Co,Fe)O3- (LBCF) are more tolerant against chromium poisoning. LNF and LBCF revealed extraordinary high tolerance against chromium poisoning. The highest tolerance against the effects of chromium under SOFC operating conditions combined with high electrical conductivity has been reported recently for (La,Ni)FeO3- [71,72], which makes this material a promising candidate for a steady long-term SOFC performance. All these perovskites are mixed electronic-ionic conductors; particularly LSCF and LNF show rather high ionic contributions to the total electrical conductivity. Effects of Cr upon the degradation of La1-xSrxCo1-yFeyO3- (LSCF)[26,71,73], La(Ni,Fe)O3- (LNF)[71,74], and (La,Ba)(Co,Fe)O3- (LBCF)[71,75] cathodes were investigated using impedance spectroscopy. As for LSM these authors concluded that the mechanism of Cr poisoning can be explained by chemical dissociation of CrO3(g) to the perovskite-structured materials and nuclei formation in the cases of LSCF and LNF, whereas no proper nuclei were reported for LBCF. In all these cathodes Cr-deposition was observed throughout the cathode both under polarization and without polarization, contrary to LSM. The amount of Cr-
37
poisoning of LSCF was considerable. The amount of deposited Cr in LSCF was even larger without polarization than under polarization, which was explained by a removing effect of nuclei for the chromium deposition under polarization conditions. As an alternative to the complicated nuclei mechanism, the following considerations can be made using the reduction model: For the reduction reaction of CrO3(g) the presence of both an electron donor and oxygen ion acceptor is necessary, and a typical mixed ionic-electronic conductor such as LSCF can take over both functions. Thus reduction of CrO3(g) takes place inside the whole cathode even without being promoted electrochemically by polarization of the cell. However under strong polarization one can expect that LSCF gets more and more ionic conducting towards the electrode-electrolyte interface, that is towards lower oxygen partial pressures, most likely resulting in retarding or inhibiting of the reduction reaction. Even if the reduction reaction is considered to be the dominant mechanism of chromium poisoning, nuclei might form in addition, but their influence on the Cr deposition compared to the reduction of CrO3(g) cannot be decided yet. Opposite to the case of LSM no driving force for CrO3(g) to migrate to the triple phase boundary exists due to the mixed ionic-electronic conducting behaviour of the regarding cathodes. The higher the contribution of the ionic conduction the less complete reduction is expected due to prolonged lack of an electron donator. Improved inhibition of the reduction of CrO3(g) is predicted for LNF, as this phase has a particularly high ionic conductivity. In recent years research activities for LaCrO3-base ceramic interconnector materials were revitalized by several groups[76-78] to circumvent the problems of chromium poisoning. However, despite rapidly developing processing techniques it is not clear at the moment if the obstacles of sinterability and low mechanical strength as well as difficult manufacturing correlated with high costs can be coped.
References 1. J.W. Fergus, Lanthanum chromite-based materials for solid oxide fuel cell interconnects, Solid State Ionics, 2004, 171, pp. 1-15. 2. H. Tu, U. Stimming, Advances, aging mechanisms and lifetime in solid-oxide fuel cells, J. Power Sources, 2004, 127, pp. 284-93. 3. J.W. Fergus, Metallic interconnects for solid oxide fuel cells, Materials Science and Engineering A, 2005, 397, pp. 271-83. 38
4. J.E. Hammer, The oxidation of materials for interconnects in solid oxide fuel cells, Thesis, University of Pittsburgh, 2005, 108 pp. 5. Z. Yang, Recent advances in metallic interconnects for solid oxide fuel cells, Int. J. Mater. Res., 2008, 53, pp. 284-93, pp. 39-54. 6. G. Bandel, Special high-temperature corrosion features of heat-resistant steels and alloys, Arch. Eisenhttenwesen, 1941, 15, pp. 271-284 (in German). 7. D. Caplan and M. Cohen, High temperature oxidation of some iron-chromium alloys, J. Metals, 1952, 4, pp. 1057-65. 8. I. Warshaw and M.L. Keith, Solid solution and chromium oxide loss in part of the system MgO-Al2O3-Cr2O3-SiO2, J. Am. Ceram. Soc., 1954, 37, pp. 161-8. 9. G.R. Wilms and T.W. Rea, Loss of metal from chromium at elevated temperatures in air, J. Less Common Metals, 1959, 1, pp. 411-12. 10. D. Caplan and M. Cohen, The volatilization of chromium oxide, J. Electrochem. Soc., 1961, 108, pp. 438-442. 11. E. Povoden, A.N. Grundy, L.J. Gauckler, Thermodynamic reassessment of the Cr-O system in the framework of solid oxide fuel cell (SOFC) research, J. Phase Equilib. Diff., 2006, 27, pp. 353-62. 12. A.T. Dinsdale, SGTE data for pure elements, Calphad, 1991, 15(4), pp. 317-425. 13. B.B. Ebbinghaus, Thermodynamics of gas phase chromium species: the chromium oxides, the chromium oxyhydroxides, and volatility calculations in waste incineration processes, Combust. Flame, 1993, 93, pp. 119-37. 14. Y.-W. Kim and G.R. Belton, Thermodynamics of volatilization of chromic oxide. 1. Species CrO3 and CrO2(OH), Met. Trans., 1974, 5, pp. 1811-16. 15. E.M. Bulewicz and P.J. Padley, Photometric investigations of behaviour of chromium additives in premixed H2+O2+N2 flames, Proc. R. Soc. Lond. A, 1971, 323, pp. 377-400. 16. O. Glemser and A. Mller, Gaseous hydroxide. 8. On a gaseous hydroxide of chromium, Z. Anorg. Allgem. Chem., 1964, 334, pp. 151-4 (in German). 17. M. Farber and R.D. Srivastava, Mass spectrometric investigation of reactions involving vanadium and chromium with potassium-seeded H2-O2 flames, Combust. Flame, 1973, 20, pp. 43-9. 18. E.J. Opila, D.L. Myers, N.S. Jacobson, I.M.B. Nielsen, D.F. Johnson, J.K. Olminsky, M.D. Allendorf, Theroretical and experimental investigation of the thermochemistry of CrO2(OH)2(g), J. Phys. Chem. A, 2007, 111, pp. 1971-80.
39
19. K. Hilpert, D. Das, M. Miller, D.-H. Peck, R. Wei, Chromium vapor species over solid oxide fuel cell interconnect materials and their potential for degradation processes, J. Electrochem. Soc., 1996, 143, pp. 3642-7. 20. C. Gindorf, L. Singheiser, K. Hilpert, Chromium vaporisation from Fe,Cr base alloys used as interconnect in fuel cells, Steel Res., 2001, 72, pp. 528-33. 21. E. Konysheva, E. Penkalla, E. Wessel, J. Mertens, U. Seeling, L. Singheiser, K. Hilpert, Chromium poisoning of perovskite cathodes by the ODS alloy Cr5Fe1Y2O3 and the high chromium ferritic steel Crofer22APU, J. Electrochem. Soc., 2006, 153, pp. A765-73. 22. S.P.S. Badwal, R. Deller, K. Foger, Y. Ramprakash, J.P. Zhang, Interaction between chromia forming alloy interconnects and air electrode of solid oxide fuel cells, Solid State Ionics, 1997, 99, pp. 297-310. 23. S. Taniguchi, M. Kadowaki, H. Kawamura, T. Yasuo, Y. Akiyama, Y. Miyake, T. Saitoh, Degradation phenomena in the cathode of a solid oxide fuel cell with an alloy separator, J. Power Sources 55 (1995) 73-79. 24. S.P. Jiang, J.P. Zhang, L. Apateanu, K. Foger, Deposition of chromium species at Srdoped LaMnO3 electrodes in solid oxide fuel cells. II. Effect on O2 reduction reaction, J. Electrochem. Soc., 2000, 147(9), pp. 3195-3205. 25. S.P. Jiang, Use of gaseous Cr species to diagnose surface and bulk process for O2 reduction in solid oxide fuel cells, J. Applied Electrochem., 2001, 31, pp. 181-192. 26. S.P. Jiang, J.P. Zhang, X.G. Zheng, A comparative investigation of chromium deposition at air electrodes of solid oxide fuel cells, J. Europ. Ceram. Soc., 2002, 22, pp. 361-73. 27. Y.D. Zhen, J. Li, S.P. Jiang, Oxygen reduction on strontium-doped LaMnO3 cathodes in the absence and presence of an iron-chromium alloy interconnect, J. Power Sources, 2006, 162, pp. 1043-1052. 28. Y. Matsuzaki, I. Yasuda, Electrochemical properties of a SOFC cathode in contact with a chromium-containing alloy separator, Solid State Ionics, 2000, 132, pp. 271-278. 29. S. Simner, M. Anderson, J. Stevenson, SOFC cathode materials development at PNNL, FY 2004 Annual Report, III.A.13, Office of Fossil Energy Fuel Cell Program, 2004. pp. 93-6. 30. E. Konysheva, J. Mertens, H. Penkalla, L. Singheiser, K. Hilpert, Chromium poisoning of the porous composite cathode. Effect of cathode thickness and current density, J. Electrochem. Soc., 2007, 154 (12), pp. B1252-B64.
40
31. Y. Matsuzaki, I. Yasuda, Dependence of SOFC cathode degradation by chromiumcontaining alloy on compositions of electrodes and electrolytes, J. Electrochem. Soc., 2001, 148, pp. A126-31. 32. S.P. Jiang, S. Zhang, Y.D. Zhen, Early interaction between Fe-Cr alloy metallic interconnect and Sr-doped LaMnO3 cathodes of solid oxide fuel cells, J. Mater. Res., 2005, 20, pp. 747-58. 33. S.P. Jiang, A comparison of O2 reduction reactions on porous (La,Sr)MnO3 and (La,Sr)(Co,Fe)O3 electrodes, Solid State Ionics, 2002, 146, pp. 1-22. 34. S.C. Paulson, V.I. Birss, Chromium poisoning of LSM-YSZ SOFC cathodes, J. Electrochem. Soc., 2004, 151(11), pp. A1961-68. 35. M. Krumpelt, T. Kaun, T.A. Cruse, M.C. Hash, SOFC research and development, FY 2004 Annual Report, III.A.1, Office of Fossil Energy Fuel Cell Program, 2004. pp. 39-43. 36. J. Zheng and P. Wu, 4th international symposium on solid oxide fuel cells 37. S.P. Jiang, J.P. Zhang, L. Apateanu, K. Foger, Deposition of chromium species at Srdoped LaMnO3 electrodes in solid oxide fuel cells. I. Mechanisms and kinetics, J. Electrochem. Soc., 2000, 147, pp. 4013-22. 38. J.W. Fergus, Effect of cathode and electrolyte transport properties on chromium poisoning in solid oxide fuel cells, Int. J. Hydrogen Energy, 2007, 32, pp. 3664-71. 39. E.A. Filonova, A.N. Demina, A.N. Petrov, Phase equilibria in the system LaMnO3SrMnO3-SrCrO4-LaCrO3, Russ. J. Inorg. Chem., 2007, 52, pp. 771-74. 40. A.N. Grundy, B. Hallstedt, L.J. Gauckler, Assessment of the La-Sr-Mn-O system, Calphad, 2004, 28, pp. 191-201. 41. M. Mogensen, S. Skaarup, Kinetic and geometric aspects of solid oxide fuel cell electrodes, Solid State Ionics, 1996, 86-88, pp. 1151-60. 42. S. Carter, A. Selcuk, R.J. Chater, J. Kajda, J.A. Kilner, B.C.H. Steele, Oxygen-transport in selected nonstoichiometric perovskite-structure oxides, Solid State Ionics, 1992, 53, pp. 597-605. 43. W. Huang, S. Gopalan, U. Pal, Measurement of partial oxygen ion conductivity of Srdoped lanthanum manganite, J. Power Sources, 2007, 173, pp. 887-90. 44. I. Yasuda, K. Ogasawara, M. Hishinuma, T. Kawada, M. Dokiya, Oxygen tracer diffusion coefficient of (La,Sr)MnO3, Solid State Ionics, 1996, 86-88, pp. 1197-1201. 45. S.P. Jiang, Development of lanthanum strontium manganite perovskite cathode materials of solid oxide fuel cells: a review, J. Mater. Sci., 2008, 43, pp. 6799-6833.
41
46. T. Horita, K. Yamaji, N. Sakai, H. Yokokawa, T. Kawada, T. Kato, Oxygen reduction sites and diffusion paths at La0.9Sr0.1MnO3-x/yttria-stabilized zirconia interface for different cathodic overvoltages by secondary-ion mass spectrometry, Solid State Ionics, 2000, 127, pp. 55-65. 47. A. Hammouche, E. Siebert, A. Hammou, M. Kleitz, A. Caneiro, Electrocatalytic properties and nonstoichiometry of the high-temperature air electrode La1-xSrxMnO3, J. Electrochem. Soc., 1991, 138, pp. 1212-16. 48. T. Matsui, K. Naito, Electrical conductivity anomaly of nonstoichiometric chromium sesquioxide, J. Nucl. Mater., 1984, 120, pp. 115-118. 49. X. Deng, P. Wei, M.R. Bateni, A. Petric, Cobalt plating of high temperature stainless steel interconnects, J. Power Sources, 2006, 160, pp. 1225-29. 50. T.J. Armstrong, A.V. Virkar, M. Smith, M. Homel, FY 2004 Annual Report, Office of Fossil Energy Fuel Cell Program, pp. 289-93. 51. P. Wei, X. Deng, M.R. Bateni, A. Petric, Oxidation and electrical conductivity behavior of spinel coatings for metallic interconnects of solid oxide fuel cells, Corrosion 63 (2007) pp. 529-36. 52. M. Stanislowski, J. Froitzheim, L. Niewolak, W.J. Quadakkers, K. Hilpert, T. Markus, L. Singheiser, Reduction of chromium vaporization from SOFC interconnectors by highly effective coatings, J. Power Sources, 2007, 164, pp. 578-89. 53. M. Bertoldi, T. Zanonella, D. Montinaro, V.M. Sglavo, A. Fossati, A. Lavacchi, C. Giolli, U. Bardi, Protective coatings of metallic interconnects for IT-SOFC application, J. Fuel Cell Sci. Tech., 2008, 5, pp. 011001-1 011001-5 54. M.R. Bateni, P. Wei, X. Deng, A. Petric, Spinel coatings for UNS 430 stainless steel interconnects, Surf. Coat. Tech., 2007, 201, pp. 4677-84. 55. X. Chen, P.Y. Hou, C.P. Jacobson, S.J. Visco, L.C. De Jonghe, Protective coatings on stainless steel interconnect for SOFCs: oxidation kinetics and electrical properties, Solid State Ionics, 2005, 176, pp. 425-33. 56. C. Collins, J. Lucas, T.L. Buchanan, M. Kopczyk, A. Kayani, P.E. Gannon, M.C. Deibert, R.J. Smith, D.-S. Choi, V.I. Gorokhovsky, Chromium volatility of coated and uncoated steel interconnects for SOFCs, Surf. Coat. Tech., 2006, 201, pp. 4467-70. 57. M.-J. Garcia-Vargas, M. Zahid, F. Tietz, T. Kiefer, P. Stevens, R. Vaen, Protection of SOFC interconnects against Cr-evaporation using atmospheric plasma-sprayed spinel layers, 7th European SOFC Forum, Session B08, File No. B081.
42
58. H. Kurokawa, C. P. Jacobson, L. C. Dejonghe, S. J. Visco, Chromium vaporization of bare and of coated iron-chromium alloys at 1073 K, Solid State Ionics, 2007, 178, pp. 287-96. 59. Y. Larring, T. Norby, Spinel and perovskite functional layers between Plansee metallic interconnect (Cr-5 wt.% Fe-1 wt.% Y2O3) and ceramic (La0.85Sr0.15)0.91MnO3 cathode materials for solid oxide fuel cells, J. Electrochem. Soc., 2000, 147, pp. 3251-56. 60. Z. Yang, G.-G. Xia, G.D. Maupin, J.W. Stevenson, Conductive protection layers on oxidation resistant alloys for SOFC interconnect applications, Surf. Coat. Tech., 2006, 201, pp. 4476-83. 61. Z. Yang, G.-G. Xia, X.-H. Li, J.W. Stevenson, (Mn,Co)3O4 spinel coatings on ferritic stainless steels for SOFC interconnect applications, Int. J. Hydrogen Energ., 2007, 32, pp. 3648-54. 62. Z. Yang, G.-G. Xia, Z. Nie, J. Templeton, J.W. Stevenson, Ce-modified (Mn,Co)3O4 spinel coatings on ferritic stainless steels for SOFC interconnect applications, Electrochem. Solid St., 2008, 11, pp. B140-43. 63. K. Fujita, K. Ogasawara, Y. Matsuzaki, T. Sakurai, Prevention of SOFC cathode degradation in contact with Cr-containing alloy, J. Power Sources, 2004, 131, pp. 261-69. 64. C.-L. Chu, J.-Y. Wang, S. Lee, Effects of La0.67Sr0.33MnO3 protective coating on SOFC interconnect by plasma-sputtering, Int. J. Hydrogen Energ., 2008, 33, pp. 2536-46. 65. E. Konysheva, J. Laatsch, E. Wessel, F. Tietz, N. Christiansen, L. Singheiser, K. Hilpert, Influence of different perovskite interlayers on the electrical conductivity between La0.65Sr0.3MnO3 and Fe/Cr-based steels, Solid State Ionics, 2006, 177, pp. 923-30. 66. N. Dekker, B. Rietveld, J. Laatsch, F. Tietz, Evaluation of interconnect alloys and cathode contact coatings for SOFC stacks, Proceedings of the 6th European Solid oxide Fuel Cell Forum, p. 319, Lucerne/Switzerland, 2004. 67. C.J. Fu, K.N. Sun, N.Q. Zhang, X.B. Chen, D.R. Zhou, Evaluation of lanthanum ferrite coated interconnect for intermediate temperature solid oxide fuel cells, Thin Solid Films, 2008, 516, pp. 1857-63. 68. P.E. Gannon, V.I. Gorokhovsky, M.C. Deibert, R.J. Smith, A. Kayani, P.T. White, S. Sofie, Z. Yang, D. McCready, S. Visco, C. Jacobson, H. Kurokawa, Enabling inexpensive metallic alloys as SOFC interconnects: An investigation into hybrid coating technologies to deposit nanocomposite functional coatings on ferritic stainless steels, Int. J. Hydrogen Energ., 2007, 32, pp. 3672-81.
43
69. X. Liu, C. Johnson, C. Li, J. Xu, C. Cross, Developing TiAlN coatings for intermediate temperature solid oxide fuel cell interconnect applications, Int. J. Hydrogen Energ., 2008, 33, pp. 189-96. 70. V.V. Kharton, E.V. Tsipis, I.P. Marozau, A.P. Viskup, J.R. Frade, J.T.S. Irvine, Mixed conductivity and electrochemical behavior of (La0.75Sr0.25)0.95Cr0.5Mn0.5O3-, Solid State Ionics, 2007, 178, pp. 101-13. 71. Y.D. Zhen, A.I.Y. Tok, F.Y.C. Boey, S.P. Jiang, Development of Cr-tolerant cathodes for solid oxide fuel cells, Electrochem. Solid St., 2008, 11, pp. B42-6. 72. T. Komatsu, H. Arai, R. Chiba, K. Nozawa, M. Arakawa, K. Sato, Cr Poisoning suppression in solid oxide fuel cells using LaNi(Fe)O3 electrodes, Electrochem. SolidState Lett., 2006, 9, pp. A9-12. 73. S.P. Jiang, S. Zhang, Y.D. Zhen, Deposition of Cr species at (La,Sr)(Co,Fe)O3 cathodes of solid oxide fuel cells, J. Electrochem. Soc., 2006, 153, pp. A127-A134. 74. Y.D. Zhen, A.I.Y. Tok, S.P. Jiang, F.Y.C. Boey, La(Ni,Fe)O3 as a cathode material with high tolerance to chromium poisoning for solid oxide fuel cells, J. Power Sources, 2007, 170, pp. 61-6. 75. Y. Zhen, S.P. Jiang, Characterization and performance of (La,Ba)(Co,Fe)O3 cathode for solid oxide fuel cells with iron-chromium metallic interconnect, J. Power Sources, 2008, 180, pp. 695-703. 76. G.-Y. Lee, R.-H. Song, J.-H. Kim, D.-H. Peck, T.-H. Lim, Y.-G. Shul, and D.-R. Shin, Properties of Cu, Ni, and V doped-LaCrO3 interconnect materials prepared by Pechini, ultrasonic spray pyrolysis and glycine nitrate processes for SOFC, J. Electroceram., 2006, 17, pp. 723-27. 77. M. Liu, L. Zhao, D. Dong, S. Wang, J. Diwu, X. Liu, G. Meng, High sintering ability and electrical conductivity of Zn doped La(Ca)CrO3 based interconnect ceramics for SOFCs, J. Power Sources, 2008, 177, pp. 451-56. 78. S.P. Jiang, L. Liu, K.P. Ong, P. Wu, J. Li, J. Pu, Electrical conductivity and performance of doped LaCrO3 perovskite oxides for solid oxide fuel cells, J. Power Sources, 2008, 176, pp. 82-89.
44
Aim of study
Aim of study
Chromium poisoning of planar SOFC with LSM cathodes and Cr-alloy interconnects is a complex process consisting of several steps that may occur simultaneously inside the cell. Mn2+ in LSM plays an important role for the adsorption of gaseous CrO3(g) and Croxyhydroxide on LSM resulting in blocked oxygen transport from the cathode to the electrolyte. Reduction of CrO3(g) at the TPB leads to the formation of electrically low conducting Cr2O3, which further retards the diffusion process of oxygen into the electrolyte. The causes and consequences of chromium poisoning are clear, and some strategies against cell degradation caused by chromium have already been successfully applied. However, it seems that strategies against the cell degradation have been mostly established in a rather random way so far. For a more systematic and thus more efficient combination of strategies a strong knowledge about the mechanisms of chromium poisoning of SOFC is required. Previous experiments have shown that the following factors: -) High temperature -) Decrease of oxygen partial pressure at the TPB under current load of SOFC and processes: -) Interaction of chromium with LSM leading to Mn-Cr-O nuclei and/or spinel formation -) Reduction of CrO3(g) to Cr2O3(s) at the TPB -) Blocking of pores at the TPB by Cr2O3 and/or spinel govern the degradation of SOFC caused by chromium. So far it was not possible to define unambiguously, which of these processes play a dominant role for the degradation and which dont. In fact it was shown that without sufficient protection against the diffusion of chromium into the cathode the degradation of SOFC caused by chromium is not a long-term phenomenon; severe degradation has been observed after several hours of testing under current load at state-of-the-art SOFC operating temperatures: from the literature findings it is obvious that the degradation of SOFC caused by chromium starts immediately after starting SOFC tests under current load. If the process of chromium poisoning were completely governed by thermodynamics, this behaviour would not be expected, but the effects of chromium would be observed only after thermodynamic equilibrium is obtained. This means that the kinetic control on the mechanisms of chromium
45
Aim of study
is high, and a degrading cell is in a non-equilibrium state particularly at the early stages of the degradation. The following questions have remained unsolved so far: -) Does spinel form by a solid (LSM)-solid (Cr2O3) reaction or directly in a solid (LSM)-gas (gaseous Cr) reaction? -) Can the concentration of deposits at the cathode-electrolyte interface under current load be explained by thermodynamics? -) How does the LSM phase chemically change due to the interaction with chromium, and can this change be explained by thermodynamics? -) Which of the phases observed in LSM contaminated by chromium form under thermodynamic control, and what are the conditions that favour their formation? This work aims to answer these questions by the application of thermodynamic calculations. Therefore the thermodynamic La-Sr-Mn-Cr-O oxide database needs to be established based on the assessments of low-order subsystems. In recent times many materials have been tested for SOFC cathodes. In particular LSM cathodes have been intensively investigated over the last decade, and several studies can be found regarding the degradation of LSM cathodes caused by chromium. Thus, and as LSM cathodes are still considered to serve as promising cathodes due to their high electrical conductivity and stability at SOFC operating conditions, in this study the author focuses on the effects of Cr on the degradation of SOFC with LSM cathodes.
Method
3.1
Benefits of the thermodynamic La-Sr-Mn-Cr-O oxide database for the understanding of Cr-poisoning of SOFC
A thermodynamic La-Sr-Mn-Cr-O oxide database is highly desirable to enable fundamental understanding of the mechanisms of chromium poisoning of LSM cathodes for SOFC. As a degrading cell is in a non-equilibrium state, the obvious question why the results of thermodynamic calculations should be feasible for a deeper understanding of the mechanisms 46
Method
of chromium poisoning needs some explanation: from the conditions of the non-equilibrium state at the beginning of the degradation process, including the operating temperature, composition of LSM, and the rate of chromium diffusion the equilibrium state of chromium poisoning, can be calculated using the thermodynamic La-Sr-Mn-Cr-O oxide database:
C non - equilibrium (A) equilibrium (B)
(3.1.1)
By calculating thermodynamic equilibria for a LSM cathode that is affected by chromium (A) in the relation above, the theoretical final state of chromium poisoning after a very long time is found by thermodynamic equilibrium calculations (B). For instance, by choosing the starting conditions composition of LSM and defined amount of Cr at a specific temperature, using the thermodynamic database one can calculate the expected thermodynamic equilibrium, for instance under reducing oxygen partial pressures reflecting the situation at the TPB under current load. Over time the system LSM + Cr will change from its nonequilibrium state at the beginning of the Cr-poisoning process towards the calculated equilibrium state. C in Eq. 3.1.1 reflects the path the system takes towards its equilibrium state. From A and B, C can be predicted for changing cathode compositions, temperatures, and oxygen partial pressures. Hence, taking into account experimental data on the chromium poisoning of SOFC and using a thermodynamic La-Sr-Mn-Cr-O oxide database, one can draw conclusions on the evolution of the phase chemistry of degraded LSM cathodes. The presented thermodynamic database of the La-Sr-Mn-Cr-O oxide system is constructed using the CALPHAD approach[1]. It contains the optimized Gibbs energy functions of solid oxide phases: for stoichiometric phases as a function of temperature, and for solid solution phases as a function of temperature and composition. The optimization of model parameters is based on the accurate assessment of experimental thermodynamic and phase diagram data of oxide subsystems.
3.2
Thermodynamic modeling
3.2.1
Stoichiometric solid oxides
The stoichiometric ternary phase , containing m and n moles of two different sorts of cations, a with the positive electrical charge r and b with the positive electrical charge q,
47
Method
respectively, and p moles of one sort of anions, c with the negative electrical charge s, the three types of ions sitting in three distinctive crystallographic sublattices, can be described by the sublattice formula (a r )m (b q )n (c s ) p . For oxides c = O and charge s = 2. To account for the charge neutrality criterion, Eq. 3.2.1 is true.
mr + nq + 2 p = 0
(3.2.1)
The molar Gibbs energy of , Gm at constant pressure is given by
Gm = A + BT + CT ln T + DT 2 + ET 3 + FT ( 1)
(3.2.2)
A, B, C, D, E, and F are model parameters to be optimized by thermodynamic and phase diagram data. As Cp() is defined by
C p = C 2 DT 6 ET 2 2 FT 2
(3.2.3)
C, D, E and F are optimized to heat capacity data only. Gm can be based on the molar Gibbs
energies of existing binary oxides Ox1: (a r )t (O2- )u and Ox2: (b q )v (O2- ) w ( t , u, v, w N ), if it is assumed that the heat capacity of the ternary oxide composed by the two binary oxides is simply the sum of the heat capacities of the composing oxides as shown in Eq. 3.2.4:
m ( a )t (O2- )u n (b G + Gm t m v
r
Gm = Gm
( a r )m (bq )n (O2- ) p
)v (O2- )w
ptv muv nwt O (g) Gm + A + BT 2tv

2
(3.2.4)
Gm is the Gibbs energy of formation of the phase relative to the oxide components. A and B
are optimized by thermodynamic and phase diagram data.
3.2.2 Solid solution phases the Compound Energy Formalism (CEF) If in the binary oxide Ox1: (a r )t (O2- )u containing cation a with the positive charge r, another sort of cation with the positive charge q, bq can sit in the same sublattice as a, the sublattice formula of the resulting solid solution phase (ss) reads (a r , b q )t (O2- )u . Eq. 3.2.5 is the criterion for charge neutrality:
48
Method
u=
t ( ya q + yb q ) 2
(3.2.5)
Using the Compound Energy Formalism (CEF)[2-4], the molar Gibbs energy of the solid solution phase contains the Gibbs energies of the compounds. For (ss) the two compounds read (a r )t (O2- )u and (bq )t (O2- )u . The Gibbs energy of (ss) at constant pressure reads
Gmss = yar G ( a
r) t
(O2- )u
+ ybq G (b
)t (O2- )u
+ tRT yar ln yar + ybq ln yar + EGmss
(3.2.6)
where yar is the site fraction of cation a on the cation sublattice, and ybq is the site fraction cation b on the cation sublattice. R=8.31451 J mol-1 K-1. The second-last term accounts for the configurational entropy of mixing of t moles of a and b. The last term describes the excess Gibbs energy of mixing due to interactions of ions in the mixture that can be accounted for by introducing interaction parameters.
3.2.3
Vacancies and the concept of reciprocal reactions
Let us consider the case of a binary oxide phase (A)2(B)3, A standing for the cation sublattice, and B denoting the anion sublattice, with only one cation a accepting the charge 3+ or 2+ in the cation sublattice. If the cation is reduced, the charge neutrality criterion is no longer obeyed by an anionic sublattice that is completely filled with oxygen. Charge neutrality under such reducing conditions can be remained by the formation of zero-charged vacancies (Va) in the anionic sublattice resulting in the phase becoming oxygen-nonstoichiometric. In the sublattice form the phase can be written as (a3+ , a 2+ )2 (O2- , Va)3 . The oxygen nonstoichiometry is denoted O3-. The molar Gibbs energy of the phase at constant pressure reads
3+
A ( Gm 2 O3 = Gma
2+
, a 2+ )2 (O2- ,Va )3
( = ya3+ yO2- Gma
3+
)2 (O2- )3
( + ya2+ yO2- Gma
2+
)2 (O2- )3
( + ya3+ yVa Gma
3+
)2 (Va )3
3+
( + ya2+ yVa Gma
)2 (Va )3
( + 2 RT ya3+ ln ya3+ + ya2+ ln ya2+ + 3RT yO2- ln yO2- + yVa ln yVa + EGma
, a 2+ )2 (O2- ,Va )3
(3.2.7)
Once again the molar Gibbs energies of all the 4 endmember compounds (a3+ )2 (O2- )3 ,
(a 2 + )2 (O 2- )3 , (a 3+ )2 (Va)3 , and (a 2 + )2 (Va)3 of the phase are required for the molar Gibbs energy
49
Method
of the phase. However, the only neutral endmember is (a3+ )2 (O2- )3 . It thus can exist, and its molar Gibbs energy can be defined by optimization of model parameters by experiments. The three other endmembers are charged and cannot exist, but a line of neutral compositions connects (a3+ )2 (O2- )3 with the reduced compound (a 2+ )2 (2 3O2- Va)3 , and its Gibbs energy can be optimized with experiments that are related to the reduction of the phase, for instance oxygen nonstoichiometry data. The composition square of the phase can be seen in Fig. 3.2.1 that is redrawn from Hillert[4], with the neutral line and the reduced compound, denoted with R, included. The 2+ charged center composition of the square, (a3+ a 2+ )2 (3 2O2- 3 2 Va)3 , denoted with A in Fig. 3.2.1, is theoretically obtained by mixing equal amounts of either
(a3+ )2 (O 2- )3 and (a 2 + )2 (Va)3 or (a3+ )2 (Va)3 and (a 2 + )2 (O 2- )3 .
Fig. 3.2.1 The surface of reference for the Gibbs energy of the reciprocal phase (a3+ , a 2+ )2 (O2- , Va)3 approximating its overall Gibbs energy for Grec > 0 and Grec = 0, plotted above the composition square. A system that obeys this relation is called a reciprocal system, and (a3+ , a 2+ )2 (O2- , Va)3 is a reciprocal phase[4]. For an unambiguous definition of the molar Gibbs energy of the reciprocal phase it is necessary to give an arbitrary molar Gibbs energy to a reference. As the chosen molar Gibbs energy of the reference is unlikely the true value, the reference should favorably be a highly charged compound, thus far off neutral compositions that can really exist. For the example of the reciprocal solid solution phase (a3+ , a 2+ )2 (O2- , Va)3 the 6+ charged compound (a3+ )2 (Va)3 is chosen as reference. 50
Method
The surface of reference for the Gibbs energy of the reciprocal phase (a3+ , a 2+ )2 (O2- , Va)3 at very low temperatures (to make the configurational entropic contribution negligible), and without excess terms for the Gibbs energy is visualized in Fig. 3.2.1 (page 50) and approximates the whole Gibbs energy of the phase. The morphology of the Gibbs energy surface depends on G of the reciprocal reaction
(a3+ )2 (Va)3 + (a 2 + )2 (O 2- )3 (a3+ )2 (O 2- )3 (a 2 + )2 (Va)3 :
G rec = G ( a
3+ )
2 (Va)3
+ G ( a
2+ )
2 (O
2- ) 3
G ( a
3+ )
2 (O
2- ) 3
G ( a
2+ )
2 (Va)3
(3.2.8)
If G of the reciprocal reaction, Grec is positive, the Gibbs energy surface is curved, and the theoretic compound A will tend to demix to (a3+ )2 (O2- )3 and (a 2+ )2 (Va)3 by only slightly oxidizing or reducing it. On the other hand, if Grec is zero, the Gibbs energy surface is flat and no tendency of demixing of A exists. In Fig. 3.2.1 (page 50) only the edge of this plane is seen as bold line. Note that in order to obtain the same Gibbs energy of the reduced compound R for Grec > 0 and Grec = 0, when the Gibbs energies of the endmember
(a 3+ )2 (O 2- )3 and the reference (a 3+ )2 (Va)3 are fixed, the Gibbs energies of the remaining
endmembers are significantly different for Grec > 0 and Grec = 0. This is not a problem for the description of a reciprocal oxide phase, as long as these endmembers are charged and away from the existing composition range of the phase. Anyway, the true surface shape of a reciprocal oxide phase with charged endmembers is not known. As no tendency of demixing was reported for the nonstoichiometric oxide solid solutions that are treated in this study, and no experiments define a proper value of the reciprocal reaction parameter, it is legitimate to define Grec = 0. 3.2.4 Calculation of defect chemistry using the Calphad approach
The Calphad approach is very powerful for the calculation of the defect chemistry of highorder nonstoichiometric reciprocal solid solution oxide phases[5] such as (A)(B)O3- perovskite with a complex sublattice formula, for instance
(a 3+ , b 2 + , Va)(c 2 + , c3+ , c 4 + , d 3+ , d 4+ , Va)(O 2- , Va)3 for a Cr-doped LSM perovskite as a function of
composition, temperature, and oxygen partial pressure. For this purpose model parameters of the reduced and oxidized compounds are optimized with experimental information on charge carriers, site fractions and oxygen content.
51
Method
3.3
Optimization of model parameters
The optimization of the thermodynamic parameters was performed using the PARROT module of the Thermo Calc[6] database system. PARROT can take into account all sorts of thermodynamic and phase diagram data simultaneously. The program minimizes the sum of squared errors between calculated and experimentally determined phase diagram and thermodynamic data. As the use of all the experimental data in a simultaneous least square calculation often leads to divergence, the authors selectively adjusted the relative weight of each experimental data point and excluded data that were inconsistent with the majority of the data points during the optimization procedure. This weighting process is based on the accurate assessment of experimental thermodynamic and phase diagram data.
References 1. N. Saunders, A.P. Miodownik, Calphad Calculation of Phase Diagrams, Pergamon Materials Series, Vol. 1. Elsevier Science Ltd., 1998, 479 p. 2. J.-O. Andersson, A.F. Guillermet, M. Hillert, B. Jansson, B. Sundman, A CompoundEnergy Model of Ordering in a Phase with Sites of Different Coordination Numbers, Acta Metall., 1986, 34, pp. 437-445. 3. M. Hillert, B. Jansson, B. Sundman, Application of the Compound-Energy Model to Oxide Systems, Z. Metallkd., 1988, 79(2), pp. 81-87. 4. M. Hillert, The Compound Energy Formalism, J. Alloy. Cmpd., 2001, 320, pp. 161-76. 5. A.N. Grundy, E. Povoden, T. Ivas, L.J. Gauckler, Calculation of Defect Chemistry Using the CALPHAD Approach, Calphad, 2006, 30, pp. 33-41. 6. B. Sundman, B. Jansson, J.O. Andersson, The Thermo-Calc databank system, Calphad, 1985, 9(2), pp. 153-90.
52
4.1
Thermodynamic reassessment of the Cr-O system in the framework of solid oxide fuel cell (SOFC) research
E. Povoden, A.N. Grundy, and L.J. Gauckler J. Phase Equilib. Diff., 2006, 27, pp. 353-62. A comprehensive compilation and evaluation of experimental and thermodynamic data for the Cr-O system is presented and, by application of the CALPHAD method, a consistent set of thermodynamic model parameters is optimized based on new experiments. Nonstoichiometry of eskolaite (Cr2+xO3) is described using the compound energy model, and the liquid is described using the two-sublattice model for ionic liquids. Cr3O4 is described as a stoichiometric compound. Also the magnetic transition in Cr2O3 and the oxygen solubility in Cr are modeled.
4.1.1
Technology
SOFC offers high fuel conversion efficiencies and, due to the high operating temperature (>1173 K), combined heat- and power-generation capability. For the planar design SOFC, which offers low fabrication costs, ceramic and metal interconnect materials have been tested and evaluated over the years. Meanwhile the use of Cr-based alloy interconnect materials has gained popularity due to their relative ease of fabrication, low manufacturing costs and high thermal conductivity[1]. Namely a Cr5Fe1Y2O3 oxide dispersion strengthened alloy with the composition 94 wt.% Cr, 5 wt.% Fe and 1 wt.% Y2O3 developed jointly by Plansee and Siemens with satisfying material properties has been promoted as a suitable alternative to the earth alkaline-doped LaCrO3 ceramics interconnect. However, the use of this alloy as an interconnect material in SOFC leads to the degradation of the fuel cell performance especially on the cathode side of the fuel cell. Loss of performance caused by the migration of Cr originating from the alloy interconnect is well documented by several investigators. Microstructural analyses of the cathode of SOFC show the formation of Cr2O3 and (CrMn)3O4, which block active sites as well as pores, thus substantially reducing the triple-
53
phase boundary area for the normal oxygen reduction reaction at the cathode/electrolyte interface. The influence of Cr from the interconnect alloy on the strontium-doped lanthanum manganite (LSM) cathode can be modelled in terms of an equilibrium thermodynamic view to contribute to strategies for reducing the SOFC chromium poisoning process by optimizing SOFC operating conditions and refining SOFC material compositions.
4.1.2
Experimental data
Phase diagram data: Experimental investigations of phase diagrams in the Cr-O system were made by Olshanskii and Shlepov[2] and Toker[3]. These authors document the existence of a large miscibility gap between the metallic and the oxide melt. Eskolaite (Cr2O3) is the dominating stable oxide phase over a wide temperature range. Results of special points in the Cr-O phase diagram from several studies are summarized in Table 4.1.1.
Table 4.1.1 Special points in the Cr-O system Eutectic Melting of Cr2O3 in Eutectic composition, X(O) ----Cr3O4 detected ----Stability range Mono-
of Cr3O4, tectic T (K) ----T (K) ----Reference Kanolt[4] experimental Wilde and Rees[5] experimental McNally et al.[6] experimental Bunting[7] experimental
air, T (K) T (K) 2257 2317 2603 ----
2543 25 --
54
--2571 ---2539
1933 1918 1941 1929 1937 1938 2 1938
0.52 0.523 0.513 0.496 0.499 0.497 0.497
no no yes yes yes yes yes
--1923 1978 1918 1974 1923 1978 1923 2 1978 3 1918 1973
2083 2083 -2160 2130 2083 2117
Olshanskii and Shlepov[2], experimental Johnson and Muan[12] experimental Degterov and Pelton[39] calculated Kowalski and Spencer[40] calculated Taylor and Dinsdale[41] calculated Toker et al.[13] experimental This work, calculated
Note: Itallicized data were used for optimization The melting temperatures of eskolaite in air reported from Kanolt[4] and Wilde and Rees[5] can be discarded as being too low. Mc Nally[6] measured a melting temperature of 2603 K in air using an induction furnace. This value significantly deviates from the result of Bunting[7], who measured T = 2543 25 K also in air. Grube and Knabe[8] found that 1 wt.% Cr2O3 lowers the melting point of metallic Cr from T = 2163 K to between T = 2043 K and 2063 K. Lam[9] reported the existence of molten chromium with oxygen impurities of 1400 ppm at T = 2133 K. The monotectic reaction of Cr (bcc) metal and liquid was found at T = 2083 K by Grube and Knabe[8] and by Olshanskii and Shlepov[2]. The question of the existence of a crystalline Cr3O4 phase was discussed controversially by several authors. Investigations made by Hilty et al.[10] and Hook and Adair[11] led them to postulate the existence of a crystalline Cr3O4 phase in the Cr-Fe-O system. Concerning the pure Cr-O system, Olshanskii and Shlepov[2] and Johnson and Muan[12] did not find Cr3O4 up to the eutectic temperature of chromium oxide, whereas Toker et al.[13] concluded from microstructural observations and a discontinuity in the slopes of the temperature-oxygen pressure curves for univariant equilibria involving metallic Cr and various chromium oxide phases that a Cr3O4 phase exists in a narrow temperature range between T = 1923 K and 1978 K. Microstructures of a quenched
55
chromium melt with maximum oxygen impurities of about 2930 ppm lately investigated by Lam[9] document an inner Cr3O4 phase and an outer Cr2O3 phase in dispersed oxides in large chromium grains and grain boundaries. This indicates that the first phase to crystallize on solidification is Cr3O4 giving strong evidence for the stability of this phase. Thus in this study the authors accept the findings of Toker et al.[13] and Lam[9].
Thermodynamic data: Oxygen Potentials: Grube and Flad[14] measured log( pO2 ) values for the Cr-Cr2O3 equilibrium between T = 1053 K and 1573 K by both oxidizing Cr to Cr2O3 and reducing Cr2O3 to Cr in a flowing H2-H2O atmosphere. At T 1573 K they were confronted with the loss of a quarter of Cr in the case of oxidation; thus at these temperatures log( pO2 ) values were determined solely from the reduction of Cr2O3. Novokhatskii and Lenev[15] studied the equilibrium of the reduction of Cr2O3 to Cr with hydrogen from T=1493 K to 1893 K. These authors used a flow method where thermal diffusion problems were suppressed by inserting corundum bushes into the reaction tube. Appreciable sublimation of metallic chromium was not observed. Davies and Smeltzer[16] determined log( pO2 ) values of Cr2O3 at T=1173 K, 1273 K, and 1373 K, using an electrochemical cell with a calcia-zirconia electrolyte and a Fe/FeO reference electrode. Toker et al.[13] measured log( pO2 ) values of Cr2O3 by equilibrating Cr and Cr2O3 in H2-CO2 mixtures of known oxygen potentials at temperatures from T = 1773 K to 2098 K. Pehlke et al.[17] used two separate series of emf measurements employing the solid oxide electrolyte galvanic cell technique from T = 1148 K to 1548 K. The reversibility and accuracy of the yttria-doped thoria electrolyte and the electrode was tested by measurements of a standard iron-chromium alloy at 1326 K. The independent results of corrected cell potentials of the two measurement series are excellent. The data are in close agreement with the gas-solid equilibrium measurements by Jeannin et al.[18]. Disagreement between the emf results from Pehlke et al.[17], Pugliese and Fitterer[19], and Tretjakow and Schmalzried[20] were assigned to possible electronic conduction in the zirconia electrolyte used by the latter authors, as well as transport of oxygen ions from the cathode to the Cr/Cr2O3 anode. Applying the same technique as Pehlke et al.[17], Holzheid and ONeill[21] noted a deviation from the well-established trend from T = 900 K to 1300 K for high-temperature data caused by finite electronic conductivity at elevated temperatures, causing transfer of oxygen through the cell, as well as the importance of sufficient time to attain equilibrium, that is, days for T < 1100 K. The obtained 56
dissociation pressures of Cr2O3 are in agreement with average values derived from emf studies using an yttria-doped thoria electrolyte worked out by Jacob[22] and a very high temperature gas-mixing study of Toker et al.[13]. Heat Capacities, heat Contents, and entropies: Andersons[23] calorimetric data set of Cp-values lacks detailed documentation of the experimental procedure. Bruce and Cannell[24] applied a two-dimensional temperature wave method using a single crystal of Cr2O3 to calculate specific heat in the temperature range 290.68 T 323.43 K, and fitted the data to the heat of diffusion equation that considers some material properties employing a least-mean-squares fit. The accuracy of this study is evident from excellent data reproduction by performing two runs in the entire temperature range. Documentation of the experiments, data presentation, and fitting procedure are worked out very carefully. Resulting Cp data correspond nicely to the most recent calorimetric results from Klemme et al.[25]. The latter authors measured a consistent data set of heat capacities of synthetic eskolaite from T = 1.5 K to 340 K with mean increments of 0.56 K. Uncertainties of 0.4 % for Cp (20 K < T < 200 K) and 0.7 % for Cp (T < 20 K) were estimated. For Cp(Cr2O3) = 120.37 J mol-1 K-1 at T = 298.15 K Chase et al.[26] relied on the calculated results from heat content measurements performed by Kelley et al.[27]. The latter authors fitted their data measured from T = 400 K to 1800 K by
H T H 298K = aT + b T 2 + cT 1 + d 2
(4.1.1)
yielding
H T H 298K = 28.53T + 1.10 103 T 2 + 3.736 105 T 1 9759
(4.1.2)
Temperature derivation of Eq. 4.1.2 results in

C p = 28.53 + 2.20 103 T 3.736 105 T 2
(4.1.3)
For S298K (Cr2O3) Chase et al.[26] relied on the results from Anderson[23], who calculated
S298K (Cr2O3) = 81.17 0.84 J K-1mol-1 by a graphical method of plotting the heat capacity
against the logarithm of the temperature and modeling the heat capacity curves with Debye
57
functions. This procedure was critically documented by other authors, for example, Klemme et al.[25]. Klemme et al.[25] recommend
S298K (Cr2O3) = 83.1 J K-1mol-1 by reevaluating emf
data from Holzheid and ONeill[21], who calculated S298K (Cr2O3) = 85.74 1.3 J K-1 mol-1 from their measurements. Dellien et al.[28] adopted their S298K (Cr2O3) value from Wagman et al.[29]. Shirokov[30] estimated S298K of a metastable CrO phase to be 46.86 J K-1 mol-1. Enthalpies of Formation: Roth and Wolf[31] found f,el H 298K (Cr2O3) = 1125.8 2.5 kJ mol-1 (el=elements) by applying a calorimetric technique. Mah[32], using a bomb calorimetric combustion technique at 1323 K and 30 atm oxygen pressure, calculated f,el H 298K (Cr2O3) =1140.98 1.7 kJ mol-1. Some difficulty caused by moisture adsorption was encountered in weighing the combustion products. This was circumvented by heating the combustion products to T = 1323 K. For the calculation of f,el H 298K (Cr2O3) the heat content data given by Kelley et al.[27] were used. Ramsey et al.[33] used heat capacity and entropy data from tabulations of Coughlin[34] to obtain f,el H 298K (Cr2O3) = 1122.06 kJ mol-1. Navrotsky[35] cited Garrels and Christ[36] for
f,el H 298K (Cr2O3) = 1128.42 kJ mol-1. Chase et al.[26] evaluated f,el H 298K (Cr2O3) = 1134.7
8.4 kJ mol-1 from several earlier studies, while Dellien et al.[28] adopted f,el H 298K (Cr2O3) = 1139.72 kJ mol-1 from Wagman et al.[29]. Klemme et al.[25] recommended f,el H 298K (Cr2O3) = 1128.2 0.4 kJ mol-1 by evaluating emf data from Holzheid and ONeill[21]. Shirokov[30] estimated f,el H 298K (CrO) = 305.4 kJ mol-1 for metastable CrO.
4.1.3
Previous assessments of the Cr-O system
Banik et al.[37] established a phase diagram for the Cr-O system based on a subregular solution model that is in good agreement with experimental data obtained by Olshanskii and Shlepov[2], thermodynamic data for Cr-Cr2O3 from Fromm and Gebhardt[38], and thermodynamic estimates for CrO from Shirokov[30]. Degterov and Pelton[39] reassessed the CrO-Cr2O3 subsystem for the molten slag database using a modified quasi-chemical model for the liquid phase. Their calculated liquidus temperature of Cr2O3 in air is T = 2571.16 K, which is in good agreement with an early finding by Bunting[7] who measured T = 2543 25 K. Kowalski and Spencer[40] used the associated solution model for the liquid phase based on the experimental data used by Taylor and Dinsdale[41]. The latter authors proposed a 58
phase diagram in good agreement with the experimental data obtained by Toker[3], using the same thermodynamic models as the authors use in this work, which are the two-sublattice ionic model for the liquid and the compound energy model for the Cr2+xO3 phase. Taylor and Dinsdale[41] fitted Cp data from Anderson[23] close to the antiferromagnetic to paramagnetic transition and data from Chase et al.[26] at elevated temperatures as a sum of magnetic and nonmagnetic contributions. Experimental information on phase relations for their assessment was taken from Olshanskii and Shlepov[2], Toker[3], Kelley et al.[27], and Grube and Knabe[8]. The heat capacity of Cr3O4 was taken as 7/5 times the nonmagnetic value for Cr2O3 according to Neumann and Kopps rule. Their calculated values for the enthalpy of formation and the entropy of Cr3O4 are in agreement with an estimate done by Chipman[42]. Their optimization of one of the charged endpoints in their compound energy model for eskolaite and the use of six interaction parameters to describe the liquid may lead to problems on extrapolation to higher-order systems, especially as their miscibility gap does not close on increasing temperature. The use of six parameters for the description of the Cr3O4 phase is somewhat incommensurate with the scanty experimental information of this phase. There is a large uncertainty concerning the exact melting point of Cr2O3, and only few thermodynamic data of the Cr3O4 phase and the liquid phase exist. This is reflected by significant variations of the position of the eutectics, the stability field of Cr3O4, and the temperature of the monotectic reaction of Cr(bcc) and liquid between the assessments of the Cr-O system.
4.1.4
Solid phases: The crystal structure of eskolaite is -Al2O3 type, space group R3c . Eskolaite shows an antiferromagnetic to paramagnetic transition at T = 305.5 K. The magnetic properties of
eskolaite can be described using a magnetic ordering model proposed by Inden[43], and simplified by Hillert and Jarl[44]. In this model, a magnetic contribution to the Gibbs energy is added to the nonmagnetic part of the Gibbs energy given as:
MAG Gm = RTln( +1) f ( )
(4.1.4)
59
where is a parameter related to the total magnetic entropy, and = T/Tc. Tc is the critical temperature for magnetic ordering. Tc and are both dependent on the composition. The magnetic parameter p equals 0.28. The defect chemistry of Cr2+xO3 with the sublattice occupation (Cr2+,Cr3+)2 (Cr3+,Va)1 (O2-)3 can be modeled using experimental data from Matsui and Naito[45]. This means that reduction is accomplished by the formation of interstitial Cr3+ and not by the formation of oxygen vacancies, which is in agreement with Young et al.[46]. When modeling nonstoichiometry in an oxide phase, it is important to submit the experimental data to a defect-chemistry analysis. In the case of Cr2+xO3 modeled with interstitial Cr3+ the defect reaction reads
x x ' x 2CrCr + Va ix + 3OO 3 2CrCr + 1 2Cri + 1 4 Va ix + 9 4OO + 3 8O 2(g)
(4.1.5)
giving the equilibrium constant

' x [CrCr ]3 2 [Cri ]1 2 [Va ix ]1 4 [OO ]9 4 pO
x x [CrCr ]2 [Va ix ][OO ]3
38 2
Kr =
(4.1.6)
' Assuming small defect concentrations all concentrations except [CrCr ] and [Cri ] are ~ 1 and ' can be ignored. Due to charge neutrality the relation [Cri ] = 3[CrCr ] must hold. Inserting this ' into Eq. 4.1.6 gives the proportionalities [Cri ] pO2 3 16 and [CrCr ] pO2 3 16 . To explain their
experimental results Matsui and Naito[45] proposed a defect reaction that leads to the same proportionality; however, their equation violates the fundamentals of defect chemistry and must be rejected in favor of the defect reaction given above (Eqs. 4.1.5 and 4.1.6). The following other interstitial defects could be assumed: Cri giving a slope of pO2 1 4 , and
Cri giving a slope of pO
3 20 2
. Both are unlikely: the former because it is unlikely to get Cr4+
on reduction, the latter because of the large size of Cr2+. Assuming the defect reaction that describes the formation of oxygen vacancies:
x x ' CrCr + 1 2OO CrCr + 1 2 Va + 1 4O2(g) O
(4.1.7)
60
' leads to a proportionality of [Va ] PO2 1 6 and [CrCr ] PO2 1 6 . This slope does not correspond O
to the experimental findings of Matsui and Naito[45]. Also the defects cannot explain the electrical properties measured by Young et al.[46]. The low nonstoichiometry data from Matsui and Naito[45] show a different slope than their higher nonstoichiometry data. In contrast to Matsui and Naito[45] who explain this assuming that neutral Cr forms interstitially, the present authors believe that the different slopes are caused by a competing defect reaction, for example charge disproportionation, Cr3+ Cr2+ + Cr4+, similar to the case of LaMnO3 perovskites[47]. The present authors didnt consider this by their defect chemistry model, as it would make the description quite complex. Fig. 4.1.1 is a graphic representation of the model the authors use to describe the oxygen nonstoichiometry of eskolaite, where each corner of the composition square represents a G parameter.
Fig. 4.1.1 Compound energy model for the Cr2+xO3 phase
The four corner compositions represent all possibilities to express the Cr2+xO3 phase according to the above formula for the sublattice occupation. The corner Cr3+:Va:O2corresponds to stoichiometric Cr2O3. The three other corner compositions present charged compounds. Only compounds along the neutral line can exist on their own. As the most reasonable way to model reduction is to use the reduced neutral endpoint
61
(Cr2+)2 (Cr3+2/3Va1/3)1 (O2-)3, labeled A in Fig. 4.1.1, one has to find functions of G of three charged corners expressed solely in terms of the neutral compositions. This is done by using the two equations for the stoichiometric and the reduced endpoints, by choosing an arbitrary reference, and by defining a reciprocal reaction giving four equations with which all Gs at the corners can be expressed. The function to model the reduction then reads
GCr2+
3+ 2 2 (Cr 2 3Va1 3 )(O )3
= GCr O + 2 3 GCrSER + A + BT + RT ( 2 3ln 2 3+1 3ln1 3)

2 3
(4.1.8)
The last term describes the configurational entropy due to mixing of Cr3+ and Va on the interstitial site. G of the 3+ charged endmember (Cr3+)2 (Cr3+)1 (O2-)3 is chosen as reference and given the value GCr3+ :Cr3+ . Then the reciprocal relation reads
Cr3+ :Cr3+
+ G
Cr 2+ :Va
= G
Cr3+ :Va
+ G
Cr 2+ :Cr3+
= Gr
(4.1.9)
In order to avoid the inevitable formation of miscibility gaps if the energy of the reciprocal relation is large we set this energy zero, which leads to
Cr3+ :Cr3+
+ G
Cr 2+ :Va
Cr3+ :Va
Cr 2+ :Cr3+
=0
(4.1.10)
This means that without introducing interaction parameters one gets an ideal description between Cr2O3 and Cr2+xO3. The expressions for all Gs at the corners resulting from Eq. 4.1.8 to 4.1.10 are listed in Table 4.1.2, pp.63-64.
62
Table 4.1.2 Thermodynamic description of the Cr-O System Element Element Cr O Liquid
(Cr 2+ ,Cr 3+ ) p (O 2- ,Va q- )q p = 2 yO2- + qyVa , q = 2 yCr 2+ + 3 yCr3+
L SER GCr 2+ :Va q- H Cr = GCR_L[49] L SER GCr3+ :Va q- H Cr = 2GCR_L[49] + GCR2O3_L 3GCR1O1_L L SER SER GCr3+ :O2- 2 H Cr 3H O = GCR2O3_L L SER SER GCr 2+ :O2- 2 H Cr 2H O = 2GCR1O1_L
Reference Cr (bcc_A2)
1 2
Mass 51.996 15.999
H298 - H0 4050.0 4341.0
S298 23.543 102.52
mol O2
LCr 2+ :O2- ,Va q- = 0 LCr3+ :O2- ,Va q- = 121000
Solid Cr (bcc_A2)
(Cr)1 (O,Va)3
bcc SER GCr:Va H Cr = GHSERCR [49] bcc SER SER GCr:O H Cr 3H O = GHSERCR + 3GHSEROO[49] + 243T
0 bcc Cr:O,Va
= 709542 Tc = 311.5
p = 0.4
= 0.008
CrO
Cr1O SER SER GCr:O1 H Cr H O = GCR1O1
Cr2O3
(Cr 2+ , Cr 3+ ) 2 (Cr 3+ ,Va)1 (O 2- )3
Cr 3 SER SER GCr 2O:Va:O 2- 2 H Cr -3H O = GCR2O3 3+ Cr 3 SER SER GCr 2O:Cr3+ :O 2- 3H Cr 3H O = GCR2O3 + GHSERCR [49] 3+ Cr O3 SER SER GCr 22+ :Cr3+ :O 2- 3H Cr 3H O = GCRO0 + 1 GHSERCR [49] 5.2923T 3 Cr O3 SER SER GCr 22+ :Va:O 2- 3H Cr 3H O = GCRO0 2 GHSERCR [49] 5.2923T 3
p = 0.28
Tc = 308.6
= 3.0
Cr3O4
Cr3O SER SER GCr:O 4 3H Cr 4 H O = GCR3O4
Functions
63
GCR2O3 = 1164542 + 728.56T 119.8T ln T 4.97 10-3T 2 + 1050000T -1 GCR1O1 = 0.5GCR2O3 0.5GHSEROO[49] + 255269 53.82T GCR3O4 = 1.5GCR2O3 0.5GHSEROO[49] + 280045 93.76T GCRO0 = 108305 + GCR2O3 + 2 GHSERCR [49] 3 GCR2O3_L = GCR2O3 + 439078 169T GCR1O1_L = 0.5GCR2O3 0.5GHSEROO[49] + 339673 121.4T
Note: All parameters are in SI units: J, mol, K, Pa: R = 8.31451 J mol-1 K-1. Parameters for solid Cr, liquid Cr, and gaseous O are from Dinsdale[49]
In contrast to Taylor and Dinsdale[41], who needed 4 parameters to model the Cr2O3 phase and had to arbitrarily equate the G of (Cr2+)2 (Va)1 (O2-)3 to stoichiometric Cr2O3, the latter constraint is not needed in the model, and one can reduce the number of parameters to only two. The oxygen solubility in solid Cr(bcc) can be described by an interstitial solution model of the form (Cr)1(O,Va)3. For the optimization of model parameters, literature data from Caplan and Fraser[48] are used. It was not possible to model the oxygen solubility using the endmember GCr:O as this endmember turned out to be too stable and CrO3 appeared in the stability diagram at high oxygen partial pressures. Therefore a large value is given to the endmember GCr:O (in this case 0 was a large number) and the oxygen stability is modeled with the temperature dependence of GCr:O and a regular interaction parameter 0 LCr:O,Va that must of course be negative. The Cr3O4 phase is based on the eskolaite phase. Its heat capacity is given by Neumann and Kopps rule. Metastable CrO is described in the same way. The descriptions for solid and liquid chromium metal and gaseous O2 are from Dinsdale[49].
Ionic liquid:
The two-sublattice ionic liquid model[50,51] is selected to describe the ionic liquid. As the experimental data on the liquid phase are scarce, the number of parameters is kept as low as possible. The sublattice occupation (Cr3+,Cr2+)p(O2-,Vaq-)q is chosen. With this expression one is able to obtain reasonable results for the liquid phase using the positive interaction parameters, 0 LCr2+ :O2- ,Va and 0 LCr3+ :O2- ,Va that are required to give the miscibility gap. Fig. 4.1.2 (next page) is a graphic expression of the model, where each corner of the composition square represents a G parameter of the liquid phase.
64
Fig. 4.1.2 Two-sublattice ionic liquid model for the Cr-O system
The four corner compositions represent all possibilities to express the liquid phase according to the above formula. The liquid composition changes along the hyperbolic curves in Fig. 4.1.2. A special feature of the Cr-O system is the occurrence of a eutectic very close to the composition of CrO. The eutectic temperature is mainly determined by the value of the corner Cr2+:O2-. One derives the GL functions of the oxide compositions (Cr3+:O2-) and (Cr2+:O2-) from the eskolaite phase. The GL of liquid Cr is taken from Dinsdale[49]. In this model description of the liquid phase metallic Cr-liquid can be described by both the corners Cr2+:Va and Cr3+:Va. Cr3+:Va must be metastable compared to Cr2+:Va. One way of doing this would be to simply say that Cr3+:Va equals Cr2+:Va plus a large positive term, for example +600000 as given to GCu2+ :Va by Hallstedt et al.[52] in his original assessment of the Cu-O system. This is however problematic for reciprocal systems. If the reciprocal energy of the system is large there will be a tendency to form miscibility gaps as pointed out by Hillert and Sundman[53]. Hallstedt and Gauckler[54] recently reoptimized the Cu-O liquid, obtaining the parameter GCu2+ :Va from the reciprocal relation and giving it a reciprocal energy of 0. This considerably improved the description of the Cu-O liquid and removed the inverted
65
miscibility gap found at high temperatures in the original assessment[52]. An identical strategy is employed here. Thus metallic liquid is given by the corner
L parameter GCr3+ :Va is obtained by the reciprocal reaction given as
G LCr2+ :Va .The
G L 3+
Cr
:Va
= 2 G L 2+
Cr
:Va
+ G L 3+
Cr
:O2-
3 G L 2+
Cr
:O2-
+ Gr ; Gr = 0
(4.1.11)
4.1.5
Optimization of parameters
The complete set of optimized thermodynamic parameters describing the Cr-O system is given in Table 4.1.2 (pp. 63-64). The optimization of the thermodynamic parameters was performed using the PARROT module of the Thermo Calc[55] database system. In principle, PARROT can take into account all sorts of thermodynamic and phase diagram data simultaneously. The program minimizes the sum of squared errors between the calculated and experimentally determined phase diagram and thermodynamic data. As the use of all the experimental data in a simultaneous least square calculation often leads to divergence, the authors selectively adjusted the relative weight of each experimental data point and excluded data that were inconsistent with the majority of the data points during the optimization procedure. The first parameters to be optimized were the Cp-parameters of Cr2O3. These parameters were then kept fixed during the rest of the optimization. The data used were heat content data from Kelley et al.[27] and Cp data from Klemme et al.[25] at T = 290 K and from T = 335 K to 338 K with a low relative weight. The authors optimized Tc and using Cp data from Klemme et
al.[25] close to the antiferromagnetic to paramagnetic transition temperature. To determine the
parameters describing the enthalpy and entropy of Cr2O3 log( pO2 ) data from Jeannin et al.[18] and Toker et al.[13], high temperature emf data from Holzheid and ONeill
[21]
, and, with low
relative weight, f,el H 298.16 and S298.16 data from Holzheid and ONeill[21] were used. In the next step the authors optimized the nonstoichiometry of Cr2+xO3 using data from Matsui and Naito[45]. They assessed Cr3O4 and the liquid simultaneously, using experimental phase equilibria data from Toker et al.[13], experimental data on the liquidus at the oxygen poor side from Toker et al.[13], and experimental data on the liquidus at the oxygen rich side of the miscibility gap from Olshanskii and Shlepov[2]. The melting temperature of eskolaite in air was taken from Bunting[7]. Finally the solubility of O in solid Cr was optimized using data
66
from Caplan and Fraser[48]. In Table 4.1.1 (p. 54-55) values that were used for our optimization are written in italic letters.
4.1.6
Results and discussion
Phase diagram:
The calculated phase diagram with oxygen isobars is shown in Fig. 4.1.3.
Fig. 4.1.3 Calculated Cr-O phase diagram with oxygen isobars (Pa, logarithmic) given.
The gas phase was not included in the calculation An enlargement of the phase diagram close to the CrO composition is presented in Fig. 4.1.4 (next page).
67
Fig. 4.1.4 Enlargement of the calculated Cr-O phase diagram close to the
CrO composition, with experimental data and oxygen isobars (Pa, logarithmic) included The shape of the liquidus at the oxygen poor side of the miscibility gap resulting from the authors optimization relying on a single experimental datum from Toker et al.[13] and an earlier experiment from Olshanskii and Shlepov[2] is slightly deviating from former optimizations. The calculated liquidus temperature of eskolaite in air is T = 2539 K, in good agreement with the measurement from Bunting[7]. For the monotectic temperature of the reaction of Cr (bcc) and liquid the present authors calculate T = 2117 K, and for the eutectic one gets T=1938 K at a mole fraction of oxygen of 0.497. Cr3O4 is formed at T = 1918 K by the eutectoid reaction Cr2 O3 + Cr + 1 2O2 Cr3O4 . At a mole fraction of oxygen > 0.497 it decomposes in a peritectic reaction at T = 1973 K forming Cr2O3 and liquid. Fig. 4.1.5 (next page) shows the calculated oxygen potential phase diagram of the Cr-O system with experimental log( pO ) data included.
2
68
Fig. 4.1.5 Calculated oxygen potential phase diagram of the Cr-O system,
with experimental log( pO2 ) data as a function of temperature from different studies The experimentally determined phase stabilities from Toker et al.[13] are particularly well reproduced by the authors optimization. The shape and size of the miscibility gap is speculative due to the lack of experimental data. The stability of Cr3O4 is shown in the log( pO2 ) versus temperature diagram in Fig. 4.1.6.
Fig. 4.1.6 Stability of Cr3O4 in the log( pO2 ) versus temperature diagram
69
The solubility of oxygen in Cr(bcc) is shown in Fig. 4.1.7.
Fig. 4.1.7 Calculated oxygen solubility in Cr(bcc)
with experimental data and oxygen isobars (Pa, logarithmic) included For the maximum solubility of oxygen in Cr(bcc) one calculates 0.08 at.% at T=1938 K. If the commonly used however grubby notation Cr2O3- is applied, the total charge of Cr is given by 6+2. The maximum calculated = 0.098 at T = 1918 K. The cation overstoichiometry resulting from the presented optimization might seem somewhat high, but it results simply from the extrapolation of experimental data from Matsui and Naito[45] on excess Cr as a function of pO down to the oxygen partial pressure at the Cr-Cr2O3
2
equilibrium following the proportionality given by the defect chemistry analysis in section 4.1.4. The comparison of the calculated nonstoichiometry in Cr2+xO3 with the experimental data by Matsui and Naito[45] is given in Fig. 4.1.8 (next page).
70
Fig. 4.1.8 Optimized nonstoichiometry of Cr2+xO3 with the only available experimental data
from Matsui and Naito[45] included. Optimization of a temperature dependence is represented by dotted lines. Solid lines result from our accepted optimization without considering temperature dependence. The low nonstoichiometry data show a different slope than the higher nonstoichiometry data. The solid lines correspond to the optimization that is accepted in this work. Obviously the calculated results show a temperature dependence that is significantly stronger compared to the experiments. Considering a temperature dependence for the reduced neutral endpoint of the phase Cr2+xO3 gives values of GCRO0 = 202130 + 235T + GCR2O3 + 2 3GHSERCR (dotted lines in Fig. 4.1.8) and leads to the reduced neutral endpoint being too stable at low temperatures. Therefore, and due to existing data at only three different temperatures from a single author it was decided not to optimize a temperature dependence giving
GCRO0 = 108305 + GCR2O3 + 2 3GHSERCR . The data at low oxygen nonstoichiometries were
not used, as the introduction of an additional defect species would be required to reproduce these.
71
Thermodynamic Data:
The heat capacities, Cp, of the solid Cr2O3 phase (Fig. 4.1.9) are well represented by our assessment.
Fig. 4.1.9 Comparison of calculated heat capacities of Cr2O3 with experimental data
For the magnetic parameter we calculate 3.0, and for Tc we get 308.6. For f,el H 298K (Cr2O3) we calculate 1123 kJ mol-1, which is in particularly good agreement with the data from Ramsey et al.[33], and for S298K (Cr2O3) we get 85 J K-1mol-1, which is very close to the results from Holzheid and ONeill[21]. For f,el H 298K (Cr3O4) we calculate 1402 kJ mol-1, and for S298K (Cr3O4) we get 175 J K-1mol-1. These values for Cr3O4 deviate significantly from the results of Taylor and Dinsdale[41] who calculated f H 298K (Cr3O4) = 1447.685 kJ mol-1, and
S298K (Cr3O4) = 150.555 J K-1 mol-1. For a metastable CrO phase we calculate
f,el H 298K = 306 kJ mol-1 , and S298K = 79 J K -1 mol-1 based on the estimates of Shirokov[30].
72
4.1.7
Conclusions
With the presented reassessment of the Cr-O system the authors are able to excellently describe available thermodynamic and phase diagram data with as few optimizing parameters as possible. However, it must be kept in mind that experimental data on the liquid miscibility gap are largely missing, and that a large variation of the measured melting points of eskolaite exists.
References
1. S.P.S. Badwal, R. Deller, K. Foger, Y. Ramprakash, J.P. Zhang, Interaction between chromia forming alloy interconnects and air electrode of solid oxide fuel cells, Solid State Ionics, 1997, 99, pp. 297-310. 2. Y.I. Olshanskii, V.K. Shlepov, Sistema Cr-Cr2O3, Dokl. Akad. Nauk. SSSR, 1953, 91(3), pp. 561-64. 3. N.Y. Toker, Equilibrium phase relations and thermodynamics for the systems Cr-O and Fe-Cr-O in the temperature range 1500 to 1825 C, Thesis, 1978, Pennsylvania State University. 4. C.W. Kanolt, Melting points of some refractory oxides, J. Wash. Acad. Sci., 1913, 3, pp. 315-18. 5. W.T. Wilde, W.J. Rees, The ternary system MgO-Al2O3-Cr2O3, Brit. Ceram. Trans. J., 1943, 42(7), pp. 123-55. 6. R.N. McNally, F.I. Peters, P.H. Ribbe, Laboratory furnace for studies in controlled atmospheres; melting points of MgO in a N2 atmosphere and of Cr2O3 in N2 and in air atmospheres, J. Am. Ceram. Soc., 1961, 44(10), pp. 491-93. 7. E.N. Bunting, Phase equilibria in the system Cr2O3-Al2O3, J. Res. Nat. Bur. Stand., 6(6), 1931, pp. 947-49. 8. G. Grube, R. Knabe, The system palladiumchromium, Z. Elektrochem., 1936, 42(11), pp. 793-804 (in German). 9. R.K.F. Lam, Melting and casting of high purity chromium with controlled oxygen content, 2000, United States Patent 6039788. 10. D.C. Hilty, W.D. Forgeng, R.L. Folkman, Oxygen solubility and oxide phases in the FeCr-O system, T. I. Min. Metall. Eng., 1955, 203(2), pp. 253-68.
73
11. R.E. Hook, A.M. Adair, The formation and dissolution of chromium oxides in chromium,
T. Metall. Soc. Aime, 1964, 230(6), pp. 1278-83.

12. R.E. Johnson, A. Muan, Phase diagrams for the systems Si-O and Cr-O, J. Am. Ceram.
Soc., 1968, 51(8), pp. 430-33.

13. N.Y. Toker, L.S. Darken, A. Muan, Equilibrium phase relations and thermodynamics of the Cr-O system in the temperature range of 1500 C to 1825 C, Metall. Trans. B, 1991,
22(2), pp. 225-32.

14. G. Grube, M. Flad, Affinity and enthalpy of the solid solution in the system Cr-Ni, Z.
Elektrochem., 1942, 48(7), pp. 377-89 (in German).

15. A. Novokhatskii, L.M. Lenev, Thermodynamic properties of Cr2O3 and FeCr2O4 at high temperatures, Russ. J. Inorg. Chem., 1966, 11(9), pp. 1078-80. 16. H. Davies, W.W. Smeltzer, Oxygen and metal activities of the chromiumnickeloxygen system between 900 and 1100C, J. Electrochem. Soc., 1974, 121(4), pp. 543-49. 17. R.D. Pehlke, F.N. Mazandarany, R.H. Radzilowski, Solid oxide electrolyte emf cell determination of the standard free energy of Cr2O3 and applications to chromiumbearing mineral systems, Geochim. Cosmochim. Ac., 1975, 39, pp. 833-45. 18. Y. Jeannin, C. Mannerskantz, F.D. Richardson, Activities in ironchromium alloys, T.
Metall. Soc. Aime, 1963, 227(2), pp. 300-5.

19. L.A. Pugliese, G.R. Fitterer, Activities and phase boundaries in the CrNi system using a solid electrolyte technique, Metall. Trans., 1970, 1(7), pp. 1997-2002. 20. J.D. Tretjakow, H. Schmalzried, The thermodynamics of spinel phases (chromite, ferrite, aluminate), Berich. Bunsen Gesell., 1965, 69(5), pp. 396-402 (in German). 21. A. Holzheid, H.S. ONeill, The Cr-Cr2O3 oxygen buffer and the free energy of formation of Cr2O3 from high-temperature electrochemical measurements, Geochim. Cosmochim.
Ac., 1995, 59(3), pp. 475-79.

22. K.T. Jacob, Potentiometric determination of the Gibbs free energy of formation of cadmium and magnesium chromites, J. Electrochem. Soc., 1977, 124, pp. 1827-31. 23. C.T. Anderson, The heat capacities of chromium, chromic oxide, chromous chloride and chromic chloride at low temperatures, J. Am. Ceram. Soc., 1937, 59, pp. 488-91. 24. R.H. Bruce, D.S. Cannell, Specific heat of Cr2O3 near the Neel temperature, Phys. Rev. B, 1977, 15(9), pp. 4451-59. 25. S. Klemme, H.S. ONeill, W. Schnelle, E. Gmelin, The heat capacity of MgCr2O4, FeCr2O4, and Cr2O3 at low temperatures and derived thermodynamic properties, Am.
Mineral., 2000, 85, pp. 1686-93.

74
26. M.W. Chase, C.A. Davies, J.R. Jr. Downey, D.J. Flurip, R.A. McDonald, A.N. Syverud, Janaf thermochemical tables - 3rd ed., J. Phys. Chem. Ref. Data, 1985, 14(Suppl. 1): pp. 940-42. 27. K.K. Kelley, F.S. Boericke, E.H. Huffman, W.M. Bangert, Thermodynamic properties of carbides of chromium, Bur. Mines Tech. Paper, 1944, 662, 43 pp. 28. I. Dellien, F.M. Hall, L.G. Hepler, Chromium, molybdenum, and tungsten: thermodynamic properties, chemical equilibria, and standard potentials, Chem. Rev., 1976, 76(3), pp. 283-310. 29. D.D. Wagman, W.H. Evans, V.B. Parker, I. Halow, S.M. Bailey, R.H. Schumm, Selected values of chemical thermodynamic properties. Tables for elements 35 through 53 in the standard order of arrangement, NBS Tech. Notes, 1969, 270(4). 30. N.I. Shirokov, Thermodynamic properties of chromous oxide, Dokl. Akad. Nauk. SSSR,
Metal., 1973, 2, p. 102.

31. W.A. Roth, U. Wolf, The heat of formation of chromium oxide, Z. Elektrochem., 1940,
46, pp. 45-46 (in German).

32. D. Mah, Heats of formation of chromium oxide and cadmium oxide from combustion calorimetry, J. Am. Chem. Soc., 1954, 76(13), pp. 3363-65. 33. J. N. Ramsey, D. Caplan, A.A. Burr, Thermodynamics of the oxidation of chromium, J.
Electrochem. Soc., 1956, 103(2), pp. 135-38.

34. J.P. Coughlin, Contributions to the data on theoretical metallurgy, Bur. Mines Bull., 1954,
542, 80 pp.
35. A. Navrotsky, Thermochemistry of chromium compounds, especially oxides at high temperature, Geochim. Cosmochim. Ac., 1975, 39, pp. 819-32. 36. R.M. Garrels and C.L. Christ: Solutions, Minerals and Equilibria, Harper & Row, 1965, p. 410. 37. G. Banik, T. Schmitt, P. Ettmayer, B. Lux, Thermodynamic consideration on the system Cr-Cr2O3, Z. Metallkd., 1980, 71(10): pp. 644-45. 38. E. Fromm, E. Gebhardt: Gases and Carbon in Metals, Springer Verlag, Berlin, Heidelberg, New York, 1976, pp. 521-34 (in German). 39. S. Degterov, A.D. Pelton, Critical evaluation and optimization of the thermodynamic properties and phase diagrams of the CrO-Cr2O3, CrO-Cr2O3-Al2O3, and CrO-Cr2O3-CaO systems, J. Phase Equilib., 1996, 17(6), pp. 476-87.
75
40. M. Kowalski, P.J. Spencer, Thermodynamic reevaluation of the Cr-O, Fe-O and Ni-O systems: 229-43. 41. J.R. Taylor, A.T. Dinsdale, A thermodynamic assessment of the Ni-O, Cr-O and Cr-Ni-O systems using the ionic liquid and compound energy models, Z. Metallkd., 1990, 81(5), pp. 354-66. 42. J. Chipman, Atomic interactions in molten alloy steels, J. Iron Steel Inst., 1955, 180, pp. 97-106. 43. G. Inden, Determination of chemical and magnetic interchange energies in bcc alloys. I. General treatment, Z. Metallkd., 1975, 66(10), pp. 577-82. 44. M. Hillert, M. Jarl, A model of alloying effects in ferromagnetic metals, Calphad, 1978, Remodelling of the liquid, bcc and fcc phases, Calphad, 1995, 19(3), pp.
2(3), pp. 227-38.

45. T. Matsui, K. Naito, Existence of hypostoichiometric chromium sesquioxide at low oxygen partial pressures, J. Nucl. Mater., 1985, 136, pp. 78-82. 46. E.W.A. Young, J.H. Gerretson, J.H.W. de Witt, The oxygen partial-pressure dependence of the defect structure of chromium(III)oxide, J. Electrochem. Soc., 1987, 134(9), pp. 2257-60. 47. A.N. Grundy, E. Povoden, T. Ivas, L.J. Gauckler, Calculation of defect chemistry using the Calphad approach, Calphad, 2005, 30, pp. 33-41. 48. D. Caplan, M.J. Fraser, A.A. Burr: in Ductile Chromium, ASM, Cleveland, Ohio, 1957, p. 196. 49. A.T. Dinsdale, SGTE data for pure elements, Calphad, 1991, 15(4), pp. 317-425. 50. M. Hillert, B. Jansson, B. Sundman, J. gren, A two-sublattice model of molten solutions with different tendency of ionization, Metall. Trans. A, 1985, 16A, pp. 261-66. 51. B. Sundman, Modification of the two-sublattice model for liquids, Calphad, 1991, 15, pp. 109-19. 52. B. Hallstedt, D. Risold, L.J. Gauckler, Thermodynamic assessment of the copper-oxygen system, 1994, J. Phase Equilib., 1994, 15(5), pp. 483-99. 53. M. Hillert, B. Sundman, Predicting miscibility gaps in reciprocal liquids, Calphad, 2001,
25(4), pp. 599-605.

54. B. Hallstedt, L.J. Gauckler, Revision of the thermodynamic descriptions of the Cu-O, AgO, Ag-Cu-O, Bi-Sr-O, Bi-Cu-O, Sr-Cu-O, Ca-Cu-O and Sr-Ca-Cu-O systems, Calphad, 2003, 27(2), pp. 177-91.
76
55. B. Sundman, B. Jansson, J.O. Andersson, The Thermo-Calc databank system, Calphad, 1985, 9(2), pp. 153-90.
4.2
Thermodynamic assessment of the Mn-Cr-O system for SOFC
materials
E. Povoden, A.N. Grundy, and L.J. Gauckler
Int. J. Mater. Res., 2006, 97, pp. 569-78.

By application of the CALPHAD method, a consistent set of thermodynamic model parameters is optimized for the Cr-Mn-O system based on experimental data. Chromium manganese spinel MnyCr3-yO4 and its tetragonally distorted polymorph are described using the compound energy model, and the liquid is described using the two-sublattice model for ionic liquids. Also solid solutions of the phases (Cr1-yMny)2+xO3, Mn2-yCryO3, and (Mn1-yCry)1-xO are considered. Relevance for solid oxide fuel cells is discussed.
4.2.1
Introduction
For the planar design of SOFC the use of heat-resistant high chromium alloys has been promoted as a suitable alternative to earth alkaline doped LaCrO3 ceramic interconnect materials[1,2]. However, mobilization predominantly via the gas phase[3] of Cr originating from the alloy interconnect leads to the formation of Cr2+xO3 (eskolaite) and chromium manganese spinel MnyCr3-yO4 which block catalytically active sites as well as pores, thus substantially diminishing the triple phase boundary area for the normal oxygen reduction reaction at the cathode/electrolyte interface[4]. Simner et al.[5] observed that the formation of chromium manganese spinel layers on top of a Cr2O3 oxide scale on the surface of a Mn-containing ferritic stainless steel (Crofer22 APU) interconnect with 76.6 wt.% Fe, 22.8 wt.% Cr, and 0.45 wt.% Mn resulted in an improvement of short-term SOFC operation. The processes by which these protective oxide scales reduce the chromium poisoning and their effect on cell degradation during long-term SOFC operation are not well known yet. We are contributing to the understanding of the underlying thermodynamics of these processes by assessing the MnCr-O system using the CALPHAD approach. The thermodynamic data of the pure elements are taken from Dinsdale[6], and the data for the Mn-O, Cr-O, and Mn-Cr binaries from Grundy et al.[7], Povoden et al.[8], and Lee
[9]
77
respectively. No new ternary phases are found in the Mn-Cr-O system, however all the binary oxides except pyrolusite (prl), MnO2, show varying degrees of mutual solid solubility. The most prominent oxide phase in the Mn-Cr-O system is cubic chromium manganese spinel with the formula MnyCr3-yO4. Normal spinel is given by the formula [A2+](B3+)2O4, whereas spinel of the formula [B3+](A2+B3+)O4 with half of B on the tetrahedral sites marked with angular brackets in the above formulas is called inverse spinel. In the case of cubic MnyCr3-yO4 both the trivalent cations of manganese and chromium show a remarkable preference to fill the octahedral sites marked with round brackets in above formulas[10]. Spinel containing a large amount of Mn3+ becomes tetragonally distorted on lowering the temperature as a consequence of the macroscopic Jahn-Teller effect that is caused by the distortion of the octahedral sites occupied by Mn3+[11]. In this work we use the following abbreviations: -spl for cubic chromium manganese spinel solid solution, -spl for tetragonally distorted polymorph spinel solid solution, -hsm (hausmannite) for the cubic and -hsm (-hausmannite) for the tetragonally distorted Mn3O4 endmember of the spinel solid solution, bxb for Mn2O3 (bixbyite) with dissolved Cr, esk for Cr2+xO3 (eskolaite) with dissolved Mn, mgs for Mn1-xO (manganosite) with dissolved Cr, bcc for chromium manganese alloy with bcc A2 structure, and liq for the liquid phase.
4.2.2
Experimental
Phase diagram data:
Our calculated phase diagram of the MnOx-Cr2O3 system in air is shown in Figs. 4.2.1, 4.2.2 (p. 79), and 4.2.3 (p. 80). Fig. 4.2.4 (p. 80) shows the calculated phase diagram at
pO2 = 110-4 Pa .
78
Fig. 4.2.1 Calculated pseudo-binary phase diagram of the system MnOx-Cr2O3 in air. The gas
phase was not included in the calculation.
Fig. 4.2.2 Calculated pseudo-binary phase diagram of the system MnOx-Cr2O3 in air, with
experimental data. The gas phase was not included in the calculation. 79
Fig. 4.2.3 Mn rich part of the calculated pseudo-binary phase diagram
of the system MnOx-Cr2O3 in air, with experimental data.
Fig. 4.2.4 Calculated pseudo-binary phase diagram of the system MnOx-Cr2O3 under strongly
reducing conditions ( PO2 = 110-4 Pa ), showing the expanded stability field of -spl + mgs.
80
Speidel and Muan[12] present a phase diagram of the MnOx-Cr2O3 system in air in the temperature range 873 K to 2253 K resulting from the determination of phase equilibria using quenching techniques and X-ray and microscopic examination (Fig. 4.2.2, p. 79). They find -spl + -spl + bxb coexisting in equilibrium at T = 1183 5 K, and -spl + esk + liq in equilibrium at T = 2243 20 K. They estimate a minimum temperature of T = 773 K for the stability of -spl. The Mn solubility in (Cr2-yMny)1+xO3 reported from Speidel and Muan[12] is significantly higher than it is found by Golikov et al.[13] and Pollert et al.[14]. Golikov et al.[13] studied the MnOx-Cr2O3 system using quenching techniques and high temperature X-ray diffractometry in air in the temperature range from T = 973 K to 1673 K. Their resulting phase diagram is in considerable disagreement with the findings of Speidel and Muan[12]. They report a minimum temperature of -spl stability of T = 973 K and lower solubility of Cr in tetragonally distorted MnyCr3-yO4 and of Cr in Mn2-yCryO3. They consider the solubility limit of Mn in (Cr1-yMny)2+xO3 to be negligible. Pollert et al.[14,15] studied phase stabilities in the MnOx-Cr2O3 system in the temperature range from T = 1100 K to 1620 K in air by means of X-ray measurement of annealed samples. Their data are shown in Figs. 4.2.2, p. 79 and 4.2.3, p. 80. The solubility limit of Cr in Mn2-yCryO3 is measured by these authors to be y = 0.14 at T = 1105 K in air. This value is in agreement with the result from Geller and Espinosa[16], but it is lower than the findings from Speidel and Muan[12]. Pollert et al.[14] report a solubility limit of
y = 1.42 at oxygen partial pressure >> 20000 Pa. From the absence of changes of the lattice
parameters of esk in equilibrium with -spl annealed at T = 1105 K and 1620 K in air compared to pure Cr2O3 they conclude that the solubility of Mn in (Cr2-yMny)1+xO3 is low and does not depend significantly on temperature. Tanahashi et al.[17] investigated the compositions of coexisting -spl + mgs and -spl + esk from pO = 210-6 to 2102 Pa at T = 1873 K thus determining the range of solid solubility of
2
-spl by quenching techniques under controlled CO-CO2 atmosphere (Fig. 4.2.5, next page).
81
Fig. 4.2.5 Cr contents of -spl in equilibrium with mgs, -spl in equilibrium with esk, esk in
equilibrium with -spl, and mgs in equilibrium with -spl in the pseudo-binary MnOx-Cr2O3 system as a function of oxygen partial pressures at T = 1073 K, 1473 K, and 1873 K. Experimental data are included. Phase relations were verified using X-ray diffraction. In order to identify the equilibrium compositions, each phase in the quenched specimens was subjected to electron probe microanalysis (EPMA). They found increasing Mn solubility in cubic MnyCr3-yO4 at oxygen partial pressure higher than 2 10-2 Pa , and significantly increasing Cr solubility in cubic MnyCr3-yO4 with decreasing oxygen partial pressure (Fig. 4.2.5). The compositions of -spl are located on a line connecting MnCr2O4 with -hsm in the ternary plot. From this result Tanahashi et al.[17] conclude that Mn is dissolved in cubic MnyCr3-yO4 in the form of Mn3O4. They report small solubility of Mn in (Cr1-yMny)2+xO3 at pO2 = 210-6 , which increases slightly with increasing oxygen partial pressure. They found almost unchanging solubility of Cr in (Mn1-yCry)1-xO from pO = 210-6 to 2 102 Pa at T = 1873 K. As the compositions of mgs
2
solid solution are located on the line connecting Mn1-xO with CrO in the ternary phase
82
diagram, these authors conclude that chromium oxide dissolves in (Mn1-yCry)1-xO in the form of CrO. Bobov et al.[18] investigated the composition changes of the two phase equilibrium -spl + mgs at T = 1073 K, 1173 K, and 1273 5 K in the oxygen partial pressure range from 0.1 to 10-13 Pa (Fig. 4.2.5, p. 82). For the invariant three phase equilibrium mgs + -spl + bcc Ranganathan and Hajra[19] measured the Mn content in bcc to be 25.2 cat.% at T = 1323 K performing an isopiestic experiment (Fig. 4.2.6).
Fig. 4.2.6 Ternary phase diagram of the system Cr-Mn-O with stoichiometric single phase
equilibria (points), single solid solution phase equilibria (heavy lines), two-phase fields and three-phase fields. Dotted lines are tie lines. Three-phase field boundaries are denoted with thin solid lines. Also the experimental result on the three phase equilibrium MnO + MnCr2O4 + bcc from Ranganathan[19] is plotted. There are no data on oxygen nonstoichiometry of MnyCr3-yO4.
83
Holba et al.[11] investigated the martensitic -spl -spl transition temperatures, enthalpies and entropies of MnyCr3-yO4 samples annealed at T = 1723 K using X-ray analysis and DTA measurement. With decreasing Mn-content the temperature for the transition decreases (Fig. 4.2.2, p. 79). Samples with y < 1.8 do not show JahnTeller distortion at room temperatures, and remain -spl. Giving -spl the formula [Mn2+](Mn3+,Cr3+)2O4 this corresponds to a minimum concentration of [Mn3+] = 0.4 for the formation of -spl.
Thermodynamic data:
Cubic spinel (-spl): Only values for the standard Gibbs energy of formation of -spl of the composition MnCr2O4
-spl are published. Tanahashi et al.[20] derive f GMnCr2O4 = 958 8 kJ mol-1 from liquid Mn,
solid Cr and oxygen at T = 1873 K in the pO range from 210-6 to 1.510-4 Pa from the
2
standard Gibbs free energy changes of the reactions Mn (in molten Fe) + 2 Cr (in molten Fe) +2 O2(g) = MnCr2O4 and Mn (in molten Cu) + 2 Cr (s) + 2O2(g) = MnCr2O4 (4.2.2) (4.2.1)
-spl Using compiled f GMnO and f GCr2O3 [21] they calculate f GMnCr2O4 from its oxides to be
59 8 kJ mol-1. We recalculate this value using the most recently assessed f GMnO [7] and
f GCr2O3
[8]
-spl values at T = 1873 K giving f GMnCr2O4 = 66 8 kJ mol-1. Tsai and Muan[22]
experimentally determined compositions of coexisting MnyCr3-yO4 - MnyAl3-yO4 solid solutions formed from Cr2O3-Al2O3 mixtures at T = 1873 K. From these data and the activities of CrO1.5 in Mn0.5AlO2 and AlO1.5 in Mn0.5CrO2 obtained from a previous study[23] they derive
G of the reaction
CrO1.5 + Mn0.5AlO2 = AlO1.5 + Mn0.5CrO2
(4.2.3)
-spl to be 10 kJ mol-1 at T = 1873 K, which is equivalent to f G of 1 2( f GMnCr2O4 f GMnAl2O4 ) .
This means that the Gibbs energy of formation of -spl of the composition MnCr2O4 from its 84
oxides using this calculation technique depends on the value of f GMnAl2O4 . Tsai and Muan[22] chose the value determined by Lenev and Novokhatskiy[24], f GMnAl2O4 = 32.6 kJ mol-1
-spl giving f GMnCr2O4 = 52.6 kJ mol-1 at T = 1873 K. Using other values for f GMnAl2O4 reported -spl in the literature[25-27] leads to deviating f GMnCr2O4 of 34.4 10 kJ mol-1, 46.1 kJ mol-1, and
36.1 kJ mol-1. Biggers[28] by using the same technique as Tsai and Muan
-spl MnO-Cr2O3 system found f GMnCr2O4 = 59.0 kJ mol-1 at T=1523 K.
[22]
in the CoO-
-spl The spread of f GMnCr2O4 values resulting from different studies and our recalculations is
shown in Fig. 4.2.7.
Fig. 4.2.7 Calculated Gibbs energy of formation of -spl of the composition MnCr2O4 as a
function of temperature, with experimental data and error bars. Filled symbols correspond to originally reported literature data, unfilled symbols correspond to recalculated values. Tetragonally distorted spinel (-spl): Pollert et al.[15] present thermodynamic data on the transition of -spl to -spl. According to these authors the transition of pure -hsm to -hsm takes place at T=1445 K.

= 18810 J mol-1, and S
= 13 J K-1 mol-1.
85
4.2.3
Cubic spinel (-spl):
There is experimental evidence on the presence of Cr2+ in -spl as the Cr endmember of -spl was found to be a stable phase in a small temperature range[29,30]. As the degree of inversity of
-spl is very low[10], -spl can therefore be described by the simple formula
[Mn2+,Cr2+](Cr3+,Mn3+)2O4 using the compound energy model[3133]. Lu et al.[34] measured the electrical conductivity of -spl. They propose small polaron hopping between Mn3+ and Mn4+ on the octahedral sites as mechanism for the electrical conductivity. To maintain electroneutrality Mn2+ is formed on the octahedral sites resulting in a charge disproportionation reaction. Considering these findings an alternative description of -spl would read [Mn2+,Cr2+](Cr3+,Mn2+,Mn3+,Mn4+)2O4. However there is no experimental data quantifying the amount of Mn4+ in -spl, so we stick to the less complex description without Mn2+ and Mn4+ on the octahedral sites. In our description we further go along with the presumption that the amount of oxygen vacancies may be neglected. All endmembers of our model -spl are neutral. In our CALPHAD assessment the G values of all compounds are given relative to the enthalpy of selected reference states for the elements at T = 298.15 K and p = 105 Pa[6]. This state is denoted SER (Stable Element Reference). The Gibbs energies of the endmembers [Mn2+](Mn3+)2O4 that corresponds to
-hsm, and [Cr2+](Cr3+)2O4 that corresponds to Cr3O4 are taken from the assessed binaries[7,8].
The Gibbs energy of the endmember of the formula [Mn2+](Cr3+)2O4 is given by the expression
Cr O -spl -hsm G[Mn 2+ ](Cr3+ ) O = 2 3 G[Cr32+ 4](Cr3+ ) O + 1 3 G[Mn 2+ ](Mn3+ ) O + A-spl + B-splT
2 4 2 4 2 4
(4.2.4)
This endmember is considerably more stable than the endmember of the formula [Cr2+](Mn3+)2O4. We define this last endmember using a reciprocal relation
Cr O -spl -hsm -spl G[Cr2+ ](Mn3+ ) O = G[Cr32+ 4](Cr3+ ) O + 1 3 G[Mn 2+ ](Mn3+ ) O G[Mn2+ ](Cr3+ ) O
2 4 2 4 2 4 2
(4.2.5)
The Gibbs energy of the reciprocal reaction is taken to be zero. Thus the endmember
-spl -spl G[Cr 2+ ](Mn3+ ) O becomes less stable the more stable G[Mn 2+ ](Cr3+ ) O becomes.
2 4 2 4
86
Tetragonally distorted spinel (-spl):
The transformation of -spl to -spl is due to the Jahn-Teller distortion of the octahedral sites occupied by trivalent Mn ions leading to the tetragonal structure of -spl[11,15]. Experiments[11,14,15,34] show that -spl is stabilized at high Mn contents, and the Cr solubility in tetragonally distorted MnyCr3-yO4 does not extend beyond MnCr2O4. It is very unlikely that trivalent cations are incorporated into the tetrahedral sites of -spl, as the degree of inversity is increasing with higher temperature, whereas for ordering due to JahnTeller distortion the opposite holds. Due to these considerations we may write [Mn2+](Cr3+,Mn3+)2O4 to describe
-spl. Hence, the two endmembers of -spl read [Mn2+](Mn3+)2O4 and [Mn2+](Cr3+)2O4. The
Gibbs energy of [Mn2+](Mn3+)2O4 is equal to -hsm. The Gibbs energy of [Mn2+](Cr3+)2O4 is given by
Cr O -spl -hsm G[Cr2+ ](Mn3+ ) O = 2 3 G[Cr32+ 4](Cr3+ ) O + 1 3 G[Mn 2+ ](Mn3+ ) O + A-spl + B -splT
2 4 2 4 2 4
(4.2.6)
Bixbyite (bxb):
Geller and Espinosa[16] postulate the incorporation of Cr into Mn2-yCryO3 by a simple substitution mechanism between Cr3+ and Mn3+. This is a reasonable assumption as the radii of these ions are very similar[35]. The incorporation of chromium of valencies other than three is mentioned nowhere in literature. Thus we may describe bxb as (Mn3+,Cr3+)2(O2-)3. The Gibbs energy of the endmember (Mn3+)2(O2-)3 is taken from Grundy et al.[7], and the Gibbs energy of (Cr3+)2(O2-)3 is given by
bxb esk G(Cr3+ ) O = G(Cr3+ )
2 3
2 (Va)1 (O
2- ) 3
+ Abxb
(4.2.7)
The experimental data could be reproduced without the optimization of a temperature dependent parameter.
Manganosite (mgs):
Based on the proposed incorporation of Cr into (Mn1-yCry)1-xO in the form of CrO[15] we tested a description of mgs given by (Mn2+,Mn3+,Cr2+,Va)(O2-). Using this description the solubility of Cr in function of oxygen partial pressure experimentally determined by Tanahashi[17] could not be reproduced correctly. (Mn2+,Mn3+,Cr3+,Va)(O2-) leads to far more satisfactory
87
reproduction of these data. With this description we are in agreement with OKeefe and Valigi[36] who observe a decrease in the lattice parameter of mgs compared to undoped Mn1-xO providing strong support for the assumption that it is Cr3+ that is substituting the much larger Mn2+ ion and which is forcing the lattice to contract. These experiments also provide evidence against Cr being incorporated interstitially. The model description of mgs is shown in Fig. 4.2.8.
Fig. 4.2.8 Geometrical representation of the mgs phase described
using the compound energy model. The Gibbs energy of the neutral endpoint
( Cr
3+ 2/3
Va1/3 ) O
is
given
by
2 3 G mgs + 1 3 G mgs + RT (1 3ln1 3 + 2 3ln 2 3) and based on the Gibbs energy of 1 mole of (Va )O1 3 (Cr3+ )O1
mgs Gas esk. Using G(Va)O1 = 1 GO2 the following expression for the parameter G mgs is obtained 3+
(Cr
)O1
G mgs = 1 2 G esk 3+ 3+
(Cr )O1 (Cr
)2 (Va)1 (O2- )3
Gas 1 3 GO 3 2 RT (1 3ln1 3 + 2 3ln 2 3) + Amgs

2
(4.2.8)
The Gibbs energies of the three other endmembers are taken from Grundy et al.[7].
Eskolaite (esk):
Pollert et al.[14] postulate the incorporation of trivalent Mn ions into (Cr1-yMny)2+xO3. Agreeing with these authors we model the solubility of Mn by (Cr3+,Cr2+,Mn3+)2(Cr3+,Va)1(O2-)3. The Gibbs energy of (Mn3+)2(Cr3+)1(O2-)3 is given by 88
G esk3+
(Mn
)2 (Cr3+ )1O3
= G bxb3+
(Mn
)2 (O2- )3
+ GCr SER + Aesk
(4.2.9)
and the Gibbs energy of (Mn3+)2(Va)1(O2-)3 by
G esk3+
(Mn
)2 (Va)1O3
= G bxb3+
(Mn
)2 (O2- )3
+ Aesk
(4.2.10)
We take the Gibbs energies of the other endmembers from Povoden et al.[8].
Cr-Mn alloys:
We describe the oxygen solubility in bcc by an interstitial solution model of the form (Cr,Mn)1(O,Va)3. Experimental data on the oxygen solubility in pure bcc A2 chromium metal[37] were used for the description of (Cr)1(O,Va)3[8]. No data are reported for the oxygen solubility in pure bcc A2 manganese metal. Assuming low oxygen solubility in bcc
bcc manganese metal we give a large value to the endmember GMn:O .
The descriptions of further alloy phases are taken from[9].
Liquid:
We model the liquid phase as (Cr3+,Cr2+, Mn3+,Mn2+)p(O2-,Vaq-)q using the two-sublattice model for ionic liquids parameter 0 Lliq3+
Cr ,Mn3+ :O2-
[38,39]
. The liquidus temperature is optimized using the interaction
.The binary interaction parameters are taken from Grundy et al.[7],
Povoden et al.[8], and Lee[9].
4.2.4
Optimization of parameters
The complete set of optimized thermodynamic parameters describing the Mn-Cr-O system is given in Table 4.2.1 (pp. 90-92).
89
Table 4.2.1 Thermodynamic description of the Cr-Mn-O system a) Element Element Reference Mass H298 - H0 S298
Cr Mn O
Liquid (liq)
BCC_A2 CBCC_A12 mol O2
51.996 54.938 15.999
4050.0 4996.0 4341.0
23.543 32.008 102.52
(Cr 2+ ,Cr 3+ ,Mn 2+ ,Mn 3+ ) p (O 2- ,Va q- )q p = 2 yO2- + qyVa , q = 2 yCr 2+ + 2 yMn 2+ + 3 yCr3+ + 3 yMn3+
liq SER GCr 2+ :Va q- H Cr = GCR_L[9] liq SER GCr3+ :Va q- H Cr = 2GCR_L + GCR2O3_L[9] 3GCR1O1_L[9] liq SER SER GCr3+ :O2- 2 H Cr 3H O = GCR2O3_L[9] liq SER SER GCr 2+ :O2- 2 H Cr 2H O = 2GCR1O1_L[9] liq SER GMn 2+ :Va q- H Mn = GMN_L[7] liq SER GMn3+ :Va q- H Mn = 2GMN_L[7] + GMN2O3_L[8] 3GMN1O1_L[8] liq SER SER GMn3+ :O2- 2 H Mn 3H O = GMN2O3_L[8] liq SER SER GMn 2+ :O2- 2 H Mn 2H O = 2GMN1O1_L[8]
0 liq Cr 2+ :O2- ,Va q-
L L
= 121000[9] = 129519[8] = 45459[8] = 33859 [8] = 15009 + 13.6587T = 504 + 0.9479T = 188487.6
0 liq Mn 2+ :O 2- ,Va q1 liq Mn 2+ :O 2- ,Va q-
0 liq Mn 2+ ,Mn 3+ :O 2- ,Va q-
L L
0 liq Cr 2+ ,Mn 2+ :Va q1 liq Cr 2+ ,Mn 2+ :Va q-
0 liq Cr 3+ ,Mn 3+ :O 2-
Bcc A2 alloy (bcc)

(Mn,Cr)1 (Va,O)3
90
bcc SER GCr:Va H Cr = GHSERCR [7] bcc SER GMn:Va H Mn = GHSERMN[7] bcc SER SER GCr:O H Cr 3H O = GHSERCR [7] + 3GHSEROO[7] + 243T bcc SER SER GMn:O H Mn 3H O = GHSERMN[7] + 3GHSEROO[7]
0 bcc Cr,Mn:Va
= 20328 + 18.7339T [10] = 9162 + 4.4183T [10]
1 bcc Cr,Mn:Va
0 bcc Cr:O,Va
= 709542[9]
p = 0.4 Tcbcc = 311.5 yCr 580 yMn + yCr yMn [1325 1133( yCr yMn ) 2 10294( yCr yMn ) 4 + 26706( yCr yMn ) 6 28.117( yCr yMn )8 ][10]
bcc
= 0.008 yCr 0.27 yMn + yCr yMn [0.48643 0.72035( yCr yMn ) 2 1.93265( yCr yMn ) 4 ][10]
Manganosite (mgs)
(Mn 2+ ,Mn 3+ , Cr 3+ ,Va)1 (O 2- )1
mgs SER SER GMn 2+ :O 2- H Mn H O = GMN1O1[8] mgs SER SER GCr 3+ :O 2- H Cr H O = 0.5GCR2O3[9] + 71549.3 7.93845T mgs SER SER GMn 3+ :O2- H Mn H O = GMN1O1[8] 21883.5213 22.1853365T
0 mgs Mn 2+ ,Mn 3+ :O 2-
= 42104.8766[8] = 46513.1533[8]
1 mgs Mn 2+ ,Mn 3+ :O 2-
Bixbyite (bxb)
(Mn 3+ ,Cr 3+ ) 2 (O 2- )3
bxb SER SER GMn 3+ :O2- 2 H Mn 3H O = GMN2O3[8] bxb SER SER GCr 3+ :O 2- 2 H Cr 3H O = GCR2O3[9] + 3459
Eskolaite (esk)
(Cr 2+ , Cr 3+ ,Mn 3+ ) 2 (Cr 3+ ,Va)1 (O 2- )3
esk SER SER GCr 3+ :Va:O 2- 2 H Cr 3H O = GCR2O3[9] esk SER SER GCr 3+ :Cr3+ :O 2- 3H Cr 3H O = GCR2O3[9] + GHSERCR [7] esk SER SER GCr 2+ :Cr3+ :O 2- 3H Cr 3H O = GCRO0[9] + 1 GHSERCR [7] 5.2923T 3 esk SER SER GCr 2+ :Va:O 2- 2 H Cr 3H O = GCRO0[9] 2 GHSERCR [7] 5.2923T 3 esk SER SER SER GMn 3+ :Cr3+ :O2- 2 H Mn H Cr 3H O = GMN2O3[8] + GHSERCR [7] + 39503 esk SER SER GMn 3+ :Va:O 2- 2 H Mn 3H O = GMN2O3[8] + 39503
Magnetic contribution[9] p = 0.28

esk Tcesk = 308.6 yCr3+ :Cr3+ :O2esk esk = 3 yCr :Cr
3+ 3+ 3+
:O 22-
esk c
= 308.6 y
esk Cr 3+ :Va:O 2-
esk esk = 3 yCr :Va:O esk esk = 3 yCr :Cr

2+ 2+ 3+
esk Tcesk = 308.6 yCr 2+ :Cr3+ :O2esk Tcesk = 308.6 yCr 2+ :Va:O2-
:O 22-
esk esk = 3 yCr :Va:O
Cubic spinel (-spl)
91
(Mn 2+ ,Cr 2+ )1 (Cr 3+ ,Mn 3+ ) 2 (O 2- ) 4

-spl SER SER SER GMn 2+ :Cr 3+ :O2- H Mn 2 H Cr 4H O = GSPINEL -spl SER SER GCr 2+ :Cr 3+ :O 2- 3H Cr 4H O = GCR3O4[9] -spl SER SER GMn 2+ :Mn 3+ :O 2- 3H Mn 4H O = GMN3O4B[8] -spl SER SER SER GCr 2+ :Mn 3+ :O 2- H Cr 2 H Mn 4H O = GCR3O4+GMN3O4B[8] -GSPINEL
Tetragonally distorted spinel (-spl)

(Mn 2+ )1 (Cr 3+ ,Mn 3+ ) 2 (O 2- ) 4
-spl SER SER SER GMn 2+ :Cr 3+ :O 2- H Mn 2 H Cr 4H O = GTSPINEL -spl SER SER GMn 2+ :Mn 3+ :O 2- 3H Mn 4H O = GMN3O4[8]
Functions
GSPINEL = 2 GCR3O4+ 1 GMN3O4B[8] 3 3 210795.5+61.69T GTSPINEL = 2 GCR3O4+ 1 GMN3O4[8] 3 3 200941.9+75.1T a)
Note: All parameters are in SI units: J, mol, K, Pa: R = 8.31451 J mol-1 K-1.
The optimization of the thermodynamic parameters is performed using the PARROT module of the Thermo Calc[40] database system. PARROT takes into account all sorts of thermodynamic and phase diagram data simultaneously. The program minimizes the sum of squared errors between the experimentally determined phase diagram and thermodynamic data and the corresponding calculated data. As the use of all the experimental data in a simultaneous least square calculation often leads to divergence, we selectively adjust the relative weight of each experimental data point and exclude data that are inconsistent with the majority of the data points during the optimization procedure.
-spl To optimize the parameters A-spl and B-spl in Eq. 4.2.4 we use the f GMnCr2O4 value derived
from Tsai and Muan[23] using f GMnAl2O4 from Kim and McLean[25] and the composition of bcc in equilibrium with -spl and mgs reported by Ranganathan and Hajra[19]. Further the melting temperature of -spl in air found by Speidel and Muan[12] is used to optimize A-spl and B-spl in Eq. 4.2.4 and 0 Lliq3+ ,Mn3+ :O2- . The temperature found from Speidel and Muan[12] for Cr the two phase equilibrium of Mn rich -spl (Mn = 94.7 cat.%) and -spl is used to optimize
A-spl and B-spl in Eq. 4.2.4 and A-spl and B-spl in Eq. 4.2.4. All these data are given a high
weight. Further we use with lower weights the temperature of the two phase equilibrium -spl + liq at X(Cr) = 0.105 from Speidel and Muan[12] to optimize A-spl and B-spl in Eq. 4.2.4 and
0 liq
Cr3+ ,Mn3+ :O2-
, and we use data on the solubility of Cr in MnyCr3-yO4 at T=1873 K under varying
92
oxygen partial pressures from Tanahashi et al.[17] shown in Fig. 4.2.5, p. 82, and data from Pollert et al.[14] on the two phase equilibrium -spl + bxb in air (Fig. 4.2.3, p. 80) to optimize
A-spl and B-spl in Eq. 4.2.4. Temperature data from Speidel and Muan[12] and Pollert et al.[14]
on the phase equilibria -spl + -spl and -spl + -spl + bxb are used to optimize A-spl and
B-spl in Eq. 4.2.4, and A-spl and B-spl in Eq. 4.2.6. We use data from Holba et al.[11] on the
temperature dependence of the diffusionless transformation of -spl to -spl shown in Fig. 4.2.2, p. 79 to optimize A-spl and B-spl in Eq. 4.2.4, and A-spl and B-spl in Eq. 4.2.6. Abxb in Eq. 4.2.7 is optimized using data on the solubility of Cr in Mn2-yCryO3 from Pollert[14], for
Amgs in Eq. 4.2.8 data on the solubility of Cr in (Mn1-yCry)1-xO from Tanahashi et al.[17] are
used (Fig. 4.2.5, p. 82), and for Aesk in Eq. 4.2.9 and 4.2.10 data on the solubility of Mn in (Cr1-yMny)2+xO3 from Pollert et al.[14] shown in Fig. 4.2.2, p. 79 and Fig. 4.2.5, p. 82 are used.
4.2.5
Results
Phase diagram data:
The calculated phase diagram of the pseudo-binary system MnOx-Cr2O3 in air is shown in Figs. 4.2.1 and 4.2.2, p. 79. In Fig. 4.2.3, p. 80 the Mn rich part of the diagram is shown in detail. -spl is stable in a large temperature range from T = 513 K to 2243 K and from X(Cr) = 0 to X(Cr) = 0.67 in air. -spl and esk coexist from X(Cr) = 0.66 to 0.992 at 1203 K in air. The maximum Mn solubility in (Cr1-yMny)2+xO3 is 0.2 cat.% at 2243 K in air. Single phase -spl is stable in a small TX(Cr) range from T = 1153 K to 1441 K and X(Cr) = 0 to 0.054 in air. spl coexists with -spl from X(Cr) = 0 at T = 1441 K to X(Cr) = 0.175 at T = 1156 K. The dotted line in Fig. 4.2.2, p. 79 shows the temperature dependence of the diffusionless transformation of -spl to -spl. bxb is stable from T = 694 K to 1154 K in air. The maximum Cr solubility in Mn2-yCryO3 is 23 cat.% at T = 668 K in air. The two-phase field bxb + -spl is only found in a very small area at about 1150 K. prl coexists with esk at T < 513 K in air. From T = 513 K to 668 K and maximum X(Cr) = 0.65 prl and -spl coexist in air. prl + bxb is stable in a small area from T = 668 K to 694 K and X(Cr) = 0 to 0.23 in air. mgs is not stable in air. At pO2 = 400 Pa it starts to form in equilibrium with -spl in a small area at the Mn-rich side of the MnOx-Cr2O3 system around T = 1840 K. This two-phase field expands under more reducing conditions which can be seen in the calculated phase diagram of
93
the pseudo-binary MnOx-Cr2O3 system at an oxygen partial pressure of 110-4 Pa in Fig. 4.2.4, p. 80. In Fig. 4.2.5, p. 82 experimental and calculated solubility data of Cr in mgs + -spl, and Mn in esk + -spl are presented. Experimental data from Tanahashi et al.[17] on the solubility of Cr in the phases of the two phase equilibria mgs + -spl and esk + -spl at 1873 K, and from Bobov et al.[18] on the solubility of Cr in the phases of the two phase equilibria mgs + -spl at
T = 1073 K, 1173 K, and 1273 K in function of log(pO2 ) are compared to the calculated
results from this work. The data from Bobov et al.[18] were not used for the optimization. In the isothermal phase diagram of the Mn-Cr-O system at T = 1323 K of Fig. 4.2.6, p. 83 the stable alloy phases of the system are plotted in addition to the oxides based on the assessment of the binary Cr-Mn system from Lee[9]. Isothermal sections of the Cr2O3-MnO-MnO2 system at different temperatures are plotted in Fig. 4.2.9 (next page).
94
Fig. 4.2.9 Isothermal sections of the ternary system Cr2O3-MnO-MnO2 showing oxide and
liquid evolution as a function of temperature and composition. Stoichiometric single phase equilibria are points, and single solid solution phase equilibria are heavy lines. Dotted lines are tielines. Three-phase field boundaries are denoted with thin lines.
95
Thermodynamic data:
The calculated enthalpy, entropy and Gibbs energy of formation of -spl of the composition
-spl -spl MnCr2O4 from the elements is f H MnCr2O4 = 1599421 J mol-1, SMnCr2O4 = 116 J mol-1 K-1, and -spl f GMnCr2O4 = 1634017 J mol at T = 298.15 K. The calculated temperature dependence of the
-1
Gibbs energy of formation of of -spl of the composition MnCr2O4 from the oxides MnO and
-spl Cr2O3 is shown in Fig. 4.2.7, p. 85. At T = 1873 K we get f GMnCr2O4 = 34388 J mol-1. In the -spl temperature range from 1050 to 1800 K f GMnCr2O4 from the oxides is given by the term
89167 + 29.338 T with an error of 0.21 %. For -spl of the composition MnCr2O4 we calculate f H -spl = 1596517 J mol-1,
-spl
= 98 J mol-1K-1, and f G -spl = 1625681 J mol-1 at T = 298.15 K.
4.2.6
Discussion
Phase diagram data:
Our assessed phase diagram is in rough agreement with the findings from Speidel and Muan[12]. Large deviations of our calculated phase diagram from the phase diagram presented by these authors concern the stability of the liquid and phase stabilities at low temperatures. For both cases Speidel and Muan[12] mention the speculative character of their phase diagram due to the lack of experimental data. Our assessed phase diagram is in excellent agreement with the findings of Pollert et al.[14,15] and Ranganathan et al.[19] as shown in Figs. 4.2.2 (p. 79), 4.2.3 (p. 80), and Fig. 4.2.6 (p. 83). Our calculated T0 line for the diffusionless transformation of -spl -spl is in perfect agreement with experiments by Holba et al.[11] (Fig. 4.2.2, p. 79). The calculated dependence of -spl solid solubility on oxygen partial pressures shown in Fig. 4.2.5, p. 82 agrees well with the results from Tanahashi et al.[17]. The results from Bobov et al.[18] on the other hand cannot be reproduced. Fig. 4.2.9 (p. 95) represents the phase relations of the oxides and the evolution of liquid formation. The three-phase regions prl + esk + bxb and bxb + esk + -spl, and the two-phase fields prl + bxb, -spl + bxb, and mgs + -spl dominate the system in a wide temperature range from T=1200 K to 1900 K (Figs. 4.2.9 a to c). At T=1700 K -spl is no longer stable and the three-phase regions mgs + -spl + -spl and bxb + -spl + -spl (Fig. 4.2.9 a) disappear. At this temperature small oxygen nonstoichiometry of esk is apparent. The oxygen nonstoichiometry of mgs is yet insignificant (Fig. 4.2.9 b), but it increases at elevated 96
temperatures becoming apparent at T = 2000 K (Fig. 4.2.9 d). At T = 1900 K phase relations become more complex due to incipient melting (Fig. 4.2.9 c). At T = 1900 K liquid is formed at the Mn-rich side of the system, and the-three phase fields bxb + -spl + liq and mgs + -spl + liq emerge (Fig. 4.2.9 c). Even more three phase regions due to increasing liquid formation start to exist at T = 2000 K (Fig. 4.2.9 d). At T = 2400 K the only remaining stable solid phases are esk and prl (Fig. 4.2.9 f).
Thermodynamic data:
-spl -spl Our calculated f GMnCr2O4 value at T = 1873 K is in agreement with the f GMnCr2O4 value spl derived from Tsai and Muan[22] using f GMnAl2O4 from Kim and McLean[25].
Our assessed H
and S
values for the transition of -hsm to -hsm compare
favorably with the values reported by Holba et al.[12]. The calculated H needed for the transformation to take place.
and S
values
for the transformation of -spl to -spl are very small, indicating that only very little energy is
4.2.7
Applications on SOFC
Due to the large stability range of -spl and esk in air it is not realistic to prevent the formation of these unwanted phases under oxidizing conditions at the cathode side of SOFC operated with high Cr alloy interconnects and LSM cathode. The composition of Crofer22 APU alloy is close to the Cr-corner of the Mn-Cr-O phase diagram. In a thermodynamic view the formation of -sp with the composition MnCr2O4 (Point A in Fig. 4.2.6, p. 84) on Crofer22 APU alloy is expected under SOFC operating conditions. Hence, the formation of a protective Cr2O3 single phase layer followed by a chromium manganese spinel on Mn bearing interconnects as it is observed by Simner et al.[5] must be kinetically controlled. The occurrence of other Cr-Mn phases in the protective scales formed during thermal exposure of Crofer 22 APU interconnects is not expected thermodynamically. This is obvious from Fig. 4.2.6, p. 83. Recently Qu et al.[42] found that the electrical conductivity of chromium manganese spinel increases with increasing Mn-content. The problem of the application of synthesized Mn-rich
-spl on the interconnect for the purpose of combining a decrease of Cr evaporation with
enhanced electrical conductivity is that -spl will with time tend towards its stable composition of MnCr2O4, which is at point A in Fig. 4.2.6, p. 83 associated with decreasing 97
electrical conductivity. Also -spl will then transform to -spl and on thermal cycling of the SOFC back to -spl leading to mechanical stresses that might result in the appearance of cracks.
4.2.8
Conclusions
Due to the lack of inversity and oxygen nonstoichiometry of spinel we chose a model description of -spl and -spl without the introduction of vacancies into the spinel structure. All features of the system are well described with the optimization of only 8 additional optimization parameters. There is a surprising lack of thermodynamic data on -spl, and the only available
-1 -spl f GMnCr2O4 values are spread over a range of 31 kJ mol . Recalculating old experiments using
new thermodynamic data together with phase diagram data we achieved a description, which is very close to the experimental findings of several authors, and we present well-founded
fH, S, and fG data for -spl and -spl.
Our Thermo Calc[40] dataset resulting from the presented CALPHAD modeling of the Mn-CrO system allows the calculation of phase stabilities, compositions and transformations of unwanted MnxCr3-xO4 spinel solid solution and eskolaite phases in solid oxide fuel cells under any desired temperature and oxygen partial pressure conditions.
References
1. K. Hilpert, W.J. Quadakkers, L. Singheiser, in: W. Vielstich, A. Lamm, H.A. Gasteiger (Eds.), Handbook of Fuel Cells Fundamentals, Technology and Applications, John Wiley & Sons, Chichester, 2003, p. 1037. 2. J.W. Fergus, Lanthanum chromite-based materials for solid oxide fuel cell interconnects, Solid State Ionics, 2004, 171, pp. 1-15. 3. D. Das, M. Miller, H. Nickel, K. Hilpert, in: U. Bossel (Ed.), First European Solid Oxide Fuel Cell Forum Proceedings, Vol. 2, Druckerei J. Kinzel, Gttingen, 1994, p. 703. 4. S.P.S. Badwal, R. Deller, K. Foger, Y. Ramprakash, J.P. Zhang, Interaction between chromia forming alloy interconnects and air electrode of solid oxide fuel cells, Solid State
Ionics, 1997, 99, pp. 297-310.
98
5. S.P. Simner, M.D. Anderson, G.-G. Xia, Z. Yang, L.R. Pederson, J.W. Stevenson, SOFC performance with Fe-Cr-Mn alloy interconnect, J. Electrochem. Soc., 2005, 152, pp. A740-45. 6. A. T. Dinsdale, SGTE Data for Pure Elements, Calphad, 1991, 15(4), pp. 317-425. 7. A.N. Grundy, B. Hallstedt, L.J. Gauckler, Assessment of the Mn-O system, J. Phase
Equilib., 2003, 24, pp. 21-39.

8. E. Povoden, A.N. Grundy, L.J. Gauckler, Thermodynamic reassessment of the Cr-O system in the framework of solid oxide fuel cell (SOFC) research, J. Phase Equilib. Diff., 2006, 27, pp. 353-62. 9. B-J. Lee, A thermodynamic evaluation of the Cr-Mn and Fe-Cr-Mn systems, Metall. Trans., 1993, A 24, pp. 1919-1933. 10. J.M. Hastings, L.M. Corliss, Magnetic structure of manganese chromite, Phys. Rev., 1962,
126, pp. 556-65.

11. P. Holba, M. Neviva, E. Pollert, Tetragonal distortion of spinel solid-solutions MnCr2O4Mn3O4, Mater. Res. Bull., 1975, 10, pp. 853-60. 12. D.H. Speidel, A. Muan, The system manganese oxide-Cr2O3 in air, J. Am. Ceram. Soc., 1963, 46, pp. 577-78. 13. Y.V. Golikov, V.F. Balakirev, Phase equilibrium diagram of the system Mn-Cr-O, J. Solid
State Chem., 1987, 71, pp. 562-65.

14. E. Pollert, M. Nevriva, J. Novak, Phase diagram of the Mn2O3-Cr2O3 system in air, Mater.
Res. Bull., 1980, 15, pp. 1453-56.

15. E. Pollert, M. Nevriva, J. Novak, Miscibility gap of MnxCr3-xO4 spinels, J. Phys. Chem.
Solids, 1977, 38, pp. 1145-47.

16. S. Geller, G.P. Espinosa, Magnetic and crystallographic transitions in Sc3+, Cr3+, and Ga3+ substituted Mn2O3, Phys. Rev., 1970, B 1, pp. 3763-69. 17. M. Tanahashi, N. Furuta, C. Yamauchi, T. Fujisawa, Phase equilibria of the MnO-SiO2CrOx system at 1873 K under controlled oxygen partial pressure, ISIJ Int., 2001, 41, pp. 1309-1315. 18. A.P. Bobov, A.G. Zalazinsky, V.F. Balakirev, Y.V. Golikov, G.I. Chufarov, Peculiarities of phase-diagram in the reduction of Me0.25Mn2.75O4 solid-solutions, Zh. Fiz. Khim., 1984,
58, pp. 750-751 (in Russian).

19. S. Ranganathan, J.P. Hajra, Alloy oxide equilibria in the Cr-Mn-O system, Bull. Mater.
Sci., 1987, 9, pp. 149-58.
99
20. M. Tanahashi, N. Furuta, T. Taniguchi, C. Yamauchi, T. Fujisawa, Standard Gibbs free energy of formation of MnO-saturated MnO.Cr2O3 solid solution at 1873 K, ISIJ Int., 2003, 43, pp. 7-13. 21. I. Barin: Thermochemical Data of Pure Substances, 2nd Ed., Parts I and II, VCH Verlagsgesellschaft mbH, Weinheim, 1993. 22. H.T. Tsai, A. Muan, Activity composition relations in FeCr2O4-FeAl2O4 and MnCr2O4MnAl2O4 solid-solutions at 1500C and 1600 C, J. Am. Ceram. Soc., 1992, 75, pp. 1407-11. 23. H.T. Tsai, A. Muan, Activity composition relations in refractory oxide solid-solutions at high-temperatures the system Cr2O3-Al2O3, J. Am. Ceram. Soc., 1992, 75, pp. 1412-15. 24. L.M. Lenev, I.A. Novokhatskiy, Phase diagram of system MnO-Al2O3 and thermodynamic properties of MnAl2O4, Izv. Akad. Nauk SSSR, Met., 1966, 3, pp. 73- (in Russian). 25. C.K. Kim, A. McLean, Thermodynamics of iron-manganese aluminate spinel inclusions in steel, Metall. Trans. B, 1979, 10B, pp. 575-84. 26. K.T. Jacob, Revision of thermodynamic data on MnO-Al2O3 melts, Can. Metall. Q., 1981,
20, pp. 89-92.

27. S. Dimitrov, A. Weyl, D. Janke, Control of the manganese-oxygen reaction in pure iron melts, Steel Res., 1995, 66, pp. 87-92. 28. J.V. Biggers: Ph.D. Thesis, Pennsylvania State University, University Park, PA, 1966. 29. N.Y. Toker, L.S. Darken, A. Muan, Equilibrium phase-relations and thermodynamics of the Cr-O system in the temperature-range of 1500C to 1825C, Metall. Trans. B, 1991,
22, pp. 225-232.

30. R.K.F. Lam: United States Patent 6039788, 2000. 31. J.-O. Andersson, A.F. Guillermet, M. Hillert, B. Jansson, B. Sundman, A CompoundEnergy Model of Ordering in a Phase with Sites of Different Coordination Numbers, Acta
Metall., 1986, 34, pp. 437-445.

32. M. Hillert, B. Jansson, B. Sundman, Application of the Compound-Energy Model to Oxide Systems, Z. Metallkd., 1988, 79(2), pp. 81-87. 33. M. Hillert, The Compound Energy Formalism, J. Alloy. Cmpd., 2001, 320, pp. 161-76. 34. Z. Lu, J. Zhu, E.A. Payzant, M.P. Paranthaman, Electrical conductivity of the manganese chromite spinel solid solution, J. Am. Ceram. Soc., 2005, 88, pp. 1050-53. 35. R.D. Shannon, Revised effective ionic-radii and systematic studies of interatomic distances in halides and chalcogenides, Acta Crystallogr., 1976, A 32, pp. 751-67. 100
36. M. OKeefe, M. Valigi, The electrical properties and defect structure of pure and chromium-doped MnO, J. Phys. Chem. Solids, 1970, 31, pp. 947-962. 37. D. Caplan, M.J. Fraser, A.A. Burr, in: Ductile Chromium, ASM, Cleveland, Ohio, 1957, p. 196. 38. M. Hillert, B. Jansson, B. Sundman, J. gren, A Two-Sublattice Model of Molten Solutions with Different Tendency of Ionization, Metall. Trans. A, 1985, 16A, pp. 261-66. 39. B. Sundman, Modification of the Two-sublattice Model for Liquids, Calphad, 1991, 15, pp. 109-19. 40. B. Sundman, B. Jansson, J.-O. Andersson, The Thermo-Calc Databank System, Calphad, 1985, 9(2), pp. 153-90. 41. V.M. Eremenko, G.M. Lukashenko, V.R. Sidorko, Thermodynamic properties of alloys of manganese with transition elements of fourth period (Cr Fe Co Ni) and with copper, Russ.
J. Phys. Chem., 1968, 42, pp. 343-.

42. W. Qu, L. Jian, J.M. Hill, D.G. Ivey, Electrical and microstructural characterization of spinel phases as potential coatings for SOFC metallic interconnects, J. Power sources, 2006, 153, pp. 114-24.
4.3
Thermodynamic assessment of the La-Cr-O system
E. Povoden, M. Chen , A.N. Grundy, T. Ivas, and L.J. Gauckler
J. Phase Equilib. Diff. (accepted)

The La-Cr and the La-Cr-O systems are assessed using the Calphad approach. The calculated La-Cr phase diagram as well as LaO1.5-CrO1.5 phase diagrams in pure oxygen, air, and under reducing conditions are presented. Phase equilibria of the La-Cr-O system are calculated at T = 1273 K as a function of oxygen partial pressure. In the La-Cr system reported solubility of lanthanum in bcc chromium is considered in the modeling. In the La-Cr-O system the Gibbs energy functions of La2CrO6, La2(CrO4)3, and perovskitestructured LaCrO3 are presented, and oxygen solubilities in bcc and fcc metals are modeled. Emphasis is placed on a detailed description of the perovskite phase: the orthorhombic to rhombohedral transformation and the contribution to the Gibbs energy due to a magnetic order-disorder transition are considered in the model. The following standard data of
101
stoichiometric
perovskite
are The
calculated: f,oxides H 298 K (LaCrO3 ) = 73.7 kJ mol-1 , Gibbs energy of formation from the
and S 298 K (LaCrO3 ) = 109.2 J mol-1 K -1 .
oxides, f,oxides G(LaCrO3 ) = 72.4030.0034 T (kJ mol-1) (T = 1273 K to 2673 K) is calculated. The decomposition of the perovskite phase by the reaction
1 3 LaCrO3 La 2 O3 + Cr + O2 (g) is calculated as a function of temperature and oxygen partial 2 4
pressure: at 1273 K the oxygen partial pressure of the decomposition, pO2 (decomp) = 10-20.97 Pa. Cation nonstoichiometry of La1-xCrO3 perovskite is described using the compound energy formalism (CEF), and the model is submitted to a defect chemistry analysis. The liquid phase is modeled using the two-sublattice model for ionic liquids.
4.3.1
Introduction
In SOFC, the thermodynamic stability of the cathode is particularly important for efficient long-term operation. Sr-doped lanthanum manganites (LSM) with the perovskite structure are used as cathode materials in SOFC. Diffusion of chromium from the metallic interconnect with high chromium content into the cathode leads to the formation of Mn(Cr,Mn)2O4 spinel and Cr2O3 along with a severe cell voltage decrease[1-4]. As the thermal expansions of LaCrO3-based interconnect and conventional perovskite cathode materials are similar, and Crdiffusion into the cathode from LaCrO3-based interconnects is significantly lower than from Cr-containing metallic interconnects, recently Sr-, V-doped[5] and Zn-doped[6] La1-xCaxCrO3- have been considered as promising alternative interconnect materials for SOFC. Furthermore earth alkaline-containing LaCrO3 has been proposed as a cathode material in a recent study by Jiang et al[7]. The presented thermodynamic assessment of the La-Cr-O system is laying the grounding for extensions to the thermodynamic La-Sr-Mn-Cr-O oxide database that is required to understand the thermodynamics of SOFC degradation by chromium. It is also a starting point for extensions to thermodynamic databases with additional components serving as dopants in LaCrO3 for SOFC interconnect and cathode applications. The assessment of the La-Cr-O system using the Calphad approach is based on the recently reassessed La-O[8] and Cr-O subsystems[9]. The lattice stabilities of elements are adopted from Dinsdale[10]. All available experimental phase diagram, thermodynamic, and structurechemical data are critically assessed, aiming on minimizing the squared errors between
102
experiments and calculation during the optimization of model parameters using the PARROT module of the Thermocalc[11] software.
4.3.2
Literature review of the La-Cr system
The La-Cr system has a eutectic at T = 1138 K[12,13] and 3.4 at.% Cr[13] and a monotectic at
T = 1983 K[12] or T = 2103 K[14] and 96 at.%[12] or 99.1 at.%[14] Cr, as well as a large liquidliquid miscibility gap[12,13]. No intermetallic phases were found in the La-Cr system[12,13]. Berezutskii et al.[15] determined the partial enthalpy of mixing in La-Cr liquid with infinite dilution of Cr, H Cr at T = 1700 K using high-temperature calorimetry. As small additions of rare-earth metals essentially increase the high-temperature corrosion resistance of chromium[16], modeling of the La-solubility in bcc-structured Cr, denoted as
Crss , is of technological interest. Small solubility of La in Crss was reported
[12,14,17]
, whereas
Cr is almost insoluble in La[13]. The solubility of La in Crss was determined in investigations by Savitskii et al.[12] from T = 1073 K up to the melting of Cr using metallographic and microhardness techniques to be 2.5 at.% at 1983 K. Svechnikov et al.[14] reported a La solubility of 0.68 at.% at T = 2103 K, and Epstein et al.[17] found La < 0.04 at.% in Crss at T = 1533 K. The solubility of La in Crss decreases towards lower temperatures.
4.3.3
Literature review of the La-Cr-O system
In the LaO1.5- CrO1.5 system two eutectics were found at 19 at.% Cr2O3 (T = 2243 K)[18] or 12 at.% Cr2O3 (T = 232320 K)[19], and at 84 at.% Cr2O3 (T = 2248 K)[18] or 80 at.% Cr2O3 (T = 247320 K)[19] in argon atmosphere on either side of the congruently melting perovskitestructured lanthanum chromite[18-20] (in this study oxides containing Cr(III) and Cr with higher valencies than three are denoted as chromite and chromate respectively). The melting temperature of lanthanum chromite in air, Tm(air) = 2773 K was determined by Fox[21] and by Coutures[20] using a thermal analysis technique described in more detail in earlier publications.[22-24] The melting temperature was measured with optical pyrometers. Tm(argon) = 2703 K was reported by Tresvjatskiy et al.[18], but in the graphic presentation of the phase diagram in the same paper Tm(argon) 2600 K, and the exact value of the oxygen partial pressure was not specified. Experimentally determined special points in the LaO1.5-CrO1.5 quasibinary system reveal a considerable spread. This is not surprising as experiments are complicated due to the high investigation temperatures and evaporation of predominantly Cr[25,26]. Furthermore deviations between the data from Tresvjatskiy et al.[18] and Berjoan[19]
103
may partly originate from differences of the oxygen partial pressure, which in both studies was not exactly specified. The peritectic phase diagram proposed by Cassedanne[27] is in gross conflict with the phase diagram data from the other groups. Experimental oxygen solubilities in pure Cr and La were considered in thermodynamic assessments by Povoden et al.[9] and Grundy et al. solubilities in Crss are missing.
[28]
, whereas experiments on oxygen
Lanthanum chromates:
The following lanthanum chromates were documented: Berjoan[19] reported that orthorhombic La2CrO6 forms at T > 923 K. Using differential scanning calorimetry (DSC) he determined the enthalpy change of the reaction
3 2La 2 O3 + Cr2 O3 + O2(g) 2La 2 CrO6 2
(4.3.1)
at T=1055 K and pO2 =83000 Pa to be 1518 kJ mol-1. The enthalpy of formation of La2(CrO4)3 from the elements at T = 298 K was proposed by Tsyrenova et al.[29] to be 396111.7 kJ mol-1. La2(CrO4)3 decomposes by
890 1030 K La 2 (CrO4 )3 2LaCrO3 + 0.5Cr2 O3 + 2.25O2(g)
(4.3.2)
An enthalpy change of 231 kJ mol-1 was determined for this reaction at the average temperature of T = 960 K.[30] LaCrO4 has been interpreted as a mixed-valent intermediate decomposition product of La2(CrO4)3[30,31]. Stoichiometries and thermal stability ranges of lanthanum chromates with complex formulae were reported by Berjoan et al.[32]. However these were in significant disagreement with later results obtained by the same author[19].
The perovskite phase:
Existing
experimental
data
of
lanthanum
chromite
perovskite
structure[33-45],
thermodynamics[30,33-35,43,46-53], phase stability[54], and nonstoichiometry[55-56] along with the investigation techniques used are listed in Table 4.3.1 (next page).
104
Table 4.3.1 Calculated and experimental thermodynamic data of La-Cr oxides

Rhombohedral LaCrO3 Standard enthalpy f,elements H 298K f,oxides H 298K f,oxides H 975K
LaCrO3
= 1368.2 kJmol-1 this work, calculated = 73.7 kJmol-1 this work, calculated = 62.35 kJmol-1 this work, calculated = 73.06 2.79 kJmol-1 Drop solution calorimetry in 2PbO B2 O3[46]
LaCrO3 LaCrO3
f,oxides H 1078K
LaCrO3
Standard entropy S 298K

LaCrO3
= 109.2 Jmol-1K -1 this work, calculated
Gibbs energy of formation by 1 3 La 2 O3 + Cr2 O3 LaCrO3 2 4 T = 1273 K G = 76.75 kJmol1 this work, calculated T = 1273 K T = 1273 K T = 2100 K T = 2100 K
G = 30.1 1.5 kJmol 1[49] solid oxide electrolyte - emf G = 42.29 0.38 kJmol 1[52] CaF2 - based emf G = 79.52 kJmol 1 this work, calculated G = 78.9 1.1 kJmol 1[53] Knudsen mass spectrometry
G = 72.403 0.0034T (kJmol1 ), 1273 - 2673 K this work, calculated G = 44.45 + 0.002115T 0.4(kJmol 1 ), 855 -1073 K[50] CaF2 - based emf G = 94.758 + 0.08530T (kJmol1 ), 700 - 885 K[51] CaF2 - based emf
Enthalpy increments H H 298 K , (kJmol1 ) T = 1090 K 98.19, this work, calculated 94.4[30] HT (high temperature) - calorimetry T = 1350 K 133.05 this work, calculated 139.2[30] HT - calorimetry Activity of Cr2 O3 in LaCrO3 T = 2100 K T = 2100 K La 2 CrO6 Enthalpy of the formation reaction La 2 O3 + 0.5Cr2 O3 +1.5O 2(g) La 2 CrO6
2 f,oxides H 298K
aCr2O3 = 1.11 104 this work, calculated aCr2O3 = 1.1 104 1.1 105[53] Knudsen mass spectrometry
La CrO6
= 73.0 kJmol1 this work, calculated
2 S 298K
La CrO6
= 330 Jmol-1K -1 this work, calculated

La (CrO4 )3 La (CrO4 )3
La 2 (CrO 4 )3
2 f,elements H 298K
= 3845 kJmol1 this work, calculated = 3961 11.7 kJmol1 , calculated[29]
2 f,elements H 298K
2 S 298K
La (CrO4 )3
= 516 Jmol1K 1 this work, calculated
Enthalpy of the formation reaction La 2 O3 +1.5Cr2 O3 + 2.25O 2(g) La 2 (CrO 4 )3

2 f,oxides H 298K
La (CrO4 )3
= 372 kJmol1 this work, calculated
Enthalpy of the dissociation reaction La 2 (CrO 4 )3 2LaCrO3 + 0.5Cr2 O3 + 2.25O 2(g)

2 H 298K
La (CrO4 )3
= 231 kJmol1[30] and this work, fitted
105
Crystal and magnetic structure: LaCrO3 is orthorhombic (o-prv) at room temperature and transforms to rhombohedral structure (r-prv) at higher temperatures[20,33-42]. The temperatures, enthalpy and entropy changes of this first-order[44] transition taken from the literature are listed in Table 4.3.2 along with the investigation techniques used.
Table 4.3.2 Calculated and experimental data of the orthorhombic to rhombohedral transition of LaCrO3
Transition temperature (K) 540, this work, calculated 503 583[33] adiabatic calorimetry 544 1[34] a) DTA, DSC, thermogravimetry, dilatometry 536 [35] a) adiabatic shield calorimetry, HT - XRD (air and vacuum) 563 5[36] DTA, dilatometry, HT - XRD, HT - microscopy, HT - x - ray photography 550[37] HT - XRD 528 533[38] a) HT - XRD 533 3[38] a) DTA 543[39] XRD 533[20] HT - XRD 540 2[40] a) HT - XRD, DSC 533 5[40] a) HT - XRD, dilatometry 545[41] heating, DSC 535[41] cooling, DSC 550[41] HT - XRD 523[42] a) starting transition, simultaneous DSC - XRD 541[42] a) completed transition, simultaneous DSC - XRD 533[43] estimated from neutron powder diffraction 509[44] DSC, XRD Enthalpy change of transition (Jmol1 ) 340, this work, calculated 502.08 41.84 at 503 583 K [33] calculated from adiabatic calorimetry 277 at 544 1 K[34] a) DSC 403.25 at 536 K[35] a) calculated from adiabatic shield calorimetry 340 (10 - 40) at 533 5[40] a) DSC 380 at 550 K[41] DSC 310 at 509 K[44] DSC Entropy change of transition (Jmol-1K 1 ) 0.63, this work, calculated 0.96 at 503 583K[33] calculated from adiabatic calorimetry 0.5[34] a) calculated from DSC 0.75[35] a) calculated from adiabatic shield calorimetry
a)
used for optimization
The reported transformation temperatures lie between T = 503 K and 583 K. The determined enthalpy and entropy changes vary from 277 J mol-1 to 502.08 J mol-1 and 0.5 J mol-1 to 106
0.96 J mol-1 K-1. A transformation from rhombohedral to cubic structure at a temperature close to T = 1300 K was reported by Ruiz et al.[37] and Momin et al.[41], whereas Coutures et al.[20] reported T = 1923 K using high-temperature x-ray diffraction (HT-XRD), in agreement with Berjoan[19] (T = 1923 20 K) using dilatometry. Berjoan[19] further reported prevailing of the cubic structure at T = 2173 K. On the other hand Geller and Raccah[38] as well as Hfer and Kock[34] did not observe the rhombohedral to cubic transition up to T = 1873 K and
T = 1823 K respectively using differential thermal analysis (DTA).

A magnetic order-disorder transition was documented to occur at T 287 K[35], 289 K[45], or 295 K[46]. Enthalpy of formation: Cheng and Navrotsky[47] determined the enthalpy of formation of LaCrO3 by oxide melt solution calorimetry at T = 1078 K. Heat capacity and enthalpy increment data: the heat capacities of LaCrO3 were measured by Korobeinikova and Reznitskii[33] from T = 340 K to 900 K using adiabatic calorimetry, Hfer and Kock[34] (480 to 610 K) and Satoh et al.[45] (150 to 450 K) using DSC, Satoh et al.[45] (T = 77 K to 280 K) using alternating current calorimetry, Sakai et al.[35] (T=100 K to 1000 K) using laser-flash calorimetry, and Sakai and Stlen[43] (T = 272 K to 1000 K) using adiabatic shield calorimetry. Enthalpy increments of LaCrO3 at T = 1090 K and 1350 K were measured by Suponitskii[30] using a high-temperature heat-conducting calorimeter. Gibbs energy of formation: in order to obtain the Gibbs energy of formation of LaCrO3, Chen
et al.[49] measured electromotive force (emf) of the solid oxide galvanic cell Pt/Cr, La2O3,
LaCrO3/MgO-stabilized ZrO2/Cr2O3, Cr/Pt at 1273 K. Azad et al.[50], Chen et al.[51], and Dudek et al.[52] measured emf of Pt, O2/La2O3, LaF3/CaF2/LaF3, LaCrO3, Cr2O3/O2, Pt in pure oxygen from T = 855 K to 1073 K, T = 700 to 885 K, and T = 1273 K respectively. Peck et
al.[53] derived the Gibbs energy of formation of LaCrO3 from the determination of the
thermodynamic activity of Cr2O3 in LaCrO3 for the Cr2O3-poor phase boundary of LaCrO3 in the temperature range from T = 1887 K to 2333 K using Knudsen effusion mass spectrometry. Chemical stability: Nakamura et al.[54] reported no weight loss of lanthanum chromite at
T=1273 K from pure oxygen to pO2 =10-16.1 Pa using thermogravimetry combined with X-ray
diffraction (XRD). This means that the perovskite phase does not decompose within this oxygen partial pressure range, and its oxgen nonstoichiometry is negligible. Cation nonstoichiometry and defect chemistry: Maximum excess Cr in single-phase La1-xCrO3 of 0.38 cat.% in furnace-cooled LaCrO3 annealed at T = 1773 K in air was reported from Khattak and Cox[55]. Single phase lanthanum chromite with 0.76 cat.% to 1.28 cat.% excess
107
Cr was prepared at T = 1773 K in a pure oxygen atmosphere[56]. Iliev et al.[56] observed an intensity decrease of the high frequency band in a Raman spectrum of lanthanum chromite measured after annealing the phase in vacuum at T = 1273 K. This feature was assigned to a reduced number of Cr4+ due to partial removal of oxygen during the annealing of the originally lanthanum-deficient perovskite phase. Interpretations of the defect chemistry of the perovskite phase were made from electrical conductivity measurements: the electrical conduction in lanthanum chromite is almost purely electronic[37,57], affirming the lack of oxygen vacancies in the structure, in line with the results from thermogravimetry[54]. Ruiz et al.[37] reported that the ionic transport number in lanthanum chromite is less than 0.05 % up to T = 1250 K. Akashi et al.[58] measured the isothermal electrical conductivity of an equilibrated La1-xCrO3-Cr2O3 mixture with 5 vol.% excess Cr2O3 from T = 1573 K to 1673 K from pO2 = 1.0 103 Pa to pO2 = 2.0 104 Pa . They observed an extraordinarily slow equilibration of the samples: More than four months were required to measure the electrical conductivity at equilibrium state. The conductivity was proportional to pO2 3 16 , the same as reported in an earlier study[25]. On the other hand a slope of pO2 1 4 from T=700 K to 1300 K and purely intrinsic conductivity > 1600 K stated by Shvaiko-Shvaikovskii et al.[57] is inconsistent with the findings from Akashi et al.[58] ShvaikoShvaikovskii et al.[57] deduced n-type conductivity from measurements of transport number, resistivity and thermo-emf at pO2 = 1Pa and pO = 102 Pa , the electrical conductivity being
2
proportional to pO2 3 8 . The transition from reduced to stoichiometric chromite was accompanied by a decrease of about 0.1% in weight, thus the presence of interstitial Cr in reduced chromite was proposed. However n-type conductivity was not approved by any further study. Several groups[58,59] agree that the electrical neutrality is maintained by holes and lanthanum vacancies, and that the carrier is the hole in lanthanum chromite[25,58-60]. Akashi et al.[58] reported that concentrations of lanthanum vacancies and holes slightly increase from T = 1550 K to 1700 K. In contrast to the other authors Shvaiko-Shvaikovskii et al.[57] and Meadowcroft[25] proposed the occurrence of chromium vacancies instead of lanthanum vacancies.
108
4.3.4
Thermodynamic modeling and optimization
Metal phases:
In order to account for the solubility of La in Crss , the zeroth-order, compositionindependent interaction parameter[61] 0 Lbcc Cr,La:Va was given a positive value. We chose the solubility values from Svechnikov et al.[14] for its optimization, as these data are more comparable to solubilities in other rare earths-transition elements systems. Povoden et al.[9] described the solubility of oxygen in Cr(bcc) using the model Cr(Va,O)3. For the reasons discussed recently[62], we reassess the oxygen-solubility in Cr(bcc) using the model (Cr)(O,Va)1.5, and
Crss
is then given by the two-sublattice description
(La,Cr)(Va,O)1.5. The Gibbs energy of the end-member (Cr)(O)1.5 is defined as

3 2 3 gas[10] G + A + BT 4 O2
SER SER G(Cr)(O)1.5 H Cr H O = GCr(bcc)[10] +
(4.3.3)
SER H x is the standard enthalpy of the stable state of element x at 298.15 K and 10 Pa.
[10]
A and
B are adjustable parameters; using the PARROT module of the Thermocalc software[11] A was
given the fix value 0 for the reasons discussed in an earlier paper[9], and B and a regular interaction parameter 0 LCr:O,Va were optimized with the same experimental data[9]. Due to the lack of experimental data the oxygen solubility in Crss was modeled as an ideal extension of the oxygen solubilities in pure La and Cr.
Solid oxides:
Lanthanum chromates: The Gibbs energy function of La2CrO6 was based on the sum of the Gibbs energy functions of La2O3, Cr2O3, and O2 in proper stoichiometries and A + BT parameters that were fitted to the enthalpy of formation from the oxides, Eq. 4.3.1 as well as thermal stability data. The thermal stability of La2CrO6 is slightly influenced by the thermodynamics of the intermediate, mixedvalent chromates mentioned above. In order to refine the model parameters of La2CrO6, it was thus necessary to consider these mixed-valent chromates in a provisional version of the thermodynamic La-Cr-O database in spite of their arguable stoichiometries, and to optimize their model parameters with phase diagram data[19,32]. The formation of chromates that contain mixed Cr valences may be explained by gradual reduction of Cr6+ in La2CrO6 as the temperature increases. These chromates can be interpreted as intermediate products in the
109
scope of a sluggish decomposition of La2CrO6, which starts at T = 1153 K[19,32] and is completed at T = 1473 K[32] or 1523 K[19]. The simplified decomposition reaction reads
1+ x 1.5 - (1.5 x) La 2 O3 + La1- x CrO3 + O2 (g) 2 2
mixed-valent chromates La 2 CrO6
[19,32]
(4.3.4)
Slight differences of the oxygen partial pressure during experiments may be reflected by a variable extent of Cr-reduction, and consequently ambiguous stoichiometries of mixed-valent intermediate chromates. These lanthanum chromates with conflicting stoichiometries[19,32] are not included in the presented thermodynamic database. The Gibbs energy function of La2(CrO4)3 was formulated using the same strategy as for La2CrO6. The model parameters were fitted to the experimental enthalpy and temperature of decomposition[30]. The enthalpy of formation from the elements[29] was not used as it is a calculated value. We go along with the interpretation of LaCrO4 being an intermediate reaction product during the decomposition of La2(CrO4)3 by Eq. 4.3.2 and do not include this phase in the modeling. The perovskite phase: The following denotations are used in this section: the superscript prv is written if the regarding Gibbs energy expression is the same for both orthorhombic and rhombohedral perovskite. The superscripts o-prv and r-prv stand for Gibbs energy expressions that have different values for orthorhombic and rhombohedral perovskite. GRPRV denotes the Gibbs energy function for stoichiometric rhombohedral perovskite. GVCR4O and GLCR4O stand for the Gibbs energy functions of the completely oxidized neutral endmember. GVCR4O and GLCR4O are set equal for orthorhombic and rhombohedral perovskite. Stoichiometric perovskite: The Gibbs energy function of stoichiometric rhombohedral
r-prv LaCrO3 with the sublattice formula (La3+)(Cr3+)(O2-), GLaCrO3 is given by
r-prv SER SER SER GLaCrO3 HLa H Cr 3H O = GLa3+ :Cr3+ :O2 = GRPRV
1 1 GCr2O3 [9] + GLa2O3 [8] + Gmag + A + BT + CT ln T 2 2
(4.3.5)
The parameters A, B, and C are optimized using the enthalpy of formation from Cheng and Navrotsky[47], activity-data of Cr2O3 in LaCrO3 from Peck et al.[53], heat capacity-data obtained by adiabatic calorimetry from Sakai and Stlen,[35] and enthalpy increment-data
110
measured at high temperatures[30]. A phase diagram with congruent melting of lanthanum chromite and two eutectics[18,19] cannot be reproduced by using the emf-experiments[49-52]. Thus these data were excluded from the optimization.
A+BT parameters of the low-temperature orthorhombic perovskite phase were optimized with
those temperatures[34,35,38,40,42], enthalpies[34,35,40] and entropies[34,35] of transition having been obtained by combined investigation techniques and being internally most consistent. The rhombohedral to cubic transformation at high temperatures is not considered in the model, as there is no existing thermodynamic data for this transition. Cation-nonstoichiometric perovskite: To choose a proper model for nonstoichiometric perovskite the following considerations are made: the formation of interstitial Cr in lanthanum chromite proposed by Shvaiko-Shvaikovskii et al.[57] is unlikely due to the densely-packed perovskite structure, and oxygen nonstoichiometry can be excluded from thermogravimetry[54] and electrical conductivity[37,58] measurements. Thus the defects in n-type conducting[57] lanthanum chromite are ambiguous and were not considered in the model. B-site vacancies are energetically less favored than A-site vacancies in the perovskite structure[63,64]. This means that the simplest sublattice model to describe cation nonstoichiometric La1-xCrO3 reads (La3+,Va)(Cr3+,Cr4+)(O2-)3. While this model results in a satisfying reproduction of experimental data, irreconcilable trouble is encountered at the extension to the LaO1.5-MnO1.5-CrO1.5 system required for SOFC applications due to diversities between the model descriptions of lanthanum chromite and lanthanum manganite[65]. These are solved by allowing Va on the B-site and the anion sublattice of lanthanum chromite just like in lanthanum manganite[65] leading to the appropriate sublattice formula (La3+,Va)(Cr3+,Cr4+,Va)(O2-,Va)3. The optimization of selective model parameters listed in Table 4.3.3 (pp. 112-114) resulted in negligible concentrations of Va on the B-site and the anion sublattice, and the perovskite formula essentially remains La1-xCrO3.
111
Table 4.3.3 Model descriptions and Gibbs energy functionsa)
Liquid (liq) (La 3+ , Cr 2+ , Cr 3+ ) p (O 2- , Va q- )q p = 2 yO2- + qyVa , q = 3 yLa3+ + 2 yCr 2+ + 3 yCr3+

liq SER GLa3+:Vaq- HLa = GLALIQ[8] liq SER SER GLa3+:O2- 2 HLa 3H O = GLA2O3LIQ[8] liq SER GCr 2+ :Vaq- H Cr = GCRLIQ[10] liq SER GCr3+:Vaq- H Cr = 2GCRLIQ[10] GCR2O3_L[9] 3GCR1O1_L[9] liq SER SER GCr3+:O2- 2 H Cr 3H O = 5GCROLIQ 179638 + 79.923T [9] liq SER SER GCr 2+ :O2- 2 H Cr 2H O = 2GCR1O1_L[9]
Interaction terms Lliq 2+ :O2- ,Vaq- = Lliq 3+ :O2- ,Vaq- = 121000[9] Cr Cr Lliq 2+ ,La3+ :Vaq- = 61397 5.23T + (65393 23T )( yCr 2+ yLa3+ ) Cr Lliq 2+ ,La3+ :O2- = 101850 39016( yCr 2+ yLa3+ ) Cr Lliq 3+,La3+:O2- = 101850 39016( yCr3+ yLa3+ ) Cr Bcc A2 phase (La, Cr)(Va, O)1.5

bcc SER GCr:Va HFe = GHSERCR [10] bcc SER GLa:Va HLa = GLABCC[10]
3 SER 3 bcc SER GCr:O H Cr H O = GHSERCR [10] + G (O 2(g) )[10] + 113.17755T 2 4 3 SER 3 SER bcc GLa:O HLa H O = GLABCC[10] + G (O 2(g) )[10] 855000 + 142.5T [28] 2 4 Lbcc Cr:Va,O = 355151.422
Lbcc Cr,La:Va = 83500 p = 0.4[9] Tcbcc = 311.5 yCr [9]
bcc = 0.008 yCr [9]
112
La1-xCrO3 perovskite (La 3+ ,Va)(Cr 4+ ,Cr3+ , Va)(O2- , Va)3

o-prv SER SER SER GLa3+ :Cr3+:O2- HLa HCr 3HO = GOPRV + Gmag r-prv SER SER SER GLa3+ :Cr3+:O2- HLa HCr 3HO = GRPRV + Gmag o-prv SER SER SER GLa3+ :Cr4+:O2- HLa HCr 3HO = 5 6 GS4O[72]
+GOPRV GS3V[72] + 1 6GS4V[72] + Gmag
r-prv SER SER SER GLa3+ :Cr4+:O2- HLa HCr 3HO = 5 6GS4O[72]
+GRPRV GS3V[72] + 1 6GS4V[72] + Gmag

o-prv SER SER GLa3+:Cr3+ :Va HLa HCr = GOPRV 1.5 G(O2(g) )[10] + Gmag r-prv SER SER GLa3+:Cr3+ :Va HLa HCr = GRPRV 1.5 G(O2(g) )[10] + Gmag
o-prv SER SER GLa3+:Cr4+ :Va HLa HCr = 5 6GS4O[72] GS3V[72] +1 6GS4V[72]
+GOPRV 1.5 G(O2(g) )[10] + Gmag

r-prv SER SER GLa3+:Cr4+ :Va HLa HCr = 5 6GS4O[72] GS3V[72] +1 6GS4V[72]
+GRPRV 1.5 G(O2(g) )[10] + Gmag
o-prv SER SER GVa:Cr3+:O2- HCr 3HO = GOPRV + 1.5GVCR4O
+0.5GVVV[65] 2GLCR4O + 0.75 G(O2(g) )[10] 1.41263T + Gmag
r-prv SER SER GVa:Cr3+:O2- HCr 3HO = GRPRV + 1.5GVCR4O
+0.5GVVV[65] 2GLCR4O + 0.75 G(O2(g) )[10] 1.41263T + Gmag
o-prv SER GVa:Cr3+:Va HCr = GOPRV + 1.5GVCR4O
+0.5GVVV[65] 2GLCR4O 0.75 G(O2(g) )[10] 1.41263T + Gmag
r-prv SER GVa:Cr3+:Va HCr = GRPRV + 1.5GVCR4O
+0.5GVVV[65] 2GLCR4O 0.75 G(O2(g) )[10] 1.41263T + Gmag
o-prv SER SER r-prv SER SER GVa:Cr4+:O2- HCr 3HO = GVa:Cr4+:O2- HCr 3HO =
= 2GVCR4O + 1 3GVVV[65] 4 3GLCR4O + 0.5 G(O2(g) )[10] + 4.35056T + Gmag
o-prv SER r-prv SER GVa:Cr4+:Va HCr = GVa:Cr4+:Va HCr =
= 2GVCR4O + 1 3GVVV[65] 4 3GLCR4O G(O2(g) )[10] + 4.35056T + Gmag Interaction term Lprv3+:Cr3+,Va:O2- = Lprv3+:Cr4+,Va:O2- = 250000 La La Magnetic contribution Tco-prv = Tcr-prv = 291.35 yi:j:k i = La 3+ ,Va j = Cr 4+ ,Cr3+ k = O-2, Va La - Chromates La 2CrO6 (La 3+ )2 (Cr 6+ )(O2- )6
prv o-prv = r-prv = 0.894 yi:j:k
SER SER SER 2 GLa3+ :Cr6 :O2- 2HLa HCr 6HO = GLA2CRO6 6+
La CrO
La 2 (CrO4 )3 (La 3+ )2 (Cr 6+ )3 (O2- )12
SER SER SER 4 3 2 GLa3+ :Cr6+:O2- 2HLa 3HCr 12HO = GLA2CR3O12
La (CrO )
113
Functions Perovskite Stoichiometric orthorhombic perovskite GOPRV = 0.5GLA2O3A[8] + 0.5GCR2O3[9] 73931 + 3.01T 0.68T ln T Stoichiometric rhombohedral perovskite GRPRV = 0.5GLA2O3A[8] + 0.5GCR2O3[9] 73591 + 2.38T 0.68T ln T Neutral nonstoichiometric perovskite endmembers GS4O[72] = 597648 + 213.38T 47.56T ln(T ) 0.00307T 2 + 190000T 1 +0.5GCR2O3[9] + 0.25 G(O2(g) )[10] GS3O[72] = 472704 + 191.7186T 47.56T ln(T ) 0.00307T 2 + 190000T 1 +0.5GCR2O3[9] GVCR4O = 0.5GCR2O3[9] + 0.25 G(O2(g) )[10] 291802 250T GLCR4O =1 3GLA2O3A[8] + 0.5GCR2O3[9] + 0.25 G(O2(g) )[10] 200000 Perovskite reference[72] GS4V[72] = 607870 + 268.9T 47.56T ln(T ) 0.00307T 2 + 190000T 1 +0.5GCR2O3 1.25 G(O2(g) )[10] La 2CrO6 GLA2CRO6 = GLA2O3A[8] + 0.5GCR2O3[9] + 0.75 G(O2(g) )[10] 72615 4.5T La 2 (CrO4 )3 GLA2CR3O12 = GLA2O3A[8] + 1.5GCR2O3[9] + 2.25 G(O2(g) )[10] 371557 + 205T
a)
All parameters are in SI units : J, mol, K. R = 8.31451 Jmol K
Using the compound energy formalism (CEF)[66-68] the molar Gibbs energy of La1-xCrO3 reads

prv prv Gm = yi y j yk Gi: j:k + RT yi ln yi + y j ln y j + yk ln yk + EGm + Gmag
(4.3.6)
where yi is the site fraction of Va and La3+ on the A-sublattice, yj is the site fraction of Cr3+, Cr4+ and Va on the B-sublattice, and yk is the site fraction of O2- and Va on the anion sublattice of the perovskite A1-xBO3. R = 8.31451 J mol-1 K-1. The third-last term accounts for the configurational entropy of mixing. The second-last term describes the excess Gibbs energy of mixing due to interactions of ions in the mixture. These are accounted for by the
114
optimization of interaction parameters. The last term designates the magnetic contribution to the Gibbs energy. For the magnetic part of the Gibbs energy a magnetic ordering-model proposed by Inden[69] and simplified by Hillert and Jarl[70] was used. A short summary of this model can be found in Chen et al.[71] The magnetic parameters Tc and were fitted to the C p data around the magnetic transition temperature from Sakai and Stlen[35]. Fig. 4.3.1 is a visualization of the Cr-containing part of the model the authors use to describe the cation nonstoichiometry of lanthanum chromite.
Fig. 4.3.1
Representation of the Cr-containing part of the model for nonstoichiometric
lanthanum chromite. The thin lines margin the neutral plane. The neutral compounds used for the optimization are marked by the black spots. The parameters of the compound energy formalism are the Gibbs energies of the not necessarily neutral 12 end-member compounds Gi: j:k , with the 8 Cr-containing compounds being the corners of the cube. Only compounds inside the neutral plane can exist on their own.
prv prv G(La3+ )(Cr 4+ )(O2- ) and G(La3+ )(Cr 4+ )(Va) are given in Table 4.3.3 (pp. 112-114). These endmembers of
3 3
nonstoichiometric perovskite have been fixed firmly by a sufficient number of consistent experiments in
3
the
LaO1.5-SrO-CrO1.5
3
system
[72]
Thus
the
authors
prv prv adopted G(La3+ )(Cr 4+ )(O2- ) and G(La3+ )(Cr 4+ )(Va) from Povoden et al.[72].
115
The neutral Cr4+ -containing endmembers

2 1 2 2 1 1 GVa:Cr 4+ :O2 + GVa:Cr 4+ :Va + 3RT ln + ln 3 3 3 3 3 3 1 1 gas = GVCR4O = GCr2O3 [9] + GO2 [10] + Gmag + A + BT 2 4
prv SER SER GVaCrO2 H Cr 2 H O =
(4.3.7)
and
2 SER 2 1 2 2 1 1 SER SER H = GLa3+ :Cr 4+ :O2 + GVa:Cr 4+ :O2 + RT ln + ln H Cr 3H O 3 La 3 3 3 3 3 3 (4.3.8) 1 1 1 gas = GLCR4O = GCr2O3 [9] + GLa2O3 [8] + GO2 [10] + Gmag + A 2 3 4
prv GLa2 3CrO3
and reciprocal relations which were set zero in analogy to Grundy et al.[65] were used to obtain
prv prv prv prv G(Va)(Cr3+ )(O2- ) , G(Va)(Cr 4+ )(O2- ) , G(Va)(Cr3+ )(Va) , and G(Va)(Cr 4+ )(Va) . The configurational entropy3 3 3 3
term in Eq. 4.3.7 describes random mixing of O2- with Va on the anion sublattice. In Eq. 4.3.8 it describes random mixing of La3+ and Va on the A-site. The parameters A and B of Eq. 4.3.7 and A of Eq. 4.3.8 are optimized using experimental data of excess Cr in perovskite[56]. Furthermore the temperature dependence of lanthanum vancancy and hole concentrations from Akashi et al.[58] was considered in the optimization. As cation diffusion in La1-xCrO3 is extremely slow even at high temperatures, the Croverstoichiometry in a furnace-cooled specimen reported by Khattak and Cox[55] does most likely not represent the overstoichiometry at an intermediate temperature and was not used for the optimization.
prv G(La3+ )(Cr3+ )(Va) results from a reciprocal relation which was set zero in analogy to Grundy et
3
al.
[65]
:
3 G (O2(g) )[10] 2
o-prv,r-prv SER SER o-prv,r-prv GLaCrVa3 HLa H Cr = GLa3+ :Cr3+ :Va = GLaCrO3
(4.3.9)
Using Eqs. 4.3.5 to 4.3.8 and adopting the Gibbs energies of the remaining endmembers
prv prv prv prv G(La3+ )(Va)(O2- ) , G(La3+ )(Va)(Va) , G(Va)(Va)(O2- ) , and G(Va)(Va)(Va)3 from Grundy et al.
3 3 3
[65]
, the 12
endmembers of the compound energy formalism of the perovskite phase are defined. The
116
introduction of positive interaction parameters 0Lprv3+ ,Va:Cr3+ :O2- and 0Lprv3+ ,Va:Cr4+ :O2- that were given La La the same values circumvents too high Cr4+ contents at low temperatures that would be in conflict with the experiments.
The liquid phase:
The two-sublattice model for ionic liquids[73,74] was used for the description of the liquid phase of the La-Cr-O system. It was based on the liquid descriptions of the binary subsystems. The chromium species considered in the liquid are Cr2+ and Cr3+. Higher oxidation states are unlikely to exist in the liquid at normal oxygen partial pressures. The liquid is thus given by the model description (La3+,Cr2+,Cr3+)p(O2-,Vaq-)q. The experimentally determined temperatures and liquid compositions[13,14] at the eutectic and monotectic in the metallic La-Cr system and the partial enthalpy of mixing of Cr, H Cr [15] in La-Cr liquid were used to optimize the temperature-dependent regular 0 Lliq 2+ ,La3+ :Va and subregular 1Lliq 2+ ,La3+ :Va Cr Cr interaction parameters to account for interactions between La and Cr. Furthermore the two regular interaction parameters 0 Lliq 3+ ,La3+ :O2- = 0 Lliq 2+ ,La3+ :O2- and the two subregular 1Lliq 3+ ,La3+ :O2- = Cr Cr Cr
1 liq Cr 2+ ,La3+ :O2-
were optimized. It was assumed that the interactions between Cr2+-La3+ and
Cr3+-La3+ are of the same order of magnitude in the oxide melt, thus the two regular interaction parameters were set equal to each other, as were the two sub-regular interaction parameters. Using the following data for their optimization led to the lowest error between experiments and calculation: the composition and temperature of the eutectic at the La-rich side and the composition of the eutectic at the Cr-rich side in the oxide LaO1.5- CrO1.5 system from Tresvjatskiy et al.[18], the temperature of the eutectic at the Cr-rich side from Berjoan[19], and the congruent melting temperature of the perovskite phase from Coutures et al.[20] and Fox[21]. Berjoan[32] and Tresvjatskiy et al.[18] did not specify the value of the prevailing oxygen partial pressure during their phase diagram experiments conducted in an argon atmosphere. As a value of the oxygen partial pressure is required for the optimization, we defined pO2 = 1 Pa.
4.3.5
Results and Discussion
The La-Cr system:
The calculated phase diagram of the La-Cr system is presented in Fig. 4.3.2 (next page), together with experimental phase diagram data[12,13,14,17].
117
Fig. 4.3.2 Calculated phase diagram of the La-Cr system with data
from the literature included (symbols). The positive value of 0Lbcc Cr,La:Va used to model the bcc phase results in a large miscibility gap between the La-rich and Cr-rich metals, which is tantamount to a small solubility of La in
Crss in agreement with the experiments
[14,17]
. The model description of the bcc phase results
in a tiny solubility of Cr in La(bcc), denoted as La ss , of 2 10-3 at.% at 1134 K, the lowest temperature of stable La ss , which further decreases as a function of increasing temperature. The calculated enthalpies of mixing are shown in Fig. 4.3.3 (next page) together with the experimentally determined value[15] that is well reproduced by the calculation.
118
Fig. 4.3.3 Calculated partial enthalpies of mixing of La and Cr in La-Cr liquid, and integral
enthalpies of mixing as a function of composition, with the experiment from Berezutskii et
al.[15] at T = 1700 K included (symbol with error-bar).

Considerable deviations of the calculated liquidus from experiments at the Cr-rich side of the system can be ascribed to the problem of two different melting temperatures for Cr cited in the literature, which are T = 2180 K and 2130 K. The higher value was favored by Dinsdale[10] and is adopted in this study, whereas the lower melting temperature was chosen by Savitskii et al.[12] and Svechnikov et al.[14]. A satisfying reproduction of the experimental data was obtained by considering a moderate temperature dependence of
0 liq Cr 2+ ,La3+ :Va
and 1Lliq 2+ ,La3+ :Va . This is unfortunately associated with Cr
an inverse liquid-liquid miscibility gap with a minimum at X(Cr) = 0.25 and T 5000 K that is of course unphysical.
The La-Cr-O system:
Phase equilibria: Calculated LaO1.5- CrO1.5 phase diagrams in pure oxygen at pO2 =105 Pa, in air at
119
pO2 = 21278 Pa, and under reducing conditions at pO2 = 1 Pa representing the typical oxygen
partial pressure in argon atmosphere are shown in Fig. 4.3.4 together with experimental data[18-21].
Fig. 4.3.4
Calculated phase diagrams of the LaO1.5-CrO1.5 system in pure oxygen, air
atmosphere, and under reducing conditions representing argon atmosphere at pO2 = 1 Pa with experimental data included (symbols). Excess Cr in lanthanum chromite is favored at high oxygen partial pressures. A decrease of Cr4+ during annealing of an originally lanthanum-deficient perovskite phase under reducing conditions is predicted by the model, reflected by the disappearance of Cr overstoichiometry. This is in line with the interpretations of Raman spectra from Iliev et al.[56] Be it that the reported thermodynamic data of La2CrO6[19] and La2(CrO4)3[30] are correct, lanthanum chromite is expected to be metastable at room temperature, and orthorhombic perovskite is stable only at pO2 102 Pa. La2CrO6 is stable within a wide temperature-range in pure oxygen, whereas it does not form in air and argon atmosphere. Due to the ambiguous oxygen partial pressure of phase diagram experiments[18,19] and the conflicting data on the melting temperature of lanthanum chromite in argon atmosphere[18] the presented liquid description is rather tentative. Under oxidizing conditions Cr3+ is favored over Cr2+ in the liquid. Analogous to Fe in the La-Fe-O system[62] this oxidation of Cr2+ to Cr3+ governs shifts of eutectic compositions and temperatures and the increase of the melting temperature of the perovskite phase on increasing the oxygen partial pressure. On the other hand a significant amount of Cr3+ in the ionic liquid is reduced to Cr2+ under reducing conditions, and the liquid stability increases considerably at the Cr-rich part of the system
120
leading to a considerably lowered eutectic temperature. The liquid description using the twosublattice model for ionic liquids also resulted in a significantly larger decrease of the melting temperature of lanthanum chromite at pO2 1 Pa than the given values in argon atmosphere[18]. Despite this discrepancy we did not go for an alternative liquid model for the sake of consistency with our previously assessed systems. In Fig. 4.3.5 calculated phase equilibria of the La-Cr-O system at T = 1273 K are shown as a function of oxygen partial pressure.
Fig. 4.3.5 Calculated phase equilibria of the La-Cr-O system at T = 1273 K
as a function of oxygen partial pressure. It is obvious that no mutual solubilities of La and Cr in bcc metal in equilibrium with oxides are expected. The same oxygen solubility in Cr as in the assessment by Povoden et al.[9] was obtained using the new model description (Cr)(O,Va)1.5. At pO2 = 10-34.04 Pa metallic liquid forms at the lanthanum-rich side of the phase diagram. Thermodynamic data: Calculated thermodynamic data of solid oxides are listed together with experimental data from the literature in Table 4.3.1 (p. 105). Calculated and experimental data on the orthorhombic to rhombohedral transition of LaCrO3 are listed in Table 4.3.2 (p. 106). Table 4.3.3 (pp. 105-107) is a compilation of the Gibbs energy functions and model descriptions of the phases in the La-Cr-O system obtained in this study. 121
Lanthanum chromates: Testing an optimization of model parameters of La2(CrO4)3 by using all available thermodynamic data[29,31] resulted in gross disagreement between optimized and reported values. The considerable error might be explained by experimental difficulties to reach equilibrium at the low investigation temperatures, and/or by significant deviations between the thermodynamic standard data used for the calculation of the enthalpy of formation from the elements[29] and assessed values[8-10]. Anyway the model parameters were fitted to the experimental data[30], whereas the calculated standard enthalpy of formation from the elements[29] was rejected, bearing in mind the high degree of uncertainty of the resulting description. The perovskite phase: the calculated heat capacities of LaCrO3 are compared with experiments from the literature in Fig. 4.3.6.
Fig. 4.3.6
Calculated heat capacities of LaCrO3 (solid curve) as a function of T with
experimental data included (symbols). The dashed line marks the temperature of the o-prv r-prv transition. The calculated C p -curve extrapolates well to high temperatures. The use of C p -data from Sakai and Stlen[35] along with enthalpy increment-data from Suponitskii[30] to optimize the
122
parameter CTlnT of the Gibbs energy of stoichiometric perovskite resulted in the lowest error between experiments and calculation. As CTlnT was set equal for o-prv and r-prv, their C p is the same. The experimentally determined C p -peak around 545 K caused by the first-order transition o-prv r-prv is in fact a discontinuity which cannot be implemented in the model. The calculated transition temperature of T = 540 K is shown by the broken line in Fig. 4.3.6. The calculated C p -peak at T = 290 K reflects the temperature of the magnetic order-disorder transition, the transition temperature being in agreement with the experiments. Two values for the magnetic parameter p are possible depending on the crystal structure, p=0.28 and p=0.4, whereby the proper p-value for structures other than bcc, fcc, and hcp is not available in the literature. The C p -anomaly is equally well reproduced by the model[69,70] using p = 0.28 or p = 0.4. For the sake of compatibility with the recent assessment of the La-Fe-O system[62] we chose p = 0.28. Experimental enthalpy increments[30] are well reproduced by the calculation (see Table 4.3.1, p. 105). Due to the consistency between both groups of calorimetric experiments[30,35] the term CTlnT is fixed firmly. A small peak which was found around 855 K can be explained most likely by the decomposition of an undetected impurity phase[35]. The calculated Gibbs energies of the formation of LaCrO3 from the oxides
1 La O + 1 Cr O LaCrO 3 2 2 3 2 2 3
(4.3.10)
are listed as a function of temperature together with data from the literature
[49-53]
in Table
4.3.1, p. 105. The resulting Gibbs energies of formation from emf-measurements are remarkably less negative than the Gibbs energies of formation derived from Knudsen mass spectrometry[53]. Only the use of the latter data for the optimization resulted in the proper phase diagram with congruent melting of the perovskite phase and two eutectics. It needs to be clarified why all of the emf-measurements are problematic: Azad et al.[50] stated that the Gibbs energy of formation of LaCrO3 cannot be studied properly using the solid oxide electrolyte method due to experimental difficulties in measuring the low oxygen potentials encountered in a mixture of coexisting LaCrO3-La2O3-Cr. Yet it is obvious that the CaF2based emf-technique is neither suitable for the determination of thermodynamic data of lanthanum chromite, as it unavoidably leads to emf that are too low. A possible explanation is found in a study by Akila and Jacob[75]: Fine precipitates of CaO can form on the surface of CaF2 in water- or oxygen-containing atmosphere. In this case the emf depends on the activity
123
of CaO at the electrode/electrolyte interface, and changing activity of CaO at the electrode/electrolyte interface can alter the chemical potential of fluorine at this electrode and thus the emf across the electrolyte. Chemical stability of the perovskite phase: The calculated oxygen partial pressure for the decomposition of lanthanum chromite by the reaction
1 3 LaCrO3 La 2 O3 + Cr + O2 (g) 2 4
(4.3.11)
is pO2 = 10-20.97 at 1273 K. The calculated decomposition of the perovskite phase by Eq. 4.3.11 is plotted as a function of temperature and oxygen partial pressure in Fig. 4.3.7.
Fig. 4.3.7 Calculated decomposition of lanthanum chromite
as a function of temperature and oxygen partial pressure. Defect chemistry of the perovskite phase:
124
Applying a defect chemistry analysis of La1-xCrO3 in equilibrium with Cr2O3 the following defect reaction for its oxidation can be written in the sublattice form, if [Va ] and [Va ] are O Cr assumed to be negligible according to Akashi et al.[58]:
1 (La 3+ )(Cr 3+ )(O2- )3 + O2(g) (La 3+3 Va1 3 )(Cr 4+ )(O2- )3 2 4
(4.3.12)
Using Krger-Vink notation this defect reaction reads

1 2 x 1 x x x x La La +CrCr +3OO + O2(g) La La + Va + CrCr +3OO 4 3 3 La
(4.3.13)
and the equilibrium constant of the oxidation reaction is

x x [Va ]1 3[La La ]2 3[CrCr ][OO ]3 La x x x 3 14 [La La ][CrCr ][OO ] pO2
K ox =
(4.3.14)
x x x For small oxidation extent [La La ] , [CrCr ] , and [OO ] can be considered to be ~ 1, and charge
neutrality is maintained by
1 [Va ]= [CrCr ] La 3
(4.3.15)
Substituting this into Eq. 4.3.14 gives the proportionalities [Va ], [CrCr ] PO2 16 . La
x x The concentrations of the defects LaLa , Va La , CrCr , and CrCr in La1-xCrO3 correspond to the site
fractions yprv 3+ , yprv , yprv 3+ , and yprv 4+ in the compound energy formalism. yprv 3+ , yprv , yprv 3+ , A La A Va B Cr B Cr A La A Va B Cr
y prv 4+ and the tiny fractions y prv and y prv are plotted logarithmically as a function of B Cr B Va O Va
log pO2 at T = 1073 K and 1673 K in Fig. 4.3.8 (next page) for lanthanum chromite in
equilibrium with Cr2O3. The line for yprv 3+ at 1073 K cannot be seen as it is very close to 1. A La
125
Fig. 4.3.8 Calculated site fractions of species in La1-xCrO3 in thermodynamic equilibrium
with Cr2O3 logarithmically plotted at T = 1073 K and 1673 K as a function of of 3/16 of the calculated defect concentrations is indicated in the triangle.
pO
. The slope
At T=1073 K a constant slope of 3/16 of the defect concentrations [Va ] and [CrCr ] shown in La
the triangle, is calculated from very high to very low oxygen partial pressures. This slope is
fixed by the defect reaction Eq. 4.3.12. At T = 1673 K the slope of 3/16 of [Va ] and [CrCr ] is La
reproduced by the calculated slope using the compound energy formalism at 105 Pa > pO >
2
10-8 Pa; hence oxidation of LaCrO3 to La1-xCrO3 governs the electrical conductivity of perovskite with fixed activity of Cr2O3 at unity between pO2 = 105 Pa and 10-8 Pa at this
temperature. The calculated slopes of [Va ] and [CrCr ] are equal to the slope of the electrical La
conductivity from 1573 to 1673 K between pO2 = 1.0 103 Pa and pO2 = 2.0 104 Pa determined by Akashi et al.[58]. The conflicting data from Shvaiko-Shvaikovskii et al.[57] may be explained by problems of reaching equilibrium due to extraordinarily slow cation diffusion in
lanthanum chromite. In Fig. 4.3.9 (next page) the calculated slopes of VaLa and CrCr are
126
compared with slopes of [Va ] and [CrCr ] determined by Akashi et al.[58] as a function of La
reciprocal temperatures.
Fig. 4.3.9 Calculated defect concentrations in La1-xCrO3 in thermodynamic equilibrium with
Cr2O3 (solid lines) logarithmically plotted as a function of reciprocal temperature along with the data from Akashi et al.[58] derived from electrical conductivity measurements (symbols with error-bars, broken lines). The calculated concentrations agree well with the data derived from electrical conductivity
measurements[58]. The calculated amount of [Va ] relative to [CrCr ] is fixed by the criterion for La
charge neutrality, Eq. 4.3.15, as calculated [Va ] and [Va ] are very small. The calculated O Cr relative defect concentrations are in line with those proposed by Akashi et al.[58]. The presented defect chemistry calculations are still rather tentative, as the temperature and oxygen partial pressure dependence of excess Cr in La1-xCrO3 has not been investigated systematically so far.
127
4.3.6
Conclusions
Model parameters of the presented thermodynamic La-Cr-O database were optimized with assessed thermodynamic and phase diagram data. The thermodynamic descriptions of lanthanum chromates and the liquid phase are rather tentative due to humble or sketchy experimental information. The thermodynamic modeling of lanthanum chromite was based on experimental thermodynamic data reported by Peck et al.[53] and Cheng and Navrotsky[46], as the use of these data for the optimization of model parameters resulted in a proper reproduction of the phase equilibria derived from experiments. The orthorhombic to rhombohedral transition in lanthanum chromite and the magnetic order-disorder transformation are well reproduced by the model. Using the new database the stability limits of lanthanum chromite in function of temperature and oxygen partial pressure can be quantified. The proposed existence of lanthanum vacancies and holes to maintain charge neutrality in lanthanum chromite with excess Cr is reproduced by the model, and the calculated slopes of defect concentrations in function of oxygen partial pressure and temperature are in line with the slopes derived from electrical conductivity measurements. However the amounts of excess Cr in La1-xCrO3 used for the optimization of the cation nonstoichiometry are preliminary, and further work on the temperature dependence of excess Cr as a function of temperature and oxygen partial pressure would allow a more accurate quantification of the defect chemistry of lanthanum chromite.
References
1. D. Das, M. Miller, H. Nickel, K. Hilpert, First European Solid Oxide Fuel Cell Forum Proceedings, Vol. 2, U. Bossel (Ed.), Druckerei J. Kinzel, Gttingen, 1994, p. 703. 2. S. P. S. Badwal, R. Deller, K. Foger, Y. Ramprakash and J. P. Zhang, Interaction between chromia forming alloy interconnects and air electrode of solid oxide fuel cells, Solid State
Ionics, 1997, 99, pp. 297-310.

3. S.P. Jiang, J.P. Zhang, and K. Foger, Deposition of chromium species at Sr-doped LaMnO3 electrodes in solid oxide fuel cells. I. Mechanism and kinetics, J. Electrochem.
Soc., 2000, 147(11), pp. 4013-22.

4. A. Carlsson, A.-M.H. Kjaer and J.G. Larsen, Cr-Mn deposition at the three-phase boundary observed by TEM, Proceedings of the 26th Risoe International Syposium on 128
Materials Science: Solid State Electrochemistry, S. Linderoth, A. Smith et al. (Eds.), Ris National Laboratory, Roskilde, Denmark, 2005, pp. 155-62. 5. G.-Y. Lee, R.-H. Song, J.-H. Kim, D.-H. Peck, T.-H. Lim, Y.-G. Shul, and D.-R. Shin, Properties of Cu, Ni, and V doped-LaCrO3 interconnect materials prepared by Pechini, ultrasonic spray pyrolysis and glycine nitrate processes for SOFC, J. Electroceram., 2006,
17, pp. 723-27.

6. M. Liu, L. Zhao, D. Dong, S. Wang, J. Diwu, X. Liu, G. Meng, High sintering ability and electrical conductivity of Zn doped La(Ca)CrO3 based interconnect ceramics for SOFCs,
J. Power Sources, 2008, 177, pp. 451-56.

7. S.P. Jiang, L. Liu, K.P. Ong, P. Wu, J. Li, J. Pu, Electrical conductivity and performance of doped LaCrO3 perovskite oxides for solid oxide fuel cells, J. Power Sources, 2008,
176, pp. 82-89.

8. M. Zinkevich, S. Geupel, F. Aldinger, A. Durygin, S.K. Saxena, M. Yang, Z.-K. Liu, Phase diagram and thermodynamics of the La2O3-Ga2O3 system revisited, J. Phys. Chem.
Solids, 2006, 67, pp. 1901-07.

9. E. Povoden, A.N. Grundy, L.J. Gauckler, Thermodynamic reassessment of the Cr-O system in the framework of solid oxide fuel cell (SOFC) research, J. Phase Equilib. Diff., 2006, 27(4), pp. 353-362. 10. A.T. Dinsdale, SGTE data for pure elements, Calphad, 1991, 15(4), pp. 317-425. 11. B. Sundman, B. Jansson, J.-O. Andersson, The Thermo-Calc databank system, Calphad, 1985, 9(2), pp. 153-90. 12. E.M. Savitskii, V.F. Terekhova, A.V. Kholopov, Phase diagram for alloys in the chromium-lanthanum system, Russian J. Inorg. Chem., 1960, 5(3), pp. 362-63. 13. O.M. Barabash and Y.N. Kobal, Crystal structures of metals and alloys, Naukova Dumka, Kiev, 1986, p. 264 (in Russian). 14. V.M. Svechnikov, G.F. Kobzenko, M.V. Kireev, V.G. Ivancenko, Yttrium and lanthanum solubility in chromium, Dopov. Akad. Nauk. A, 1973, 2, pp. 167-71 (in Ukrainian). 15. V.V. Berezutskii, N.I. Usenko, M.I. Ivanov, Thermochemistry of binary alloys of lanthanum with 3d-transition metals, Powder Metall Met. Ceram., 2006, 45(5-6), pp. 266271. 16. F.N. Tavadze, O.I. Mikadze, N.P. Keshelava, B.P. Bulia, High-temperature corrosion of dilute chromium-lanthanum alloys, Oxid. Met., 1986, 25(5-6), pp. 335-350.
129
17. S.G. Epstein, A.A. Bauer, R.F. Dickerson, Solubility limits of yttrium and the lanthanide rare-earth elements in chromium and chromium-iron alloys, U.S. At. Energy Comm., 1959, BMI-1376. 18. S.G. Tresvjatskiy, L.M. Lopato, V.N. Pavlivov, A.V. Shevchenco, Phase correlation in the systems formed by oxides of rare earths and by oxide of chromium, Rev. Int. Hautes.
Temp., 1968, 5, pp. 45-50.

19. R. Berjoan, Contribution to the study of interactions between oxygen and mixed lanthanum oxide and chromium oxide(III), or lanthanum chromite, Rev. Int. Hautes.
Temp., 1976, 13, pp. 119-35 (in French).

20. J.-P. Coutures, J.M. Badie, R. Berjoan, J. Coutures, R. Flamand, A. Rouanet, Stability and thermodynamic properties of rare earth perovskites, High Temp. Sci., 1980, 13, pp. 33136. 21. M. Fox, Examination of freezing points of rare earth oxides and chromites, C.R. Acad.
Sc. Paris, Gr. 8, 1965, 260, pp. 6389-92 (in French).

22. M. Fox, Measurement of the solidification points of several refractory oxides, Sol.
Energy, 1965, 9(1), pp. 61-67.

23. M. Fox, Comments on the allotropic transformations of the rare earth sesquioxides, Z.
Anorg. Allg. Chem., 1963, 337, pp. 313-24 (in German).

24. M. Fox, Measurement of the freezing points of refractory oxides, Rev. Int. Hautes.
Temp., 1966, 3, pp. 309-26 (in French).

25. D.B. Meadowcroft, Some properties of strontium-doped lanthanum chromite, Brit. J.
Appl. Phys., 1969, 2(2), pp. 1225-33.

26. S.A. Suvorov, A.P. Shevchik, Chemical equilibria involving lanthanum chromite, Refract.
Ind. Ceram., 2004, 45(2), pp. 94-99.

27. J. O. Cassedanne, Examination of the equilibrium diagram La2O3-Cr2O3, An. Acad. Bras.
Cienc., 1968, 40(1), pp. 57-60 (in French).

28. A.N. Grundy, B. Hallstedt, L.J. Gauckler, Thermodynamic assessment of the lanthanumoxygen system, J. Phase Equilib., 2001, 22(2), pp. 105-13. 29. S.B. Tsyrenova, Y.L. Suponitskii, M.K. Karapetyants, Comparison of thermal properties of oxygen-containing compounds of rare-earth elements, scandium and yttrium. 2. Thermochemical properties of lanthanum chromates(VI), Zh. Fiz. Khim., 1974, 48(11), pp. 2705-07. 30. Yu.L. Suponitskii, Thermal properties and thermochemistry of lanthanide chromates,
Russ. Chem. B., Int. Ed., 2005, 54(2), pp. 294-99.

130
31. H. Schwarz, On the chromates(V) of the rare earths. I lanthanum chromate(V), LaCrO4, Z.
Anorg. Allg. Chem., 1963, 322(1-2), pp. 1-14 (in German).

32. R. Berjoan, J.P. Traverse, J.P. Coutures, Structural study and thermal decomposition of lanthanum based chromate, Rev. Chim. Miner., 1973, 10, pp. 309-16 (in French). 33. A.V. Korobeinikova, A. Reznitskii, Specific Heat, Enthalpy, and Entropy of Lanthanum Chromite between 298 and 900 K, High Temp., 1976, 14(4), pp. 805-06. 34. H.E. Hfer, W.F. Kock, Crystal Chemistry and Thermal Behavior in the La(Cr, Ni)O3 Perovskite System, J. Electrochem. Soc., 1993, 140(10), pp. 2889-94. 35. N. Sakai, S. Stlen, Heat Capacity and Thermodynamic Properties of Lanthanum(III) Chromate(III): LaCrO3, at Temperatures from 298.15 K. Evaluation of the Thermal Conductivity, J. Chem. Thermodyn., 1995, 27(5), pp. 493-506. 36. V.N. Pavlikov, L.M. Lopato, S.G. Tresvyatskii, Phase Transformations of Certain Chromites of the Rare Earth Elements, Izv. Akad. Nauk SSSR, Neorg. Mater., 1966, 2(4), pp. 679-82 (in Russian). 37. J.S. Ruiz, A.-M. Anthony, M. Fox, On the Semiconducting Properties of Lanthanum Chromite, C.R. Acad. Sc. Paris B, 1967, 264, pp. 1271-74 (in French). 38. S. Geller, P.M. Raccah, Phase Transitions in Perovskitelike Compounds of the Rare Earths, Phys. Rev. B, 1970, 2(4), pp. 1167-72. 39. N. Terao, Some Observations on the Formation and Structure of Lanthanum Chromite, LaCrO3 at Room Temperature, C.R. Acad. Sc. Paris C, 1973, 276, pp. 5-8 (in French). 40. S.P. Tolochko, I.F. Kononyuk, V.A. Lyutsko, Yu.G. Zonov, Phase Transitions in Solid Solutions Based on Lanthanum Chromite, Izv. Akad. Nauk SSSR, Neorg. Mater., 1985,
23(9), pp. 1520-24 (in Russian).

41. A.C. Momin, E.B. Mirza, D. Mathews, High-Temperature X-ray Diffractometric Studies of LaCrO3, J. Mater. Sci. Lett., 1991, 10, pp. 1246-48. 42. T. Hashimoto, N. Tsuzuki, A. Kishi, K. Takagi, K. Tsuda, M. Tanaka, K. Oikawa, T. Kamiyama, K. Yoshida, H. Tagawa, M. Dokiya, Analysis of Crystal Structure and Phase Transition of LaCrO3 by Various Diffraction Measurements, Solid State Ionics, 2000,
132, pp. 183-90.

43. N. Sakai K. Oikawa, T. Kamiyama, T. Hashimoto, Y. Shimojyo, Y. Morii, Structural Phase Transition of Orthorhombic LaCrO3 Studied by Neutron Powder Diffraction, J.
Solid State Chem., 2000, 154, pp. 524-29.
131
44. T. Hashimoto, N. Matsushita, Y. Murakami, N. Kojima, K. Yoshida, H. Tagawa, M. Dokiya, T. Kikegawa, Pressure-Induced Structural Phase Transition of LaCrO3, Solid
State Commun., 1998, 108(9), pp. 691-94.

45. I. Weinberg, P. Larssen, Electron Paramagnetic Resonance and Antiferromagnetism in LaCrO3, Nature, 1961, 192(4), pp. 445-46. 46. J. Cheng, A. Navrotsky, Enthalpies of Formation of LaMO3 Perovskites (M=Cr, Fe, Co, and Ni), J. Mater. Res., 2005, 20(1), pp. 191-200. 47. H. Satoh, S. Koseki, M. Takagi, W.Y. Chung, N. Kamegashira, Heat Capacities of LnCrO3 (Ln=Rare Earth), J. Alloys Compt., 1997, 259, pp. 176-82. 48. N. Sakai, T. Kawada, H. Yokokawa, M. Dokiya, Y. Takahashi, Heat Capacity Measurement of Lanthanum Chromite by Laser Flash Method, Abstracts of the 27th
Symposium on Thermal Analysis and Calorimetry, Kyoto, Japan, November 1991, pp. 7677 (in Japanese). 49. S. Chen, Z. Hao, F. Li, G. Zhou, Determination of the Standard Free Energy of Formation of LaCrO3 at 1273 K, Zhongguo Xitu Xuebao, 1987, 5(3), pp. 19-24 (in Chinese). 50. A.M. Azad, R. Sudha, O. M. Sreedharan, Thermodynamic Stability of LaCrO3 by a CaF2Based E.M.F. Method, J. Less-Common Met., 1990, 166(1), pp. 57-62. 51. Z. Chen, X. Xing, W. Yuan, X. Huang, H. Li, Thermodynamic Properties of LaCrO3,
Rare Metals, 2006, 25(5), pp. 562-66.

52. M. Dudek, G. Rg, A. Kozlowska-Rg, Thermodynamic Stability of LaCrO3 and CaZrO3 Using a Solid-State Galvanic Cell Method, Polish J. Chem., 2007, 81, pp. 23-29. 53. D.-H. Peck, M. Miller, D. Kobertz, H. Nickel, K. Hilpert, Vaporization of LaCrO3: Partial and Integral Thermodynamic Properties, J. Am. Ceram. Soc., 1996, 79(12), pp. 3266-72. 54. T. Nakamura, G. Petzow, L.J. Gauckler, Stability of the Perovskite Phase LaBO3 (B=V, Cr, Mn, Fe, Co, Ni) in Reducing Atmosphere I. Experimental Results, Mater. Res. Bull., 1979, 14, pp. 649-59. 55. C.P. Khattack and D.E. Cox, Structural Studies of the (La,Sr)CrO3 System, Mat. Res.
Bull., 1977, 12, pp. 463-72.

56. M.N. Iliev, A.P. Litvinchuk, V.G. Hadjiev, Y.-Q. Wang, J. Cmaidalka, R.-L. Meng, Y.-Y. Sun, N. Kolev, M.V. Abrashev, Raman Spectroscopy of Low-Temperature (Pnma) and High-Temperature (R 3 c) Phases of LaCrO3, Phys. Rev. B, 2006, 74, pp. 214301-1214301-7.
132
57. V.E. Shvaiko-Shvaikovskii, V.P. Popov, V.G. Gordon, Influence of the Synthesis Conditions on the Electrical Properties of LaCrO3, Izv. Akad. Nauk SSSR, Neorg. Mater., 1979, 15(8), pp. 1441-45 (in Russian). 58. T. Akashi, T. Maruyama, T. Goto, Transport of Lanthanum Ion and Hole in LaCrO3 Determined by Electrical Conductivity Measurements, Solid State Ionics, 2003, 164, pp. 177-83. 59. G.V.S. Rao, B.M. Wanklyn, C.N.R. Rao, Electrical Transport in Rare Earth OrthoChromites, -Manganites and -Ferrites, J. Phys. Chem. Solids, 1971, 32, pp. 345-58. 60. A.K. Tripathi, H.B. Lal, Electrical Transport in Light Rare-Earth Orthochromites, J.
Mater. Sci., 1982, 17, pp. 1595-1609.

61. N. Saunders, A.P. Miodownik, Calphad Calculation of Phase Diagrams, Pergamon Materials Series, Vol. 1. Elsevier Science Ltd., 1998, pp. 94-96. 62. E. Povoden-Karadeniz, A.N. Grundy, M. Chen, T. Ivas, L.J. Gauckler, Thermodynamic Assessment of the La-Fe-O System, J. Phase Equilib. Diff., accepted 63. C.N.R. Rao, J. Gopalakrishnan, K. Vidyasagar, Superstructures, Ordered Defects & Nonstoichiometry in Metal Oxides of Perovskite & Related Structures, Indian J. Chem., 1984, 23A, pp. 265-84. 64. L.G. Tejuca, J.L.G. Fierro, Structure and Reactivity of Perovskite-Type Oxides, Advances in Catalysis, 1989, 36, pp. 243-. 65. A.N. Grundy, E. Povoden, T. Ivas, L.J. Gauckler, Calculation of Defect Chemistry Using the CALPHAD Approach, Calphad, 2006, 30, pp. 33-41. 66. J.-O. Andersson, A.F. Guillermet, M. Hillert, B. Jansson, B. Sundman, A CompoundEnergy Model of Ordering in a Phase with Sites of Different Coordination Numbers, Acta
Metall., 1986, 34, pp. 437-45.

67. M. Hillert, B. Jansson, B. Sundman, Application of the Compound-Energy Model to Oxide Systems, Z. Metallkd., 1988, 79(2), pp. 81-87. 68. M. Hillert, The Compound Energy Formalism, J. Alloy. Cmpd., 2001, 320, p 161-76. 69. G. Inden, Determination of Chemical and Magnetic Interchange Energies in BCC Alloys. I. General Treatment, Z. Metallkd., 1975, 66(10), pp. 577-82. 70. M. Hillert, M. Jarl, A Model of Alloying Effects in Ferromagnetic Metals, Calphad, 1978,
2(3), p 227-38.
71. M. Chen, B. Hallstedt, L.J. Gauckler, Thermodynamic Assessment of the Co-O System, J.
Phase Equilib., 2003, 24(3), pp. 212-27.
133
72. E. Povoden, M. Chen, A.N. Grundy, L.J. Gauckler, Thermodynamic La-Sr-Mn-Cr-O oxide database for solid oxide fuel cell applications, submitted to Scripta Mater. 73. M. Hillert, B. Jansson, B. Sundman, J. gren, A Two-Sublattice Model of Molten Solutions with Different Tendency of Ionization, Metall. Trans. A, 1985, 16A, pp. 261-66. 74. B. Sundman, Modification of the Two-sublattice Model for Liquids, Calphad, 1991, 15, pp. 109-19. 75. R. Akila, K.T. Jacob, The Mobility of Oxygen Ions in CaF2, J. Appl. Electrochem., 1990,
20, pp. 294-300.
4.4
Thermodynamic
La-Sr-Mn-Cr-O
oxide
database
for
SOFC
applications
E. Povoden, M. Chen, A.N. Grundy, and L.J. Gauckler to be submitted The thermodynamic La-Sr-Mn-Cr-O oxide database is obtained as an extension of thermodynamic assessments of oxide subsystems using the Calphad approach. Gibbs energy functions of SrCrO4, Sr2.67Cr2O8, Sr2CrO4, and SrCr2O4 are presented. Experimental solid solubilities and nonstoichiometries in La1-xSrxCrO3- and LaMn1-xCrxO3- are reproduced by the model.
4.4.1
Introduction
Sr-doped lanthanum manganite (LSM) with the perovskite structure ABO3- is used as cathode materials in SOFC. However diffusion of chromium from the metallic interconnects into the cathode leads to a severe cell voltage decrease that was linked to the formation of Crcontaining phases[1,2]. A thermodynamic La-Sr-Mn-Cr-O oxide database is highly desirable for the development of endurable SOFC: thermodynamic calculations set an important base for the optimization of cathodes aiming to avoid long-term degradation due to chromium poisoning. The database should meet the demand to calculate stable and metastable phase equilibria, thermodynamic driving forces and activities, as well as defect concentrations of the cathode contaminated by Cr at different temperatures and oxygen partial pressures. These requirements are conformed by using the CALPHAD approach. For the construction of the
134
La-Sr-Mn-Cr-O oxide database La-Mn-Cr-O oxide and La-Sr-Cr-O oxide systems are assessed. Sr-Mn-Cr-O oxide is treated as ideal extension from the subsystems.
4.4.2
Assessment of data from the literature
Previous assessments of the La-O, Cr-O, and La-Cr-O databases are adopted[3-5], and the LaSr-Mn-O oxide database is taken from Grundy et al.[6] with a slight modification: Grundy et
al.[6] allowed Mn3+ on the A-site of LSM to reproduce experimental oxygen

nonstoichiometries under low oxygen partial pressures. Due to large differences between the ionic radii of La3+ and Mn3+ and possible coordination numbers (1.5 for 12-fold coordinated La3+, 0.785 for at maximum 6-fold coordinated Mn3+)[7] we omit Mn3+ on the A-site. Calculation of the oxygen nonstoichiometry of perovskite + MnO instead of metastable single phase perovskite[6] leads to a good agreement between experimental and calculated nonstoichiometries. No quaternary phases or solid solutions were found in the Sr-Mn-Cr-O oxide system[8].
Sr-Cr-O oxide:
Thermodynamic functions for Sr-Cr-oxides in the SSUB database[9] are based on estimates[10]. We propose optimized thermodynamic functions for oxide phases of the Sr-Cr-O oxide system resulting from the assessment of all available experimental data: agreement exists between Gibbs energies of formation of SrCrO4 determined by emf technique using a Y2O3 stabilized ZrO2 electrolyte[11,12], whereas emf measurements using CaF2-based emftechnique[13] led to conflicting results likely caused by competing reactions[14]. Differences concern the reported stabilities of further compounds[11,12,15-19]: for the stabilities of SrCr2O4 and Sr2CrO4 we trust the accurate study of Jacob[11], which is in agreement with Negas and Roth[15]. On the other hand the conflicting phase equilibria presented by Kisil[16] lack experimental details. Sr3Cr2O7[12] was approved as high pressure phase only[17]. The stoichiometry of a phase defined as Sr3Cr2O8[15] was later corrected to be essentially Sr2.67Cr2O8 by using microprobe analysis[18], in agreement with Hartl and Braungart[19].
La-Sr-Cr-O oxide:
In the La-Sr-Cr-O oxide and La-Mn-Cr-O oxide systems no quaternary stoichiometric compounds were reported. Phase equilibria in the La-Sr-Cr-O oxide system in air at 1223 K and under vacuum at 1873 K were determined by using solid state technique[18]. Limited solution of Sr in La1-xSrxCrO3- perovskite[18] was confirmed by a later investigation[20]. The 135
existence of several Ruddlesden-Popper phases is restricted to reducing conditions; solely Sr(La,Sr)CrO4 showed reproducible stoichiometry[18]. In contrast to Peck et al.[18] it was proposed earlier that Sr(La,Sr)CrO4 were stable in air[10]. The exact temperature and oxygen partial pressure range of Sr(La,Sr)CrO4 is ambiguous, thermodynamic data are missing, and the solubility of Cr is unknown. Thus its extension to the quinary database would not be reliable. As Ruddlesden-Popper phases have not been reported to form during SOFC operation with LSM cathodes, Sr(La,Sr)CrO4 is omitted in the modeling. Myoshi et al.[20] investigated the single phase region of La1-xSrxCrO3 with x = 0.1, 0.2, and 0.3 as a function of temperature and oxygen partial pressure using XRD analysis. Peck et al.[21] determined the Gibbs energy of formation of La1-xSrxCrO3 with x = 0.1, 0.2, and 0.3 using Knudsen mass spectrometry. Cheng and Navrotsky[22] measured enthalpies of formation of La1-xSrxCrO3- with x = 0.1, 0.2. and 0.3, and = 0, 0.04, 0.09, and 0.11 using drop calorimetry at
T = 1080 K. Positive in the perovskite formula reflects oxygen deficiency, whereas negative
essentially stands for cation nonstoichiometriy. Nonstoichiometry data for La1-xSrxCrO3-

with x=0.1, 0.2. and 0.3 at T = 1273K, 1373 K, 1473K, and 1573 K[23], and La0.8Sr0.2CrO3- at 1273 K[24] were measured as a function of oxygen partial pressure using thermogravimetry. Cr4+ and oxygen vacancies are regarded as the major defects[23,24].
La-Mn-Cr-O oxide:
In the La-Mn-Cr-O oxide system no quaternary stoichiometric compounds were reported. An isothermal section of the La-Mn-Cr-O oxide system at 1073 K in air and pure oxygen has been published without further commenting of experimental evidences[25]. Complete solid solution between the LaMnO3 and the LaCrO3 perovskites was affirmed[8].
of LaMn0.9Cr0.1O3- was measured using thermogravimetry[26].

La-Sr-Mn-Cr-O oxide:
In the La-Sr-Mn-Cr-O oxide system complete solid solubility of Mn and Cr is reported for La1-xSrxMn1-yCryO3- perovskite[8]. Plint et al.[27] concluded from the similarity between X-ray absorbtion spectra of Cr K of LaCrO3 and La1-xSrxMn0.5Cr0.5O3- with x = 0.2, 0.25, and 0.3 at
T = 1173 K that Cr4+ were absent in the latter. Perovskite+MnCr2O4 spinel equilibrium of a
powdered mixture of La0.8Sr0.2MnO3 and Cr2O3 at 1073 K was reported after 1000 h of heat treatment in air[28].
136
Magnetic and structural transitions of La1-xSrxCrO3- [2935], LaMn1-xCrxO3- [26,3640], and La1-xSrxMn1-yCryO3- [8,41] were reported. Transitions of LaMn1-xCrxO3- are complex as they depend on temperature, composition and oxygen partial pressure. Consistency among transition data for La1-xSrxCrO3- and LaMn1-xCrxO3- prevails, whereas diversities exist regarding the transitions in La1-xSrxMn1-yCryO3- . Thus, in terms of the applicability of the new database for SOFC the authors omit structural transitions in the modeling. Magnetic transitions have been well reproduced by an ordering-model[42,43] for LaCrO3[5]. However we did not obtain satisfying results in higher-order perovskites, most likely due to interactions that cannot be reproduced by the model. As the magnetic transitions are low temperature features, their modeling was omitted without consequences for the applicability of the database for SOFC.
4.4.3
Modeling and optimization
Sr-Cr-O oxide:
The sublattice models (Sr2+)(Cr6+)(O2-)4 and (Sr2+)(Cr3+)2(O2-)4 are employed for the descriptions of SrCrO4 and SrCr2O4. (Sr2+)8/3(Va)1/3(Cr6+)2/3(O2-)8/3(Cr5+)4/3(O2-)16/3 and (Sr2+)(O2-)1(Sr2+)(Cr4+)(O2-)3 were chosen for Sr2.67Cr2O8 following the proposed formula[19] and for the Ruddlesden-Popper phase Sr2CrO4, accounting for the structural feature of alternating rocksalt- and perovskite layers of the latter. Gibbs energy functions of Sr-Croxides were formulated as
1 2
SER SER SER gas G(Sr)x(Cr)y(O)z xH Sr yH Cr zH O = x GSrO[44] + y GCr2O3[4] + v GO2 [45] + A + BT
(4.4.1)
SER v = 0.75, 0, 7/6, and 0.25 for SrCrO4, SrCr2O4, Sr2.67Cr2O8, and Sr2CrO4 respectively. H a is
the standard enthalpy of the stable state of element a at 298.15 K and 105 Pa[45]. A and B are adjustable parameters; their optimization with the following experimental phase diagram and thermodynamic data using the PARROT module of the Thermocalc software[46] resulted in the lowest error between model and experiments: Gibbs energies of formation of SrCrO4[11,12] and phase stabilities of SrCr2O4 and Sr2CrO4 investigated by equilibration experiments of different mixed oxide compositions under controlled atmospheres[11], and the equilibrium Sr2.67Cr2O8+SrCrO4+Cr2O3 as a function of temperature and oxygen partial pressure[11,12]. All
137
reported phase equilibria[11] are correctly reproduced by the model. Optimized parameters and calculated and experimental thermodynamic data are listed in Table 4.4.1 and 4.4.2.
Table 4.4.1 Optimized model parameters

Sr - Cr oxides ASrCrO4 = 273771 J; BSrCrO4 = 131.6 J ASr2CrO4 = 145000 J; BSr2CrO4 = 50 J ASrCr2O4 = 98000 J; BSrCr2O4 = 95.5 J ASr2.67Cr2O8 = 508507 J; BSr2.67Cr2O8 = 219 J La1 xSrxCrO3
gas GLa3+ :Cr 4+ :Va = 5 6GS4O GS3V + 1 6GS4V + GRPRV[5] 1.5 GO2 [45]
GLa3+ :Cr 4+ :O2- = 5 6GS4O GS3V + 1 6GS4V + GRPRV[5] GSr 2+ :Cr 3+ :O2- = GS3V + 1 6GS4O 1 6GS4V GSr 2+ :Cr 3+ :Va = GS3V 5 6GS4O + 5 6GS4V
Table 4.4.2 Calculated and experimental thermodynamic data

SrO +1 2 Cr2 O3 + 3 4 O 2 = SrCrO 4 G = 273.774 + 0.13152T kJmol-1 this work, calc. G = 213.050 + 0.106904T kJmol-1 ,851 1116 K[12] G = 273.825 0.2 + 0.13157T kJmol-1 ,950 1280 K[10] Sr2.67 Cr2 O8 + SrCrO 4 + Cr2 O3 O 2 = 265.859 + 0.15832T kJmol-1 this work, calc. O 2 = 262.340 + 0.15553T kJmol-1 ,1073 1473 K[11] O 2 = 276.767 + 0.166T kJmol-1 ,1080 1380 K[10] (1 x) 2 La 2 O3 + xSrO + 1 2 Cr2 O3 + x 4 O 2 2 O2 = La1 xSrx CrO3 x = 0.1, = 0, T = 2000 K, G = 93.3 kJmol-1 this work, calc. G = 85.7 kJmol-1[20]
x = 0.2, = 0, T = 2000 K, G = 102.4 kJmol-1 this work, calc. G = 88.7 kJmol-1[20] x = 0.3, = 0, T = 2000 K, G = 109.4 kJmol-1 this work, calc. G = 93.5 kJmol-1[20] x = 0.1, = 0, T = 298 K, H = 65.2 kJmol-1 this work, calc.
H = 67.88 kJmol-1[21] x = 0.1, = 0.04, T = 298 K, H = 55.1 kJmol-1 this work, calc. H = 59.15 kJmol-1[21] x = 0.2, = 0, T = 298 K, H = 56.8 kJmol-1 this work, calc. H = 50.54 kJmol-1[21] x = 0.2, = 0.09, T = 298 K, H = 34.0 kJmol-1 this work, calc. H = 34.76 kJmol -1[21] x = 0.3, = 0, T = 298 K, H = 48.3 kJmol-1 this work, calc.
H = 36.72 kJmol-1[21] x = 0.3, = 0.11, T = 298 K, H = 20.6 kJmol-1 this work, calc. H = 20.48 kJmol -1[21]
138
It is essential for a consistent description of the perovskite phase that defects that occur in the structure in low-order systems remain on the same sites at the extension to higher order; this is achieved by using the same model. We adopt the description (A,Va)(B,Va)(O-2,Va)3 with A, B = cations and Va = vacancies[6] using the compound energy formalism[47]. The molar Gibbs energy of the perovskite phase then reads

prv prv Gm = yi y j yk Gi: j:k + RT yi ln yi + y j ln y j + 3 yk ln yk + EGm

i j k i j k
(4.4.2)
where yi is the site fraction of each cation and Va on the A-sublattice, yj is the site fraction of each cation and Va on the B-sublattice, and yk is the site fraction of O2- and Va on the anion sublattice. R = 8.31451 J mol-1 K-1. The second-last term accounts for the configurational entropy of mixing. The last term describes the excess Gibbs energy of mixing. It can be accounted for by introducing interaction parameters. The parameters of the compound energy formalism are the Gibbs energies of the end-member compounds Gi: j:k . Typical compositions of Sr-doped lanthanum manganites used for SOFC cathodes, e.g. La0.8Sr0.2MnO3- are rhombohedral at SOFC operating temperatures (T=1073 K to 1273 K), and small amounts of Cr brought into the cathode unlikely lead to a change of the structure. Thus it is reliable to take the Gibbs energies of the compounds of rhombohedral perovskite from [5] for the model. Using the above model and the proposed defect chemistry[22-24] the sublattice formula for La1-xSrxCrO3- reads (La3+,Sr2+,Va)(Cr3+,Cr4+,Va)(O2-,Va)3. The molar Gibbs energy G of La1-xSrxCrO3- is uniquely defined as follows: Gs of the endmembers (La3+)(Cr3+)(O2-)3, (La3+)(Cr3+)(Va)3, (La3+)(Va)(O2-)3, (La3+)(Va)(Va)3, (Sr2+)(Va)(O2-)3, (Sr2+)(Va)(Va)3, (Va)(Va)(O2-)3, and (Va)(Va)(Va)3 and ternary interaction parameters are adopted[5,6,48],
G(Sr2+)(Cr4+)(Va)3
1 5 gas GCr2O3 [4] GO2 [45] 2 4
SER SER GSrCrVa3 H Sr H Cr = GSr2+ :Cr 4+ :Va = GS4V = GSrO[44] +
(4.4.3)
is chosen as reference, and A and B parameters of G of two neutral compounds
139
SER SER SER GSrCrO3 H Sr H Cr 3H O = GSr2+ :Cr 4+ :O2 = GS4O
= GSrO[44] +
1 1 gas [4] G + GO2 [45] + A + BT 2 Cr2O3 4
(4.4.4)
SER SER SER GSrCrO2.5Va0.5 H Sr H Cr 2.5H O = GS3V
5 1 1 5 5 1 1 GSr2+ :Cr3+ :O2 + GSr2+ :Cr3+ :Va + RT ln + ln = GSrO[44] + GCr2O3 [4] + A + BT 6 6 2 6 6 6 6
(4.4.5)
are optimized with all available experimental data of the perovskite phase. Eq. 4.4.4 denotes
G (Sr2+)(Cr4+)(O2-)3, with A = 27027 and B = 69.6. A = 136453 and B = 91.2 for
GSrCrO2.5Va0.5 in Eq. 4.4.5. Gs of the remaining endmembers (La )(Cr )(Va)3,
3+
4+
(La3+)(Cr4+)(O2-)3, (Sr2+)(Cr3+)(O2-)3, and (Sr2+)(Cr3+)(Va)3 are obtained by conversions of reciprocal equations that are set zero[48] and are listed in Table 4.4.1 (p. 138). Though structure-chemical information of site occupancies in LaMn1-xCrxO3- perovskite is missing, it is reliable to allow Cr4+ on the B-site: as Cr4+ exists in nonstoichiometric lanthanum chromite perovskite[5], it is expected that it is not removed from the structure if the phase is doped. Thus for LaMn1-xCrxO3- we propose the sublattice formula (La3+,Va)(Mn2+,Mn3+,Mn4+,Cr3+,Cr4+,Va)(O2-,Va)3. All endmember compounds have been defined in the assessed subsystems. The regular interaction parameter 0L(La3+:Cr3+,Mn3+:O2-) accounting for interactions between Cr and Mn cations is fitted to experimental nonstoichiometries[26]; 0L(La3+:Cr3+,Mn3+:O2-) = +9421 J. The
3+ 2+
sublattice
2+ 3+
formula
4+ 3+
of
4+
the
2-
quinary
perovskite
reads
(La ,Sr ,Va)(Mn ,Mn ,Mn ,Cr ,Cr ,Va)(O ,Va)3. All endmembers have been defined in the assessed subsystems.
4.4.4
Results and discussion
The reproduction of experimentally determined Gibbs energies[21] and enthalpies of formation[22], solid solubilities[18,20], and nonstoichiometries[23,24] of La1-xSrxCrO3-, and phase equilibria in the La-Sr-Cr-O oxide system by the modeling is satisfying as shown in Table 4.4.2 (p. 138), and in Figs. 4.4.1 and 4.4.2 (next page).
140
Fig. 4.4.1 LaO1.5-SrO-CrO1.5 system calculated at T = 1223 K in air atmosphere (solid lines)
with experimental data[17] included (symbols). prv = La1-xSrxCrO3-. Calculated phase equilibria are the same as in[17]. Filled circles, blank circles, and circles with crosses denote single phase, two phase, and three phase equilibria.
141
Fig. 4.4.2, p. 141 Calculated (lines) and experimental (symbols)[22,23] nonstoichiometries of
La1-xSrxCrO3- at different temperatures for x = 0.1, 0.2, 0.25, and 0.3 as a function of oxygen partial pressure. The calculated isothermal section of the La-Mn-Cr-O oxide system at T = 1273 K in air is presented in Fig. 4.4.3.
Fig. 4.4.3 LaO1.5-MnOx-CrO1.5 system calculated at T = 1273 K in air atmosphere. -spl =
tetragonally distorted Cr-Mn-spinel, -spl = cubic Cr-Mn-spinel, prv = LaMn1-xCrxO3-. The calculated nonstoichiometries of La1-xSrxCrO3- are in good agreement with the experimental values at higher temperatures, as shown in Fig. 4.4.4 (next page). However it was not possible to reproduce the nonstoichiometries at T = 1073 K and 973 K. Deducing from the change of from T = 1273 K to 1173 K the measured increase of from T = 1173 K to 1073 K might be too small, possibly caused by equilibration difficulties due to slow diffusion.
142
Fig. 4.4.4 Calculated (lines) and experimental (symbols)[27] nonstoichiometries of
LaMn0.9Cr0.1O3- at different temperatures as a function of oxygen partial pressure. To approximate the absence of Cr4+[27] in quinary perovskite, it would be necessary to give large positive values to the regular interaction parameters
0 0
L(Sr2+:Cr3+,Mn3+:O2-) and
L(Sr2+:Cr4+,Mn3+:O2-). Experimentally determined nonstoichiometry of LaCrO3 indicates the
existence of some Cr4+, and the conclusion of missing Cr4+[27] is not based on a direct chemical analysis of Cr valencies. We believe that complete removal of Cr4+ from the perovskite structure is unlikely. Thus we stick to a model without interaction parameters. Experimental findings[8,28] are in line with our calculations.
4.4.5
Conclusions
The thermodynamic La-Sr-Mn-Cr-O oxide database has been obtained by combining thermodynamic assessments of oxide subsystems. We propose the model (La3+,Sr2+,Va)(Mn2+,Mn3+,Mn4+,Cr3+,Cr4+,Va)(O2-,Va)3 for the quinary perovskite phase. Optimized by experiments in pseudoternary and pseudoquaternary oxide subsystems, this model allows the quantitative calculation of defects as a function of composition, temperature, and oxygen partial pressure. The new database is adapted for quantitative calculations of
143
phase equilibria and defect chemistry in a Sr-doped lanthanum manganite SOFC cathode poisoned by chromium.
References
1. S. P. S. Badwal, R. Deller, K. Foger, Y. Ramprakash and J. P. Zhang, Interaction between chromia forming alloy interconnects and air electrode of solid oxide fuel cells, Solid State Ionics, 1997, 99, pp. 297-310. 2. S.P. Jiang, J.P. Zhang, K. Foger, Deposition of chromium species at Sr-doped LaMnO3 electrodes in solid oxide fuel cells I. Mechanism and kinetics, J. Electrochem. Soc., 2000,
147, pp. 4013-4022.

3. M. Zinkevich, S. Geupel, F. Aldinger, A. Durygin, S.K. Saxena, M. Yang, Z.-K. Liu, Phase diagram and thermodynamics of the La2O3-Ga2O3 system revisited, J. Phys. Chem.
Solids, 2006, 67, pp. 1901-07.

4. E. Povoden, A.N. Grundy, and L.J. Gauckler, Thermodynamic reassessment of the Cr-O system in the framework of solid oxide fuel cell (SOFC) research, J. Phase Equilib. Diff., 2006, 27(4), pp. 353-62. 5. E. Povoden, M. Chen, A.N. Grundy, T. Ivas, L.J. Gauckler, Thermodynamic assessment of the La-Cr-O system, J. Phase Equilib. Diff., accepted 6. A.N. Grundy, B. Hallstedt, L.J. Gauckler, La1-xMn1-yO3-z perovskites modelled with and without antisite defects using the Calphad approach, Solid State Ionics, 2004, 173, pp. 1721. 7. R.D. Shannon, C.T. Prewitt, Effective ionic radii in oxides and fluorides, Acta Crystall. B-
Stru., 1969, B25, pp. 925-.

8. E.A. Filonova, A.N. Demina, A.N. Petrov, Phase equilibria in the system LaMnO3SrMnO3-SrCrO4-LaCrO3, Russ. J. Inorg. Chem., 2007, 52, pp. 771-774. 9. K. Hack (ed.), The SGTE casebook: thermodynamics at work, Institute of Materials, London, 1996. 10. H. Yokokawa, N. Sakai, T. Kawada, M. Dokiya, Chemical thermodynamic considerations in sintering of LaCrO3-based perovskites, J. Electrochem Soc., 1991, 138, pp. 1018-27. 11. K.T. Jacob, K.P. Abraham, Phase relations in the system Sr-Cr-O and thermodynamic properties of SrCrO4 and Sr3Cr2O8, J. Phase Equilib., 2000, 21, pp. 46-53.
144
12. T. Maruyama, T. Inoue, T. Akashi, Standard Gibbs energies of formation of SrCrO4 and Sr3Cr2O8, Mater. T. Jim., 1998, 39, pp. 1158-61. 13. A.M. Azad, R. Sudha, O.M. Sreedharan, The standard Gibbs energies of formation of ACrO4 (A=Ca, Sr or Ba) from emf-measurements, Thermochim. Acta, 1992, 194, pp. 129-36. 14. R. Akila and K.T. Jacob, The mobility of oxygen ions in CaF2, J. Appl. Electrochem., 1990, 20, pp. 294-300. 15. T. Negas, R.S. Roth, System SrO-chromium oxide in air and oxygen, J. Res. NBS A Phys.
Ch., 1969, 73A, pp. 431-42.

16. Y.K. Kisil, N.G. Sharova, B.V. Slobodin, Phase formation in the system SrO-CrO3-Cr2O3,
Inorg. Mater., 1989, 25, pp. 1490-91.

17. E. Castillo-Martnez, M.A. Alario-Franco, Revisiting the Sr-Cr(IV)-O system at high pressure and temperature with special reference to Sr3Cr2O7, Solid State Sci., 2007, 9, pp. 564-73. 18. D.H. Peck, M. Miller, K. Hilpert, Phase diagram studies in the SrO-Cr2O3-La2O3 system in air and under low oxygen pressure, Solid State Ionics, 1999, 123, pp. 59-65. 19. K. Hartl, R. Braungart, Strontiumchromate(V, VI), Sr2.67Va0.33(CrO4)1.33(CrO4)0.67, a hightemperature compound with defect-bariumphosphate-structure, Z. Naturforsch. Pt. B, 1978, 33, pp. 952-953 (in German). 20. S. Myoshi, S. Onuma, A. Kaimai, H. Matsumoto, K. Yashiro, T. Kawada, J. Mizusaki, H. Yokokawa, J. Solid State Chem., 2004, 177, pp. 4112-. 21. D.-H. Peck, M. Miller, K. Hilpert, Vaporization and thermodynamics of La1-xSrxCrO3- investigated by Knudsen effusion mass spectrometry, Solid State Ionics, 2001, 143, pp. 401-12. 22. J. Cheng, A. Navrotsky, Energetics of La1-xAxCrO3- perovskites (A = Ca or Sr), J. Solid
State Chem., 2005, 178, pp. 234-44.

23. J. Mizusaki, S. Yamauchi, K. Fueki, A. Ishikawa, Nonstoichiometry of the perovskitetype oxide La1-xSrxCrO3-, Solid State Ionics, 1984, 12, pp. 119-24. 24. K. Hilpert, R.W. Steinbrech, F. Boroomand, E. Wessel, F. Meschke, A. Zuev, O. Teller, H. Nickel, L. Singheiser, Defect formation and mechanical stability of perovskites based on LaCrO3 for solid oxide fuel cells (SOFC), J. Europ. Ceram. Soc., 2003, 23, pp. 300920.
145
25. H. Yokokawa, T. Horita, N. Sakai, K. Yamaji, M.E. Brito, Y.-P. Xiong, H. Kishimoto, Thermodynamic considerations on Cr poisoning in SOFC cathodes, Solid State Ionics, 2006, 177, pp. 3193-98. 26. L. Morales, A. Caneiro, Evolution of crystal structure with the oxygen content in the LaMn0.9Cr0.1O3+ (3.00 3+ 3.12) compound, J. Solid State Chem., 2003, 170, pp. 404-10. 27. S.M. Plint, P.A. Connor, S. Tao, J.T.S. Irvine, Electronic transport in the novel SOFC anode material La1-xSrxCr0.5Mn0.5O3, Solid State Ionics, 2006, 177, pp. 2005-08. 28. T. Komatsu, R. Chiba, H. Arai, K. Sato, J. Power Sources, 2008, 176, pp. 132-. 29. C.P. Khattak, D.E. Cox, Structural studies of (La,Sr)CrO3 system, Mat. Res. Bull., 1977,
12, 463-71.
30. K. Tezuka, Y. Hinatsu, A. Nakamura, T. Inami, Y. Shimojo, Y. Morii, Magnetic and neutron diffraction study on perovskites La1-xSrxCrO3, J. Solid State Chem., 1998, 141, pp. 404-10. 31. K. Tezuka, Y. Hinatsu, K. Oikawa, Y. Shimojo, Y. Morii, Studies on magnetic properties of La0.95Sr0.05CrO3 and La0.85Sr0.15CrO3 by means of powder neutron diffraction, J. Phys.:
Condens. Matter, 2000, 12, pp. 4151-60.

32. F. Nakamura, Y. Matsunaga, N. Ohba, K. Arai, H. Matsubara, H. Takahashi, T. Hashimoto,
xSrxCrO3
Analysis of magnetic and structural phase transition behaviors of La1-
for preparation of phase diagram, Thermochim. Acta, 2005, 435, 222-29.
33. K.R. Chakraborty, S.M. Yusuf, P.S.R. Krishna, M. Ramanadham, A.K. Tyagi, V. Pomjakushin, Structural study of La0.75Sr0.25CrO3 at high temperatures, J. Phys.: Condens.
Matter, 2006, 18, 8661-72.

34. Y. Matsunaga, F. Nakamura, H. Takahashi, T. Hashimoto, Analysis of relationship between magnetic property and crystal structure of La1-xSrxCrO3 (x=0.13, 0.15), Solid
State Comm., 2008, 145, 502-06.

35. Y. Matsunaga, H. Kawaji, T. Atake, H. Takahashi, T. Hashimoto, Thermochim. Acta, 2008, 474, pp. 57-. 36. O. Cabeza, M. Long, M.A. Bari, C.M. Muirhead, M.G. Francesconi, C. Greaves, Magnetization and resistivity in chromium doped manganites, J. Phys.: Condens. Matter, 1999, 11, pp. 2569-78.
146
37. Z. El-Fadli, M.R. Metni, F. Sapia, E. Martinez, J.V. Folgado, D. Beltrn, A. Beltrn, Structural effects of Co and Cr substitution in LaMnO3+, J. Mater. Chem., 2000, 10, pp. 437-43. 38. M. Tseggai, P. Nordblad, R. Tellgren, H. Rundlf, G. Andr, F. Boure, Synthesis, nuclear structure, and magnetic properties of LaCr1-yMnyO3 (y=0, 0.1, 0.2, and 0.3), J.
Alloy Compd., 2008, 457, 532-40.

39. S.A. Howard, J.-K. Yau, H.U. Anderson, X-ray-powder diffraction structural phasetransition study of La(Cr1-xMnx)O3 (x=0 to 0.25) using the Rietveld method of analysis, J.
Am. Ceram. Soc., 1992, 75, pp. 1685-87.

40. M. Hrovat, S. Bernik, J. Holc, D. Kuscer, D. Kolar, Preliminary data on solid solubility between LaCrO3 and LaFeO3 or LaMnO3, J. Mater. Sci. Lett., 1997, 16, pp. 143-46. 41. N. Kallel, J. Dhahri, S. Zemni, E. Dhahri, M. Oumezzine, M. Ghedira, H. Vincent, Effect of Cr doping in La0.7Sr0.3Mn1-xCrxO3 with 0 x 0.5, Phys. Stat. Sol. (a), 2001, 184, pp. 319-25. 42. G. Inden, Determination of chemical and magnetic interchange energies in bcc alloys. I. General treatment, Z. Metallkd., 1975, 66(10), pp. 577-82. 43. M. Hillert and M. Jarl, A model of alloying effects in ferromagnetic metals, Calphad, 1978, 2(3), pp. 227-38. 44. D. Risold, B. Hallstedt, L.J. Gauckler, The strontium-oxygen system, Calphad, 1996, 20, pp. 353-61. 45. A.T. Dinsdale, SGTE data for pure elements, Calphad, 1991, 15(4), pp. 317-425. 46. B. Sundman, B. Jansson, J.-O. Andersson, The thermo-calc databank system, Calphad, 1985, 9, 153-90. 47. M. Hillert, The compound energy formalism, J. Alloy. Cmpd., 2001, 320, pp. 161-76. 48. A.N. Grundy, M. Chen, B. Hallstedt, L.J. Gauckler, Assessment of the La-Mn-O system,
J. Phase Equilib. Diff., 2005, 26, 131-51
147
Thermodynamic calculations of impacts of chromium on LSM cathodes
5 Thermodynamic calculations of impacts of chromium on Srdoped lanthanum manganite (LSM) cathodes for solid oxide fuel cells (SOFC)
E. Povoden, T. Ivas, M. Chen, and L.J. Gauckler, to be submitted A new thermodynamic database is used for thermodynamic equilibrium calculations in a Srdoped lanthanum manganite cathode (LSM) affected by chromium at typical operation temperatures of 1073 K and 1273 K as a function of oxygen partial pressure. From the results of these calculations it is concluded that the processes of chromium poisoning of solid oxide fuel cells (SOFC) are partly explicable by thermodynamics, and partly they occur under kinetic control: at the cathode/electrolyte interface of a Cr-poisoned cell Cr-Mn spinel exists in thermodynamic equilibrium with LSM, whereas Cr2O3 is metastable. The spinel formation goes along with increasing Mn2+ in LSM under decreasing oxygen partial pressures. From the thermodynamic calculations structural chemical changes in the cathode perovskite caused by the interaction with chromium can be predicted: it is shown that the interaction of chromium with the LSM cathode leads to a change of the defect chemistry of the perovskite phase. In particular the concentrations of cation and oxygen vacancies are smaller than in an LSM without chromium under decreased oxygen partial pressure at 1273 K. This has consequences for the electrochemical properties of the cell: the electronic conductivity of the cathode will decrease, and the contribution of a vacancy mechanism for the oxygen diffusion in LSM is thermodynamically hampered in the presence of chromium at high temperature and high current loads. Even though the chromium problem cannot be solved satisfactorily by varying the cathode composition or the SOFC operating conditions, the deterioration of the cell performance is expected to be less pronounced when the cell is operated at lower temperatures and current loads. Proper strategies to prevent the problem of chromium poisoning are proposed.
148
Thermodynamic calculations of impacts of Cr on LSM cathodes
5.1
Introduction
Chromium-containing metallic interconnects are commonly used in planar-design solid oxide fuel cells (SOFC) due to their high oxidation resistance, thermal stability, mechanical strength, good electronic and negligible ionic conductivity, as well as low fabrication costs. However high-valent gaseous Cr-oxide and chromium-oxyhydroxides can diffuse under fuelcell operation conditions from the interconnect into the cathode up to the cathode-electrolyte interface, where they cause the degradation of the cell by detrimentally affecting the O2adsorbtion, -reduction, and -diffusion process[1]. In the last decade a lot of efforts were made to elucidate the degradation mechanisms, though partly with conflicting results. Consequences of Cr poisoning have been investigated specifically in (La1-xSrx)MnO3- (LSM) perovskite-structured cathodes. For the mechanism of chromium poisoning two models have been proposed: 1) reduction of gaseous CrO3(g) in dry atmosphere or chromium oxyhydroxide(g) in wet atmosphere under polarization[2-6] and 2) chemical dissociation of Crspecies on the LSM surface[7-14]. Ad 1) In an LSM cathode the reduction of CrO3(g) is expected to be localized at the triple phase boundary, where the reaction partners for the reduction, electron-donating LSM and oxygen-accepting yttrium-stabilized zirconia (YSZ) are available[15]. This reduction reaction would compete with the oxygen reduction and lead to blocking of the active sites at the triple phase boundary (TPB). Ad 2) In contrast to 1) it was proposed that gaseous Cr-species would be chemically dissociated to LSM under the polarization of the cell. This affinity would be linked to the creation of free Mn2+ on the surface of LSM due to the oxygen partial pressure gradient caused by the polarization. Mn2+ would serve as agent for the formation of Cr-Mn-O nuclei that would be able to migrate to the triple phase boundary and further into the electrolyte. Consequently Cr-Mn spinel and Cr2O3(s) would form, associated to these nuclei. The chemical dissociation approach is coherently based on the interpretation of a large number of impedance spectra. Both groups of researchers agree that without polarization Cr is randomly deposited inside the cathode, and no Cr2O3(s) is formed. On the other hand the electrochemical reduction of CrO3(g) was rejected by the authors favoring the chemical dissociation approach.
149
In the critical assessment in chapter 1.3.6 it was concluded that doubtless reasons to reject the reduction approach do not exist. One critical point concerns the extension of dense Cr2O3layers into the YSZ electrolyte[6]: this phenomenon can be explained best by continuous feeding of an initial Cr2O3-layer with CrO3(g), the latter becoming reduced at a new TPB consisting of YSZ and electron-donating Cr2O3(s). On the other hand this process cannot be explained satisfactorily by using the chemical dissociation approach. Even though particularly the early stages of chromium poisoning occur in thermodynamic non-equilibrium, the system SOFC develops towards thermodynamic equilibrium by time. This is reflected by a flattening of the curves that reflect the performance deterioration as a function of time, such as the curves of voltage drop and overpotential loss. Thus thermodynamic calculations allow interpretations of the phase equilibria that result from the interactions between LSM and chromium, as well as changes of the phase chemistry that are associated with the chromium contamination of LSM cathodes.
5.2
Method
The La-Sr-Mn-Cr-O oxide database is used for the following thermodynamic calculations: phase equilibria in Cr-contaminated LSM (in the following denoted as LSM(Cr)), phase compositions of LSM(Cr) and Cr-Mn spinel, defect concentrations of LSM(Cr), as well as driving forces for the formation of Cr2O3 were calculated with the poly-module of the ThermoCalc software[16]. The following model descriptions were used: for the Cr-contaminated cathode perovskite with the
3+
general
2+
formula
4+ 3+
ABO3
2+ 4+ 3+
proper
2-
sublattice for cubic
description spinel,
reads AB2O4,
(La ,Sr ,Va)(Mn ,Mn ,Mn ,Cr ,Cr ,Va)(O ,Va)3, for tetragonally distorted spinel (Mn2+)(Mn3+,Cr3+)2(O2-)4 was chosen[17], (Mn2+,Cr2+)(Mn3+,Cr3+)2(O2-)4 was used[18], and for Cr2O3 (Cr2+,Cr3+)2(Cr3+,Va)(O2-)3 was taken[18]. Uptake of Cr in LSM is expected, as a complete solid solubility of Cr in LSM has been shown experimentally[19]. For proper thermodynamic calculations of phase equilibria thermodynamic conditions need to be set that reflect the conditions of the chromium contamination of SOFC: the bulk pressure (room pressure, 101325 Pa), the operation temperature (typically from T = 1073 K to 1273 K), the oxygen partial pressure, the cathode composition, and the amount of chromium.
150
The oxygen partial pressure at the interconnect-cathode interface is air. Under current load it is expected that the oxygen partial pressure will strongly decrease close to the cathodeelectrolyte interface in the triple phase boundary (TPB) region where the oxygen reduction in LSM takes place: the oxygen partial pressure at the cathode-electrolyte interface, pO2(i) can be approximated from the measured cell voltage of a Pt/LSM/YSZ/Ni-Cermet/Pt solid oxide cell and the fuel composition by using the equation for the overall electromotive force E of the cell:
pO RT 2(i) ln 4 F pO
E=
(5.2.1)
2(an)
R = 8.31451 J mol-1 K-1, F = 96485.309 C mol-1 and pO2(an) is given by the ratio of H2-H2O in
the fuel. From a measured cell voltage of 0.7 V[2] at T = 1173 K (fuel: 97 vol.% H2, 3 vol.% H2O) and a high current load of 300 mA cm-2 a strong decrease of the oxygen partial pressure at the oxygen reduction sites is expected, pO2(i) 0.01 Pa. As we are interested in the influences of chromium throughout a cathode under realistic operation conditions of SOFC, results of the thermodynamic calculations are presented for pO 21278 Pa 0.01 Pa.
2
Several LSM cathode compositions can be found in the literature. Part of them is cation stoichiometric, and part of them has excess Mn that is known to prevent unwanted formation of electrochemically isolating zirconate at the cathode/electrolyte interface. In this study two cathode compositions are used for the thermodynamic calculations: La0.9Sr0.1MnO3- and (La0.8Sr0.2)0.9MnO3-. The sublattice model for this perovskite phase[20] allows the formation of vacancies on each site and changing valencies of Mn as a function of temperature and oxygen partial pressure. The amount of chromium in the system is defined by the partial pressure of the Cr-gas phase:
Cr RT
pCr = exp
(5.2.2)
This means that by knowing the partial pressure of the Cr-gas phase in the TPB region, it is possible to calculate the thermodynamics of the chromium contamination. The problem is that the definite amount of gas that contributes to the degradation phenomena is not known
151
exactly, as only a fraction of the Cr-gas that evaporates from the Cr2O3 scale on the Cr-alloy interconnect interacts with LSM or is reduced. Fortunately the amount of deposited Cr in a degraded LSM cathode has been analyzed as a function of distance from the cathode/YSZ electrolyte interface[21], and the combined data of X(Cr) and the oxygen partial pressure at the TPB fix the chemical potential of Cr. The amount of deposited Cr close to the LSM(Cr)/YSZ interface was about 3 wt.% after a long cell test of 300 h at T = 1073 K. If one assumes that the pO2 under the test conditions was 1 Pa at the the LSM(Cr)/YSZ interface (normal cell performance), the chemical potential of the Cr-gas phase can be calculated. Even though it is clear that the chemical potential of Cr will change if the amount of evaporated Cr from different interconnect materials is different, the Cr-gas reservoir is assumed to be in a saturated state due to unlimited supply from the interconnect during the cell performance, and thus its chemical potential is fixed in the thermodynamic calculations. This simplification is reasonable, as in all investigated cell tests with LSM and Cr-alloy interconnects the degradation was similar, so that changing chromium amounts due to different interconnect alloys are obviously not detrimental for the cell degradation. H2O (operation of SOFC in humid air) is not considered in the calculations, as neither hydroxides nor solubilities of hydrogen or OH were included in the La-Sr-Mn-Cr-O oxide database.
5.3
5.3.1
Results
Thermodynamic calulcations of La0.9Sr0.1MnO3 contaminated by chromium
Fig. 5.3.1 (next page) shows phase fractions in Cr-poisoned La0.9Sr0.1MnO3- at constant chemical potential of CrO3, (CrO3) = 300000 J mol-1 referred to 100000 Pa of CrO3(g) as a function of oxygen partial pressure at T = 1273 K and 1073 K, and in Figs. 5.3.2 (next page) and 5.3.3 (p. 153) phase equilibria are indicated: the cathode remains single phase at pO2 > 102.75 Pa. By decreasing the oxygen partial pressure, tetragonally distorted Mn3O4 spinel (t-sp), the manganese endmember of the Cr-Mn spinel solid solution phase forms.
152
Fig. 5.3.1 phase fractions in Cr-poisoned La0.9Sr0.1MnO3- as a function of oxygen partial
pressure at T=1273 K and 1073 K at (CrO3) = 300000 J mol-1
Fig. 5.3.2 Phase equilibria in Cr-poisoned La0.9Sr0.1MnO3- and defect concentrations of
La0.9Sr0.1(Mn,Cr)O3- as a function of oxygen partial pressure at T = 1273 K and
(CrO3) = 300000 J mol-1. A, B, and C denote sublattices of the perovskite phase, with A
and B standing for the cation sublattices and C standing for the oxygen sublattice. Vertical lines indicate boundaries between different phase equilibria 153
Fig. 5.3.3 Phase equilibria in Cr-poisoned La0.9Sr0.1MnO3- and defect concentrations of
La0.9Sr0.1(Mn,Cr)O3- as a function of oxygen partial pressure at T = 1073 K and

(CrO3)= 300000 J mol-1. The vertical line indicates the boundary between phase equilibria
At T = 1273 K (Figs. 5.3.1, p. 153 and 5.3.2, p. 153), tetragonally distorted spinel remains stable to pO2 = 10-0.4 Pa. At lower pO2 cubic Mn-Cr spinel forms. At 1073 K (Figs. 5.3.1, p. 152 and 5.3.3), tetragonally distorted spinel remains stable to pO2 = 100.75 Pa, followed by the formation of cubic spinel at lower pO2 . This means that by decreasing the oxygen partial pressure from pO2 = 104.3, the pressure of air, to 10-1.5 Pa, the amount of spinel in the contaminated cell increases. At 1073 K Cr-Mn spinel formation is less pronounced, and CrMn spinel formation starts at lower pO2 than at 1273 K. To find out about the structural chemical changes in the cathode perovskite caused by reaction with chromium, the fractions of species in a specific sublattice (site fractions) are calculated at
T=1273 K and 1073 K (plots in Figs. 5.3.2, p. 152 and 5.3.3) as a function of pO2 The results .
are compared with the calculated site fractions in a cathode with a very small chemical
154
potential of Cr, (CrO3) = 106 J mol-1 that means with practically no Cr (Figs. 5.3.4 to 5.3.5).
Fig. 5.3.4 Defect concentrations in La0.9Sr0.1MnO3-
as a function of oxygen partial pressure at T=1273 K.
Fig. 5.3.5 Defect concentrations in La0.9Sr0.1MnO3-
as a function of oxygen partial pressure at T = 1073 K. 155
In general defect concentrations of the Cr-contaminated LSM differ from the defect concentrations in LSM without Cr at a high temperature of 1273 K: with Cr the concentrations of vacancies on the A- and B-sublattices decrease stronger by decreasing the oxygen partial pressure. The increase of oxygen vacancies by decreasing the oxygen partial pressure on the other hand is weaker when chromium is present. At T = 1273 K and pO2 = 1 Pa, which is the expected pO2 at the LSM/YSZ interface at 250 mA cm-2 current load, the site fractions of cation vacancies on the A- and B-sublattices for LSM(Cr) are y(Va)A = 1.98x10-6, y(Va)B=4.3x10-6, whereas in LSM y(Va)A = 3.086x10-6 and y(Va)B = 7.096x10-6 are calculated. The concentration of oxygen vacancies at 1 Pa and T = 1273 K in LSM(Cr) is slightly higher than in LSM, y(Va)C = 3.01x10-5 in LSM(Cr),compared to y(Va)C = 2.57x10-5 in LSM. A pronounced drop of cation and oxygen vacancies is calculated at 1273 K and pO2 = 10-1 Pa, the expected oxygen partial pressure at the TPB under a high current load of 300 mA cm-2: the concentration of oxygen vacancies in LSM(Cr) is y(Va)C = 3.39x10-5, compared to
y(Va)C = 9.48x10-5 in LSM. This means that if the oxygen partial pressure at the LSM/YSZ
interface strongly decreases the vacancy concentrations will drop significantly. The concentrations of Cr3+ and Cr4+ in LSM(Cr) increase when the temperature increases and the oxygen partial pressure decreases. The calculated compositions of tetragonally distorted spinel (Fig. 5.3.6 a, next page) and cubic spinel (Fig. 5.3.6 b) formed during chromium poisoning show a strong dependence upon the oxygen partial pressure: only under low oxygen partial pressures a significant amount of chromium is found in the spinel phase, whereas at higher oxygen partial pressures the spinel phase has a composition close to Mn3O4. At T = 1073 K the spinel phase contains less chromium than at T = 1273 K.
156
Fig. 5.3.6 Calculated site fractions of ions in cubic spinel (6 a) and tetragonally distorted
spinel (6 b) formed during chromium poisoning at T = 1273 K and 1073 K
5.3.2
Thermodynamic calculations of (La0.8Sr0.2)0.9MnO3- contaminated by chromium
From Fig. 5.3.7 it is obvious that in this widely used LSM composition Cr-poisoning leads to the formation of additional phases already at high oxygen partial pressures: A small amount of about 5 mol% of the pure spinel endmember, tetragonally distorted Mn3O4 (t-sp) is expected to form. At T = 1073 K Mn2O3 is stable in a Cr-contaminated LSM cathode with excess Mn in air.
Fig. 5.3.7 phase fractions in Cr-poisoned (La0.8Sr0.2)0.9MnO3- as a function of
oxygen partial pressure at T = 1273 K and 1073 K and (CrO3) = 300000 J mol-1 157
In Fig. 5.3.8 the compositional changes of cubic spinel are plotted as a function of oxygen partial pressure at T = 1273 K. In general, the compositions of the spinel phases formed during chromium poisoning become richer in Cr under more reducing conditions, as in the case of cation-stoichiometric LSM.
Fig. 5.3.8 Calculated site fractions of ions in cubic spinel formed
during chromium poisoning at T = 1273 K It is interesting whether the consequences of chromium for the concentrations of defects in LSM(Cr) with excess Mn are more or less pronounced than in cation-stoichiometric LSM(Cr): Phase equilibria and defect concentrations in a (La0.8Sr0.2)0.9MnO3- cathode are shown in Fig. 5.3.9 (next page).
158
Fig. 5.3.9
Phase equilibria in Cr-poisoned (La0.8Sr0.2)0.9(Mn,Cr)O3- and defect
concentrations in (La0.8Sr0.2)0.9(Mn,Cr)O3- as a function of oxygen partial pressure at T = 1273 and (CrO3) = 300000 J mol-1. The vertical line indicates the boundary between different phase equilibria Fig. 5.3.10 (next page) is a comparison of defect concentrations of (La0.8Sr0.2)0.9MnO3- with Cr (broken lines in Fig. 5.3.10) and without Cr (solid lines in Fig. 5.3.10) at 1273 K. The vacancy concentrations on the A-sites and B-sites of the Cr-contaminated perovskite phase are basically in the middle between these vacancy concentrations in LSM. In LSM(Cr) the concentrations of these cation vacancies drop strongly at low pO2 . Mn2+ is higher in LSM(Cr) than in LSM at higher pO2 , and the concentration of oxygen vacancies is lower in LSM(Cr) than in LSM at relatively high and low pO2 .
159
Fig. 5.3.10 Defect concentrations in (La0.8Sr0.2)0.9(Mn,Cr)O3- (dashed lines) and
(La0.8Sr0.2)0.9MnO3- (solid lines) as a function of oxygen partial pressure at T = 1273. Calculated concentrations of all species in LSM(Cr) and tetragonally distorted spinel in equilibrium are listed in Table 5.3.1.
Table 5.3.1 Compositions of LSM(Cr) and spinel in equilibrium at different T at pO2=1 Pa with and without Cr.
5.3.3
Thermodynamic testing of LSM with Mn-deficiency
Only in a cathode with Mn-deficiency it is possible to push the formation of additional phases towards a lower oxygen partial pressure: for the case of La0.871Sr0.148Mn0.947O3- spinel formation becomes important only at pO2 < 0.1 Pa, as it is illustrated in Fig. 5.3.11, next page.
160
Fig. 5.3.11 Phase fractions in a Cr-poisoned Mn-deficient LSM as a function of
oxygen partial pressure at T = 1273 and 1073 K and (CrO3) = 300000 J mol-1. The influence of chromium on defect concentrations in La0.871Sr0.148(Mn,Cr)0.947O3- is illustrated in Fig. 5.3.12, next page: The concentration of oxygen vacancies in La0.871Sr0.148(Mn,Cr)0.947O3- is half of an order of magnitude higher than in
(La0.8Sr0.2)0.9MnO3- at high oxygen partial pressures. However, after reaching a plateau at

pO2 = 103 Pa, y(Va)C even decrease slightly towards lower pO2 , and the concentration of
oxygen vacancies is almost an order of magnitude lower then in (La0.8Sr0.2)0.9MnO3- at pO2 = 10-1 Pa.
161
Fig. 5.3.12 Phase equilibria in Cr-poisoned Mn-deficient LSM and defect concentrations in
La0.871Sr0.148(Mn,Cr)0.947O3- as a function of oxygen partial pressure at T = 1273 K and

(CrO3) = 300000 J mol-1 (solid lines). Dashed lines indicate the defect concentrations in
(La0.8Sr0.2)0.9MnO3 without chromium. Vertical lines indicate boundaries between different phase equilibria
5.3.4
Formation of Cr2O3
This phase was not found in the thermodynamic calculations, and thus its formation is kinetically controlled. One can get an idea about the degree of metastability of a phase by calculating its thermodynamic driving force. This is the amount of energy that is needed to bring the phase to its stable state. The more negative the driving force, the more energy is needed to stabilize the phase, and the driving force for the formation of the phase is low. If the driving force is 0, the phase is thermodynamically stable. In Fig. 5.3.13 (next page) the driving force of Cr2O3 is plotted as a function of temperature at two different pO2 in a LSM cathode with excess Mn under Cr-poisoning.
162
Fig. 5.3.13 Driving force of Cr2O3 as a function of temperature
at different pO2 at (CrO3) = 300000 J mol-1. The driving force for the formation of Cr2O3 is less negative at higher oxygen partial pressures.
5.4
Discussion
In the following the results of the thermodynamic calculations are compared to experimental findings on chromium poisoning from the literature. Interpretations are given, which of the chromium poisoning mechanisms occur under thermodynamic control. By carrying out equilibrium calculations of state-of-the-art LSM cathodes with the compositions La0.9Sr0.1MnO3- and (La0.8Sr0.2)0.9MnO3- at constant chromium in the gas phase it was tested if spinel formation would be favored thermodynamically under low oxygen partial pressure, i.e. close to the electrode-electrolyte interface under polarization conditions. The calculations showed that this is indeed the case. As the A-sublattice of the spinel is completely filled by Mn2+ under the cell operation conditions, and the only source for this species is LSM, it is obvious that spinel formation will be associated with increasing Mn2+ in LSM. Thus, as Mn2+ in LSM increases as a function of decreasing pO2 , also the amount of
163
spinel formed is higher at low oxygen partial pressure. Cr-gas reveals increasing affinity to LSM towards the electrode-electrolyte interface: both Cr solid solution in LSM and spinel formation increase under decreasing the oxygen partial pressure. From the calculation it is interpreted that the spinel phase that forms under Cr-poisoning of the cathode will contain a high amount of Mn, if the oxygen partial pressure at the cathode/electrolyte interface is about 1 Pa. Only at lower pO2 significant Cr is incorporated in the spinel phase, with the stoichiometric MnCr2O4 phase forming at about pO2 = 10-1 Pa. Though spinel formation is thermodynamically driven in Cr-contaminated SOFC, it seems that spinel formation per se is not one of the keys of severe cell degradation due to chromium, but the affinity of Cr-gas to the LSM surface, as even very small Cr contamination in the ppm range apparently leads to a dramatic decrease of the oxygen diffusion in LSM[22]. From impedance spectroscopy analyses it was consistently concluded that the oxygen diffusion is severely influenced by chromium. The thermodynamic calculations showed that Cr interacting with LSM leads to a change of the defect chemistry of the perovskite phase, and particularly to a decreasing amount of oxygen vacancies at high temperatures and low oxygen partial pressures. As the formation of oxygen vacancies in LSM is inhibited, oxygen diffusion to the triple phase boundary is retarded. The results of the thermodynamic defect chemistry calculations of LSM(Cr) thus indicate that the deterioration of the oxygen diffusion is higher under at decreased oxygen partial pressures reflecting high current loads. Cr2O3 is found in degraded SOFC, particularly under high current load. However this phase was not found in the thermodynamic calculations, and its driving force remains negative under SOFC operating conditions. This means that its formation is kinetically controlled. Even though Cr2O3 is not a thermodynamically stable phase in Cr-contaminated SOFC, a strong tendency exists for CrO3(g) to be reduced to Cr2O3(s) at the TPB, as the reduction reaction has a large negative G. It was also mentioned earlier that a high tendency for the precipitation of Cr2O3(s) from CrO3(g) exists[23]. In addition to the adsorption process it is expected that a great many of Cr-gas molecules are driven further to the energy valley for their reduction, namely the TPB. Thus it is non-contradictory that coupling of Cr-gas to LSM and subsequent spinel formation at the LSM surface, and the reduction of CrO3 (g) at the TPB leading to the metastable reduction product Cr2O3(s) occur simultaneously. An alternative way to form Cr2O3 was discussed by Konysheva et al.[24]: The kinetic decomposition of the spinel phase may occur in an oxgen partial pressure gradient due to different mobilities of Mn2+ and Cr3+.
164
In addition to the inhibiting of oxygen diffusion to the TPB and blocking of active triple phase boundary sites by the thermodynamically controlled formation of spinel and the kinetically driven formation of Cr2O3(s) and thus retarded diffusion of oxygen ions into the electrolyte, further unwanted consequences of chromium poisoning can be ascribed to the low electrical conductivity of Cr2O3[25]. Also Cr-Mn-spinel and Cr-doped LSM are significantly less conductive than pure LSM[26-29]. The electrical conductivity of Cr-Mn spinel decreases as its Cr-content increases. From the thermodynamic calculations it can be predicted that increasing the current load will lead to the formation of a spinel phase with a low electrical conductivity. The ohmic resistance of spinel will also increase due to more Cr dissolved in spinel as the amount of deposits of chromium in the cathode increases as a function of time. Furthermore it is expected that the electrical conductivity of LSM is influenced by chromium particularly under high current loads, as chromium leads to decreased concentrations of cation and oxygen vacancies in LSM(Cr) relative to LSM under such operating conditions of SOFC.
5.5
Conclusions
Thermodynamic calculations of LSM contaminated by chromium showed that the formation of spinel is thermodynamically driven, whereas Cr2O3 is a metastable phase that forms under kinetic control in degraded SOFC. The formation of spinel is favored under low oxygen partial pressures, thus in an SOFC under current load this phase is found predominantly at the LSM/YSZ interface. The interaction between chromium and LSM leads to changes of the defect chemistry of the LSM perovskite phase. Particularly diminished concentration of oxygen vacancies relative to LSM without chromium may be a reason for the inhibited oxygen diffusion in degraded SOFC at high temperatures and high current loads . This is also true for Mn-deficient LSM compositions, though the formation of spinel can be restricted to lower oxygen partial pressures. Its defect chemistry is particularly problematic at low oxygen partial pressures, the concentration of oxygen vacancies being strongly diminished relative to LSM with excess Mn. Anyway the use of a Mn-deficient LSM cathode for SOFC is not recommended due to the formation of electrically isolating zirconate.
165
By lowering the operation temperature of SOFC additional phases are expected to form at lower oxygen partial pressures. Thus it is expected that in this case the consequences of Crpoisoning will persist particularly at high current loads. From the thermodynamic point of view it can be summarized that there are neither optimized SOFC operating conditions nor optimized LSM compositions that eliminate the chromium problem in SOFC with LSM cathode and Cr-alloy interconnect. Even though the deterioration of the cell performance due to chromium is expected to be less pronounced when the operation temperature and current load is decreased, chromium poisoning of SOFC with an LSM cathode and Cr-alloy interconnect can only be prevented by applying effective coatings that act as diffusion barrier in combination with additional functional layers. Furthermore, interaction between Mn from LSM with Cr may be cumbered by proper doping of the perovskite with further B-site cations.
References
1. Y. Matsuzaki, I. Yasuda, Electrochemical properties of a SOFC cathode in contact with a chromium-containing alloy separator, Solid State Ionics, 2000, 132, pp. 271-278. 2. S. Taniguchi, M. Kadowaki, H. Kawamura, T. Yasuo, Y. Akiyama, Y. Miyake, T. Saitoh, Degradation phenomena in the cathode of a solid oxide fuel cell with an alloy separator,
J. Power Sources, 1995, 55, pp. 73-9.

3. S.P.S. Badwal, R. Deller, K. Foger, Y. Ramprakash, J.P. Zhang, Interaction between chromia forming alloy interconnects and air electrode of solid oxide fuel cells, Solid State
Ionics, 1997, 99, pp. 297-310.

4. Y. Matsuzaki, I. Yasuda, Dependence of SOFC cathode degradation by chromiumcontaining alloy on compositions of electrodes and electrolytes, J. Electrochem. Soc., 2001, 148, pp. A126-31. 5. H. Tu, U. Stimming, Advances, aging mechanisms and lifetime in solid-oxide fuel cells, J.
Power Sources, 2004, 127, pp. 284-93.

6. S.C. Paulson, V.I. Birss, Chromium poisoning of LSM-YSZ SOFC cathodes, J.
Electrochem. Soc., 2004, 151(11), pp. A1961-68.

7. S.P. Jiang, J.P. Zhang, L. Apateanu, K. Foger, Deposition of chromium species at Srdoped LaMnO3 electrodes in solid oxide fuel cells I. Mechanisms and kinetics, J.
Electrochem. Soc., 2000, 147(9), pp. 3195-3205.
166
8. S.P. Jiang, J.P. Zhang, L. Apateanu, K. Foger, Deposition of chromium species at Srdoped LaMnO3 electrodes in solid oxide fuel cells. I. Mechanisms and kinetics, J.
Electrochem. Soc., 2000, 147, pp. 4013-22.

9. S.P. Jiang, Use of gaseous Cr species to diagnose surface and bulk process for O2reduction in solid oxide fuel cells, J. Applied Electrochem., 2001, 31, pp. 181-192. 10. S.P. Jiang, A comparison of O2 reduction reactions on porous (La,Sr)MnO3 and (La,Sr)(Co,Fe)O3 electrodes, Solid State Ionics, 2002, 146, pp. 1-22. 11. S.P. Jiang, J.P. Zhang, X.G. Zheng, A comparative investigation of chromium deposition at air electrodes of solid oxide fuel cells, J. Europ. Ceam. Soc., 2002, 22, pp. 361-73. 12. S.P. Jiang, S. Zhang, Y.D. Zhen, Mater. Sci. Eng. B (119) (2005) 80-86. 13. Y.D. Zhen, J. Li, S.P. Jiang, Oxygen reduction on strontium-doped LaMnO3 cathodes in the absence and presence of an iron-chromium alloy interconnect, J. Power Sources, 2006, 162, pp. 1043-1052. 14. Y. Zhen, S.P. Jiang, Characterization and performance of (La,Ba)(Co,Fe)O3 cathode for solid oxide fuel cells with iron-chromium metallic interconnect, J. Power Sources, 2008,
180, pp. 695-703.

15. J.W. Fergus, Effect of cathode and electrolyte transport properties on chromium poisoning in solid oxide fuel cells, Int. J. Hydrogen Energy, 2007, 32, pp. 3664-71. 16. B. Sundman, B. Jansson, J.-O. Andersson, The Thermo-Calc databank system, Calphad, 1985, 9, pp. 153-190. 17. E. Povoden, A.N. Grundy, L.J. Gauckler, Thermodynamic assessment of the Mn-Cr-O system for solid oxide fuel cell (SOFC) materials, Int. J. Mater. Res., 2006, 97, pp. 56978. 18. E. Povoden, A.N. Grundy, L.J. Gauckler, Thermodynamic reassessment of the Cr-O system in the framework of solid oxide fuel cell (SOFC) research, J. Phase Equilib. Diff., 2006, 27, pp. 353-62. 19. E.A. Filonova, A.N. Demina, A.N. Petrov, Phase equilibria in the system LaMnO3SrMnO3-SrCrO4-LaCrO3, Russ. J. Inorg. Chem., 2007, 52, pp. 771-774. 20. A.N. Grundy, B. Hallstedt, L.J. Gauckler, Assessment of the La-Sr-Mn-O system,
Calphad, 2004, 28, pp. 191-201.

21. M. Krumpelt et al., FY Annual report, 2004, pp. 39-43. 22. J. Zheng and P. Wu, 4th international symposium on solid oxide fuel cells and pers. comm. 23. D. Caplan and M. Cohen, The volatilization of chromium oxide, J. Electrochem. Soc., 1961, 108, pp. 438-442. 167
24. E. Konysheva, J. Mertens, H. Penkalla, L. Singheiser, K. Hilpert, Chromium poisoning of the porous composite cathode. Effect of cathode thickness and current density, J.
Electrochem. Soc., 2007, 154 (12), pp. B1252-B64.

25. A. Hammouche, E. Siebert, A. Hammou, M. Kleitz, A. Caneiro, Electrocatalytic properties and nonstoichiometry of the high-temperature air electrode La1-xSrxMnO3, J.
Electrochem. Soc., 1991, 138, pp. 1212-16.

26. W. Qu, L. Jian, J.M. Hill, D.G. Ivey, Electrical and microstructural characterization of spinel phases as potential coatings for SOFC metallic interconnects, J. Power sources, 2006, 153, pp. 114-24. 27. R. Koc, H.U. Anderson, S.A. Howard, in SOFC-1, S.C. Singhal, Ed., PV 89-11, p. 220, The Electrochemical Society Proceedings Series, Pennington, NJ (1989). 28. N. Sakai, T. Horita, Y.P. Xiong, K. Yamaji, H. Kishimoto, M.E. Brito, H. Yokokawa, T. Maruyama, Structure and transport property of manganese-chromium-iron oxide as main compound in oxide scales of alloy interconnects for SOFCs, Solid State Ionics, 2005, 176, pp. 681-686. 29. N.M. Sammes, M.B. Phillips, in SOFC-IV, M. Dokiya, O. Yamamoto, H. Tagawa, S.C. Singhal, Eds., PV 95-01, p. 472, The Electrochemical Society Proceedings Series, Pennington, NJ (1995).
168
Appendix
Appendix
La-Cr databasea)
@@ Database La-Cr, Povoden-Karadeniz 2008 @@ GO G ENTER-ELEMENT LA CR VA @@ELEMENT NAME REF. STATE ATOMIC MASS AMEND-ELEMENT LA DOUBLE_HCP(ABAC) 1.3891E+02 5.6902E+01,, AMEND-ELEMENT CR BCC_A2 5.1996E+01 2.35429E+01,, AMEND-ELEMENT VA VACUUM 0
H0 6.6651E+03 4.0500E+03 0
S0
0,,
@@ ---------------------------------------------------------@@ Functions @@ ---------------------------------------------------------@@ Standard data for elements, Dinsdale 1991 @@ La, double hcp ENTER-SYMBO FUNCTION GHSERLA 298.15 -7968.403+120.284604*T-26.34*T*LN(T)-.001295165*T**2; 550 Y -3381.413+59.06113*T-17.1659411*T*LN(T)-.008371705*T**2 +6.8932E-07*T**3-399448*T**(-1); 2000 Y -15608.882+181.390071*T-34.3088*T*LN(T); 4000 N @@ Cr, bcc ENTER-SYMBO FUNCTION GHSERCR 298.15 -8856.94+157.48*T-26.908*t*LN(T)+0.00189435*T**2 -1.47721E-06*T**3+139250*T**(-1); 2180 Y -34869.344+344.18*T-50.0*T*LN(T)-2.88526E+32*T**(-9); 6000 N @@------------------------------------------------------------------@@ Solid metals, Dinsdale 1991 @@ La, bcc ENTER-SYMBO FUNCTION GLABCC 298.15 -3952.161+88.072353*T-21.7919*T*LN(T)-0.004045175*T**2 -5.25865E-07*T**3; 800 Y +321682.673-3565.08252*T+513.440708*T*LN(T)-0.387295093*T**2 +4.9547989E-05*T**3-36581228*T**(-1); 1134 Y -16377.894+218.492988*T-39.5388*T*LN(T); 1193 Y -136609.91+1123.34397*T-163.413074*T*LN(T)+0.053968535*T**2 -4.056395E-06*T**3+21167204*T**(-1); 2000 Y -8205.988+174.836315*T-34.3088*T*LN(T); 4000 N @@ La, fcc ENTER-SYMBO FUNCTION GLAFCC 298.15
169
Appendix
-6109.797+89.878761*T-21.7919*T*LN(T)-0.004045175*T**2 -5.25865E-07*T**3; 1134 Y -124598.976+955.878375*T-139.346741*T*LN(T)+0.042032405*T**2 -3.066199E-06*T**3+20994153*T**(-1); 2000 Y -12599.386+178.54399*T-34.3088*T*LN(T); 4000 N @@ ---------------------------------------------------------------------@@ Liquid metal functions, Dinsdale 1991 @@ La ENTER-SYMBO FUNCTION GLALIQ 298.15 +5332.653+18.23012*T-11.0188191*T*LN(T)-0.020171603*T**2 +2.93775E-06*T**3-133541*T**(-1); 1134 Y -3942.004+171.018431*T-34.3088*T*LN(T); 4000 N @@ Cr ENTER-SYMBO FUNCTION GCR_L 298.15 +15483.015+146.059775*T-26.908*T*LN(T)+.00189435*T**2 -1.47721E-06*T**3+139250*T**(-1)+2.37615E-21*T**7; 2180 Y -16459.984+335.616317*T-50*T*LN(T); 6000 N @@-----------------------------------------------------------@@ Gas functions @@ La gas, from SGTE @@ La(g) ENTER-SYMBO FUNCTION F12026T 298.15 +422273.955-30.3347881*T-22.06299*T*LN(T)-0.005444405*T**2 +4.71447833E-07*T**3+102710.1*T**(-1); 600 Y +426628.905-85.4786162*T-13.83676*T*LN(T)-0.011938995*T**2 +1.33826017E-06*T**3-312130.2*T**(-1); 1300 Y +404460.17+114.016725*T-42.00406*T*LN(T)+0.0037094435*T**2 -2.70261E-07*T**3+2891891*T**(-1); 3200 Y +497751.747-246.085237*T+2.791973*T*LN(T)-0.006002155*T**2 +1.30043383E-07*T**3-34158815*T**(-1); 8200 Y -92343.0441+773.338363*T-111.0188*T*LN(T)+0.0037862445*T**2 -2.82257667E-08*T**3+5.418475E+08*T**(-1); 10000 N @@ Cr gas, from SGTE ENTER-SYMBO FUNCTION F7491T 298.15 +390765.331-31.5192158*T-21.36083*T*LN(T)+7.253215E-04*T**2 -1.588679E-07*T**3+10285.15*T**(-1); 1100 Y +393886.928-44.1074654*T-19.96003*T*LN(T)+.001513089*T**2 -4.23648333E-07*T**3-722515*T**(-1); 6000 N @@ Cr2 gas, from SGTE ENTER-SYMBOL FUNCTION F7763T 298.15 +598511.403+41.5353212*T-40.56798*T*LN(T)+.004961847*T**2 -1.61216717E-06*T**3+154422.85*T**(-1); 800 Y +613345.232-104.207991*T-19.7643*T*LN(T)-.007085085*T**2 -4.69883E-07*T**3-1738066.5*T**(-1); 1400 Y +642608.843-369.28626*T+17.64743*T*LN(T)-.02767321*T**2
170
Appendix
+1.605906E-06*T**3-5831655*T**(-1); 2300 Y +553119.895+159.188555*T-52.07969*T*LN(T)-.004229401*T**2 +1.5939925E-07*T**3+14793625*T**(-1); 3900 Y +347492.34+623.137623*T-105.0428*T*LN(T)+3.9699545E-04*T**2 +1.51783483E-07*T**3+1.4843765E+08*T**(-1); 5800 Y -484185.055+2598.25559*T-334.7145*T*LN(T)+.028597625*T**2 -4.97520167E-07*T**3+7.135805E+08*T**(-1); 6000 N @@ -------------------------------------------------------------------@@ Phases @@ ------------------------------------------------------------------@@ Metals @@ La, dhcp ENTER-PHASE LADHCP,, 2 1 0.5 LA ; VA;,,, ENTER-PAR G(LADHCP,LA:VA;0) 298.15 +GHSERLA;,,, @@ La, fcc ENTER-PHASE LAFCC,, 2 1 1 LA; VA;,,, ENTER-PAR G(LAFCC,LA:VA;0) 298.15 +GLAFCC;,,, @@ ----------------------------------------------------------------@@ Alloys @@ BCC_A2 ALLOY ENTER-PHASE BCC,, 2 1 3 LA CR; VA;,,, AMEND-PHASE-DESC BCC MAGN -1 0.4 ENTER-PAR TC(BCC,CR:VA;0) 298.15 -311.5;,,, ENTER-PAR BMAGN(BCC,CR:VA;0) 298.15 -0.008;,,, ENTER-PAR G(BCC,LA:VA;0) 298.15 +GLABCC;,,, ENTER-PAR G(BCC,CR:VA;0) 298.15 +GHSERCR;,,, ENTER-PAR L(BCC,LA,CR:VA;0) 298.15 83500;,,, @@ ------------------------------------------------------------------------@@ Liquid, ideal extension from lower-order systems ENTER-PHASE LIQ,, 2 1 1 LA CR; VA;,,, AMEND-PHASE LIQ COMP 2,,,,,,,, ENTER-PAR G(LIQ,LA:VA;0) 298.15 +GLALIQ;,,, ENTER-PAR G(LIQ,CR:VA;0) 298.15 +GCR_L;,,, @@ Interaction parameters from binaries ENTER-PAR L(LIQ,LA,CR:VA;0) 298.15 60713-5.49*t;,,, ENTER-PAR L(LIQ,LA,CR:VA;1) 298.15 64573-23*t;,,, @@-----------------------------------------------------------@@ Gas ENTER-PHASE GAS G 1 LA CR CR2;,,, ENTER-PAR G(GAS,LA;0) 298.15 +F12026T+RTLNP;,,, ENTER-PAR G(GAS,CR;0) 298.15 +F7491T+RTLNP;,,,, ENTER-PAR G(GAS,CR2;0) 298.15 +F7763T+RTLNP;,,, @@ GO PAR SET-OUT-LEVEL,,,,, N,, set-interactive
a)
databases scripts can be used in Thermocalc with the extension .tcm
171
Appendix
La-Sr-Mn-Cr-O-(H) oxide database
@@ LA-SR-Mn-CR-O-(H) oxide, Povoden-Karadeniz @@ @@ Database La-Sr-Mn-Cr-O-(H), first version: Feb2006 by Povoden. @@ Actual version: Dec2008 by Povoden-Karadeniz @@ @@ COMMENTS @@ @@ Sr-Cr-O liquid: simple description; associate at composition SrCrO4 can @@ help for better fit to experiments future work @@ @@ No data exist for Sr-Mn-Cr-O. Solubility of Cr in SrMnO3 not known @@ --> Subsystem Sr-Mn-Cr-O is a purely ideal extention @@ @@ no oxygen solubility in La-oxide description (taken from Zinkevich et @@ al.) considered @@ @@ Quinary Ruddlesden popper phase is very tentative, as only few phase @@ diagram data exist! @@ -------------------------------------------------------------------GO G ENTER-ELEMENT LA SR MN CR O VA H @@ELEMENT NAME REF. STATE ATOMIC MASS H0 S0 AMEND-ELEMENT LA DOUBLE_HCP(ABAC) 1.3891E+02 6.6651E+03 5.6902E+01,, AMEND-ELEMENT SR SR_FCC_A1 8.7620E+01 6.5680E+03 5.5694E+01,, AMEND-ELEMENT MN CBCC_A12 5.4938E+01 4.9960E+03 3.2008E+01,, AMEND-ELEMENT CR BCC_A2 5.1996E+01 4.0500E+03 2.35429E+01,, AMEND-ELEMENT O 1/2_MOLE_O2(G) 1.5999E+01 4.3410E+03 1.0252E+02,, AMEND-ELEMENT H 1/2_MOLE_H2(G) 0.1008E+01 0 0.65340E+02,, AMEND-ELEMENT VA VACUUM 0 0 0,, @@ @@ -------------------------------------------------------------------@@ Species @@ -------------------------------------------------------------------ENTER-SPECIES LA+2 LA/+2 ENTER-SPECIES LA+3 LA/+3 ENTER-SPECIES SR+2 SR/+2 ENTER-SPECIES MN+2 MN/+2 ENTER-SPECIES MN+3 MN/+3 ENTER-SPECIES MN+4 MN/+4 ENTER-SPECIES O2 O2 ENTER-SPECIES O3 O3 ENTER-SPECIES O-2 O/-2 ENTER-SPECIES SRO SRO ENTER-SPECIES SRO2 SRO2 ENTER-SPECIES CR+2 CR/+2 ENTER-SPECIES CR+3 CR/+3 ENTER-SPECIES CR+4 CR/+4 ENTER-SPECIES CR+6 CR/+6 ENTER-SPECIES CR1O1 CR1O1 ENTER-SPECIES CR1O2 CR1O2 ENTER-SPECIES CR1O3 CR1O3
172
Appendix
ENTER-SPECIES CR2O3 CR2O3 ENTER-SPECIES CR3O4 CR3O4 ENTER-SPECIES CRH1 CRH1 ENTER-SPECIES CRH1O1 CRH1O1 ENTER-SPECIES CRH1O2 CRH1O2 ENTER-SPECIES CRH1O3 CRH1O3 ENTER-SPECIES CRH2O2 CRH2O2 ENTER-SPECIES CRH2O3 CRH2O3 ENTER-SPECIES CRH2O4 CRH2O4 ENTER-SPECIES CRH3O3 CRH3O3 ENTER-SPECIES CRH3O4 CRH3O4 ENTER-SPECIES CRH4O4 CRH4O4 ENTER-SPECIES CRH4O5 CRH4O5 ENTER-SPECIES CRH5O5 CRH5O5 ENTER-SPECIES CRH6O6 CRH6O6 ENTER-SPECIES H2 H2 ENTER-SPECIES H2O1 H2O1 ENTER-SPECIES H1O1 H1O1 ENTER-SPECIES H1O2 H1O2 ENTER-SPECIES H2O2 H2O2 @@ @@ ---------------------------------------------------------@@ Functions @@ ---------------------------------------------------------@@ SER Lattice stabilities, Dinsdale 1991 @@ La, double hcp ENTER-SYMBO FUNCTION GHSERLA 298.15 -7968.403+120.284604*T-26.34*T*LN(T)-.001295165*T**2; 550 Y -3381.413+59.06113*T-17.1659411*T*LN(T)-.008371705*T**2 +6.8932E-07*T**3-399448*T**(-1); 2000 Y -15608.882+181.390071*T-34.3088*T*LN(T); 4000 N @@ Sr, fcc ENTER-SYMBO FUNCTION GHSERSR 298.15 -7532.367+107.183879*T-23.905*T*LN(T)-4.61225E-3*T**2 -1.67477E-07*T**3-2055*T**(-1); 820 Y -13380.102+153.196104*T-30.0905432*T*LN(T)-3.251266E-3*T**2 +1.84189E-07*T**3+850134*T**(-1); 3000 N @@ Mn, cbcca12 ENTER-SYMBO FUNCTION GHSERMN 298.15 -8115.27966+130.059572*T-23.4582*T*LN(T)-0.00734768*T**2 +69827.1*T**(-1); 1519 Y -28733.41+312.2648*T-48*T*LN(T)+1.656847E30*T**(-9); 2000 N @@ Cr, bcc ENTER-SYMBO FUNCTION GHSERCR 298.15 -8856.94+157.48*T-26.908*t*LN(T)+0.00189435*T**2 -1.47721E-06*T**3+139250*T**(-1); 2180 Y -34869.344+344.18*T-50.0*T*LN(T)-2.88526E+32*T**(-9); 6000 N @@ O1, (1/2 O2) ENTER-SYMBO FUNCTION GHSEROO 298.15 -3480.872255-25.5028601*T-11.1355068*T*LN(T)-0.005098873*T**2 +6.6184604E-07*T**3-38364.8742*T**(-1); 1000 Y
173
Appendix
-6568.76015+12.66000166*T-16.8138015*T*LN(T)-5.9579637E-04*T**2 +6.78055555E-09*T**3+262904.778*T**(-1); 3300 Y -13986.728+31.259625*T-18.9536*T*LN(T)-4.25243E-04*T**2 +1.0721E-08*T**3+4383200*T**(-1); 6000 N @@-----------------------------------------------------------@@ Binary oxides, optimized @@ La-oxides, Zinkevich 2006 ENTER-SYMBO FUNCTION GLA2O3D 298.15 -1833257+692.9664*T-120.629*T*LN(T)-0.006854*T**2 +808000*T**(-1)-1E7*T**(-2);,,, ENTER-SYMBO FUNCTION GLA2O3H 298.15 32350-13.986*T+GLA2O3D;,,, ENTER-SYMBO FUNCTION GLA2O3X 298.15 43192-18.555*T+GLA2O3D;,,, @@ Sr-oxides, Risold 1996 @@ SrO ENTER-SYMBO FUNCTION GSROSOL 298.15 -607870+268.9*T-47.56*T*LN(T)-0.00307*T**2 +190000*T**(-1);,,, @@ SrO2 ENTER-SYMBO FUNCTION GSRO2SOL 298.15 +GSROSOL+GHSEROO-43740+70*T;,,, @@ Mn-oxides, Grundy 2003 @@ MANGANOSITE, MNO ENTER-SYMBO FUNCTION GMN1O1 298.15 -4.02477557E+05+2.59355626E+02*T-4.68352649E+01*T*LN(T) -3.85001409E-03*T**2+2.12922234E+05*T**(-1);,,, @@ PYROLYSITE, MN1O2 ENTER-SYMBO FUNCTION GMN1O2 298.15 -5.45091278E+05+3.95379396E+02*T-6.52766201E+01*T*LN(T) -7.80284521E-03*T**2+6.64955386E+05*T**(-1);,,, @@ @@ MN2O3-FUNCTION, MODIFIED, Grundy 2006 ENTER-SYMBO FUNCTION GMN2O3 298.15 -9.96393E+05+5.6846E+02*T-9.911E+01*T*LN(T)-2.056E-02*T**2 +6.0822E+05*T**(-1);,,, @@ ALPHA-HAUSMANNITE, ALPHA-MN3O4 (DISTORTED) ENTER-SYMBO FUNCTION GTMN3O4 298.15 -1.43703676E+06+8.89567858E+02*T-1.54747566E+02*T*LN(T) -1.74079033E-02*T**2+9.86138663E+05*T**(-1);,,, @@ BETA-HAUSMANNITE, BETA-MN3O4 (CUBIC) ENTER-SYMBO FUNCTION GCMN3O4 298.15 -1.41618912E+06+8.75120338E+02*T-1.54747566E+02*T*LN(T) -1.74079033E-02*T**2+9.86138663E+05*T**(-1);,,, @@ Cr-O oxides, Povoden 2005 @@ METASTABLE CRO ENTER-SYMBO FUNCTION GCR1O1 298.15 +0.5*GCR2O3-0.5*GHSEROO+255269-53.82*T;,,, @@ ESKOLAITE, Cr2O3 ENTER-SYMBO FUNCTION GCR2O3 298.15 -1.164542E+06+7.2856E+02*T-119.8*T*LN(T)-4.97E-03*T**2 +1.05E+06*T**(-1);,,, @@ reduced neutral endmember of CR2O3 ENTER-SYMBO FUNCTION GCRO0 298.15 +108305+GCR2O3+0.66666666667*GHSERCR;,,, @@ CR-SPINEL CR3O4 ENTER-SYMBO FUNCTION GCR3O4 298.15 +1.5*GCR2O3-0.5*GHSEROO+280045-93.76*T;,,, @@ --------------------------------------------------------------
174
Appendix
@@ Ternary oxides, optimized (except perovskite functions) @@ @@ Functions of the La-Sr-O system, Grundy 2004 ENTER-SYMBO FUNCTION SR_ALPHA 298.15 +2*GSROSOL+2.5E+04;,,, ENTER-SYMBO FUNCTION SRH_ALPH 298.15 +2*GSROSOL+2.5E+04;,,, ENTER-SYMBO FUNCTION SRX_ALPH 298.15 +2*GSROSOL+2.5E+04;,,, ENTER-SYMBO FUNCTION LA_BETA 298.15 +GLA2O3D+2.158E+04;,,, ENTER-SYMBO FUNCTION RE_ALPHA 298.15 0;,,, ENTER-SYMBO FUNCTION RE_BETA 298.15 0;,,, @@ @@ Functions of the La-Mn-O system, Grundy et al. 2005 ENTER-SYMBO FUNCTION GL2MNO4 298.15 +GLA2O3D+GMN1O1+6.26029731E+04-3.71704891E+01*T;,,, @@ @@ Functions of the La-Cr-O system, Povoden 2008 @@ LA2CRO6 ENTER-SYMBO FUNCTION GLA2CRO6 298.15 GLA2O3D+0.5*GCR2O3+1.5*GHSEROO-73045-4.14*T;,,, @@ intermediate La-chromates @@ LA16 @@ ENTER-SYMBO FUNCTION GLA16 298.15 @@ 8*GLA2O3D+3.5*GCR2O3+9.5*GHSEROO-540404-9.55*T;,,, @@ LA7 @@ ENTER-SYMBO FUNCTION GLA7 298.15 @@ 3.5*GLA2O3D+GCR2O3+2.5*GHSEROO-154101-2.799*T;,,, @@ LA2CR3O12 ENTER-SYMBO FUNCTION GLA2CR3 298.15 GLA2O3D+1.5*GCR2O3+4.5*GHSEROO-371557+205*T;,,, @@ @@ Functions of the Sr-Mn-O system, Grundy 2004 @@ HEX Phase ENTER-SYMBO FUNCTION GSM4_HEX 298.15 +GSROSOL+GMN1O2-1.11300000E+05;,,, ENTER-SYMBO FUNCTION GSM3_HEX 298.15 +GSROSOL+0.5*GMN2O3-7.73000000E+03 -1.70000000E+01*T;,,, @@ SrMn3Oz as SrMnO3_Mn2O3 ENTER-SYMBO FUNCTION GSM4OZ 298.15 +GMN2O3+GSM4_HEX-8.79100000E+03;,,, ENTER-SYMBO FUNCTION GSM3OZ 298.15 +GMN2O3+GSM3_HEX-2.19200000E+04;,,, @@ RP1 ENTER-SYMBO FUNCTION GS4O_RP1 298.15 +2*GSROSOL+GMN1O2-1.32830000E+05;,,, ENTER-SYMBO FUNCTION GL3O_RP1 298.15 +GSROSOL+0.5*GLA2O3D+0.5*GMN2O3-68300;,,, @@ RP2 ENTER-SYMBO FUNCTION GS4O_RP2 298.15 +3*GSROSOL+2*GMN1O2-8.99100000E+04-90*T;,,, ENTER-SYMBO FUNCTION GL3O_RP2 298.15 +GSROSOL+GLA2O3D+GMN2O3-137400;,,, @@ RP3 ENTER-SYMBO FUNCTION GSM4_RP3 298.15 +4*GSROSOL+3*GMN1O2-3.78500000E+05;,,, @@ Sr7Mn4O1
175
Appendix
ENTER-SYMBO FUNCTION GS7M4 298.15 +7*GSROSOL+4*GMN1O2-6.12450000E+05+50*T;,,, @@ @@ Functions of the Sr-Cr-O system, Povoden 2008 @@ SRCR2O4 ENTER-SYMBO FUNCTION GSC2O4 298.15 +GSROSOL+GCR2O3+98000-95.5*T;,,, @@ SR2CRO4 ENTER-SYMBO FUNCTION GS2CO4 298.15 2*GSROSOL+0.5*GCR2O3+0.5*GHSEROO-145000+50*T;,,, @@ SR3CR2O8 ENTER-SYMBO FUNCTION GS3C2O8 298.15 +2.66666667*GSROSOL+GCR2O3+2.333333333333*GHSEROO -508507+219*T;,,, @@ SRCRO4 ENTER-SYMBO FUNCTION GSCO4 298.15 +GSROSOL+0.5*GCR2O3+1.5*GHSEROO-273771 +131.6*T;,,, @@ SRCR2O7 ENTER-SYMBO FUNCTION GSC2O7 298.15 +GSROSOL+GCR2O3+3*GHSEROO-325047+196*T;,,, @@ @@ Functions of the Mn-Cr-O system, Povoden 2005 @@ CUBIC SPINEL ENTER-SYMBO FUNCTION GSPINEL 298.15 0.666666667*GCR3O4+.33333333334*GCMN3O4-210795.3+61.69*T;,,, @@ TETRAGONALLY DISTORTED SPINEL ENTER-SYMBO FUNCTION GTSPINEL 298.15 0.666666667*GCR3O4+.33333333334*GTMN3O4-200942+75.1*T;,,, @@ @@ Functions of the La-Sr-Cr-O oxide system, Povoden 2008 @@ Ruddlesden Popper phase ENTER-SYMBO FUNCTION GREFRP 298.15 +2*GSROSOL+0.5*GCR2O3+0.5*GHSEROO;,,, ENTER-SYMBO FUNCTION GLCR3O_RP1 298.15 +GLACRO3+GSROSOL+7000-25*t;,,, @@ ---------------------------------------------------------------@@ Perovskite functions @@ Grundy 2005 @@ Charge compensated by Mn+4 (correct) ENTER-SYMBO FUNCTION GL3O 298.15 +0.5*GLA2O3D+0.5*GMN2O3-63367+51.77*T-7.19*T*LN(T) +232934*T**(-1);,,, @@ ENTER-SYMBO FUNCTION GL3OL 298.15 +0.5*GLA2O3D+0.5*GMN2O3-63367+51.77*T-7.19*T*LN(T) +232934*T**(-1)-3429+4.72*t;,,, @@ ENTER-SYMBO FUNCTION GL3OR 298.15 +0.5*GLA2O3D+0.5*GMN2O3-63367+51.77*T-7.19*T*LN(T) +232934*T**(-1)+400-0.4*t;,,, ENTER-SYMBO FUNCTION GL2O 298.15 +0.5*GLA2O3D+GMN1O1+27672;,,, ENTER-SYMBO FUNCTION GL4O 298.15 +0.5*GLA2O3D+0.75*GMN1O2-91857+20.31*T;,,, ENTER-SYMBO FUNCTION GV4O 298.15 +0.333333*GLA2O3D+GMN1O2-53760;,,, ENTER-SYMBO FUNCTION GVVV 298.15 +6*GL2O+4*GL4O+3*GV4O-12*GL3O-254212;,,, ENTER-SYMBO FUNCTION GMS3O 298.15 +GSROSOL+0.5*GMN2O3-7.73000000E+03-1.44550000E+04
176
Appendix
-1.70000000E+01*T;,,, ENTER-SYMBO FUNCTION GMS4O 298.15 +GSROSOL+GMN1O2-1.11300000E+05+2.26500000E+04 -7.69000000E+00*T;,,, ENTER-SYMBO FUNCTION ANTI 298.15 +547422;,,, @@ Reciprocals: all 0!!!!! @@ @@ Povoden 2008 @@ LaCrO3-PEROVSKITE ENTER-SYMBO FUNCTION GLACRO3 298.15 0.5*GLA2O3D+0.5*GCR2O3-73591+2.38*T-0.68*T* LN(T);,,, ENTER-SYMBO FUNCTION GALACRO3 298.15 +GLACRO3-340+0.63*t;,,, @@ (LaSr)CrO3+/-delta-Perovskite @@ Reference SrCrVa3 ENTER-SYMBO FUNCTION GS4V 298.15 GSROSOL+0.5*GCR2O3-2.5*GHSEROO;,,, @@ Function for neutral endmember SrCrO3 ENTER-SYMBO FUNCTION GS4O 298.15 GSROSOL+0.5*GCR2O3+0.5*GHSEROO+10222-55.52*t;,,, @@ Function for neutral endmember SR(CR+3,VA)O3 ENTER-SYMBO FUNCTION GN 298.15 +GSROSOL+0.5*GCR2O3+11.2386*T+135166-88.42*t;,,, @@ Functions for defect chemistry ENTER-SYMBO FUNCTION GVCR4O 298.15 0.5*GCR2O3+0.5*GHSEROO-291802-250*t;,,, ENTER-SYMBO FUNCTION GLACR4O 298.15 0.33333*GLA2O3D+.5*GCR2O3+0.5*GHSEROO-200000;,,, @@ ---------------------------------------------------------------------@@ LIQUID FUNCTIONS @@ ---------------------------------------------------------------------@@ Liquid metal functions, Dinsdale 1991 @@ La ENTER-SYMBO FUNCTION GLALIQ 298.15 +5332.653+18.23012*T-11.0188191*T*LN(T)-0.020171603*T**2 +2.93775E-06*T**3-133541*T**(-1); 1134 Y -3942.004+171.018431*T-34.3088*T*LN(T); 4000 N @@ Sr ENTER-SYMBO FUNCTION GSRLIQ 298.15 +2194.997-10.118994*T-5.0668978*T*LN(T)-3.1840595E-2*T**2 +4.981237E-06*T**3-265559*T**(-1); 1050 Y -10855.29+213.406219*T-39.463*T*LN(T); 3000 N @@ Mn ENTER-SYMBO FUNCTION GMN_L 298.15 +GHSERMN+17859.91-12.6208*T-4.41929E-21*T**7; 1519 Y +GHSERMN+18739.51-13.2288*T-1.656847E30*T**(-9); 3000 N @@ Cr ENTER-SYMBO FUNCTION GCR_L 298.15 +15483.015+146.059775*T-26.908*T*LN(T)+.00189435*T**2 -1.47721E-06*T**3+139250*T**(-1)+2.37615E-21*T**7; 2180 Y -16459.984+335.616317*T-50*T*LN(T); 6000 N @@ ---------------------------------------------------------------------
177
Appendix
@@ Liquid oxide functions, optimized @@ liquid La2O3, Zinkevich 2006 ENTER-SYMBO FUNCTION GLA2O3LIQ 298.15 -1833257+692.9664*T-120.629*T*LN(T)-0.006854*T**2 +808000*T**(-1)-1E7*T**(-2)+141329-56.6220*T;,,, @@ liquid SrO, Risold ENTER-SYMBO FUNCTION GSROLIQ 298.15 -566346+449*T-73.1*T*LN(T);,,, @@ liquid Mn oxides, Grundy ENTER-SYMBO FUNCTION GMN1O1_L 298.15 GMN1O1+4.39465890E+04-2.06284295E+01*T;,,, ENTER-SYMBO FUNCTION GMN2O3_L 298.15 +2*GMN1O1+GHSEROO-6.49525609E+04+4.31437957E+01*T;,,, @@ liquid Cr oxides, Povoden ENTER-SYMBO FUNCTION GCR1O1_L 298.15 0.5*GCR2O3-0.5*GHSEROO+339673-121.4*T;,,, ENTER-SYMBO FUNCTION GCR2O3_L 298.15 GCR2O3+439078-169*T;,,, @@------------------------------------------------------------------@@ LIQUID WATER, from T.C.R.A.S. Class: 4 ENTER-SYMBO FUNCTION GH2O_L 298.15 -332319.671+1078.59563*T-186.8669*T*LN(T) +.2320948*T**2-9.14296167E-05*T**3+978019*T**(-1); 500 Y -62418.8788-3288.18729*T+495.1304*T*LN(T) -.504926*T**2+4.917665E-05*T**3-18523425*T**(-1); 540 Y -8528143.9+142414.45*T-22596.19*T*LN(T) +27.48508*T**2-.00631160667*T**3+5.63356E+08*T**(-1); 600 Y -331037.282+741.178604*T-117.41*T*LN(T); 601 N @@ -----------------------------------------------------------------------@@ GAS FUNCTIONS @@ --------------------CHROMIUM GAS---------------------------------------@@ CR Gas: SGTE v 3.0 (1998), source: Thermocenter of russian academy @@ of science (T.C.R.A.S), Class: 4 @@ SGTE=scientific group thermodata Europe @@ ENTER-SYMBO FUNCTION F7491T 298.15 +390765.331-31.5192158*T-21.36083*T* LN(T) +7.253215E-04*T**2-1.588679E-07*T**3+10285.15*T**(-1); 1100 Y +393886.928-44.1074654*T-19.96003*T*LN(T)+.001513089*T**2 -4.23648333E-07*T**3-722515*T**(-1); 6000 N @@ CR1O1 Gas (SGTE 1998; from T.C.R.A.S Class: 5) ENTER-SYMBO FUNCTION F7705T 298.15 +176483.869-31.9513659*T-30.2897*T*LN(T)-.00607059*T**2 +9.229905E-07*T**3+35263.135*T**(-1); 900 Y +170853.62+32.1684007*T-39.74749*T*LN(T)+.00119977*T**2 -1.52515733E-07*T**3+682877*T**(-1); 4000 Y +307209.502-414.237405*T+14.48744*T*LN(T)-.008463125*T**2 +1.722975E-07*T**3-64209900*T**(-1); 8400 Y -403765.708+805.224944*T-121.5329*T*LN(T)+.003139382*T**2 -1.36845867E-08*T**3+6.35563E+08*T**(-1); 10000 N @@ CR1O1 Gas, REASSESSED BY MING CHEN (2006) BASED ON EBBINGHAUS (1993) ENTER-SYMBO FUNCTION GCR1O1_G 298.15 +173449.4-33.083*T-30.097*T*LN(T)-0.0063*T**2 +31300*T**(-1)+9.5567E-7*T**3; 1000 Y +167489.3+37.31*T-40.555*T*LN(T)+0.00148*T**2 +873600*T**(-1)+1.69E-7*T**3;
178
Appendix
3000 N @@ CR1O2 Gas, reassessment Chen 2006, based on Ebbinghaus 1993 ENTER-SYMBO FUNCTION GCR1O2_G 298.15 -109942.78+10.59*T-39.526*T*LN(T)-0.0155*T**2 +245800*T**(-1)+2.43E-6*T**3; 1000 Y -118120+123.81*T-56.696*T*LN(T)-0.00012*T**2 +932050*T**(-1)-8.01667E-8*T**3; 3000 N @@ CR1O3 Gas, reassessed by Ming Chen 2006, based on Ebbinghaus 1993 ENTER-SYMBO FUNCTION GCR1O3_G 298.15 -341231.99+130.61*T-57.2*T*LN(T)-.0216*T**2 +428900*T**(-1)+3.401E-6*T**3; 1000 Y -354716.09+299.89*T-82.569*T*LN(T)-0.00016*T**2 +1814700*T**(-1)+7.33E-9*T**3; 3000 N @@-------------------OXYGEN GAS------------------------------------@@ O Gas (JANAF 1982, assessment dated 3/77 from SGTE) ENTER-SYMBO FUNCTION F13349T 298.15 +243206.494-20.8612582*T-21.01555*T*LN(T)+1.2687055E-04*T**2 -1.23131283E-08*T**3-42897.09*T**(-1); 2950 Y +252301.423-52.0847281*T-17.21188*T*LN(T)-5.413565E-04*T**2 +7.64520667E-09*T**3-3973170.5*T**(-1); 6000 N @@ O3 Gas (SGTE 1998; from T.C.R.A.S Class: 4) ENTER-SYMBO FUNCTION F14021T 298.15 +130696.944-37.9096643*T-27.58118*T*LN(T)-.02763076*T**2 +4.60539333E-06*T**3+99530.45*T**(-1); 700 Y +114760.623+176.626737*T-60.10286*T*LN(T)+.00206456*T**2 -5.17486667E-07*T**3+1572175*T**(-1); 1300 Y +49468.3956+710.09482*T-134.3696*T*LN(T)+.039707355*T**2 -4.10457667E-06*T**3+12362250*T**(-1); 2100 Y +866367.075-3566.80563*T+421.2001*T*LN(T)-.1284109*T**2 +5.44768833E-06*T**3-2.1304835E+08*T**(-1); 2800 Y +409416.383-1950.70834*T+223.4437*T*LN(T)-.0922361*T**2 +4.306855E-06*T**3-21589870*T**(-1); 3500 Y -1866338.6+6101.13383*T-764.8435*T*LN(T)+.09852775*T**2 -2.59784667E-06*T**3+9.610855E+08*T**(-1); 4900 Y +97590.043+890.798361*T-149.9608*T*LN(T)+.01283575*T**2 -3.555105E-07*T**3-2.1699975E+08*T**(-1); 6000 N @@ -------------O-H GAS----------------------------------@@ H2 Gas (JANAF THERMOCHEMICAL TABLES SGTE) ENTER-SYMBO FUNCTION H2GAS 298.15 -9522.97393+78.5273873*T-31.35707*T*LN(T)+0.0027589925*T**2 -7.46390667E-07*T**3+56582.3*T**(-1); 1000 Y +180.108664-15.6128262*T-17.84857*T*LN(T)-0.00584168*T**2 +3.14618667E-07*T**3-1280036*T**(-1); 2100 Y -18840.1663+92.3120249*T-32.05082*T*LN(T)-0.0010728235*T**2 +1.14281783E-08*T**3+3561002.5*T**(-1); 6000 N @@ H Gas (JANAF 1982, assessment dated 3/77 from SGTE) ENTER-SYMBO FUNCTION HGAS 298.15 +211801.621+24.4989821*T-20.78611*T*LN(T); 6000 N @@ H2O1 Gas (SGTE 1998; from T.C.R.A.S Class: 4) ENTER-SYMBO FUNCTION F10963T 298.15 -250423.434+4.45470312*T-28.40916*T*LN(T) -.00623741*T**2-6.01526167E-08*T**3-64163.45*T**(-1); 1100 Y -256145.879+30.1894682*T-31.43044*T*LN(T) -.007055445*T**2+3.05535833E-07*T**3+1246309.5*T**(-1); 2800 Y
179
Appendix
-268423.418+116.690197*T-42.96842*T*LN(T) -.003069987*T**2+6.97594167E-08*T**3+2458230.5*T**(-1); 8400 Y -489068.882+553.259882*T-92.4077*T*LN(T) +.0016703495*T**2-1.32333233E-08*T**3+1.765625E+08*T**(-1); 18000 Y -165728.771+239.645643*T-59.77872*T*LN(T) +2.213599E-04*T**2-1.2921095E-09*T**3-4.1931655E+08*T**(-1); 20000 N @@ H1O1 Gas (SGTE 1998; from T.C.R.A.S Class: 1) ENTER-SYMBO FUNCTION F10666T 298.15 +30698.6898+15.9096451*T-29.97699*T*LN(T) +.001713168*T**2-6.799205E-07*T**3-25503.82*T**(-1); 3000 Y +31735.5127-12.686636*T-25.42186*T*LN(T) -.003149545*T**2+1.34404917E-07*T**3+116618.65*T**(-1); 8600 Y +41016.0783-20.7343256*T-24.94216*T*LN(T) -.0023107985*T**2+5.91863E-08*T**3-6415210*T**(-1); 18000 Y -154907.953+370.326117*T-69.24542*T*LN(T) +.0019361405*T**2-1.47539017E-08*T**3+1.4391015E+08*T**(-1); +326722.277-65.0792741*T-24.2768*T*LN(T) +6.42189E-05*T**2-1.30298483E-10*T**3-8.292415E+08*T**(-1); 20000 N @@ H1O2 Gas (SGTE 1998; from T.C.R.A.S Class: 4) ENTER-SYMBO FUNCTION F10729T 298.15 +1075.64106-55.242048*T-24.45435*T*LN(T) -.018507875*T**2+2.36297E-06*T**3-29469.05*T**(-1); 800 Y -7932.99164+54.2016233*T-40.775*T*LN(T) -.00501027*T**2+2.122915E-07*T**3+925845*T**(-1); 3600 Y -67875.8961+275.406716*T-68.1173*T*LN(T) +6.12331E-04*T**2-6.573855E-09*T**3+26048030*T**(-1); 6000 N @@ H2O2 Gas (JANAF SECOND EDIT SGTE) ENTER-SYMBO FUNCTION F10983T 298.15 -147258.971-37.1497212*T-26.10636*T*LN(T) -.036948065*T**2+6.659505E-06*T**3+65357.65*T**(-1); 700 Y -156470.505+120.191295*T-50.94271*T*LN(T) -.007931945*T**2+4.29733833E-07*T**3+684985.5*T**(-1); 1500 N @@-------------------CR-O-H GAS---------------------------------------@@ CR(OH)1 Gas REASSESSED BY MING CHEN (2006) BASED ON EBBINGHAUS (1993) ENTER-SYMBO FUNCTION GCRH1O1_G 298.15 +68260+52.87*T-46.257*T*LN(T) -0.002*T**2+185600*T**(-1)-1.51E-7*T**3; 1000 Y 56684+136.53*T-57.551*T*LN(T) +0.0018*T**2+2218000*T**(-1)-2.75E-7*T**3; 3000 N @@ CRO(OH)1 Gas REASSESSED BY MING CHEN (2006) BASED ON EBBINGHAUS (1993) ENTER-SYMBO FUNCTION GCRH1O2_G 298.15 -274384.32+190.52*T-74.175*T*LN(T) +0.0031*T**2+266750*T**(-1)-9.135E-7*T**3; 1000 Y -276268.52+180.7*T-72.053*T*LN(T) -0.0015*T**2+938850*T**(-1)+5.15E-8*T**3; 3000 N @@ CRO2(OH) Gas, combined assessment Chen 2006 @@ based on Ebbinghaus 1995 @@ and from Povoden based on Kim and Belton 1974 @@ (data suggested by Opila 2007) ENTER-SYMBO FUNCTION GCRH1O3_G 298.15 -497678+273.77*T-83.724*T*LN(T) -0.0146*T**2+715900*T**(-1)+2.39E-6*T**3; 1000 Y -492562+351.47*T-96.9*T*LN(T) -.0018*T**2+1338300*T**(-1)+9.334E-8*T**3; 3000 N @@ CR(OH)2 Gas, assessment Chen 2006, based on Ebbinghaus 1995 ENTER-SYMBO FUNCTION GCRH2O2_G 298.15 -351288.4+195.86*T-75.927*T*LN(T) -0.0007*T**2+243850*T**(-1)-9.2E-7*T**3; 1000 Y
180
Appendix
-359644+228.97*T-79.455*T*LN(T) -0.004*T**2+2094950*T**(-1)+1.91E-7*T**3; 3000 N @@ CRO(OH)2 Gas, assessment Chen 2006 based on Ebbinghaus 1995 ENTER-SYMBO FUNCTION GCRH2O3_G 298.15 -578683+391.2*T-109.525*T*LN(T) -0.00056*T**2+663150*T**(-1)-5.67833E-07*T**3; 1000 Y -582354+394.6*T-109.25*T*LN(T) -0.004*T**2+1.8967E-07*T**3+1688150*T**(-1); 3000 N @@ CRO2(OH)2 Gas, combined assessment Chen 2006 @@ based on Ebbinghaus 1995 @@ and from Povoden-Karadeniz based on Opila 2007 ENTER-SYMBO FUNCTION GCRH2O4_G 298.15 -787712+400*T-107.819*T*LN(T) -0.019*T**2+513600*T**(-1)+1.9958E-06*T**3; 1000 Y -806262+567*T-131.457*T*LN(T) -0.0049*T**2+3311450*T**(-1)+2.545E-7*T**3; 3000 N @@ CR(OH)3 Gas, assessment Chen 2006, based on Ebbinghaus 1995 ENTER-SYMBO FUNCTION GCRH3O3_G 298.15 -650064.7+464.37*T-125.5*T*LN(T) +0.006*T**2+498450*T**(-1)-2.378E-6*T**3; 1000 Y -656538+448.3*T-121.16*T*LN(T) -0.006*T**2+2525450*T**(-1)+2.93E-7*T**3; 3000 N @@ CRO(OH)3 Gas, assessment Chen 2006, based on Ebbinghaus 1995 ENTER-SYMBO FUNCTION GCRH3O4_G 298.15 -851590.83+534.3*T-137.098*T*LN(T) -0.0099*T**2+770750*T**(-1)+5.8867E-7*T**3; 1000 Y -861477.76+600.74*T-146.002*T*LN(T) -0.0065*T**2+2669300*T**(-1)+3.35E-7*T**3; 3000 N @@ CR(OH)4 Gas, assessment Chen 2006, based on Ebbinghaus 1995 ENTER-SYMBO FUNCTION GCRH4O4_G 298.15 -902751.6+712.5*T-163*T*LN(T) +0.004*T**2+785800*T**(-1)-2.355E-6*T**3; 1000 Y -909897.86+694.973194012145*T-158.41*T*LN(T) -0.0086*T**2+3058600*T**(-1)+4.3E-7*T**3; 3000 N @@ CR(OH)5 Gas, assessment Chen 2006, based on Ebbinghaus 1995 ENTER-SYMBO FUNCTION GCRH5O5_G 298.15 -978211.05+820.9*T-190.524*T*LN(T) -0.0054*T**2+1023500*T**(-1)-1.11767E-6*T**3; 1000 Y -991549.1+867.6*T-195.2*T*LN(T) -0.01*T**2+4151100*T**(-1)+5.7467E-7*T**3; 3000 N @@ CR(OH)6 Gas, assessment Chen 2006, based on Ebbinghaus 1995 ENTER-SYMBO FUNCTION GCRH6O6_G 298.15 -1029121.12+967.89*T-216.86*T*LN(T) -0.008*T**2+840600*T**(-1)-1.77E-6*T**3; 1000 Y -1053466+1080.45*T-229.874*T*LN(T) -0.014*T**2+6008850*T**(-1)+7.34E-7*T**3; 3000 N @@ CRO(OH)4 Gas, assessment Chen 2006, based on Ebbinghaus 1993 ENTER-SYMBO FUNCTION G_CRH4O5 298.15 -976204+672.6*T-162.049*T*LN(T) -0.014*T**2+1.35E-07*T**3+665100*T**(-1); 1000 Y -997791.8+813.97*T-180.65*T*LN(T) -0.0096*T**2+4.945E-07*T**3+4671850*T**(-1); 3000 N @@ CRH1 Gas (SGTE 1998; from T.C.R.A.S Class: 4) ENTER-SYMBO FUNCTION F7586T 298.15 +432449.026-56.386334*T-22.37019*T*LN(T) -.00994042*T**2+1.18913267E-06*T**3-77266.05*T**(-1); 1000 Y +421602.4+51.9865812*T-37.99681*T*LN(T) +3.349852E-04*T**2-1.13133917E-07*T**3+1368173*T**(-1); 3500 Y +587860.713-424.36214*T+18.6022*T*LN(T) -.007723995*T**2+7.99444833E-08*T**3-86705500*T**(-1); 5500 Y +1270342.46-2142.60191*T+219.9391*T*LN(T)
181
Appendix
-.034109635*T**2+7.277845E-07*T**3-5.20451E+08*T**(-1); 6000 N @@ @@ @@ -------------------------------------------------------------------@@ Phases @@ ------------------------------------------------------------------@@ Binary oxides @@ @@ Sr oxides, Risold @@ ENTER-PHASE SRO2,, 1 SRO2;,,, ENTER-PAR G(SRO2,SRO2;0) 298.15 +GSRO2SOL;,,, @@ @@ Mn oxides, Grundy @@ @@ STOICHIOMETRIC PYROLYSITE, MN1O2 ENTER-PHASE MN1O2,, 2 1 2 MN; O;,,, ENTER-PAR G(MN1O2,MN:O;0) 298.15 +GMN1O2;,,, @@ ---------------------------------------------------------------------@@ Ternary oxides @@ @@ LA-SR-O, Grundy, Chen @@ ENTER-PHASE LA2O3SS,, 2 2 3 LA+2 LA+3 SR+2; O-2 VA;,,, ENTER-PAR G(LA2O3SS,LA+2:O-2;0) 298.15 +GLAO;,,, ENTER-PAR G(LA2O3SS,LA+2:VA;0),, +GLAO-GHSEROO;,,, ENTER-PAR G(LA2O3SS,LA+3:O-2;0) 298.15 +GLA2O3D;,,, ENTER-PAR G(LA2O3SS,LA+3:VA;0),, +GLA2O3D-3*GHSEROO;,,, ENTER-PAR G(LA2O3SS,SR+2:O-2;0),, +SR_ALPHA+GHSEROO +15.87691*T;,,, ENTER-PAR G(LA2O3SS,SR+2:VA;0),, +SR_ALPHA-2*GHSEROO +15.87691*T;,,, ENTER-PAR L(LA2O3SS,LA+3,SR+2:O-2;0) 298.15 +2.149E+05-7.81E+01*T;,,, ENTER-PAR L(LA2O3SS,LA+3,SR+2:VA;0) 298.15 +2.149E+05-7.81E+01*T;,,, @@ ENTER-PHASE LA2O3_HEXSS,, 2 2 3 LA+3 SR+2; O-2 VA;,,, ENTER-PAR G(LA2O3_HEXSS,LA+3:O-2;0),, +GLA2O3H;,,, ENTER-PAR G(LA2O3_HEXSS,LA+3:VA;0),, +GLA2O3H-3*GHSEROO;,,, ENTER-PAR G(LA2O3_HEXSS,SR+2:O-2;0),, +SRH_ALPH+GHSEROO+15.87691*T;,,, ENTER-PAR G(LA2O3_HEXSS,SR+2:VA;0),, +SRH_ALPH-2*GHSEROO +15.87691*T;,,, ENTER-PAR L(LA2O3_HEXSS,LA+3,SR+2:O-2;0) 298.15 +193600-78.1*T;,,, ENTER-PAR L(LA2O3_HEXSS,LA+3,SR+2:VA;0) 298.15 +193600-78.1*T;,,, @@ ENTER-PHASE LA2O3_CUBSS,, 2 2 3 LA+3 SR+2; O-2 VA;,,, ENTER-PAR G(LA2O3_CUBSS,LA+3:O-2;0),, +GLA2O3X;,,, ENTER-PAR G(LA2O3_CUBSS,LA+3:VA;0),, +GLA2O3X-3*GHSEROO;,,, ENTER-PAR G(LA2O3_CUBSS,SR+2:O-2;0),, +SRX_ALPH+GHSEROO+15.87691*T;,,, ENTER-PAR G(LA2O3_CUBSS,SR+2:VA;0),, +SRX_ALPH2*GHSEROO+15.87691*T;,,, ENTER-PAR L(LA2O3_CUBSS,LA+3,SR+2:O-2;0) 298.15 +168700-78.1*T;,,, ENTER-PAR L(LA2O3_CUBSS,LA+3,SR+2:O-2;1) 298.15 -20000;,,, ENTER-PAR L(LA2O3_CUBSS,LA+3,SR+2:VA;0) 298.15 +168700-78.1*T;,,, ENTER-PAR L(LA2O3_CUBSS,LA+3,SR+2:VA;1) 298.15 -20000;,,, @@ La4SrO7 as BETA phase ENTER-PHASE BETA,, 2 2 3 LA+3 SR+2; O-2 VA;,,, ENTER-PAR G(BETA,LA+3:O-2;0),, +LA_BETA;,,, ENTER-PAR G(BETA,LA+3:VA;0),, +LA_BETA-3*GHSEROO;,,, ENTER-PAR G(BETA,SR+2:O-2;0),, +SR_ALPHA+416100+GHSEROO -0.33333333*RE_BETA +15.87691*T;,,,
182
Appendix
ENTER-PAR G(BETA,SR+2:VA;0),,
+SR_ALPHA+416100-2*GHSEROO
+0.66666667*RE_BETA+15.87691*T;,,, ENTER-PAR L(BETA,LA+3,SR+2:O-2;0) 298.15 -121000-237.8*T;,,, ENTER-PAR L(BETA,LA+3,SR+2:VA;0) 298.15 -121000-237.8*T;,,, @@ La4Sr3O9, Stoichiometric ENTER-PHASE LA4SR3O9,, 3 4 3 9 LA+3; SR+2; O-2;,,, ENTER-PAR G(LA4SR3O9,LA+3:SR+2:O-2;0),, +2*GLA2O3D+3*GSROSOL+229800 -136.75*T;,,, @@ SrO Solid Solution ENTER-PHASE SRO,, 2 1 1 LA+3 SR+2 VA; O-2;,,, ENTER-PAR G(SRO,LA+3:O-2;0),, +0.5*GLA2O3D+113700;,,, ENTER-PAR G(SRO,SR+2:O-2;0),, +GSROSOL;,,, ENTER-PAR G(SRO,VA:O-2;0),, 0;,,, @@ @@ LA-MN-O, Grundy @@ ENTER-PHASE L2MNO4,, 3 2 1 4 LA+3; MN+2; O-2;,,, ENTER-PAR G(L2MNO4,LA+3:MN+2:O-2;0) 298.15 +GL2MNO4;,,, @@ ENTER-PHASE LMN2O5,, 4 1 1 1 5 LA+3; MN+3; MN+4; O-2;,,, ENTER-PAR G(LMN2O5,LA+3:MN+3:MN+4:O-2;0) 298.15 +GLMN2O5;,,, @@ @@ LA-CR-O, Povoden @@ @@ STOICHIOMETRIC LA2CRO6 ENTER-PHASE LA2CRO6,, 3 2 1 6 LA+3; CR+6; O-2;,,, ENTER-PAR G(LA2CRO6,LA+3:CR+6:O-2;0) 298.15 +GLA2CRO6;,,, @@ intermediate La-chromates @@ STOICHIOMETRIC LA16 @@ ENTER-PHASE LA16,, 3 16 7 44 LA; CR; O;,,, @@ ENTER-PAR G(LA16,LA:CR:O;0) 298.15 +GLA16;,,, @@ STOICHIOMETRIC LA7 @@ ENTER-PHASE LA7,, 3 7 2 16 LA; CR; O;,,, @@ ENTER-PAR G(LA7,LA:CR:O;0) 298.15 +GLA7;,,, @@ @@ STOICHIOMETRIC LA2CR3O12 ENTER-PHASE LA2CR3,, 3 2 3 12 LA+3; CR+6; O-2;,,, ENTER-PAR G(LA2CR3,LA+3:CR+6:O-2;0) 298.15 +GLA2CR3;,,, @@ @@ SR-MN-O, Grundy @@ ENTER-PHASE SR7MN4O15,, 3 7 4 15 SR+2; MN+4; O-2;,,, ENTER-PAR G(SR7MN4O15,SR+2:MN+4:O-2;0),, +GS7M4;,,, @@ ENTER-PHASE SRMN3O6,, 5 1 2 3 1 3 SR+2; MN+3; O-2; MN+3 MN+4; O-2 VA;,,, ENTER-PAR G(SRMN3O6,SR+2:MN+3:O-2:MN+3:O-2;0),, +GSM3OZ+0.5*GHSEROO +11.23859*T;,,, ENTER-PAR G(SRMN3O6,SR+2:MN+3:O-2:MN+3:VA;0),, +GSM3OZ-2.5*GHSEROO +11.23859*T;,,, ENTER-PAR G(SRMN3O6,SR+2:MN+3:O-2:MN+4:O-2;0),, +GSM4OZ;,,, ENTER-PAR G(SRMN3O6,SR+2:MN+3:O-2:MN+4:VA;0),, +GSM4OZ-3*GHSEROO;,,, @@ ENTER-PHASE SR4MN3O10,, 3 4 3 10 SR+2; MN+4; O-2;,,, ENTER-PAR G(SR4MN3O10,SR+2:MN+4:O-2;0),, +GSM4_RP3;,,, @@ ENTER-PHASE SRMNO3_HEX,, 3 1 1 3 SR+2; MN+3 MN+4; O-2 VA;,,, ENTER-PAR G(SRMNO3_HEX,SR+2:MN+3:O-2;0),, +GSM3_HEX+0.5*GHSEROO +11.23859*T;,,, ENTER-PAR G(SRMNO3_HEX,SR+2:MN+3:VA;0),, +GSM3_HEX-2.5*GHSEROO +11.23859*T;,,,
183
Appendix
ENTER-PAR G(SRMNO3_HEX,SR+2:MN+4:O-2;0),, +GSM4_HEX;,,, ENTER-PAR G(SRMNO3_HEX,SR+2:MN+4:VA;0),, +GSM4_HEX-3*GHSEROO;,,, @@ @@ RP2, Grundy modified it in LSM modeling ENTER-PHASE RP2,, 4 1 2 2 7 SR+2; SR+2; MN+3 MN+4; O-2;,,, ENTER-PAR G(RP2,SR+2:SR+2:MN+3:O-2;0) 298.15 +3*GSROSOL+GMN2O3+GHSEROO;,,, ENTER-PAR G(RP2,SR+2:SR+2:MN+4:O-2;0) 298.15 +GS4O_RP2;,,, @@ @@ SR-CR-O, Povoden @@ @@ SRCR2O4 ENTER-PHASE SC2O4,, 3 1 2 4 SR+2; CR+3; O-2;,,, ENTER-PAR G(SC2O4,SR+2:CR+3:O-2;0) 298.15 +GSC2O4;,,, @@ SR2CRO4 ENTER-PHASE S2CO4,, 3 2 1 4 SR+2; CR+4; O-2;,,, ENTER-PAR G(S2CO4,SR+2:CR+4:O-2;0) 298.15 +GS2CO4;,,, @@ @@ SR2.67CR2O8 ENTER-PHASE S3C2N,, 3 2.666667 2 8 SR; CR; O;,,, ENTER-PAR G(S3C2N,SR:CR:O;0) 298.15 +GS3C2O8;,,, @@ SRCRO4, Povoden-Karadeniz ENTER-PHASE SCO4,, 3 1 1 4 SR+2; CR+6; O-2;,,, ENTER-PAR G(SCO4,SR+2:CR+6:O-2;0) 298.15 +GSCO4;,,, @@ SRCR2O7, doubtful phase @@ ENTER-PHASE SC2O7,, 3 1 2 7 SR+2; CR+6; O-2;,,, @@ ENTER-PAR G(SC2O7,SR+2:CR+6:O-2;0) 298.15 +GSC2O7;,,, @@ @@ CR-MN-O, Povoden @@ @@ NONSTOICHIOMETRIC MANGANOSITE (MNO) SOLID SOLUTION, Grundy, Povoden-K. ENTER-PHASE MNO_HALIT,, 2 1 1 MN+2 MN+3 CR+3 VA; O-2;,,, ENTER-PAR G(MNO_HALIT,MN+2:O-2;0) 298.15 +GMN1O1;,,, ENTER-PAR G(MNO_HALIT,MN+3:O-2;0) 298.15 +GMN1O1-21883.5213 -22.1853365*T;,,, ENTER-PAR G(MNO_HALIT,CR+3:O-2;0) 298.15 +0.5*GCR2O3-7.93845*T +71549.3;,,, ENTER-PAR G(MNO_HALIT,VA:O-2;0) 298.15 0;,,, ENTER-PAR L(MNO_HALIT,MN+2,MN+3:O-2;0) 298.15 -4.21048766E+04;,,, ENTER-PAR L(MNO_HALIT,MN+2,MN+3:O-2;1) 298.15 +4.65131533E+04;,,, @@ @@ Mn2O3 (Compatible with C-Y2O3, Grundy, Chen, Povoden-Karadeniz) ENTER-PHASE MN2O3,, 3 2 3 1 MN+3 CR+3; O-2 VA; O-2 VA;,,, AMEND-PHASE MN2O3 MAGN,, .28 ENTER-PAR TC(MN2O3,MN+3:O-2:VA;0) 298.15 +309;,,, ENTER-PAR TC(MN2O3,CR+3:O-2:VA;0) 298.15 +308.6;,,, ENTER-PAR BMAGN(MN2O3,MN+3:O-2:VA;0) 298.15 +0.59;,,, ENTER-PAR BMAGN(MN2O3,CR+3:O-2:VA;0) 298.15 +3;,,, ENTER-PAR G(MN2O3,MN+3:O-2:VA;0) 298.15 +GMN2O3;,,, ENTER-PAR G(MN2O3,MN+3:O-2:O-2;0) 298.15 +GMN2O3+GHSEROO +100000+15.87691*T;,,, ENTER-PAR G(MN2O3,MN+3:VA:O-2;0) 298.15 +GMN2O3-2*GHSEROO +100000+15.87691*T;,,, ENTER-PAR G(MN2O3,MN+3:VA:VA;0) 298.15 +GMN2O3-3*GHSEROO;,,, ENTER-PAR G(MN2O3,CR+3:O-2:VA;0) 298.15 +GCR2O3+3459;,,, ENTER-PAR G(MN2O3,CR+3:O-2:O-2;0) 298.15 +GCR2O3+GHSEROO +100000+15.87691*T;,,, ENTER-PAR G(MN2O3,CR+3:VA:O-2;0) 298.15 +GCR2O3-2*GHSEROO +100000+15.87691*T;,,, ENTER-PAR G(MN2O3,CR+3:VA:VA;0) 298.15 +GCR2O3-3*GHSEROO;,,, @@ @@ ESKOLAITE, NONSTOICHIOMETRIC CR2O3 SOLID SOLUTION, Povoden-Karadeniz
184
Appendix
ENTER-PHASE CR2O3,, 3 2 1 3 MN+3 CR+2 CR+3; CR+3 VA; O-2;,,, AMEND-PHASE CR2O3 MAGN,, .28 ENTER-PAR TC(CR2O3,MN+3:CR+3:O-2;0) 298.15 +309;,,, ENTER-PAR TC(CR2O3,MN+3:VA:O-2;0) 298.15 +309;,,, ENTER-PAR TC(CR2O3,CR+2:CR+3:O-2;0) 298.15 +308.6;,,, ENTER-PAR TC(CR2O3,CR+2:VA:O-2;0) 298.15 +308.6;,,, ENTER-PAR TC(CR2O3,CR+3:CR+3:O-2;0) 298.15 +308.6;,,, ENTER-PAR TC(CR2O3,CR+3:VA:O-2;0) 298.15 +308.6;,,, ENTER-PAR BMAGN(CR2O3,MN+3:CR+3:O-2;0) 298.15 +0.59;,,, ENTER-PAR BMAGN(CR2O3,MN+3:VA:O-2;0) 298.15 +0.59;,,, ENTER-PAR BMAGN(CR2O3,CR+2:CR+3:O-2;0) 298.15 +3;,,, ENTER-PAR BMAGN(CR2O3,CR+2:VA:O-2;0) 298.15 +3;,,, ENTER-PAR BMAGN(CR2O3,CR+3:CR+3:O-2;0) 298.15 +3;,,, ENTER-PAR BMAGN(CR2O3,CR+3:VA:O-2;0) 298.15 +3;,,, ENTER-PAR G(CR2O3,MN+3:CR+3:O-2;0) 298.15 +GMN2O3+GSERCR+39503;,,, ENTER-PAR G(CR2O3,MN+3:VA:O-2;0) 298.15 +GMN2O3+39503;,,, ENTER-PAR G(CR2O3,CR+2:CR+3:O-2;0) 298.15 +GCRO0 +.33333333334*GHSERCR -5.2923*T;,,, ENTER-PAR G(CR2O3,CR+2:VA:O-2;0) 298.15 +GCRO0 -0.666666666667*GHSERCR -5.2923*T;,,, ENTER-PAR G(CR2O3,CR+3:CR+3:O-2;0) 298.15 +GCR2O3+GHSERCR;,,, ENTER-PAR G(CR2O3,CR+3:VA:O-2;0) 298.15 +GCR2O3;,,, @@ @@ CUBIC SPINEL, Grundy, Povoden-Karadeniz ENTER-PHASE CMNCR2O4,, 3 1 2 4 MN+2 CR+2; MN+3 CR+3; O-2;,,, ENTER-PAR G(CMNCR2O4,MN+2:CR+3:O-2;0) 298.15 +GSPINEL;,,, ENTER-PAR G(CMNCR2O4,CR+2:CR+3:O-2;0) 298.15 +GCR3O4;,,, ENTER-PAR G(CMNCR2O4,MN+2:MN+3:O-2;0) 298.15 +GCMN3O4;,,, ENTER-PAR G(CMNCR2O4,CR+2:MN+3:O-2;0) 298.15 +GCR3O4+GCMN3O4-GSPINEL;,,, @@ DISTORTED_SPINEL, Grundy, Povoden-Karadeniz ENTER-PHASE TMNCR2O4,, 3 1 2 4 MN+2; MN+3 CR+3; O-2;,,, ENTER-PAR G(TMNCR2O4,MN+2:MN+3:O-2;0) 298.15 +GTMN3O4;,,, ENTER-PAR G(TMNCR2O4,MN+2:CR+3:O-2;0) 298.15 +GTSPINEL;,,, @@ -----------------------------------------------------------------------@@ Quaternary oxides @@ @@ La-Sr-Mn-O, Grundy @@ ENTER-PHASE LS3MN2O7,, 4 1 2 2 7 SR+2; LA+3 SR+2; MN+3 MN+4; O-2;,,, ENTER-PAR G(LS3MN2O7,SR+2:LA+3:MN+3:O-2;0) 298.15 +GL3O_RP2;,,, ENTER-PAR G(LS3MN2O7,SR+2:LA+3:MN+4:O-2;0) 298.15 +GL3O_RP2+GS4O_RP2 -3*GSROSOL -GMN2O3-GHSEROO-R_RP2;,,, ENTER-PAR G(LS3MN2O7,SR+2:SR+2:MN+3:O-2;0) 298.15 +3*GSROSOL+GMN2O3 +GHSEROO;,,, ENTER-PAR G(LS3MN2O7,SR+2:SR+2:MN+4:O-2;0) 298.15 +GS4O_RP2;,,, @@ -----------------------------------------------------------------------@@ Quinary phases, ideal extensions from lower-order systems @@ @@ high temperature rhombohedral perovskite, Grundy, Povoden-Karadeniz ENTER-PHASE RPEROV,, 3 1 1 3 LA+3 SR+2 VA; MN+2 MN+3 MN+4 CR+3 CR+4 VA; O-2 VA;,,, @@ ENTER-PAR G(RPEROV,SR+2:MN+2:O-2;0),, +GMS3O+GL2O-GL3OR+GHSEROO +22.47717*T;,,, ENTER-PAR G(RPEROV,SR+2:MN+3:O-2;0),, +GMS3O+0.5*GHSEROO+11.23859*T;,,, ENTER-PAR G(RPEROV,SR+2:MN+4:O-2;0),, +GMS4O;,,, ENTER-PAR G(RPEROV,SR+2:MN+2:VA;0),, +GMS3O+GL2O-GL3OR2*GHSEROO+22.47717*T;,,,
185
Appendix
ENTER-PAR G(RPEROV,SR+2:MN+3:VA;0),, ENTER-PAR G(RPEROV,SR+2:MN+4:VA;0),, ENTER-PAR G(RPEROV,SR+2:CR+3:O-2;0),, ENTER-PAR G(RPEROV,SR+2:CR+3:VA;0),, ENTER-PAR G(RPEROV,SR+2:VA:O-2;0),,
ENTER-PAR G(RPEROV,SR+2:VA:VA;0),,
ENTER-PAR ENTER-PAR ENTER-PAR ENTER-PAR ENTER-PAR
G(RPEROV,LA+3:MN+2:O-2;0),, G(RPEROV,LA+3:MN+2:VA;0),, G(RPEROV,LA+3:MN+3:O-2;0),, G(RPEROV,LA+3:MN+3:VA;0),, G(RPEROV,LA+3:MN+4:O-2;0),,
ENTER-PAR G(RPEROV,LA+3:MN+4:VA;0),,
ENTER-PAR G(RPEROV,LA+3:VA:O-2;0),,
ENTER-PAR G(RPEROV,LA+3:VA:VA;0),, ENTER-PAR G(RPEROV,VA:MN+3:O-2;0),, ENTER-PAR G(RPEROV,VA:MN+4:O-2;0),, ENTER-PAR G(RPEROV,VA:MN+2:VA;0),, ENTER-PAR G(RPEROV,VA:MN+3:VA;0),, ENTER-PAR G(RPEROV,VA:MN+4:VA;0),, ENTER-PAR ENTER-PAR ENTER-PAR ENTER-PAR ENTER-PAR G(RPEROV,VA:VA:O-2;0),, G(RPEROV,VA:VA:VA;0),, G(RPEROV,LA+3:CR+3:O-2;0),, G(RPEROV,LA+3:CR+3:VA;0),, G(RPEROV,VA:CR+3:O-2;0),,
ENTER-PAR G(RPEROV,VA:CR+3:VA;0),,
ENTER-PAR G(RPEROV,LA+3:CR+4:VA;0),,
ENTER-PAR G(RPEROV,LA+3:CR+4:O-2;0),, ENTER-PAR G(RPEROV,SR+2:CR+4:O-2;0),, ENTER-PAR G(RPEROV,SR+2:CR+4:VA;0),, ENTER-PAR G(RPEROV,VA:CR+4:O-2;0),,
ENTER-PAR G(RPEROV,VA:CR+4:VA;0),,
+GMS3O-2.5*GHSEROO+11.23859*T;,,, +GMS4O-3*GHSEROO;,,, +GN+0.1666667*GS4O -0.16666667*GS4V;,,, +GN-0.8333333*GS4O +0.8333333*GS4V;,,, +GMS3O-GL3OR+2*GL4O -1.5*GV4O+0.5*GVVV +2*GHSEROO+12.62121*T;,,, +GMS3O+2*GL4O-1.5*GV4O +0.5*GVVV-GL3OR -GHSEROO+12.62121*T;,,, +GL2O+0.5*GHSEROO+11.2386*T;,,, +GL2O-2.5*GHSEROO+11.2386*T;,,, +GL3OR;,,, +GL3OR-3*GHSEROO;,,, +0.66667*GL4O+0.5*GV4O -0.166667*GVVV -0.5*GHSEROO+5.76318*T;,,, +0.66667*GL4O+0.5*GV4O -0.166667*GVVV -3.5*GHSEROO+5.76318*T;,,, +2*GL4O-1.5*GV4O+0.5*GVVV +1.5*GHSEROO +1.41263*T;,,, +2*GL4O+0.5*GVVV-1.5*GV4O -1.5*GHSEROO+1.41263*T;,,, +GL3OR+1.5*GV4O+0.5*GVVV-2*GL4O +1.5*GHSEROO-1.41263*T;,,, +2*GV4O+0.33333*GVVV-1.33333*GL4O +GHSEROO+4.35056*T;,,, +GL2O+1.5*GV4O+0.5*GVVV-2*GL4O -GHSEROO+9.82596*T;,,, +GL3OR+1.5*GV4O+0.5*GVVV-2*GL4O -1.5*GHSEROO-1.41263*T;,,, +2*GV4O+0.3333*GVVV-1.333*GL4O -2*GHSEROO+4.35057*T;,,, +GVVV+3*GHSEROO;,,, +GVVV;,,, +GLACRO3;,,, +GLACRO3-3*GHSEROO;,,, +GLACRO3+1.5*GVCR4O+0.5*GVVV -2*GLACR4O+1.5*GHSEROO -1.41263*T;,,, +GLACRO3+1.5*GVCR4O+0.5*GVVV -2*GLACR4O-1.5*GHSEROO -1.41263*T;,,, GS4O-GN-0.1666667*GS4O +0.16666667*GS4V+GLACRO3 -3*GHSEROO;,,, GS4O+GLACRO3-GN-0.1666667*GS4O +0.16666667*GS4V;,,, +GS4O;,,, +GS4V;,,, +2*GVCR4O+0.33333*GVVV -1.33333*GLACR4O +GHSEROO+4.35056*T;,,, +2*GVCR4O+0.3333*GVVV -1.333*GLACR4O -2*GHSEROO+4.35057*T;,,,
@@ Optimized interactions
186
Appendix
ENTER-PAR L(RPEROV,LA+3,SR+2:MN+2:VA;1),, -136600;,,, ENTER-PAR L(RPEROV,LA+3,SR+2:MN+2:O-2;1),, -136600;,,, ENTER-PAR L(RPEROV,LA+3,SR+2:MN+4:VA;1),, -117000;,,, ENTER-PAR L(RPEROV,LA+3,SR+2:MN+4:O-2;1),, -117000;,,, ENTER-PAR L(RPEROV,LA+3:CR+3,MN+3:O-2;0),, 9248.6;,,, ENTER-PAR L(RPEROV,LA+3,SR+2:MN+3:O-2;0) 298.15 -1.5*t;,,, ENTER-PAR L(RPEROV,LA+3:CR+4,VA:O-2;0),, 250000;,,, ENTER-PAR L(RPEROV,LA+3:CR+3,VA:O-2;0),, 250000;,,, ENTER-PAR L(RPEROV,LA+3:MN+4,MN+3:O-2;1) 298.15 185;,,, ENTER-PAR L(RPEROV,LA+3:MN+4,VA:O-2;0),, -2;,,, ENTER-PAR L(RPEROV,LA+3,VA:MN+4:O-2;0),, 20;,,, ENTER-PAR L(RPEROV,LA+3:CR+3,MN+3:O-2;0) 298.15 3766;,,, ENTER-PAR L(RPEROV,LA+3:CR+3,MN+3:O-2;1) 298.15 -1297;,,, @@ next two interaction parameters can be used to fit @@ Cr4+ amount in perovskite ENTER-PAR L(RPEROV,SR+2:CR+4,MN+3:O-2;0) 298.15 0;,,, ENTER-PAR L(RPEROV,SR+2:CR+4,MN+4:O-2;0) 298.15 0;,,, @@ @@ Ruddlesden-Popper phase, preliminary, scarce experimental data @@ ENTER-PHASE RP,, 5 1 1 1 3 1 SR+2; LA+3 SR+2; MN+3 MN+4 CR+3 CR+4; O-2; @@ O-2;,,, @@ ENTER-PAR G(RP,SR+2:SR+2:CR+4:O-2:O-2;0) 298.15 GS2CO4;,,, @@ ENTER-PAR G(RP,SR+2:LA+3:CR+3:O-2:O-2;0) 298.15 +GLCR3O_RP1;,,, @@ ENTER-PAR G(RP,SR+2:LA+3:CR+4:O-2:O-2;0) 298.15 GS2CO4+GLCR3O_RP1 @@ -GREFRP;,,, @@ ENTER-PAR G(RP,SR+2:SR+2:CR+3:O-2:O-2;0) 298.15 GREFRP;,,, @@ ENTER-PAR L(RP,SR+2:SR+2,LA+3:CR+3:O-2:O-2;0) 298.15 100000;,,, @@ ENTER-PAR L(RP,SR+2:SR+2,LA+3:CR+4:O-2:O-2;0) 298.15 100000;,,, @@ ENTER-PAR G(RP,SR+2:SR+2:MN+3:O-2:O-2;0) 298.15 +2*GSROSOL+0.5*GMN2O3 @@ +0.5*GHSEROO;,,, @@ ENTER-PAR G(RP,SR+2:SR+2:MN+4:O-2:O-2;0) 298.15 +GS4O_RP1;,,, @@ ENTER-PAR G(RP,SR+2:LA+3:MN+3:O-2:O-2;0) 298.15 +GL3O_RP1;,,, @@ ENTER-PAR G(RP,SR+2:LA+3:MN+4:O-2:O-2;0) 298.15 +GL3O_RP1+GS4O_RP1 @@ -2*GSROSOL-0.5*GMN2O3 @@ -0.5*GHSEROO;,,, @@ -----------------------------------------------------------------------@@ Liquid, ideal extension from lower-order systems, Povoden-Karadeniz ENTER-PHASE IONIC_LIQUID Y LA+3 SR+2 MN+2 MN+3 CR+2 CR+3; O-2 VA;,,, AMEND-PHASE IONIC_LIQUID COMP 2,,,,,,,, ENTER-PAR G(ION,LA+3:O-2;0) 298.15 +GLA2O3LIQ;,,, ENTER-PAR G(ION,LA+3:VA;0) 298.15 +GLALIQ;,,, ENTER-PAR G(ION,SR+2:O-2;0) 298.15 +2*GSROLIQ;,,, ENTER-PAR G(ION,SR+2:VA;0) 298.15 +GSRLIQ;,,, ENTER-PAR G(ION,MN+2:O-2;0) 298.15 +2*GMN1O1_L;,,, ENTER-PAR G(ION,MN+2:VA;0) 298.15 +GMN_L;,,, ENTER-PAR G(ION,MN+3:O-2;0) 298.15 +GMN2O3_L;,,, ENTER-PAR G(ION,MN+3:VA;0) 298.15 +2*GMN_L+GMN2O3_L-3*GMN1O1_L;,,, ENTER-PAR G(ION,CR+2:O-2;0) 298.15 2*GCR1O1_L;,,, ENTER-PAR G(ION,CR+2:VA;0) 298.15 +GCR_L;,,, ENTER-PAR G(ION,CR+3:O-2;0) 298.15 +GCR2O3_L;,,, ENTER-PAR G(ION,CR+3:VA;0) 298.15 +2*GCR_L+GCR2O3_L-3*GCR1O1_L;,,, @@ Interaction parameters from binaries ENTER-PAR L(ION,MN+2:O-2,VA;0) 298.15 1.29519000E+05;,,, ENTER-PAR L(ION,MN+2:O-2,VA;1) 298.15 -4.54590000E+04;,,, ENTER-PAR L(ION,MN+2,MN+3:O-2;0) 298.15 -3.38590000E+04;,,, ENTER-PAR L(ION,CR+2:O-2,VA;0) 298.15 +121000;,,, ENTER-PAR L(ION,CR+3:O-2,VA;0) 298.15 +121000;,,, @@ Interaction parameters from ternaries ENTER-PAR L(ION,LA+3,MN+2:O-2;0) 298.15 -119062;,,, ENTER-PAR L(ION,LA+3,MN+2:VA;0) 298.15 +11368;,,, ENTER-PAR L(ION,LA+3,MN+2:VA;1) 298.15 -11316.4;,,,
187
Appendix
ENTER-PAR L(ION,LA+3,MN+2:VA;4) 298.15 -9111;,,, ENTER-PAR L(ION,LA+3,MN+3:O-2;0) 298.15 -119062;,,, ENTER-PAR L(ION,SR+2,MN+2:O-2;0) 298.15 -176300;,,, ENTER-PAR L(ION,SR+2,MN+2:VA;0) 298.15 +46000;,,, ENTER-PAR L(ION,SR+2,MN+3:O-2;0) 298.15 -176300;,,, ENTER-PAR L(ION,MN+2,CR+2:VA;0) 298.15 -15009+13.6587*T;,,, ENTER-PAR L(ION,MN+2,CR+2:VA;1) 298.15 +504+0.9479*T;,,, ENTER-PAR L(ION,MN+3,CR+3:O-2;0) 298.15 -188487.7;,,, ENTER-PAR L(IONIC,LA+3,CR+2:O-2;0) 298.15 -101850;,,, ENTER-PAR L(IONIC,LA+3,CR+3:O-2;0) 298.15 -101850;,,, ENTER-PAR L(IONIC,LA+3,CR+2:O-2;1) 298.15 -39016;,,, ENTER-PAR L(IONIC,LA+3,CR+3:O-2;1) 298.15 -39016;,,, ENTER-PAR L(IONIC,LA+3,CR+2:VA;0) 298.15 +60713-5.49*T;,,, ENTER-PAR L(IONIC,LA+3,CR+2:VA;1) 298.15 +64573-23*T;,,, ENTER-PAR L(ION,SR+2,CR+2:VA;0) 298.15 200000;,,, ENTER-PAR L(ION,SR+2,CR+3:O-2;0) 298.15 -619869;,,, ENTER-PAR L(ION,SR+2,CR+3:O-2;1) 298.15 -179575;,,, ENTER-PAR L(ION,SR+2,CR+2:O-2;0) 298.15 -619869;,,, ENTER-PAR L(ION,SR+2,CR+2:O-2;1) 298.15 -179575;,,, @@-----------------------------------------------------------@@ Cr-GAS, SGTE and reassessments by Chen, Povoden-Karadeniz ENTER-PHASE GAS G 1 CR CR1O1 CR1O2 CR1O3 O O2 O3 H2 H H2O1 H1O1 H1O2 H2O2 CRH1 CRH1O1 CRH1O2 CRH1O3 CRH2O2 CRH2O3 CRH2O4 CRH3O3 CRH3O4 CRH4O4 CRH4O5 CRH5O5 CRH6O6;,,, ENTER-PAR G(GAS,CR1O1;0) 298.15 +GCR1O1_G+RTLNP;,,, ENTER-PAR G(GAS,CR1O2;0) 298.15 +GCR1O2_G+RTLNP;,,, ENTER-PAR G(GAS,CR1O3;0) 298.15 +GCR1O3_G+RTLNP;,,, ENTER-PAR G(GAS,O;0) 298.15 +F13349T+RTLNP; 6000 N ENTER-PAR G(GAS,O2;0) 298.15 +F13704T+RTLNP; 20000 N ENTER-PAR G(GAS,O3;0) 298.15 +F14021T+RTLNP; 6000 N ENTER-PAR G(GAS,H2;0) 298.15 +H2GAS+RTLNP; 6000 N ENTER-PAR G(GAS,H;0) 298.15 +HGAS+RTLNP; 6000 N ENTER-PAR G(GAS,H2O1;0) 298.15 +F10963T+RTLNP; 20000 N ENTER-PAR G(GAS,H1O1;0) 298.15 +F10666T+RTLNP; 20000 N ENTER-PAR G(GAS,H1O2;0) 298.15 +F10729T+RTLNP; 6000 N ENTER-PAR G(GAS,H2O2;0) 298.15 +F10983T+RTLNP; 1500 N ENTER-PAR G(GAS,CRH1;0) 298.15 +F7586T+RTLNP; 6000 N ENTER-PAR G(GAS,CRH1O1;0) 298.15 +GCRH1O1_G+RTLNP;,,, ENTER-PAR G(GAS,CRH1O2;0) 298.15 +GCRH1O2_G+RTLNP;,,, ENTER-PAR G(GAS,CRH1O3;0) 298.15 +GCRH1O3_G+RTLNP;,,, ENTER-PAR G(GAS,CRH2O2;0) 298.15 +GCRH2O2_G+RTLNP;,,, ENTER-PAR G(GAS,CRH2O3;0) 298.15 +GCRH2O3_G+RTLNP;,,, ENTER-PAR G(GAS,CRH2O4;0) 298.15 +GCRH2O4_G+RTLNP;,,, ENTER-PAR G(GAS,CRH3O3;0) 298.15 +GCRH3O3_G+RTLNP;,,, ENTER-PAR G(GAS,CRH3O4;0) 298.15 +GCRH3O4_G+RTLNP;,,, ENTER-PAR G(GAS,CRH4O4;0) 298.15 +GCRH4O4_G+RTLNP;,,, ENTER-PAR G(GAS,CRH5O5;0) 298.15 +GCRH5O5_G+RTLNP;,,, ENTER-PAR G(GAS,CRH6O6;0) 298.15 +GCRH6O6_G+RTLNP;,,, ENTER-PAR G(GAS,CRH4O5;0) 298.15 +GCRH4O5_G+RTLNP;,,, @@ @@ CRO3 reference gas ENTER-PHASE CRGAS,, 1 CR1O3;,,, ENTER-PAR G(CRGAS,CR1O3;0) 298.15 +GCR1O3_G+RTLNP;,,, @@ @@ H2O reference gas ENTER-PHASE STEAM,, 1 H2O1;,,, ENTER-PAR G(STEAM,H2O1;0) 298.15 +F10963T+RTLNP; 20000 N @@ @@ O2 reference gas ENTER-PHASE O2GAS,, 1 O2;,,,
188
Appendix
ENTER-PAR G(O2GAS,O2;0) 298.15 @@ GO PAR @@ SET-OUT-LEVEL,,,,, N,, set-interactive
+2*GHSEROO+RTLNP;,,,
189
Curriculum vitae
Curriculum Vitae
Personal data
Povoden-Karadeniz, Erwin Date and place of birth: March 18, 1973, Graz, Austria
Education
01/2005 12/2008 Ph. D. thesis: Thermodynamic Database of the La-Sr-Mn-Cr-O Oxide System and Applications to Solid Oxide Fuel Cells, Nonmetallic Inorganic Materials, Prof. Dr. Ludwig J. Gauckler, Department of Materials, ETH Zurich, Switzerland. 1992-1999 Geoscience Studies, Karl-Franzens University Graz Master thesis: Kontaktmetamorphose und Fluid-Gestein-Interaktion in der stlichen Monzoni Kontaktaureole, Prof. Dr. Georg Hoinkes and Prof. Dr. Rainer Abart 1983-1991 Realistisches Gymnasium, Pestalozzistrasse 5, Graz 1979-1983 Elementary School, Berliner Ring, Graz
Publications
R. Abart, N. Badertscher, M. Burkhard, and E. Povoden, Oxygen, carbon and strontium isotope systematics in two profiles across the Glarus thrust: implications for fluid flow, Contr.
Miner. Petrol., 2002, 143, pp. 192-208.
190
Curriculum vitae
E. Povoden, M. Horacek, and R. Abart, Contact metamorphism of siliceous dolomite and impure limestones from the Werfen formation in the eastern Monzoni contact aureole, Miner.
Petrol., 2002, 76, pp. 99-120.

A.N. Grundy, E. Povoden, T. Ivas, and L.J. Gauckler, Calculation of defect chemistry using the CALPHAD approach, Calphad, 2006, 30, pp. 33-41. E. Povoden, A.N. Grundy, and L.J. Gauckler, Thermodynamic assessment of the Mn-Cr-O system for solid oxide fuel cell (SOFC) materials, Int. J. Mater. Res., 2006, 97, pp. 569-78. E. Povoden, A.N. Grundy, and L.J. Gauckler, Thermodynamic reassessment of the Cr-O system in the framework of solid oxide fuel cell (SOFC) research, J. Phase Equilib. Diff., 2006, 27, pp. 353-62. E. Povoden, A.N. Grundy, M. Chen, and L.J. Gauckler, Thermodynamic assessment of the La-Cr-O system, J. Phase Equilib.Diff., 2009, 1, pp. 12-27. E. Povoden-Karadeniz, A.N. Grundy, M. Chen, T. Ivas, and L.J. Gauckler, Thermodynamic assessment of the La-Fe-O system, J. Phase Equilib. Diff. (accepted) E. Povoden, M. Chen, A.N. Grundy, and L.J. Gauckler, Thermodynamic La-Sr-Mn-Cr-O oxide database for solid oxide fuel cell applications, to be submitted. E. Povoden, T. Ivas, M. Chen, and L.J. Gauckler, Thermodynamic calculations of impacts of chromium on Sr-doped Lanthanum manganite (LSM) cathodes for solid oxide fuel cells (SOFC), to be submitted. E. Povoden, T. Ivas, and L.J. Gauckler, Degradation of planar solid oxide fuel cells with (La1-xSrx)1-yMnO3 cathodes and Cr-alloy interconnects, to be submitted.
Presentations
Assessment of the Cr-O system in the frame of SOFC research E. Povoden, A.N. Grundy, and L.J. Gauckler 191
Curriculum vitae
Poster Presentation CALPHAD XXXIV, Maastricht, Netherlands, May 06th -11th, 2005 Thermodynamic assessment of the Cr-Mn-O system for solid oxide fuel cell (SOFC) materials E. Povoden Oral Presentation 1st EMPA symposium for Ph.D. students, EMPA Dbendorf, Switzerland, October 20th, 2005 Thermodynamic Assessment of the Cr-Mn-O System for Solid Oxide Fuel Cell (SOFC) Materials E. Povoden, A.N. Grundy, and L.J. Gauckler Oral Presentation 3rd Fuel Cell Research Symposium, EMPA Dbendorf, Switzerland, March 16th, 2006 Thermodynamic assessment of the La-Mn-Cr-O system for applications on solid oxide fuel cell (SOFC) materials using the CALPHAD approach E. Povoden, A.N. Grundy, and L.J. Gauckler Oral Presentation Thermo 2006, Boulder, Colorado, USA, August 4th, 2006 Thermodynamic assessment of the La-Mn-Cr-O system for applications on solid oxide fel cell (SOFC) materials E. Povoden, A.N. Grundy, and L.J. Gauckler Oral Presentation HTMC XII, Vienna, Austria, September 18th 22nd, 2006 The BiO3/2-SbO3/2-ZnO Phase Diagram at 1115C in air E. Povoden, Z. Peng, and L.J. Gauckler Poster Presentation CALPHAD XXXVI, Pennsylvania State University, Pennsylvania, USA, May 6th 11th, 2007 Thermodynamic assessment of the La-Cr-O and LaO3/2-MnOx-CrO3/2 Systems 192
Curriculum vitae
E. Povoden, A.N. Grundy, and L.J. Gauckler Oral Presentation CALPHAD XXXVI, Pennsylvania State University, Pennsylvania, USA, May 6th 11th, 2007 Thermodynamic modeling for solid oxide fuel cell research - The La-Sr-Mn-Cr-O system E. Povoden Oral Presentation 2nd MRC Graduate Symposium , ETH Zurich, June 27th, 2007 The thermodynamic LaO1.5-SrO-MnO1.5-CrO1.5 and LaO1.5-SrO-FeO1.5-CrO1.5 databases for SOFC applications using the CALPHAD approach E. Povoden Oral Presentation Guest Talk, Lehrstuhl fr physikalische Chemie, Monatnuniversitt Leoben, March 28th, 2008 Thermodynamic LaO1.5-SrO-MnO1.5-CrO1.5 and LaO1.5-SrO-FeO1.5-CrO1.5 databases for solid oxide fuel cells (SOFC) applications E. Povoden, M. Chen, A.N. Grundy, and L.J. Gauckler Oral Presentation CALPHAD XXXVII, Saariselk, Finland, June 16th 21st, 2008
193
Erratum, p. 109, lines 7-14 and Table 4.3.3, p. 112
According to the latest discussion by B. Hallstedt et al., Calphad, 2007, 31(1), p 2837 the correct model is (La,Cr)(O,Va)3 and not (La,Cr)(O,Va)1.5: in case there is information of the ordering of element X between different vacant positions in bcc described by (Me)1(Va)1(Va)1(Va)1 this is taken into account, otherwise the disordered model for interstitials in bcc, (Me)1(Va,X)3 is to be used Thus the model was also corrected in the thermodynamic assessment of the La-Fe-O system, Povoden-Karadeniz et al., J. Phase Equilib. Diff., accepted. The following argumentation for modeling of the oxygen solubility in bcc using (physically wrong!) Me(O,Va)1.5 was given in a preliminary manuscript of the La-FeO system and is repeated here for the sake of clarification: as there are three octahedral interstitial sites per metal atom in the bcc unit cell located on the cube faces and cube edges Me(Va,O)3 is undoubtedly a reasonable model description for the oxygen solubility in bcc. In previous papers the authors have proposed the model Me(Va,O)1.5 based on the argument, that it is energetically very unfavorable to simultaneously occupy both the octahedral interstitial sites on the cube faces and cube edges as these lie very close to each other. A further reason why we use this description is that for SOFC applications we are primarily interested in the oxide portion of the phase diagram and we have found that the endmember Me(O)3 used to describe oxygen solubility in the metallic phase can inadvertently appear in the oxide portion of the phase diagram. For these reasons, and in order to keep this assessment compatible with our previous assessments we reassess the oxygen solubility in bcc Fe using the model (Fe)(O,Va)1.5.

Thermodynamic Understanding of Chromium Poisoning

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Thermodynamic Understanding of Chromium Poisoning

Uploaded by

Copyright:

Available Formats

Diss. ETH No.

Zurich, December 2008

3.2.4 3.3 4 4.1

Calculation of defect chemistry using the Calphad approach

4.4.5 Conclusions 5 Thermodynamic calculations of impacts of chromium on Sr-doped 5

Appendix Thermodynamic La-Cr database Thermodynamic La-Sr-Mn-Cr-O-(H) oxide database

170 169 172

Cr2 O3(s) + 1 2O2(g) + H 2 O(g) 2CrO2 (OH)(g)

The role of current load on electrical losses of degraded SOFC

Microstructures in degraded SOFC

Amounts of chromium in SOFC tested with and without current load

Critical assessment of proposed mechanisms of chromium poisoning

Proposed strategies against chromium poisoning and their effectiveness

New ways alternative materials

Stoichiometric solid oxides

The molar Gibbs energy of , Gm at constant pressure is given by

ptv muv nwt O (g) Gm + A + BT 2tv

are optimized by thermodynamic and phase diagram data.

+ tRT yar ln yar + ybq ln yar + EGmss

Vacancies and the concept of reciprocal reactions

( = ya3+ yO2- Gma

( + ya2+ yO2- Gma

( + ya3+ yVa Gma

( + ya2+ yVa Gma

(a3+ )2 (Va)3 + (a 2 + )2 (O 2- )3 (a3+ )2 (O 2- )3 (a 2 + )2 (Va)3 :

Optimization of model parameters

air, T (K) T (K) 2257 2317 2603 ----

1933 1918 1941 1929 1937 1938 2 1938

0.52 0.523 0.513 0.496 0.499 0.497 0.497

no no yes yes yes yes yes

2083 2083 -2160 2130 2083 2117

H T H 298K = 28.53T + 1.10 103 T 2 + 3.736 105 T 1 9759

Temperature derivation of Eq. 4.1.2 results in

S298K (Cr2O3) = 83.1 J K-1mol-1 by reevaluating emf

Previous assessments of the Cr-O system

giving the equilibrium constant

. Both are unlikely: the former because it is unlikely to get Cr4+

Fig. 4.1.1 Compound energy model for the Cr2+xO3 phase

3+ 2 2 (Cr 2 3Va1 3 )(O )3

= GCr O + 2 3 GCrSER + A + BT + RT ( 2 3ln 2 3+1 3ln1 3)

Mass 51.996 15.999

H298 - H0 4050.0 4341.0

S298 23.543 102.52

LCr 2+ :O2- ,Va q- = 0 LCr3+ :O2- ,Va q- = 121000

G LCr2+ :Va .The

, and, with low

Results and discussion

The solubility of oxygen in Cr(bcc) is shown in Fig. 4.1.7.

Fig. 4.1.7 Calculated oxygen solubility in Cr(bcc)

T. Metall. Soc. Aime, 1964, 230(6), pp. 1278-83.

Soc., 1968, 51(8), pp. 430-33.

22(2), pp. 225-32.

Elektrochem., 1942, 48(7), pp. 377-89 (in German).

Metall. Soc. Aime, 1963, 227(2), pp. 300-5.

Ac., 1995, 59(3), pp. 475-79.

Mineral., 2000, 85, pp. 1686-93.

Metal., 1973, 2, p. 102.

46, pp. 45-46 (in German).

Electrochem. Soc., 1956, 103(2), pp. 135-38.

2(3), pp. 227-38.

25(4), pp. 599-605.

Thermodynamic assessment of the Mn-Cr-O system for SOFC