PAUL INGE DAHL

Synthesis and characterization of ionic

conductors based on ZrO2, BaZrO3 and SrCeO3

and

Preparation of LaFeO3 and LaCoO3 thin films

IMT-Report 2006:86 IUK-Thesis 121

SEPTEMBER 2006

DEPARTMENT OF MATERIALS SCIENCE

AND ENGINEERING

NORWEGIAN UNIVERSITY OF SCIENCE AND TECHNOLOGY

NTNU

ii

iii

This thesis has been submitted to

Department of Materials Science and Engineering

Norwegian University of Science and Technology

in partial fulfillment of the requirements for

the academic degree

Doktor Ingeniør

September 2006

iv

v

PREFACE

First and foremost, I wish to express my gratitude to my supervisors Professor Mari-Ann Einarsrud and Professor Tor Grande, for the invaluable guidance in the process of completing this thesis. Their inspiration through the experimental work and efficient feedbacks during the writing of the thesis is highly appreciated.

Many additional people need to be honored for their contribution to the work presented in this thesis:

- Professor Mats Nygren, Dr. Zhe Zhao and Mats Johnsson for operating the Spark plasma sintering apparatus.

- Dr. Reidar Haugsrud, Professor Truls Norby and Christian Kjølseth at the University of Oslo, for electrical conductivity measurements on BaZrO3 and SrCeO3 materials and for assisting the preparation of manuscript to the corresponding parts of the papers.

- MSc. Øystein Andersen operating the spray pyrolysis, administrated by Associate professor Kjell Wiik.

- Dr. Ingeborg Kaus for assisting the work on YSZ.

- Dr. Hilde Lea Lein for experimental assistance on the preparation of BaZrO3 and SrCeO3 materials.

- Dr. Hasan Okuyucu and Laura Bertolo for experimental work on LaCoO3 films and Dr. Mohan Menon for his contribution to this topic.

- The technical staff at the Department of Materials Science and Technology, in particular Elin Nilsen for assisting the XRD measurements.

- Dr. Julian Tolchard for crystallographic illustrations and discussions and for feeding me while my wife was gone.

- Dr. Tommy Mokkelbost and Dr. Johann Mastin for sharing both brilliant ideas and frustrations in the process of completing this thesis. I hope we never meet again – at 2 am in the office!

- All other co-workers in Chemistry building II, friends and family for making this time enjoyable.

vi

Financial support from Norwegian University of Science and Technology and the Research Council of Norway, Grant No.1585171431, is appreciated.

Last but not least, I would like to give a special thank my dear wife Melinda and my children Isobel and Jonas, for reminding me that there is more to life than science. Without your patience and encouragement this would not be possible.

vii

TABLE OF CONTENT

1 Summary………….…………………………………………..………….1

2 Background and motivation…………………………………...…………5

2.1 Ionic conducting oxides and their applications…………….……...5

2.1.1 Oxygen ion conductors………………………….……….…...5

2.1.2 Proton conductors…...………………………….……….……5

2.1.3 Mixed ionic/electronic conductors……...………..……….….7

2.1.4 Solid oxide fule cells.....………………………..……….……8

2.1.5 Sensors…………………………………………….…….….10

2.2 Thin film technology…………….………………………….……10

2.3 Aim of work………..………………………………………….…11

3 Review of existing literature………………………………...………….13

3.1 Oxygen ion conductors – Zirconia………………………………..13

3.1.1 Crystal structure………………………………………….…..13

3.1.2 Stabilized zirconia……………………………………….…..13

3.1.3 Ionic conductivity in YSZ………………………………...…14

3.1.4 Powder synthesis…………………………………………….18

3.1.5 Densification and microstructure……………………………19

3.2 Proton conductors – Barium zirconate and strontium cerate………20

3.2.1 Crystal structure………………………………………….…..20

3.2.2 Electrical properties………………………………………….20

3.2.3 Powder synthesis and densification………………………….22

3.2.4 Mechanical properties and chemical stability……..…………23

3.3 Thin films of perovskite-type oxides………………………………26

3.3.1 Lanthanum ferrite thin films…………………………………26

3.3.2 Lanthanum cobaltite thin films………………………………27

References………………………………………………………………….28

viii

Scientific papers

PAPER I: Synthesis and characterization of nanocrystalline

YSZ powder by combustion synthesis……………………. 39

PAPER II: Densification and properties of zirconia prepared

by three different sintering techniques……………………. 50

PAPER III: Synthesis, densification and electrical properties

of strontium cerate ceramics………………………………. 71

PAPER IV: Preparation and characterization of

barium zirconate ceramics………………………………...101

PAPER V: Oriented LaFeO3 thin films grown on

NdGaO3 by spin-coating………………………………….131

PAPER VI: Crystallization and surface morphology of oriented

LaCoO3 films prepared by three different sol-gel routes....147

ix

LIST OF ACRONYMS

AFM Atomic Force Microscopy BZ Barium Zirconate CS Conventional Sintering FESEM Field Emission Scanning Electron Microscopy G/N Glycine to Nitrate (ratio) HP Hot Pressing HTXRD High Temperature X-Ray Diffraction IR Infrared Spectroscopy LC Lanthanum Cobaltite LCC 20% Calcium substituted Lanthanum Cobaltite p(O2) Partial Pressure of Oxygen p(H2O) Partial Pressure of water vapor p(CO2) Partial Pressure of carbon dioxide p(H2) Partial Pressure of Hydrogen SC Strontium Cerate SC5Yb 5% Ytterbium substituted Strontium Cerate SEM Scanning Electron Microscopy SOFC Solid Oxide Fuel Cell SPS Spark Plasma Sintering TEC Thermal Expansion Coefficient TEM Transmission Electron Microscopy TGA Thermo Gravimetrical Analysis XRD X-Ray Diffraction YBZ 10% Yttrium substituted Barium Zirconate YSZ Yttria Stabilized Zirconia

x

1

1 SUMMARY

Ceramic electrolytes that conduct oxygen ions, protons or exhibit mixed ionic and electronic conductivity at intermediate to high temperatures are important materials for use in electrochemical devices like solid oxide fuel cells (SOFCs), gas separation membranes and sensors. Special interest in SOFC technology, where further optimization is still needed, is due to better and more environmental friendly utilization of fossil fuel or hydrogen as an alternative energy carrier. One of the main topics regarding such oxide materials has been to improve in particular the ionic conductivity at suitable operation temperatures. While the research on SOFCs mainly involves bulk materials, oxide thin films are applicable for e.g. gas sensing devices and catalysts.

The electrical and chemical properties in a specific oxide system are dependent on the amount of impurities in the system and, in the case of cation substitution e.g. the valance and size of the substitutes compared to the original ions. Effects of microstructure must be considered as the electrical and chemical properties may differ with varying grain size. This leads to the overall aim of the present work, which has been to develop complete synthesis routes to bulk materials and thin films of ceramic oxides with designed microstructure. Yttria stabilized zirconia (YSZ), being the oxygen ion conductor traditionally used in oxygen sensors and SOFCs, has been prepared and characterized along with two perovskite-type oxide materials showing high proton conductivity; strontium cerate and barium zirconate. Additionally thin films of two perovskite systems, lanthanum cobaltite and lanthanum ferrite, applicable for e.g. gas sensors, catalysts as well as SOFCs, have been prepared.

The two first scientific papers presented here deals with yttria stabilized zirconia. In Paper I the synthesis of nanocrystalline YSZ powders (ZrO2 with 8 mol% Y2O3) using smoldering combustion synthesis with glycine as fuel and nitrate as oxidizer is reported. The influence of glycine to nitrate ratio (G/N) was investigated along with the dependence of calcination temperature with regards to the crystallite size, surface area and content of residual carbonates in the prepared powders. A G/N ratio of 0.23 and calcination in the range above 600ºC in oxygen flow gave high quality powders with a crystallite size less than 10 nm. The removal of residual carbonates by calcination above 600°C led to increased sinterability.

In Paper II, three different sintering techniques have been used to prepare dense YSZ materials from high-quality powder described in Paper I; spark

2

plasma sintering, hot pressing and conventional sintering. The spark plasma sintering technique was shown to be superior to the other methods giving dense materials (≥ 96%) with uniform morphology at lower temperatures (1100°C) and shorter sintering time (5 min). The lowest obtained grain size of dense specimens was 0.21, 0.37 and 12 µm using spark plasma sintering, hot pressing and conventional sintering, respectively. The total electrical conductivity of the YSZ materials, measured by the van der Pauw technique, showed no clear correlation with the grain size. The same for the hardness, measured by the Vickers indentation method. The only effect of grain size was found for the fracture toughness, where a small decrease with increasing grain size was observed.

Preparation and characterization of SrCeO3 and BaZrO3 proton conducting oxide materials are reported in Paper III and IV. In Paper III the preparation of pure and 5% ytterbium substituted strontium cerate (SrCeO3 / SrCe0.95Yb0.05O3-δ) by spray pyrolysis of nitrate salt solutions is presented. Secondary phases detected in the as-synthesized powders were removed after calcination in nitrogen atmosphere at 1100°C (SrCeO3) and 1200°C (SrCe0.95Yb0.05O3-δ). Ball milling of calcined powders in iso-propanol resulted in particle size down to 0.06 µm. Dense SrCeO3 and SrCe0.95Yb0.05O3-δ materials with homogenous microstructure were obtained by sintering at 1350 - 1400°C in air. The grain size of the sintered specimens was in the range 6 - 10 µm for SrCeO3 and 1 - 2 µm for SrCe0.95Yb0.05O3-δ. The electrical properties of SrCe0.95Yb0.05O3-δ were in good agreement with reported data, showing mixed ionic-electronic conduction. The ionic contribution was dominated by protons below 1000°C and the proton conductivity reached a maximum of ~0.005 S/cm above 900°C. In oxidizing atmosphere, the p-type electronic conduction was dominating above ~700°C, while the contribution from n-type electronic conduction in reducing atmosphere only was significant above ~1000°C.

Preparation of pure and 10% yttrium substituted barium zirconate (BaZrO3 / BaZr0.9Y0.1O2.95) powders using the same procedure as for strontium cerate is presented in Paper IV. The crystalline powders were calcined at 1000°C to remove secondary phases and agglomerates were effectively broken down by ball milling giving particle size in the range 0.09 – 0.17 µm. Despite of similar characteristics for the two powders, the densification properties were poorer for the yttrium substituted material. Pressure less sintering of BaZrO3 at 1600°C resulted in severe grain growth (up to 18 µm) and low density (< 92%). High density (~98%) was obtained by hot pressing of both BaZrO3 and BaZr0.9Y0.1O2.95 materials. This sintering technique was shown to reduce the mobility of grain boundaries, giving

3

homogenous microstructures with average grain size down to 0.42 µm. Excess of ZrO2 was observed in some hot pressed specimens and may be due to high sintering temperatures (>1500°C) and reducing atmosphere causing evaporation of BaO(g) and Ba(g). The electrical properties of the BaZrO3 and BaZr0.9Y0.1O2.95 materials were in agreement with the literature, showing high grain boundary resistance for both materials. Slightly lower conductivity was observed for BaZrO3 in wet compared to dry atmosphere. The activation energy for bulk conductivity was higher for BaZrO3 (~100 kJ/mol) than for the acceptor substituted material (~30 kJ/mol) confirming charge compensation of the protons due to substitution.

Preparation and characterization of thin films of LaFeO3 and LaCoO3 are reported in Paper V and VI. Paper V shows how LaFeO3 precursor solution with good spinning properties was prepared using nitrate salts dissolved in methanol added acetic acid with acetyl acetone as chelating agent. Thin films were prepared by spin coating on (100)- and (110)-oriented single crystalline NdGaO3 substrates as well as (0001)-oriented single crystal Al2O3 substrates. By controlled heat treatment up to 400°C with heating rate of 0.1°C/min, homogenous continuous amorphous films were obtained on all substrates. Annealing for 1 h at 500 – 1000°C caused the formation of oriented polycrystalline thin films on NdGaO3 with both (100)- and (110)-orientation. Films prepared on sapphire were polycrystalline, randomly oriented and inhomogeneous after annealing at 700°C. The LaFeO3 films grown on the NdGaO3 substrates crystallized between 400 and 500°C and the average grain size increased from 40 to 250 nm as the temperature was increased from 500 to 700°C.

Paper VI presents the synthesis of precursor solutions for preparation of LaCoO3 and La0.8Ca0.2CoO3 films by three different sol-gel routes. Precursor solutions were synthesized from acetates, alkoxides or nitrate salts of the respective cations (La, Co and Ca), using methanol or 2-methoxyethanol as solvents and suitable chelating agents. Films were prepared on (100)-oriented single crystal yttria stabilized zirconia (YSZ) and (0001)-oriented single crystal Al2O3 substrates by dip-coating or spin-coating. Oriented polycrystalline films of LaCoO3 and La0.8Ca0.2CoO3 were grown on the cubic (100)-oriented YSZ substrates. These films were (104)-oriented when indexing according to the hexagonal structure, corresponding to (110)-orientation in the cubic structure. The orientation was independent of both type of precursor solution and deposition technique. Crystallization of the films started between 800 and 900°C. Controlled heating with heating rate of 0.1°C/min to 400°C resulted in smooth homogenous surface morphology of the prepared films. Smooth surfaces were maintained by heating of these films up to 900°C. Heating to 1000°C was assisted with a

4

severe anisotropic grain growth, giving increased surface roughness. Films prepared on hexagonal (0001)-oriented sapphire substrates were polycrystalline and randomly oriented.

5

2 BACKGROUND AND MOTIVATION

In the following introduction to the present work the main focus has been put on oxide ceramics with characteristic electrical properties. To give the reader an idea of the relevance of this work for technological applications a brief overview of the different types of conducting oxide materials is presented, along with their potential application fields. As a part of the work has dealt with thin films of mixed ionic and electronic conducting oxides, an introduction to thin film technology, and relevant applications is also given. This chapter concludes with the aim of the present work.

2.1 Ionic conducting oxides and their applications

Among the many types of ionic conductors the ones included in the present work are; oxide materials exhibiting high oxygen ion conductivity, proton conductivity or mixed electronic and ionic conduction, where the ionic contribution is from oxygen ions and/or protons.

2.1.1 Oxygen ion conductors

Solid oxide materials that exhibit oxygen ion conductivity are mainly used as oxygen conducting electrodes for solid oxide fuel cells (SOFCs), electrolysers, oxygen pumps and amperometric oxygen monitors [1]. Solid electrolyte materials in such high temperature electrochemical systems are required to exhibit relatively high oxygen ion conductivities. The most well characterized oxygen ion conductors are those based on fluorite structure [2], e.g. zirconia (ZrO2) and ceria (CeO2) properly substituted with aliovalent cations such as yttrium (Y3+) and ytterbium (Yb3+), in order to introduce oxygen vacancies on which the ionic conductivity depends. Other fluorite-related materials such as pyrochlores (A2B2O7) and materials derived from bismuth oxide (δ-Bi2O3) have also been reported to exhibit high oxygen ion conductivity [2]. Oxygen ion conduction is also found in perovskite-type oxides (ABO3), however, such materials are more often associated with mixed ionic and electronic conductivity, when containing mixed valent cations of transition metals [3]. Fig. 1 presents the conductivity in some fluorite-type oxide systems with predominantly oxygen ion conductivity. Although other oxide systems than yttria stabilized zirconia (YSZ, ZrO2 with 8 – 10% Y2O3 incorporated) have been reported to exhibit higher conductivity, YSZ is still commercially used for both oxygen sensors and as oxide electrolyte material in SOFCs. This is due to the high mechanical

6

strength, chemical and thermal stability of YSZ materials compared to other candidate oxide materials, such as acceptor-substituted ceria (CeO2 with 5 – 10% substitution of Ce with e.g. Yb, Sm or Gd).

Fig. 1 Total conductivity (σ) with varying temperature for a selection

of fluorite-type oxide materials with predominantly oxygen ion

conductivity, as presented by Steele [3]

.

Development of materials with dimensions in the nanoscale range has become an important research field due to the often different properties of such materials compared to large scaled bulk materials. This research is relevant also for oxygen ion conductors, as the electrical properties may be

7

related to the microstructure. The properties of the grain boundaries compared to bulk are determining for the ionic conductivity. Higher concentration and mobility of defects in the grain boundaries may lead to enhanced conductivity while impurities such as SiO2 or subsistent segregation on the grain boundaries, will increase the grain boundary resistance compared to bulk [4-6]. Additionally, the ionic conductivity at grain boundaries is related to the effect of space charge regions in the grains adjacent to the grain boundaries [7]. If bulk defects with high mobility, e.g. oxygen vacancies, are accumulated in the space charge region, the grain boundary conductivity should increase. However, oxygen vacancy depletion due to space charge in the vicinity of grain boundaries may increase the grain boundary resistance. Considering these effects, reducing the grain size, and consequently introducing a larger amount of grain boundaries, may increase or decrease the conductivity of a material, depending on which effects are dominating.

2.1.2 Proton conductors

Although polymeric compounds such as hydrated Nafion® exhibit high proton conductivity, comparable to aqueous HCl and liquid H3PO4, these materials are not utilisable at temperatures above ~200°C. For use at intermediate to high temperatures (>500°C) substituted phosphates and oxides are applicable [8]. Fig. 2 demonstrates how the bulk proton conductivity in these materials goes through a maximum at higher temperatures.

Materials with high and pure proton conductivity, including perovskite-type oxides (ABO3) based on barium or strontium zirconates and cerates (A = Ba, Sr, B = Zr, Ce), are candidates for electrolytes in sensors, batteries, fuel cells and electrolysers. The advantage of using proton conducting oxide materials as electrolyte material in SOFCs (see chapter 2.1.4) has been a motivation for the investigation of the proton conductors in the present work.

2.1.3 Mixed ionic/electronic conductors

The group of materials exhibiting high conductivity of both ions and electrons are also applicable to fuel cells, sensors, gas pumps, batteries and electrochemical reactors [9]. Oxides with mixed electronic and oxygen ion conductivity are e.g. perovskite-type materials (ABO3) based on lanthanum (A-site) and manganese, chromium, iron, cobalt (B-site). Additional substitution (on both A- and B-site) alters the electrical properties, and

8

custom designed materials, e.g. for SOFC cathodes, oxygen separation membranes and sensors can be prepared.

Fig. 2 Bulk proton conductivities of various oxide materials as

presented by Kreuer [10]

based on data from Norby & Larring [11]

.

2.1.4 Solid oxide fuel cells

Solid oxide fuel cells (SOFCs) are energy conversion devices that produce electricity directly from a gaseous fuel by electrochemical combination of the fuel with an oxidant [12]. The advantages of SOFCs compared to conventional electric power generation systems are high conversion efficiency and environmental compatibility, which is important in a society striving to reduce the pollution. In the traditional SOFC, oxygen (from air)

9

is reduced by a porous electrode (cathode) producing oxide ions which migrate through a solid electrolyte to the porous fuel electrode (anode) and react with the fuel (H2, CO, CH4) forming H2O and/or CO2. In turn a proton conducting solid electrolyte can be used where H2 is oxidized to produce protons that subsequently react with oxygen to form water. The operation principle in such a proton conductor based fuel cell is schematically shown in Fig. 3. The main advantage of such SOFCs based on proton conductors is that water forms and leaves the system on the side exposed to air, as opposed to the traditional SOFCs based on oxygen ion conductors where water forms on the fuel side, and consequently dilutes the fuel.

Fig. 3 Illustration of how a proton conductor based SOFC operates, by

Norby et al. [13]

.

10

2.1.5 Sensors

The most common type of oxygen sensors based on oxide materials utilizes the oxygen transport through a solid electrolyte (commercially used YSZ). This principle used in the so-called lambda sensors in gasoline engines, illustrated in Fig. 4 a). The inner and outer platinum electrode, pasted onto the YSZ electrolyte, are connected to atmospheric air (or reference gas) and the exhaust gas from the engine, respectively. The different partial pressures of oxygen create a potential which gives a direct indication of the oxygen content in the exhaust gas. Oxygen partial pressure may also be monitored by amperometric devices, as illustrated by Fig. 4 b). Replacing the solid electrolyte (YSZ) in such a device with a pure proton conducting material also allows the monitoring of hydrogen gas.

a) b)

Fig. 4 Schematic illustrations of oxygen sensors based on a solid

electrolyte like YSZ (CCM Technologies [14]

).

2.2 Thin film technology

Thin film technology has become an important research field, which supplements the research on bulk materials. Application of oxide thin films (typical thickness ~100 nm or less) as protective layers on surfaces exploited to rough conditions or improving the substrate properties are two examples on use of such thin films, however, often the film itself is the functional material [15]. The properties may differ from, and in some cases be improved compared to bulk materials. Typical fields of interest for applying oxide thin films include electronics, optics and energy storage. In

11

additionally, and more importantly, for the perovskite-type oxide films included in the present work, is the use as thin electrolyte layers (with improved electrochemical properties) in SOFCs, for gas sensors and as catalysts.

For preparation of thin oxide films several physical techniques may be applied, such as; pulsed laser deposition (PLD), atomic layer epitaxial (ALE), electrochemical oxidation, ion beam sputtering and spray pyrolysis in inductively coupled plasma (spray-ICP). Chemical techniques, like chemical vapour deposition (CVD) and chemical solution deposition (CSD) may also be applied. Using these chemical techniques imply the advantage of good ability to control the chemical composition of thin films of e.g. mixed and substituted metal oxides. CSD techniques, including sol-gel synthesis, chelate and metal-organic decomposition have successfully been applied to prepare perovskite-type thin films with good properties [16]. Relatively simple procedures make CSD techniques more cost efficient and for this reason, more applicable for commercial use. This has been the motivation for using sol-gel techniques and simple film deposition methods (dip-coating and spin-coating) in the present study of preparation of oxide films.

2.3 Aim of work

The overall aim of the work presented in this thesis has been to develop complete synthesis routes to bulk materials and thin films of ceramic oxides with designed microstructure and nanostructure, and to perform a characterization of selected properties in these materials. The work can be divided in three main subjects; synthesis and characterization of oxygen ion conductors, proton conductors and oxide thin film preparation.

In the first part based on oxygen conductors the focus has been put on yttria stabilized zirconia (YSZ) materials, commonly used in oxygen sensing devices and solid oxide fuel cells. The most relevant literature reviewed in the next chapter forms the basis from which the aim of presented work on YSZ origins. Though there are many reports on powder synthesis, densification and characterization of YSZ materials, not many of them present complete studies with all parts included. The aim of the present work on YSZ has been to obtain routes for synthesis of nanocrystalline powder and densification of fine bulk materials with well defined micro- and nanostructure. In addition to thorough microstructure studies of prepared powders and bulk materials, it has been in our interest to study the

12

electrical and mechanical properties of YSZ and relate these to the micro-/nanostructure.

The second part on proton conductors focuses on two perovskite-type oxide systems, barium ziconate (BaZrO3) and strontium cerate (SrCeO3) substituted with yttrium (Y) and ytterbium (Yb) respectively. Although well studied by others, the synthesis and densification of these materials are reported to be challenging. The main aim of this work has been to obtain control over the microstructure development in the materials during densification. By use of synthesis routes and densification techniques still not reported for these materials, we have aimed to present a study of both BaZrO3 and SrCeO3 materials, which should be of importance for others working on these systems. Furthermore, it has been in our interest to study the electrical properties of both systems, focusing on proton conductivity. The electrical properties of prepared proton conductors were investigated at the University of Oslo. Results from these studies are presented here, however, more elaborate discussions are left to forthcoming publications.

The last part of this work has been somewhat different as it deals with films rather than bulk materials. However, as for the work on proton conductors, two perovskite-type oxides known to exhibit mixed ionic and electronic conductivity have been studied; lanthanum cobaltite (LaCoO3) and lanthanum ferrite (LaFeO3). The overall aim of this part of the presented work has been to obtain fundamental understanding and experience in the preparation of oxide thin films based on these materials. The reported literature on thin films from these systems is limited, in particular on films prepared by chemical routes. With this in mind, our goal has been to contribute to this research field by presenting simple chemical synthesis routes and methods for preparing homogenous films with designed microstructure/ surface morphology. Due to the superior properties of epitaxially grown films compared to randomly oriented polycrystalline films, it has been of special interest to study preferential growth orientation on various substrate materials.

13

3 REVIEW OF EXISTING LITERATURE

The disposition of this chapter based on the materials investigated in the present work, includes three main parts; oxygen ion conductors represented by YSZ, proton conductors involving SrCeO3 and BaZrO3 materials and, finally, thin films where relevant work on LaFeO3 and LaCoO3 films is reviewed.

3.1 Oxygen ion conductors - Zirconia

Intense investigations of the science and technology of zirconia (ZrO2) over the last half century have propelled it into an outstanding, versatile material. Zirconia is a remarkable material and a case study in materials science, since structure-property correlations have been extensively examined. Atomic structure and microstructure, defects, phase transformations, and processing on one hand and properties (thermal, mechanical, electrical and optical) on the other, are intimately connected. [17,18]

3.1.1 Crystal structure

Zirconia, having monoclinic crystal structure at ambient temperature, exhibits a phase transition to tetragonal at ~1170°C, further to cubic at ~2370°C [19]. The cubic phase, which is stable from ~2370°C to the melting point (2680 ±15°C), has a fluorite-type crystal structure, illustrated in Fig. 5, in which each zirconium ion is coordinated by eight equidistant oxygen ions and each oxygen is tetrahedrally coordinated by four zirconium.

3.1.2 Stabilized zirconia

Proper substitution with certain aliovalent oxides stabilizes the cubic fluorite structure of ZrO2 from ambient temperature to the melting point [20]. The cubic phase can exist in a wide range of compositions and temperatures. Fig. 6 presents the phase relations in the ZrO2 – Y2O3 system as a function of temperature. In substituted zirconia the aliovalent guest cations reside on the sites for host cations (Zr4+), generating oxygen vacancies for charge neutrality, as demonstrated by the defect equation (Eq. 1) for substitution with Y2O3.

xOOZr

ZrO32 O3vY2OY 2 ++′ →

•• (1)

14

The oxygen vacancies lead to substantial ionic conductivity over an extended oxygen partial pressure range, where the electronic conductivity is negligible [21].

Fig. 5 The cubic fluorite structure of zirconia.

3.1.3 Ionic conductivity in YSZ

The concentration of vacancies is of importance for the ionic conductivity, hence the degree of substitution is as well. For YSZ the maximum level of ionic conductivity has been found for 8 to 10% Y2O3 incorporated into the ZrO2 lattice. Beyond this optimal level the conductivity begins to decrease, as indicated in Fig. 7. At the optimal level the conductivity may be increased by raising the operating temperature to enhance the mobility of the defects. This is why most SOFCs become efficient at operating temperatures around 1000°C. Additionally the microstructure / grain size is determining for the conductivity in the material.

15

Fig. 6 Phase diagram for the ZrO2 – Y2O3 system as presented by

Stubican et al. [20]

.

The bulk resistivity of polycrystalline ZrO2 is equivalent to that of a single crystal and it is usually not affected by sintering temperature, atmosphere and heat treatment. Efforts to reduce the bulk resistivity by changing composition (adding aliovalent oxides) are now nearly exhausted and the maximum stabilizer concentrations (e.g. 8 - 10 mol% for substitution with Y2O3) have been established [22]. In order to increase the overall ionic conductivity it has therefore been natural to look for ways to decrease the grain boundary resistivity which is influenced by the impurity level, sintering temperature, atmosphere, heat treatment, etc.

16

Fig. 7 Dependence of conductivity on composition for ZrO2-based

solid solutions at 800°C [23].

A brief review of conductivity data for YSZ bulk materials presented in the literature in addition to density and grain size (when available), is given in Table 1, along with comparable data for YSZ thin films. As YSZ shows mainly ionic conductivity [24], oxygen ion conductivity has been evaluated equal to reported total conductivities. Differences in bulk and grain boundary conductivity (listed in Table 1) have not been considered in the following and some of the listed conductivity values have been extrapolated from given data, assuming Arrhenius behavior with the total conductivity, σ, given by Eq. 2.

−

=σkT

EexpAT a (2)

A is a constant, Ea is the activation energy and k is the Boltzmann constant.

17

Table 1 Reported total conductivity, σ, at 1000°C and activation energy,

Ea, in YSZ bulk materials prepared by different sintering

techniques (CS: conventional, MS: microwave and SPS: spark-

plasma sintering). D notes average grain size. Data for YSZ

thin films (TF) are added for comparison.

Mol%

Y2O3

Technique /

T (°C)

Density

(%)

D

(µm)

σ

(mS/cm)

Ea

(kJ/mol)

Reference

#

8 SPS / 1400 99 10 160 [25]

8 SPS / 1300 97 < 1 190 92 [26]

8 SPS / 1200 95 0.15 183 109 [27]

8 SPS / 1100 96 0.21 172 102 [28]

8 SPS / 1000 91 120 [25]

8 CS / 1600 99 > 5 180 92 [26]

8 CS / 1500 98 12 137 94 [28]

8 CS / 1500 99 12 156 90 b [29]

8.5 CS / 1500 97 9.5 149 107 b [29]

7.7 CS / 1500 97 8.5 160 104 b [29]

8 CS / 1500 99 169 [30]

9 CS / 1400 93 < 1 181 b 113

b [31]

8 CS / 1350 98 3.5 107 [32]

8 CS / 1150 87 1.2 85 [32]

8 MS / 1500 99 172 [30]

8 MS / 1200 97 158 [30]

8.7 TF / 1000 - 0.02 334 90 [35]

b: bulk conductivity / activation energy

From the reviewed data in Table 1, no clear conclusions can be made for the influence of grain size on the conductivity in bulk materials. Factors such as sintering technique and temperature must also be considered. There are reports of correlation between grain size and grain boundary conductivity in zirconia bulk materials, as clearly demonstrated for calcia-substituted zirconia in Fig. 8 [33].

Many studies have been carried out on conductivity of YSZ thin films, where nanoscaled grains are more easily obtained compared to bulk materials. There is strong support for the notion that the energetics for defect formation may be substantially reduced in nanocrystalline oxides [34]. Enhanced grain boundary conductivity with decreasing grain size is reported for YSZ films [35,36], similar to the correlation between grain size and

18

conductivity shown in Fig. 8. This effect would be due to greater grain boundary diffusion than corresponding “bulk” diffusion because of high defect densities and high mobilities. An additional explanation for enhanced conduction at grain boundaries is related to the formation of space charge regions in the grain adjacent to the boundaries [37]. Enhancement of ionic conductivity in nanocrystalline oxides remains unresolved due to conflicting reports and inadequate efforts to isolate the ionic from the total conductivity. No clear lines can be drawn between the grain size effects observed in thin films and bulk materials, possibly because reduced grain size in bulk materials is more challenging than for thin films, where the sintering temperature can be lowered.

Fig. 8 Grain boundary conductivity as a function of grain size for ZrO2

stabilized with 15 mol% CaO [33]

.

3.1.4 Powder synthesis

Nanosized powders are essential as a starting point for making nanocrystalline ceramics. According to conventional sintering theory, the sinterability increases with decreasing crystallite size in the starting powder [38]. There are various methods for synthesis of high quality oxide powders; among them are precipitation techniques [39,40], combustion techniques [41-43],

19

sol-gel techniques [44,45], and hydrothermal techniques [46]. All these different techniques are based on a solution type chemistry, where precursors of the various cations are dissolved in a solvent, commonly water, and then mixed in appropriate proportions.

Oxide ceramic powders prepared by a glycine-nitrate combustion technique, as described by Chick et al. [47], are superior to powders prepared by e.g. sol-gel synthesis, showing greater compositional uniformity, lower residual carbon level and smaller particle size. In this synthesis technique glycine complexes the cations and may do so on both the carboxylic acid end or by the amino group depending on the size and charge of the cation as well as the pH in the solution. The amount of glycine added must be sufficient to retain the cations completely complexed as the solution is being evaporated. In addition to being a complexing agent, the glycine works as a fuel and will, in the presence of nitrates from the metal nitrate precursor, working as an oxidizer, cause a spontaneous ignition reaction upon evaporation. The temperature of the ignition reaction depends on the fuel-to-oxidizer ratio (glycine:nitrate (G/N) ratio) and which metal cations are introduced into the solution.

Mukasyan et al. [48] have defined three combustion modes depending on the G/N ratio for La0.8Sr0.2CrO3; Smoldering combustion synthesis (SCS): G/N < 0.39, T < 600ºC; Volume combustion synthesis (VCS): 0.39 < G/N < 0.66, 1150ºC < T < 1350ºC; Self-propagating high-temperature synthesis (SHS): 0.66 < G/N < 0.88, 800ºC < T < 1100ºC. According to Chick et al. [47], the best powder quality is achieved when the combustion temperature is the highest. For La0.8Sr0.2CrO3 the highest combustion temperature was observed for a G/N ratio in the range of 0.4 – 0.7 however, Kaus et al. has indicated that the maximum crystallite size and combustion temperature in the synthesis of nanocrystalline zirconia occurred for a G/N ratio of 0.30 [49]. This report also states that the best powders were obtained for a lower G/N ration (0.23). As the reaction temperature affects the crystallite size in the final powder it is desirable to keep this at a minimum in order to prepare finer powders.

3.1.5 Densification and microstructure

Densification of zirconia materials is widely investigated [26,28-32] and it is possible to obtain dense (> 95%) specimens by conventional sintering at temperatures in the range 1300 – 1600°C. Such sintering with long sintering time (hrs), in particular in the higher end of this temperature range will result in rather large grained materials. Larger grains may be undesirable, for example with respect to mechanical and electrical

20

properties. Other sintering techniques such as spark plasma sintering (SPS) have become more popular for controlling the grain growth during sintering. This is also indicated by the density/grain size data reviewed in Table 1, showing average grain size for YSZ materials prepared by SPS in the sub-micron range. In an SPS experiment both sintering time and temperature needed to obtain dense materials can be reduced significantly compared to conventional sintering. This can be ascribed to the effective heat transfer, as the pressing die (and green body if conducting) works as the heating element. The applied pulsed electrical field may create spark discharges during the initial part of the sintering, and clean the particle surface, facilitate grain boundary diffusion and possibly contribute to increased densification rate [50].

3.2 Proton conductors - Barium zirconate and

strontium cerate

The research on proton conducting ceramics accelerated after Iwaharas discovery of relatively high proton conductivity in acceptor doped strontium cerate (SrCeO3) at high temperatures, in the presence of water vapor or hydrogen [51]. SrCeO3 materials, along with barium zirconates (BaZrO3) are reviewed here. As these materials are often compared, due to competing properties, it has been advantageous to present their properties together in the following.

3.2.1 Crystal structure

Both BaZrO3 and SrCeO3 exhibit perovskite-type (ABO3 structures) crystal structures. BaZrO3 exists in the cubic perovskite-structure, illustrated by Fig. 8 a), while SrCeO3 has a distorted version of this structure, making the material orthorhombic, as shown in Fig. 8 b).

3.2.2 Electrical properties

The available conductivity data for acceptor substituted SrCeO3 [51-65,69],

generally proves mixed ionic-electronic conduction. No major deviations from the results of the extensive work on SrCe2.95Yb0.05O3-δ, performed by Iwahara et al. [51-53,55,60-64,69], have been reported, and the total conductivity at 1000°C for this material ranges from 0.05 to 0.005 S/cm, depending on atmosphere. Kosacki and Tuller reported that the electrical properties of SrCe2.95Yb0.05O3-δ between 800 and 1000°C are dominated by holes and electrons at high p(O2) (≥ 1 atm) and low p(O2) (< 10-20 atm), respectively

21

[57]. The proton conductivity in humid air reaches a maximum of ~0.004 S/cm around 900°C [53]. The proton transport number under such conditions exceeds 0.5 below 800°C, however increases with decreasing p(O2). At 700°C and p(O2) ~5·10-4 atm, the ionic contribution to the conductivity is predominated by protons [60].

Fig. 8 a) Cubic and b) orthorhombic perovskite structure. Central

atom represent A-site cation, while the B-site cations take

octahedral coordination with oxygen.

The electrical properties of yttrium substituted BaZrO3 are well documented in the literature [10,11,66-73], and some central data are listed in Table 2. Total bulk (ionic + electronic) conductivity up to 0.01 S/cm at 800°C in wet O2, is reported by Bohn and Schober [71]. While hole conduction suppresses the transport number for protons in oxidizing atmosphere, BaZr0.9Y0.1O2.95 is close to a perfect proton conductor under reducing conditions [71]. The main issue regarding ionic conductivity in BaZrO3 materials is the high grain boundary resistance, inhibiting proton transport, however this is not yet fully understood. For 10% yttrium substituted BaCeO3, the total conductivity is about one order of magnitude higher than the corresponding

BaZrO3 material; σtot = 0.7 and 0.06 S/cm, respectively, at 1000°C in wet hydrogen atmosphere [75]. The idea of combining the improved electrical properties of BaCeO3 with the mechanical properties and chemical stability of BaZrO3 has been studied in mixed systems (BaCe0.9-xZrx Y0.1O3-δ). For these systems increasing chemical stability against CO2 was reported with increasing Zr content, however, this was accompanied by decreased proton conductivity [74,75]. Snijkers et al. has reported improved conductivity for BaZr0.9Y0.1O2.95 specimens prepared with excess of BaO, in the range 200 -

22

300°C, when extrapolating the data to ~500°C the data equals those of specimens prepared without excess of BaO (~8·10-4 S/cm) [72].

Table 2 Reported conductivity data for BaZr0.9Y0.1O2.95 for 500°C (bulk)

and 600°C (total) under various conditions. Water partial

pressure, p(H2O), is indicated when available.

Sintering

temperature

(°C)

Density

(%)

σtotal at

600°C

(S/cm)

σbulk at

500°C

(S/cm)

Atmosphere Ref. #

1715°C 97 2·10-3 Wet synthetic

air [68]

1800°C 95 4·10-3 Wet air

p(H2O) = 1.7·103 Pa

[66]

1800°C 95 1.5·10-3 Wet H2

p(H2O) = 1.7·103 Pa

[66]

1700°C 95 ~10-3 * 7·10-4 * Ambient air [72]

1715°C 97 7·10-4 Wet Ar +

4%H2 [68]

1715°C 97 4·10-3 Wet H2 and wet Ar/O2

[71]

1700°C 91 5·10-2 ** Wet

p(H2O) = 2.3·103 Pa

[10,73]

* Extrapolated values ** BaZr0.8Y0.2O2.9

3.2.3 Powder synthesis and densification

Most reports concerning preparation of ceramic BaZrO3 and SrCeO3 powders have been by solid state ceramic method at temperatures in the range 1000 – 1450°C [10,11,55,57,66-73,76-83,94,97]. This method generally produces coarse powders not suitable for preparation of bulk materials with custom designed fine grained microstructures. Fine BaZrO3 powders with

23

more defined morphology have been prepared by co-precipitation [84,85], sol-gel techniques [86,87] and hydrothermal synthesis [87]. Other reported synthesis routes for preparation of SrCeO3 powders include complexation with EDTA [88] or citric acid [89], as well as with glycine in a combustion synthesis route [90]. Powders prepared by complexation with EDTA appear phase pure after calcination at 1200°C in inert atmosphere (He).

No thorough sintering studies of BaZrO3 and SrCeO3 ceramics seem to be available. Dense (> 95%) BaZrO3 from powders made by solid state method can only be obtained by conventional sintering above 1700°C [67,68,71,72,75]. Using spark plasma sintering, dense (>95%) ceramics of pure and Y-substituted BaZrO3 can be obtained at 1500 and 1600°C respectively [91]. Reported densities of SrCeO3 and SrCe2.95Yb0.05O3-δ materials prepared from powders by the solid state ceramic method sintered in air at 1350°C for 96 hrs, are >85 and >75%, respectively [90]. Generally, sintering temperature in the range of 1500 to 1650°C [57,79,83] in air is needed to obtain materials with >95% density when prepared from such powders. Higher densities (>97%) have been obtained from commercial powders by sintering at 1450°C in air for 2 hrs [92] and from powder prepared by a complexation route by sintering at 1300°C for 12 hrs in nitrogen atmosphere [88], confirming the importance of starting powder properties.

3.2.4 Mechanical properties and chemical stability

Chemical, mechanical and electrical properties are all of importance when considering proton conducting materials for commercial use. Among the perovskite-type zirconates and cerates of barium and strontium, there has been a continuous battle arising from higher conductivity in the cerates versus improved mechanical properties and chemical stability in the zirconates [93,94]. Young’s modulus (E) and Vickers hardness (HV) for SrCeO3 (substituted with 5% Yb) are reported to be 145 GPa and 5.5 GPa

[95]. The mechanical properties of BaZrO3 are improved compared to those of SrCeO3 with reported values of E = 181 GPa and HV = 11.1 GPa [96]. It could be mentioned that Hassan reported lower bending strength, hardness and fracture toughness for SrZrO3 compared to SrCeO3

[95], in contrast to the trend with improved mechanical properties of zirconates compared to cerates. The density (crucial for the mechanical properties) for the presented SrZrO3 material appears to be lower than for SrCeO3.

The chemical instability of SrCeO3 and BaCeO3 materials when exposed to CO2 atmosphere is well documented by Scholten et al. [94]. In an extensive review, Kreuer has compared the equilibrium reactions for barium and

24

strontium cerates and zirconates in both CO2 and H2O atmosphere, as demonstrated by Eq. 3 and 4, respectively.

3 2 3 2ABO (s) CO (g) ACO (s) BO (s)+ = + (3)

3 2 2 2ABO (s) H O(g) A(OH) (s) BO (s)+ = + (4)

Additionally the decomposition of the resulting carbonates (ACO3) and hydroxides (A(OH)2) according to Eq. 5 and 6 have been considered.

3 2ACO (s) AO(s) CO (g)= + (5)

2 2A(OH) (s) AO(s) H O(g)= + (6)

The equilibrium constants for these reactions (Eq. 3 – 6), as a function of partial pressures of CO2 and H2O are presented for SrCeO3 and BaZrO3 in Fig. 9 a) and b).

In the case of SrCeO3 the equilibrium constants for Eq. 3 and 5, as a function of p(CO2), are close to identical and both equal unity around 700°C in air. Hence, heat treatment of SrCeO3 above 700°C in air is more likely to result in the formation of strontium oxide (SrO) rather than SrCO3. The onset temperature of decomposition of SrCO3 (according to Eq. 5) obtained by thermogravimetric analysis have been reported to 845°C in 1 atm N2, and 1220 – 1275°C in 1 atm CO2. The corresponding onset temperatures for synthesis of SrCeO3 from a mixture of SrCO3 and CeO2, are 800°C and 1168 - 1190°C, in 1 atm N2 and CO2, respectively [94,97]. At lower temperatures, carbonate formation may be critical, with respect to phase purity and mechanical properties of SrCeO3 materials. Reactions with water should also be considered when operating in atmospheres with high p(H2O) [93]. BaZrO3 is far more stable in CO2 atmosphere, compared to SrCeO3, and reaction in Eq. 4 is not significant for water vapour pressure < 10 bar.

For BaZrO3 the high sintering temperatures (>1500°C) needed for obtaining dense materials is more crucial than the stability against CO2 / H2O, as the material tend to decompose, according to Eq. 7, as documented by thermodynamic studies [98,99].

3 2BaZrO (s) ZrO (s) BaO(g)= + (7)

25

a)

b)

Fig. 9 Stability diagrams for SrCeO3 and BaZrO3 in CO2 or H2O

atmosphere, as presented by Kreuer [93]

.

26

Heat treatment of Y-stabilized BaZrO3 bulk materials above 1500°C have resulted in reduced lattice parameter and yttria/zirconia excess, indicating evaporation of BaO(g) above this temperature [86]. On the surface of these BaZrO3 specimens, similar effects were observed at lower temperatures (T > 1250°C) [86]. In reducing atmosphere in contact with graphite during hot pressing or spark plasma sintering, evaporation of Ba(g), according to Eq. 8, must also be considered.

3 2BaZrO (s) C(s) ZrO (s) CO(g) Ba(g)+ = + + (8)

3.3 Thin films of perovskite-type oxides

3.3.1 Lanthanum ferrite thin films

Lanthanum ferrite (LaFeO3) has an orthorhombic distortion of the cubic perovskite structure, comparable to that of SrCeO3, as shown in Fig. 8 b). LaFeO3 is an antiferromagnetic insulator at room temperature, making the material feasible for use in magnetic sensors and as read heads in computer hard drives [100]. At elevated temperatures LaFeO3 exhibits mixed ionic-electronic conductivity [101] and a linear response for log σ versus log p(O2) have been reported for LaFeO3 thin films (at 1000°C), making the material promising for oxygen sensor applications [102]. Furthermore, LaFeO3 have been reported to show excellent sensitivity towards non-flammable gases such as CO [103] and NOx

[103-107] and exhibits selective sensitivity towards flammable gases such as ethanol [108], methane [103] and volatile sulfides such as CH3SH [109]. LaFeO3 thick films are reported to be promising for detection of γ-radiation [110]. The gas sensitivity has been reported to increase with decreasing grain size [111] making the morphology an important feature in gas sensing films. In the literature, the preparation of LaFeO3 films is reported by use of sputtering techniques [102,104,112-115], screen printing of slurries from nanosized powder [105,106,110,116] and electrochemical reduction [117]. Only sputtering techniques have resulted in growth of epitaxial LaFeO3 thin films, and rhombohedral (001)-oriented LaAlO3 substrates [102,112] and cubic (001)-oriented MgO substrates [113] have been used for this purpose. This observation is not unexpected as sputtering techniques allows good control of deposition of thin film layers, making for epitaxial orientation, compared to thicker films (> 0.5 µm) prepared by screen printing and electrochemical reduction. Randomly oriented LaFeO3 films have also been prepared on quartz substrates by CSD, using a colloidal sol of hydrous LaFeO3 particles with an average particle size of 7 nm and

27

dip-coating technique [118]. These films were shown to be fully crystalline and single phase after annealing at 650°C.

3.3.2 Lanthanum cobaltite thin films

Lanthanum cobaltite, LaCoO3, has a rhombohedral distorted version of the perovskite structure (ABO3). At elevated temperatures (above ~1400°C) the material undergoes a phase transition from rhombohedral to cubic [119], and interesting ferroelastic behavior is associated with this phase transition [120-

122]. LaCoO3 materials exhibit characteristic changes in magnetic and electrical properties with varying temperature [123]. The properties of LaCoO3 may be adjusted by substitution with e.g. calcium, strontium or barium on the A-site [120-122,124], which as an example will lower the rhombohedral to cubic phase transition temperature (materials become cubic at ambient temperature for ~50 mol% substitution [119]. Mixed ionic/electronic conductivity at elevated temperatures, makes LaCoO3-materials applicable in membrane technology and as cathode materials in solid oxide fuel cells (SOFC’s) [125]. LaCoO3 materials are also investigated with respect to catalytic properties [126] and for use in gas sensors [127,128].

The literature reveals thin films of LaCoO3 (pure and Ca-/Sr-substituted) prepared by different techniques, such as pulsed laser deposition (PLD) [128], chemical vapour deposition (CVD) [129], atomic layer epitaxy (ALE) [130], electrochemical oxidation [131], ion beam sputtering [132] and by spray pyrolysis in inductively coupled plasma (spray-ICP) [133]. LaCoO3 films have also been prepared from various sol-gel precursor solutions by spin- or dip-coating [134-143]. Sol-gel precursor solutions have been synthesized by traditionally accepted routes using metal alkoxides dissolved in alcohol with proper chelating agents, such as 2-ethylacetoacetate or polyethylene glycol (PEG) [134-136]. Less expensive variants using nitrate salts or acetates as starting materials for alcohol-based precursor solutions, with e.g. butylacetate or polyvinyl alcohol (PVA) as chelating agents, have also been reported [137-141]. LaCoO3 films have also been prepared from water based routes using ethylendiaminetetraacetic acid (EDTA) or diethylenetriamineoentaacetic acid (DTPA) as complexing agents [141-144]. None of these so-called chemical solution deposition (CSD) techniques have resulted in epitaxial growth of the films. Due to the often superior properties of crystalline epitaxial films, these are more attractive for applications and fundamental studies, compared to randomly oriented polycrystalline films [145].

28

References

[1] B.C.H. Steel, Oxygen ion conductors and their technological

applications, Mater. Science and Engineering, B13, 79-87 (1992) [2] J. A. Kilner, Fast oxygen transport in acceptor doped oxides, Solid

State Ionics , 129 [1-4] , 13-23 (2000) [3] B.C.H. Steele, Oxygen ion conductors, in High Conductivity Solid

Ionic Conductors, Recent Trends and Applications (T. Takahashi, ed.), World Scientific, Singapore, 402-446 (1989)

[4] J.E. Bauerle, Study of solid electrolyte polarization by a complex

admittance method, J. Phys. Chem. Solids, 30, 2657-2670 (1969) [5] R. Gerhardt and A. S. Nowick, Grain-boundary effect in ceria doped

with trivalent cations. 1. Electrical measurements, J. Am. Ceram. Soc., 69 [9], 641-646 (1986)

[6] D. Y. Wang and A. S. Nowick, The grain-boundary effect in doped

ceria solid electrolytes, J. Solid State Chem., 35 [3] (1980) 325-333. [7] J. Maier, Ionic-conduction in-space charge regions, Prog. Sol. State

Chem., 23 [3], 171-263 (1995) [8] T. Norby, Solid-state protonic conductors: principles, properties,

progress and prospects, Solid State Ionics, 125, 1-10 (1999) [9] I. Riess, D.S. Tannhauser, Mixed Ionic Electronic Conductors, in

High Conductivity Solid Ionic Conductors, Recent Trends and Applications (T. Takahashi, ed.), World Scientific, Singapore, 402-446 (1989)

[10] K.D. Kreuer, Proton-conducting oxides, Annu. Rev. Mater. Res., 33, 333-359 (2003)

[11] T. Norby, Y. Larring, Concentration and transport of protons in

oxides, Curr. Opin. Solid State & Mater. Sci., 2, 593-599 (1997) [12] O. Yamamoto, Solid oxide fuel cells: fundamental aspects and

prospects, Electrochimica Acta, 45 [15-16], 2423-2435 (2000) [13] T. Norby, R. Haugsrud, N. Vajeeston, New Proton Conducting

Materials for Fuel Cells and Hydrogen Separation Membranes, Proc. 9th Int. Conf. Inorg. Membr., Lillehammer – Norway, 260-267 (2006)

[14] http://rtreport.ksc.nasa.gov/techreports/2003report/100/106.html [15] M. Gelfi, E. Bontempi, R. Roberti, L. Armelao, L.E. Depero,

Residual stress analysis of thin films and coatings through XRD2

experiments, Thin Solid Films, 450, 143-147 (2004) [16] R.W. Schwartz, Chemical solution deposition of perovskite thin

films, Chem. Mater., 9, 2325-2340 (1997) [17] E. C. Subbarao, Solid electrolytes and their applications, ed. by E.C.

Subbarao, Plenum Press (1980)

29

[18] D.W. Richerson, Modern Ceramic Engineering, Marcel Dekker Inc., 2nd ed. (1992)

[19] E.C. Subbarao, Zirconia - an overview, Science and Technology of Zirconia, 3, 1–23 (1981)

[20] V.S. Stubican, R.C. Hink, S.P. Ray, Phase equilibria and ordering in

the system ZrO2 – Y2O3, J. Am. Ceram. Soc., 61 [1-2], 17-21 (1978) [21] N. Q. Minh and T. Takahashi, Science and Technology of Ceramic

Fuel Cells, Elsevier Science B. V., Amsterdam, 70-92 (1995) [22] X. Guo, R.-Z. Yuan, On the grain boundaries of ZrO2-based solid

electrolyte, Solid State Ionics, 80, 159-166 (1995) [23] T. Takahashi, Physics of Electrolytes, Vol. 2, (J. Hladic ed.),

Academic Press, London, 980-1049 (1972) [24] A. Weyl, D. Janke, High-temperature ionic conduction in

multicomponent solid oxide solutions based on zirconia, J. Am. Ceram. Soc., 80 [4], 861-873 (1997)

[25] X.J. Chen, K.A. Khor, S.H. Chan, L.G. Yu, Preparation yttria-

stabilized zirconia electrolyte by spark-plasma sintering, Mater. Sci. Eng., A341, 43-48 (2003)

[26] T. Takeuchi, I. Kondoh, N. Tamari, N. Balakrishnan, K. Nomura, H. Kageyama, Y. Takeda, Improvement of mechanical strength of 8

mol% yttria-stabilized zirconia ceramics by spark-plasma sintering, J. Electrochem. Soc., 149 [4], A455-A461 (2002)

[27] U. Anselmi-Tamburini, Spark plasma sintering and characterization

of bulk nanostructured fully stabilized zirconia: Part II.

Characterization studies, J.E. Garay, Z.A. Munir, J. Mater. Res., 19

[11], 3263-69 (2004) [28] P.I. Dahl, I. Kaus, K. Wiik, Z. Zhao, M. Johnsson, M. Nygren, T.

Grande, M.-A. Einarsrud, Densification and properties of zirconia

prepared by three different sintering techniques, Ceram. Int. (2006) - Accepted

[29] I.R. Gibson, G.P. Dransfield, J.T. Irvine, Influence of yttria

concentration upon electrical properties and susceptibility to ageing

of yttria-stabilised zirconias, J. Europ. Ceram. Soc., 18, 661-667 (1998)

[30] F.T. Ciacchi, S.A. Nightingale, S.P.S. Badwal, Microwave sintering

of zirconia-yttria electrolytes and measurement of their ionic

conductivity, Solid State Ionics, 86-88, 1167-1172 (1996) [31] R. Ramamoorthy, D. Sundararaman, S. Ramasamy, Ionic

conductivity of ultrafine-grained yttria stabilized zirconia

polymorphs, Solid State Ionics, 123, 271-278 (1999) [32] X.J. Chen, K.A. Khor, S.H. Chan, L.G. Yu, Influence of

microstructure on the ionic conductivity of yttria-stabilized zirconia

electrolyte, Mater. Sci. Eng., A335, 246-252 (2002)

30

[33] M. Aoki, Y.-M. Chiang, I. Kosacki, J.-R. Lee, H.L. Tuller, Y.J. Liu, Solute segregation and grain-boundary impedance in high-purity

stabilized zirconia, J. Am. Ceram. Soc., 79, 1169-1180 (1996) [34] H.L. Tuller, Ionic conduction in nanocrystalline materials, Solid

State Ionics, 131, 143-157 (2000) [35] I. Kosacki, H.U. Anderson, Microstructure – property relationships

in nanocrystalline oxide thin films, Ionics, 6, 294-311 (2000) [36] I. Kosacki, T. Suzuki, V. Petrovsky, H.U. Anderson, Electrical

conductivity of nanocrystalline ceria and zirconia thin films, Solid State Ionics, 136-137, 1225-1233 (2000)

[37] J. Maier, Space-charge regions in solid 2-phase systems and their

conduction contribution. 3. Defect chemistry and ionic-conductivity

in thin films, Solid state ionics, 23, 59-67 (1987) [38] R.M. German, Sintering Theory and Practice, John Wiley & Sons,

Inc., New York, 104-110 (1996) [39] S.K. Tadokoro, E. N. S. Muccillo, Physical characteristics and

sintering behavior of ultrafine zirconia-ceria powders, J. Europ. Ceram. Soc., 22, 1723-1728 (2002)

[40] S.-G. Chen, Y.-S. Yin, D.-P. Wang , J. Li, Reduced activation

energy and crystalline size for yttria-stabilized zirconia nano-

crystals: an experimental and theoretical study, J. Cryst. Growth, 267 [1-2], 100-109 (2004)

[41] S. Jiang, W. A. Schulze, V. R. W. Amarakoon, G. C. Stangle, Electrical properties of ultrafine-grained yttria-stabilized zirconia

ceramics, J. Mater. Res., 12 [9], 2374-2380 (1997) [42] Y. Wu, A. Bandyopadhyay, S. Bose, Processing of alumina and

zirconia nano-powders and compacts, Mater. Sci. Eng. A, 380, 349-355 (2004)

[43] K.C. Patil, S.T. Aruna and T. Mimani, Combustion synthesis: an

update, Curr. Opin. Solid State Mat. Sci., 6, 507-512 (2002) [44] C. Laberty-Robert, F. Ansart, C. Deloget, M. Gaudon, A. Rousset,

Powder synthesis of nanocrystalline ZrO2-8%Y2O3 via a

polymerization route, Mater. Res. Bull., 36, 2083-2101 (2001) [45] C. Laberty-Robert, F. Ansart, C. Deloget, M. Gaudon, A. Rousset,

Dense yttria stabilized zirconia: sintering and microstructure, Ceram. Int., 29, 151-158 (2003)

[46] Y. B. Khollam, A. S. Deshpande, A. J. Patil, H. S. Potdar, S. B. Deshpande, S. K. Date, Synthesis of yttria stabilized cubic zirconia

(YSZ) powders by microwave-hydrothermal route, Mater. Chem. Phys., 71 [3], 235-241 (2001)

[47] L. A. Chick, L. R. Pederson, G. D. Maupin, J. L. Bates, L. E. Thomas and G. J. Exarhos, Glycine-nitrate combustion synthesis of

oxide ceramic powders, Mater. Lett., 10 [1-2], 6 (1990)

31

[48] A. S. Mukasyan, C. Costello, K. P. Sherlock, D. Lafarga, A. Varma, Perovskite membranes by aqueous combustion synthesis: synthesis

and properties, Sep. Pur. Tech., 25, 117-126 (2001) [49] I. Kaus, P.I. Dahl, J. Mastin, T. Grande, M.-A. Einarsrud, Synthesis

and characterization of nanycrystalline YSZ powder by smoldering

combustion synthesis, J. Nanomater., 2006, 1-7 (2006) [50] Z. Shen, M. Johnsson, Z. Zhao, M. Nygren, Spark plasma sintering

of alumina, J. Am. Ceram. Soc., 85 [8], 1921-27 (2002) [51] H. Iwahara, T. Esaka, H. Uchida and N. Maeda, Proton conduction

in sintered oxides and its application to steam electrolysis for

hydrogen-production, Solid State Ionics, 3-4, 359-363 (1981) [52] H. Iwahara, Technological challenges in the application of proton

conducting ceramics, Solid State Ionics, 77, 289-298 (1995) [53] T. Yajima, H. Iwahara, Studies on proton behavior in doped

perovskite-type oxides: (II) Dependence of equilibrium hydrogen

concentration and mobility on dopant content in Yb-doped SrCeO3, Solid State Ionics, 53-56, 983-988 (1992)

[54] T. Scherban, A.S. Nowick, Bulk protonic conduction in Yb-doped

SrCeO3, Solid State Ionics, 35, 189-194 (1989) [55] N. Matsunami, T. Yajima, H. Iwahara, Permeation of implanted

deuterium through SrCeO3 (5% Yb), Nucl. Instr. Meth. Phys. Res., B65, 278-281 (1992)

[56] I. Kosacki, J. Schoonman, M. Balanski, Raman scattering and ionic

transport in SrCe1-xYbxO3, Solid State Ionics, 57, 345-351 (1992) [57] I. Kosacki, H.L. Tuller, Mixed conductivity in SrCe0.95Yb0.05O3

protonic conductors, Solid State Ionics, 80, 223-229 (1995) [58] U. Reichel, R.R. Arons, W. Schilling, Investigation of n-type

electronic defects in the protonic conductor SrCe1-xYxO3-α,, Solid State Ionics, 86-88, 639-645 (1996)

[59] T. Matzeke, M. Cappadonia, Proton conductive perovskite solid

solutions with enhanced mechanical stability, Solid State Ionics, 86-

88, 659-663 (1996) [60] H. Uchida, N. Maeda, H. Iwahara, Relation between proton and hole

conduction in SrCeO3-based solid electrolytes under water-

containing atmospheres at high temperatures, Solid State Ionics, 11, 117-124 (1983)

[61] H. Uchida, H. Yoshikawa, T Eseka, S. Ohtsu, H. Iwahara, Formation

of protons in SrCeO3-based proton conducting oxides. Part II.

Evaluation of proton concentration and mobility in Yb-doped

SrCeO3, Solid State Ionics, 36, 89-95 (1989) [62] T. Yajima, H. Iwahara, H. Uchida, K. Koide, Relation between

proton conduction and concentration of oxide ion vacancy in

32

SrCeO3 based sintered oxides, Solid State Ionics, 40/41, 914-917 (1990)

[63] T. Hibino, K. Mizutani, T. Yajima, H. Iwahara, Evaluation of proton

conductivity in SrCeO3, BaCeO3, CaZrO3 and SrZrO3 by

temperature programmed desorption method, Solid State Ionics, 57, 303-306 (1992)

[64] H. Iwahara, Proton conducting ceramics and their applications, Solid State Ionics, 86-88, 9-15 (1996)

[65] T. Schober, F. Krug, W. Schilling, Criteria for the application of

high temperature proton conductors in SOFCs, Solid State Ionics, 97, 369-373 (1997)

[66] R.C.T. Slade, S.D. Flint, N. Singh, Investigation of protonic

conduction in Yb- and Y-doped barium zirconates, Solid State Ionics, 82, 135-141 (1995)

[67] K.D. Kreuer, St. Adams, W. Münch, A. Fuchs, U. Klock, J. Maier, Proton conducting alkaline eart zirconates and titanates for high

drain electrochemical applications, Solid State Ionics, 145, 295-306 (2001)

[68] T. Schober, H.G. Bohn, Water vapor solubility and electrochemical

characterization of the high temperature proton conductor

BaZr0.9Y0.1O2.95, Solid State Ionics, 127, 351-360 (2000) [69] H. Iwahara, T. Yajima, T. Hibino, K. Ozaki, H. Suzuki, Protonic

conduction in calcium, strontium and barium zirconates, Solid State Ionics, 61, 65-69 (1993)

[70] W. Wang, A.V. Virkar, Ionic and electron-hole conduction in

BaZr0.93Y0.07O3-δ by 4-probe dc measurements, J. Power Sources, 142, 1-9 (2005)

[71] H.G. Bohn, T. Schober, Electrical conductivity of the high-

temperature proton conductor BaZr0.9Y0.1O2.95, J. Am. Ceram. Soc., 83 [4], 768-772 (2000)

[72] F.M.M. Snijkers, A. Nuekenhoudt, J. Cooymans, J.J. Luyten, Proton

conductivity and phase composition in BaZr0.9Y0.1O3-δ, Scripta Materialia, 50, 655-659 (2004)

[73] K.D. Kreuer, Aspects of the formation and mobility of protonic

charge carriers and the stability of perovskite-type oxides, Solid State Ionics, 125 [1-4], 285-302 (1999)

[74] K.H. Ryu, S.M. Haile, Chemical stability and proton conductivity of

doped BaCeO3 – BaZrO3 solid solutions, Solid State Ionics, 125, 355-367 (1999)

[75] K. Katahira, Y. Kohchi, T. Shimura, H. Iwahara, Protonic

conduction in Zr-substituted BaCeO3, Solid State Ionics, 138, 91-98 (2000)

33

[76] H. Iwahara, T. Esaka, H. Uchida, N. Maeda, Proton conduction in

sintered oxides and it’s application to steam electrolysis for

hydrogen production, Solid State Ionics, 3/4, 359-363 (1981) [77] T. Scherban, A.S. Nowick, Bulk protonic conduction in Yb-doped

SrCeO3, Solid State Ionics, 35, 189-194 (1989) [78] I. Kosacki, J. Schoonman, M. Balanski, Raman scattering and ionic

transport in SrCe1-xYbxO3, Solid State Ionics, 57, 345-351 (1992) [79] H. Iwahara, T. Yajima, T. Hibino, K. Ozaki, H. Suzuki, Protonic

conduction in calcium, strontium and barium zirconates, Solid State Ionics, 61, 65-69 (1993)

[80] U. Reichel, R.R. Arons, W. Schilling, Investigation of n-type

electronic defects in the protonic conductor SrCe1-xYxO3-α,, Solid State Ionics, 86-88, 639-645 (1996)

[81] T. Matzeke, M. Cappadonia, Proton conductive perovskite solid

solutions with enhanced mechanical stability, Solid State Ionics, 86-

88, 659-663 (1996) [82] S.V. Chavan, A.K. Tyagi, Sub-solidus phase equilibria in CeO2-SrO

system, Thermochimica Acta, 390, 79-82 (2002) [83] H. Matsumoto, T. Shimura, H. Iwahara, T. Higuchi, K. Yashiro, A.

Kaimai, T. Kawada, J. Mizusaki, Hydrogen separation using proton-

conducting perovskites, J. All. Comp., 408-412, 456-462 (2006) [84] J. Brzezinska-Miecznik. K. Haberko, M.M. Bucko, Barium zirconate

ceramic powder synthesis by the coprecipitation – calcination

technique, Mater. Lett., 56, 273-278 (2002) [85] F. Boschini, B. Robertz, A. Rulmont, C. Cloots, Preparation of

nanosized barium zirconate powder by thermal decomposition of

urea in an aqueous solution containing barium and zirconium, and

by calcination of the precipitate, J. Eur. Ceram. Soc., 23, 3035-3042 (2003)

[86] A. Magrez, T. Schober, Preparation, sintering and water

incorporation of proton conducting BaZr0.9Y0.1O3-δ: comparison

between three different synthesis techniques, Solid State Ionics, 175, 585-588 (2004)

[87] P.P. Phule, D.C. Grundy, Pathways for the low temperature

synthesis of nano-sized crystalline barium zirconate, Mater. Sci. Eng., B23, 29-35 (1994)

[88] K.J. de Vires, Electrical and mechanical properties of proton

conducting SrCe0.95Yb0.05O3-α, Solid State Ionics, 100, 193-200 (1997)

[89] S. Cheng, V.K. Gupta, J.Y.S. Lin, Synthesis and hydrogen

permeation properties of asymmetric proton-conducting ceramic

membranes, Solid State Ionics, 176, 2653-2662 (2005)

34

[90] S.V. Chavan, A.K. Tyagi, Preparation of Sr0.09Ce0.91O1.91, SrCeO3,

and Sr2CeO4 by glycine-nitrate combustion: Crucial role of oxidant-

to-fuel ratio, J. Mater. Res., 19 [11], 3181-3188 (2004) [91] U. Anselmi-Tamburini, M.T. Buscaglia, M. Viviani, M. Bassoli, C.

Bottini, V. Buscaglia, P. Nanni, Z.A. Munis, Solid-state synthesis

and spark plasma sintering of submicron BaYxZr1-xO3-x/2 (x = 0, 0.08

and 0.16) ceramics, J. Eur. Ceram. Soc., 26, 2313-2318 (2006) [92] H. Taherparvar, J.A. Kilner, R. Baker, M. Sahibzada, Solid State

Ionics, 162-163, 297-303 (2003) [93] K.D. Kreuer, On the development of proton conducting materials for

technological applications, Solid State Ionics, 97, 1-15 (1997) [94] M.J. Scholten, J. Schoonman, J.C. van Miltenburg, H.A.J. Oonk,

Synthesis of strontium and barium cerate and their reaction with

carbon dioxide, Solid State Ionics, 61, 83-91 (1993) [95] D. Hassan, S. Janes, R. Clasen, Proton-conducting ceramics as

electrode/electrolyte materials for SOFC's - part 1: preparation,

mechanical and thermal properties of sintered bodies, J. Europ. Ceram. Soc., 23, 221-228 (2003)

[96] K.J. de Vires, Electrical and mechanical properties of proton

conducting SrCe0.95Yb0.05O3-α, Solid State Ionics, 100, 193-200 (1997)

[97] A.N. Shirsat, K.N.G. Kaimal, S.R. Bharadwaj, D. Das, Thermodynamic stability of SrCeO3, J. Solid State Chem., 177, 2007-2013 (2004)

[98] T. Tsuneo, S. Stølen, H. Yokoi, Thermodynamic study of barium

zirconates by mass-spectrometry, J. Nucl. Mater., 209, 174-179 (1994)

[99] T. Tsuneo, Thermodynamic properties of ternary barium oxides, Thermochimica Acta, 253, 155-165 (1995)

[100] J.B. Kortright, D.D. Awschalom, J. Stöhr, S.D. Bader, Y.U, Idzerda, S.S.P. Parkin, I.K. Schuller, H.-C. Siegmann, Research frontiers in

magnetic materials at soft X-ray synchrotron radiation facilities, J. Magnetism Magn. Mater., 207, 7-44 (1999)

[101] O. Yamamoto, Y. Takeda, R. Kanno, M. Noda, Perovskite-type

oxides as oxygen electrodes for high temperature oxide fuel cells, Solid State Ionics, 23, 241-246 (1987)

[102] I. Hole, T. Tybell, J.K. Grepstad, I. Wærnhus, T. Grande, K. Wiik, High temperature transport kinetics in heteroepitaxial LaFeO3 thin

films, Solid State Electronics, 47, 2279-2282 (2003) [103] N.N. Toan, S. Saukko, V. Lantto, Gas sensing with semiconducting

perovskite oxide LaFeO3, Physica B: Condensed Matter, 327, 279-282 (2003)

35

[104] E. Traversa, S. Matsushima, G. Okada, Y. Sadaoka, Y. Sakai, K. Watanabe, NO2 sensitive LaFe3 thin films prepared by r.f. sputtering, Sensors and Actuators B, 25, 661-664 (1995)

[105] M. Carotta, M. Butturi, G. Martinelli, Y. Sadaoka, P. Nunziante, E. Traversa, Microstructural evolution of nanosized LaFeO3 powders

from the thermal decomposition of a cyano-complex for thick film

gas sensors, Sensors and Actuators B, 44, 590-594 (1997) [106] J. Yoon, M. Grilli, E. Di Bartolomeo, R. Polini, E. Traversa, The

NO2 response of solid electrolyte sensors made using nano-sized

LaFeO3 electrodes, Sensors and Actuators B, 76, 483-488 (2001) [107] H. Aono, E. Traversa, M. Sakamoto, Y. Sadaoka, Crystallographic

characterization and NO2 gas sensing property of LnFeO3 prepared

by thermal decomposition of Ln---Fe hexacyanocomplexes,

Ln[Fe(CN)6]·nH2O, Ln = La, Nd, Sm, Gd, and Dy, Sensors and Actuators B, 94, 132-139 (2003)

[108] S. Zhao, J. Sin, B. Xu, M. Zhao, Z. Peng, H. Chai, A high

performance ethanol sensor based on field-effect transistor using a

LaFeO3 nano-crystalline thin-film as a gate electrode, Sensors and Actuators B, 64, 83-87 (2000)

[109] C. Xiangfeng, P. Siciliano, CH3SH-sensing characteristics of

LaFeO3 thick-film prepared by co-precipitation method, Sensors and Actuators B, 94, 197-200 (2003)

[110] K. Arshak, O. Korostynska, S. Clifford, Screen printed thick films of

NiO and LaFeO3 as gamma radiation sensors, Sensors and Actuators A, 110, 354-360 (2004)

[111] C. Xu, J. Tamaki, N. Miura, N. Yamazoe, Grain size effects on gas

sensitivity of porous SnO2-based elements, Sensors and Actuators B, 3, 147-155 (1991)

[112] J.K. Grepstad, Y. Takamura, A. Scholl, I. Hole, Y. Suzuki, T. Tybell, Effects of thermal annealing in oxygen on the

antiferromagnetic order and domain structure of epitaxial LaFeO3

thin films, Thin Solid Films, 486, 108-112 (2005) [113] Y.-H. Lee, J.-M. Wu, Epitaxial growth of LaFeO3 thin films by RF

magnetron sputtering, J. Crystal Growth, 263, 436-441 (2004) [114] J.P. Locquet, J. Perret, J. Fompeyrine, E. Mächler, J.W. Seo, G. Van

Tendeloo, Doubling the critical temperature of La1.9Sr0.1CuO4

using epitaxial strain, Nature, 394, 453-456 (1998) [115] A. Scholl, J. Stöhr, J. Lüning, J.W. Seo, J. Fompeyrine,H. Siegwart,

J.P. Locquet, F. Nolting, S. Anders, E.E. Fullerton, M.R. Scheinfein, H.A. Padmore, Observation of antiferromagnetic domains in

epitaxial thin films, Science, 287, 1014-1016 (2000) [116] E. Traversa, Y. Sadaoka, M. Carotta, G. Martinelli, Environmental

monitoring field tests using screen-printed thick-film sensors based

36

on semiconducting oxides, Sensors and Actuators B, 65, 181-185 (2000)

[117] Y. Mastumoto, J. Hombo, Preparation of LaFeO3 perovskite film

using electrochemical reduction, J. Electroanal. Chem., 348, 441-445 (1993)

[118] M. Rajendran, M.G. Krishna, A.K. Bhattacharya, Fabrication and

characterization of aqueous sol-gel-derived LaFeO3 thin films, Modern Physics Lett. B, 14 (22-23), 801-808 (2000)

[119] J. Mastin, M.-A. Einarsrud, T. Grande, Crystal structure and thermal

properties of La1-xCaxCoO3-delta (0 <= x <= 0.4), Chem. Mater., 18, 1680-1687 (2006)

[120] K. Kleveland, N. Orlovskaya, T. Grande, A.M.M. Moe, M.-A. Einarsrud, Ferroelastic behaviour of LaCoO3-based ceramics, J. Am. Ceram. Soc., 84, 2029-2033 (2001)

[121] S. Faaland, P.E. Wullum, R. Holmestrand, T. Grande, M.-A. Einarsrud, Stress-strain behaviour during compression of

polycrystalline La1-xCaxCoO3-based ceramics, J. Am. Ceram. Soc., 88 [3], 726-730 (2005)

[122] J. Mastin, H.L. Lein, T. Grande, M.-A. Einarsrud, Mechanical

properties of LaCoO3-based materials, to be published [123] M.A. Señarís-Rodríguez, J.B. Goodenough, Magnetic and transport

properties of the system La1-xSrxCoO3-δ (0 < x ≤ 0.50), J. Solid State Chem., 118, 323-336 (1995)

[124] M.A. Señarís-Rodríguez, M.P. Breijo, S. Castro, C. Rey, M. Sanchez, R.D. Sanchez, J. Mira, A. Fondado, J. Rivas, Peculiarities

in the electrical and magnetic properties of cobalt perovskites Ln1-

xMxCoO3 (Ln3+

: La3+

, M2+

: Ca2+

, Sr2+

, Ba2+

; Ln3+

: Nd3+

, M2+

:

Sr2+

), Int. J. Inorg. Mater., 1, 281-287 (1999) [125] V.V. Kharton, E.N. Naumovich, A.V. Kovalevsky, A.P. Viskup,

F.M. Figueiredo, I.A. Bashmakov, F.M.B Marques, Mixed electronic

and ionic conductivity of LaCo(M)O3 (M=Ga, Cr, Fe or Ni), Solid State Ionics, 138, 135-148 (2000)

[126] R.N. Singh, B. Lal, High surface area lanthanum cobaltate and its A

and B sites substituted derivatives for electrocatalysis of O2

evolution in alkaline solution, Int. J. Hydrogen Energy, 27, 45-55 (2002)

[127] E. Brosha, R. Mukundan, D.R. Brown, F. H. Garzon, J.H. Visser, M. Zanini, Z. Zhou, E.M. Logothetis, CO/HC sensors based on thin

films of LaCoO3 and La0.8Sr0.2CoO3-8 metal oxides, Sensors and Actuators B, 69, 171-182 (2000)

[128] D.T.V. Anh, W. Olthuis, P. Bergveld, Sensing properties of

perovskite oxide La0.5Sr0.5CoO3-8 obtained by using pulsed laser

deposition, Sensors and Actuators B, 103, 165-168 (2004)

37

[129] M. Losurdo, A. Sacchetti, P. Capezzuto, G. Bruno, L. Armelao, D. Batteca, G. Bottaro, A. Gasparotto, C. Maragno, E. Tondello, Optical and electrical properties of nanostructure LaCoO3 thin

films, App. Phys. Lett., 87, 0601909 (2005) [130] H. Seim, M. Nieminen, L. Niinistö, H. Fjellvåg, L.-S. Johansson,

Growth of LaCoO3 thin films from β-diketonate precursors, Appl. Surface Sci., 112, 243-250 (1997)

[131] Y. Matsumoto, T. Sesaki, J. Hombo, A new preparation method of

LaCoO3 perovskite using electronic oxidation, Inorg. Chem., 31, 738-741 (1992)

[132] T. Hattori, T.Matsui, H. Tsuda, H. Mabuchi, K. Morii, Fabrication

and electric properties of LaCoO3 thin films by ion-beam sputtering, Thin Solid Films, 388, 183-188 (2001)

[133] H. Ichinose, H. Katsuki, M. Nagano, Deposition of LaMO3(M=Co,

Cr, Al) films by spray pyrolysis in inductively coupled plasma, J. Crystal Growth, 144, 59-64 (1994)

[134] H.J. Hwang, J. Moon, M. Awano, K. Maeda, Sol-Gel Route to

porous lanthanum cobaltite (LaCoO3) thin films, J. Am. Ceram. Soc., 83, 2852-2854 (2000)

[135] H.J. Hwang, M. Awano, Preparation of LaCoO3 catalytic thin film

by the sol-gel process and its NO decomposition characteristics, J. Europ. Ceram. Soc., 21, 2103-2107 (2001)

[136] H.J. Hwang, A. Towata, M. Awano, K. Maeda, Sol-Gel route to

perovskite-type Sr-substituted LaCoO3 thin films and effects of

polyethylene glycol on microstructure evolution, Scripta Mater., 44, 2173-2177 (2001)

[137] B. Trummer, O. Fruhwirth, K. Reichmann, M. Holzinger, W. Sitte, P. Pölt, Preparation and characterisation of LaNixCo1-xO3, J. Europ. Ceram. Soc., 19, 827-829 (1999)

[138] S. Javorič, G. Dražič, M. Kosec, J. Europ. Ceram. Soc., 21, 1543-1546 (2001)

[139] E. Bontempi, L. Armelao, D. Barreca, L. Bertolo, G. Bottaro, E. Pierangelo, L.E. Depero, Structural characterized of sol-gel

lanthanum, cobaltite thin films, Crystal Eng., 5, 291-298 (2002) [140] M. Gelfi, E. Bontempi, R. Roberti, L. Armelao, L.E. Depero,

Residual stress analysis of thin films and coatings through XRD2

experiments, Thin Solid Films, 450, 143-147 (2004) [141] Y. Zhang, Y. Zhu, R. Tan, W. Yao, L. Cao, Influence of PEG

additive and precursor concentration on the preparation of LaCoO3

film with peroskite structure, Thin Solid Films, 388, 160-164 (2001) [142] Y. Zhu, R. Tan, J. Feng, S. Ji, L. Cao, The reaction and poisoning

mechanism of SO2 and perovskite LaCoO3 film model catalysts,

Appl. Catalysis A: General, 209, 71-77 (2001)

38

[143] L. Hong, X. Chen, Z. Cao, Preparation of a perovskite

La0.2Sr0.8CoO3-x membrane on a porous MgO substrate, J. Europ. Ceram. Soc., 21, 2207-2215 (2001)

[144] Y. Zhu, R. Tan, T. Yi, S. Ji, X. Ye, L. Cao, Preparation of nanosized

LaCoO3 perovskite oxide using amorphous heteronuclear complex

as a precursor at low temperature, J. Mater. Sci., 35, 5415-5420 (2000)

[145] D. P. Norton, Synthesis and properties of epitaxial electronic oxide

thin-film materials, Mater. Sci. Eng. R, 43, 139-247 (2004)

PAPER I

40

41

42

43

44

45

46

47

48

PAPER II

50

51

Densification and properties of zirconia prepared by three different sintering techniques

Paul Inge Dahl, Ingeborg Kaus, Kjell Wiik, Tor Grande, Mari-Ann Einarsrud †

Department of Materials Science and Engineering Norwegian University of Science and Technology

N-7491 Trondheim, Norway

Zhen Zhao, Mats Johnsson, Mats Nygren

Inorganic Chemistry, Arrhenius Laboratory Stockholm University

SE-106 91 Stockholm, Sweden

Abstract

Densification of nanocrystalline yttria stabilized zirconia (YSZ) powder with 8 mol% Y2O3, prepared by a glycine/nitrate smoldering combustion method, was investigated by spark plasma sintering, hot pressing and conventional sintering. The spark plasma sintering technique was shown to be superior to the other methods giving dense materials (≥ 96%) with uniform morphology at lower temperatures and shorter sintering time. The grain size of the materials was 0.21, 0.37 and 12 µm after spark plasma sintering, hot pressing and conventional sintering, respectively. Total electrical conductivity of the materials showed no clear correlation with the grain size, but the activation energy for spark plasma sintered materials was slightly higher than for materials prepared by the two other densification methods. The hardness, measured by the Vickers indentation method, was found to be independent on grain size while fracture toughness, derived by the indentation method, was slightly decreasing with increasing grain size.

Keywords: YSZ, densification, grain size, electrical conductivity, hardness, fracture toughness.

† Corresponding author. E-mail: Mari-Ann.Einarsrud@material.ntnu.no

52

Introduction

Yttria stabilized zirconia (YSZ) with its high ionic and low electronic conductivity is generally the material of choice as an electrolyte in solid oxide fuel cells (SOFC). In order to improve the conductivity and thereby lower the operating temperature alternative electrolyte materials (i.e. doped CeO2) have been suggested but also often rejected due to poor mechanical properties. Improved conductivity in the traditional YSZ materials is therefore desired. Increasing conductivity with decreasing grain size has been reported for YSZ thin films [1,2] and nanocrystalline YSZ bulk materials may therefore give the desired improvement of the ionic conductivity. It is not given, however, that the properties observed for thin films can be directly converted into nanostructured bulk material properties. The reported changes in ionic conductivity for bulk YSZ with different grain size are small (or not existing) [3-8]. Factors such as density, sintering technique and effects due to impurities must be considered before concluding with an effect of grain size.

Thin films with grain size down to ~10 nm have been prepared [2]. The preparation of nanocrystalline bulk materials is more challenging since higher temperatures and longer sintering times compared to those for films are usually needed. The grain growth normally occurs during the final stage of the sintering process were the temperature is highest. In hot pressing (HP) and spark plasma sintering (SPS) uniaxial pressure is applied during sintering allowing densification at reduced temperature and time and thereby suppressing the grain growth. The SPS technique, where a pulsed direct current is passed through an electrical conducting pressing die working as the heating element gives more rapid densification rate due to the use of pressure and rapid heating rate. The presence of a pulsed electrical field that during the initial part of the sintering might create sparks that clean the particles surface and thus facilitate grain boundary diffusion and electrical field enhanced diffusion processes might also contribute to increased densification rate [9].

The objective of the present work is to evaluate different densification methods for the preparation of dense fine grained bulk YSZ materials. Moreover, possible effects of the grain size on the total conductivity and the mechanical properties of these materials are also explored.

53

Experimental

Sample preparation

Nanocrystalline YSZ powder (8 mol% Y2O3) was prepared as previously described using a glycine/nitrate smoldering combustion method [10]. Powder synthesized with the optimal glycine/nitrate ratio of 0.23 was calcined in oxygen flow at 650°C for 24 hrs, ball milled for 12 hrs and dried at 400°C for 12 hrs. This preparation method resulted in single phase powder with crystallite size less than 10 nm and a particle size less than 50 nm calculated from surface area [10]. A small amount of residual carbonate species was present in the powder after calcination [10]. The theoretical density calculated the X-ray diffraction data was 5.96 g/cm3 [10]. The powders were sieved (250 µm) before compacting and sintering.

Green bodies for conventional sintering were prepared by uniaxial pressing at 64 MPa followed by cold isostatic pressing (CIP) at 200 MPa giving green body density around 43% of theoretical density.

Densification

Dilatometry (Netzsch, DIL 402 C) on green body cylinders was performed in ambient air with a heating rate of 120°C/h up to 1450°C. For densification studies, three different sintering techniques were used; spark plasma sintering (SPS), hot pressing (HP) and conventional sintering. The SPS (Dr. Sinter 2050, Sumitomo Coal Mining Co. Ltd., Japan) was performed in vacuum using cylindrical graphite dies also working as the heating element. A uniaxial pressure varying from 50 MPa to 110 MPa was applied at room temperature and released at the end of the holding time at the sintering temperature. The sintering temperature varied from 1100°C to 1300°C and the holding time varied from 0 to 10 min. A heating rate of 100°C/min between 600°C and 1300°C and a cooling rate of >350°C/min down to 1000°C was used. Previous studies have shown that this is a reasonable cooling rate [11].

HP was performed under nitrogen flow in a clam furnace (Entech, VSTF 40/15) using cylindrical graphite dies with an inner diameter of 15 mm. No compaction of the starting powders was made prior to the sintering. After heating to 600°C with a rate of 600°C/h, a uniaxial pressure of 25 MPa was applied before heating to sintering temperature (1150°C - 1300°C) at a rate of 300°C/h. The pressure was released after desired sintering time (varying from 0 to 6 hrs) at this temperature followed by a cooling to room temperature at 600°C/h.

54

Conventional sintering was performed in air in a muffle furnace (Entech, SF-4/17). Green bodies were sintered at temperatures from (i) 1150°C to 1300°C for 1 h with a heating rate of 300°C/h from 600°C (cooling rate 600°C/h) or (ii) 1500°C for 12 hrs with a heating/cooling rate of 200°C/h. Final density was measured by the Archimedean method (ISO 5017) in isopropanol.

Microstructure

The microstructure of the dense specimens was studied by field emission scanning electron microscopy (FE-SEM) (Hitachi S-4300SE) on polished and thermally etched surfaces. Thermal etching for 12 min at a temperature 50°C below sintering temperature was used. Average grain size was estimated from the FE-SEM micrographs using the linear intercept method [12]. For the specimens with average grain size above 1 µm the measurements were done over a minimum of 50 grains while a minimum of 100 grains were used for determining grain sizes below 1 µm.

Electrical conductivity

Electrical conductivity measurements were performed using the van der Pauw technique [13] on sintered, polished (parallel planar) specimens. With four platinum electrodes attached to the specimen surface close to the circumference, the samples were slowly heated (1°C/min) to 1000°C in a vertical tube furnace. At constant temperatures, from ~1000°C to ~500°C (and back up to ~1000°C), the specific electrical conductivity was calculated from perpendicular sets of currents (I1, I2) and voltages (U1, U2).

Mechanical properties

Hardness was calculated using the Vickers indentation method (Matsuzawa DVK-1S) [14] on polished surfaces. For each sample, 10-12 indents were made with an applied load of 2.9 N and measured by optical microscopy (Reickert MeF3 A) using a digital camera (Sony DXC-950P colour video camera). For fracture toughness (KIC) calculations [14], 10 indents with an applied load of 49 N, were made for each sample. KIC was derived from average crack length, measured by SEM (Hitachi S-3500N), as well as experimental Vickers hardness values and a modulus of elasticity (E = 218 GPa) previously reported by Donzel and Roberts [15].

55

Results

Densification

The curves in Fig. 1, derived from linear shrinkage, show how the density varied with time during SPS and conventional sintering. The sintering rate was higher for SPS as close to full density (> 96%) was achieved within minutes whereas hours were needed in order to obtain dense materials by conventional sintering. Dense materials were achieved at significantly lower temperature by SPS (1150 – 1200°C) compared to conventional sintering (1450°C). Maximum sintering rate (derived from the slope of the curves) was found between 1100°C and 1150°C for SPS and about 1250°C for the conventional sintering. As seen from Fig. 1 the densification started slightly earlier when the applied pressure during SPS was increased from 50 to 100 MPa however, the sintering rate was not affected.

Time (min)

0 2 4 6 8 200 300 400 500 600 700

De

ns

ity (

%)

30

40

50

60

70

80

90

100

SPS 50 MPa

SPS 100 MPa

Conventional

1150oC1200oC

1450oC

Fig. 1 Change in density of YSZ during SPS and conventional

sintering as a function of sintering time. The marks (x)

indicate time for which the isothermal sintering temperatures

were obtained. Densities for SPS specimens are derived

directly from linear shrinkage during sintering and the final

density measured by the Archimedean method. Densities for

conventional sintering are calculated using both green and

final density with the linear shrinkage (dL/L0) as a scaling

factor.

56

The SPS and HP specimens were grey or black depending on sintering time. The colour changed to white upon heating to temperatures above ~1000°C in ambient air. Sintering conditions as well as final density of the sintered specimens are listed in Table 1. Each specimen is given a numerical ID indicating temperature-pressure-time. Final densities achieved by isothermal SPS, HP and conventional sintering are displayed in Fig. 2 a). SPS is superior as a relative density of 96.3% was obtained at 1150°C (SPS-1150-70-5). Conventional sintering at equivalent temperatures gives significantly lower density than SPS and HP.

Table 1 Sintering parameters and densities of YSZ specimens obtained

from different sintering techniques. CS, HP and SPS notes

conventional sintering, hot pressing and spark plasma sintering,

respectively. Number code indicates temperature-pressure-time.

Specimen ID Temperature

(°°°°C)

Applied pressure

(MPa) Time

Density

(%)

CS-1500-0-12 1500 - 12 h 97.4

HP-1200-25-1 1200 25 1 h 90.8

HP-1200-25-6 1200 25 6 h 93.6

HP-1250-25-0.5 1250 25 0.5 h 93.9

HP-1250-25-1 1250 25 1 h 96.9

HP-1250-25-3 1250 25 3 h 97.0

HP-1300-25-0 1300 25 0 h 92.3

HP-1300-25-0.5 1300 25 0.5 h 97.5

HP-1300-25-1 1300 25 1 h 99.3

SPS-1150-70-5 1150 70 5 min 96.3

SPS-1200-70-5 1200 70 5 min 97.3

SPS-1250-70-5 1250 70 5 min 98.4

SPS-1300-70-5 1300 70 5 min 98.5

SPS-1200-50-0 1200 50 0 min 85.8

SPS-1200-50-1 1200 50 1 min 94.0

SPS-1200-50-5 1200 50 5 min 98.8

SPS-1200-50-10 1200 50 10 min 99.5

SPS-1100-110-8 1100 110 8 min 96.0

SPS-1150-100-5 1150 100 5 min 96.3

SPS-1150-100-3* 1150 100* 3 min 98.5

SPS-1150-100-5* 1150 100* 5 min 99.3

* Pressure applied at sintering temperature

57

a)

Temperature (oC)

1150 1200 1250 1300

Den

sit

y (

%)

50

60

70

80

90

100

SPS, 70 MPa, 5 min

HP, 25 MPa, 1 h

Conventional, 1 h

b)

SPS time (min)

0 2 4 6 8 10

Den

sit

y (

%)

75

80

85

90

95

100

HP time (hrs)

0 1 2 3 4 5 6

SPS, 1200oC, 50 MPa

HP, 1250oC, 25 MPa

HP, 1200oC, 25 MPa

Fig. 2 Densities of YSZ specimens sintered by different techniques at;

a) different temperatures and b) different time.

58

Fig. 2 b) shows how the density varies with sintering time at constant temperature. SPS shows the highest degree of densification. The density of hot pressed specimen HP-1250-25-1 was measured to 96.9%. In comparison the density of SPS-1200-50-5 was 98.8 showing that both temperature and in particular sintering time can be reduced significantly for the SPS technique compared to HP. An attempt to lower the SPS temperature to 1100°C by applying a pressure of 110 MPa (SPS-1100-110-8) resulted in a final density of 96.0%. By applying a pressure at the sintering temperature, a final density of 98.5% and 99.3% was obtained for SPS-1150-100-3* and SPS-1150-100-5*, respectively.

Microstructure and grain growth

FE-SEM micrographs of a selection of sintered YSZ materials are shown in Fig. 3. As seen in Fig. 3 a) through d), SPS gives homogenous microstructures with narrow grain size distribution. The grain growth during SPS, with increasing isothermal time at 1200°C (Fig. 3 a) and b)), is moderate compared to the grain growth with increasing sintering temperature (Fig. 3 c) and d)). The smallest grain size (0.21 µm) for spark plasma sintered YSZ was obtained for the SPS-1100-110-8 specimen (Fig. 3 c)). In contrast severe grain growth was observed when the SPS sintering temperature was increased from 1100°C to 1300°C (SPS-1300-70-5 shown in Fig. 3 d)). Grain growth during HP with increasing isothermal time at 1250°C (Fig. 3 e) and f)) is moderate, comparable to SPS (Fig. 3 a) and b)). The time scale however, is different and specimen HP-1250-25-3 (Fig. 3 f)) showed slightly smaller grains than SPS-1200-50-10 (Fig. 3 b)). The conventionally sintered specimen (CS-1500-12) in Fig. 3 g) has larger grains and with enclosed pores.

The average grain size development for YSZ materials, sintered at different temperatures and times, is displayed in Fig. 4 a) and b), respectively. The results indicate that HP is a good technique for densification of YSZ with suppressed grain growth. For relatively dense YSZ (97%), an average grain size down to 0.37 µm was obtained (HP-1250-25-1). The SPS technique show a more severe grain growth with increasing temperature (Fig 4 a)) compared to HP, but as a function of time (Fig. 4 b)) the growth is moderate for SPS as well. The average grain size of the conventionally sintered specimen (CS-1500-0-12) was 12 µm.

59

Fig. 3 FESEM micrographs of sintered YSZ specimens after oxidation

and thermal etching 50°C below sintering temperature.

f) HP-1250-25-3 500 nm

c) SPS-1100-110-8 500 nm

g) CS-1500-0-12 3 µm

e) HP-1250-25-0 500 nm

a) SPS-1200-50-0 500 nm

d) SPS-1300-70-5 3µm

b) SPS-1200-50-10 500 nm

60

a)

Temperature (oC)

1150 1200 1250 1300

Gra

in s

ize

(µ

m)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

SPS, 70 MPa, 5 min

HP, 25 MPa, 1 h

b)

SPS time (min)

0 2 4 6 8 10

Gra

in s

ize (

µm

)

0.20

0.25

0.30

0.35

0.40

0.45

0.50

0.55

HP time (hrs)

0.0 0.5 1.0 1.5 2.0 2.5 3.0

SPS, 1200oC, 50 MPa

HP, 1250oC, 25 MPa

Fig. 4 Average grain size as a function of sintering temperature (a)

and time (b) for YSZ sintered by hot pressing and spark plasma

sintering.

61

Electrical conductivity

The electrical conductivity of selected specimens is presented in Fig. 5 in the form of plots of log(σT) versus 1000/T. Table 2 summarizes the electrical conductivity at 900°C and the associated activation energy for specimens sintered by the different techniques. For YSZ specimens sintered from synthesized powder no clear trend of variation with grain size (varying from 0.21 µm to 12 µm) was observed for neither conductivity (σ) nor activation energy (EA). It can however be remarked that EA for the SPS specimens are in the slightly higher (99 – 102 kJ/mol) compared to that of the HP specimen (96 kJ/mol) and conventional sintered specimen (94 kJ/mol).

1000/T (K-1

)

0.8 0.9 1.0 1.1 1.2 1.3

log

[(σσ σσ

T)

/ (S

cm

-1K

)]

0.0

0.5

1.0

1.5

2.0

2.5

CS-1500-0-12

HP-1250-25-1

SPS-1300-70-5

SPS-1150-70-5

SPS-1100-110-8

Fig. 5 Logarithmic plot of the electrical conductivity for YSZ materials

measured by the van der Pauw technique.

62

Table 2 Properties of YSZ specimens obtained from different sintering

techniques.

Specimen ID Density

(%)

Average

grain size

(µm)

Conductivity

at 900°C

(mS/cm)

Activation

energy

(kJ/mol)

CS-1500-0-12 97.5 12 70 94

SPS-1300-70-5 98.5 3.3 83 101

HP-1250-25-1 96.9 0.37 82 96

SPS-1150-70-5 96.3 0.24 75 99

SPS-1100-110-8 96.0 0.21 82 102

Mechanical properties

The SEM micrograph in Fig. 6 a) show an intergranular crack obtained from a 49 N indent on the SPS-1300-70-5 specimen. Intergranular cracks were observed for all samples except the larger grained conventionally sintered (1500°C) YSZ specimen which showed intragranular crack behaviour (Fig. 6 b)). Vickers hardness (HV) and fracture toughness (KIC) of YSZ materials with varying grain size are shown in Fig. 7 a) and b), respectively. As seen there is no significant change in HV for specimen with grain size from 0.2 to 12 µm, while the fracture toughness is slightly increasing with decreasing grain size.

a) b)

Fig. 6 a) SEM micrograph of a intergranular crack (from 49 N indent)

in YSZ sintered by SPS at 1300°C. b) SEM micrographs of the

intragranular crack (from 49 N indent) in YSZ sintered

conventionally at 1500°C.

a) SPS-1300-70-5 4 µm b) CS-1500-0-12 4µm

63

Grain size (µm)

0.1 1 10

KIC

(M

Pa

m1

/2)

1.1

1.3

1.5

1.7

HV (

GP

a)

12

13

14

15

a)

b)

Fig. 7 Vickers hardness (a) and fracture toughness (b) for YSZ

materials with different grain size. The uncertainty bars indicate

the standard deviations.

Discussion

Densification

SPS is shown to be a highly efficient technique for densification of YSZ at temperatures 100-200°C lower than that needed by HP. For HP with an applied pressure of 25 MPa, a temperature of 1300°C is needed to obtain fully dense (>99% of theoretical value) YSZ materials. In comparison, 1150°C for 5 min (100 MPa) was sufficient when using the SPS technique. The higher pressure used in SPS can partly explain the lower sintering temperature. The main reason for both the short sintering time and reduced temperature, however, is the efficient heat transfer as well as the self-heating from spark discharge between the particles. Under these conditions

64

residual carbonates in the powder will be efficiently removed as long as the system is not closed. By not applying the load until the isothermal SPS temperature is reached, CO2 is allowed to escape the system before the main densification starts. Residual carbonates in the powder have been shown to inhibit the sintering [10]. A more efficient removal of carbonates during SPS compared to HP and conventional sintering can therefore also explain the higher sintering rate and lower sintering temperature. The densification rate by SPS has a maximum 50 – 100°C lower than reported by Anselmi-Tamburini et al. for SPS of commercial powder (TZ-8Y) [16].

Microstructure and grain growth

HP and SPS have been proved to produce dense YSZ materials with small grains compared to conventional sintering where severe grain growth is inevitable in the final stage of densification. The kinetics of grain growth during HP and SPS was investigated by calculating the growth order, n, and (for HP) the activation energy, EA, for grain growth. The growth order was calculated from the slope of the linear regression plots shown in Fig. 8 a) assuming the growth obeys a power-law relationship [17]. According to the LSW theory [18, 19] the exponent, n, should be equal to 3 when the grain growth is controlled by volume diffusion. Furthermore, n = 2 can be assigned to growth controlled by energy difference across curved interfaces while n = 4 is associated with coarsening controlled by grain boundary diffusion. Hence, grain growth for both SPS and HP, with calculated exponents n = 3.3 and n = 2.6 respectively, seem to be dominated by volume diffusion. Even though the exponent n for SPS is higher, the relative grain growth D - D0 is lower for HP. It is shown in Fig. 4 a) how the grain growth for SPS with increasing sintering temperature is far more severe than for HP. More effective decomposition of carbonate species during SPS, giving grain boundaries with higher purity and therefore higher mobility, can explain the higher grain growth rate compared to that of HP.

It should be noted that plotting ln(D) instead of ln(D - D0) as a function of ln(t) led to unreasonably high growth order values. Simplified plots like this are often reported [20-22] and will for severe grain growth (several orders of magnitude) give satisfactory estimates of n. From our observations we can state that this fails for smaller grain size changes with sintering time. The grain growth observed for SPS follow the same trend as recently reported work on YSZ [16]. The activation energy for grain growth during HP, calculated from the slope in Fig. 8 b), was 226 kJ/mol. This is comparable to reported value of 289 kJ/mol for YSZ (8% Y2O3)

[23].

65

a)

ln[t / hrs]

-4 -3 -2 -1 0 1

ln[(

D -

D0)

/ µ

m]

-3.0

-2.5

-2.0

-1.5

-1.0

n = 3.3

n = 2.6

SPS

HP

b)

1000/T (K-1)

0.76 0.78 0.80 0.82 0.84

ln[(

D-D

0)2

.6 /

t]

-3.6

-3.2

-2.8

-2.4

-2.0

-1.6

-1.2

EA = 226 kJ/mol

Fig. 8 a) Grain growth exponents, n, for HP at 1250°C and SPS at

1200°C, calculated from the slopes of the linear regression plots

of ln(D – D0) versus ln(t). b) Activation energy for grain growth

during HP, calculated from the slope of the linear regression

plot of ln((D – D0)/t) versus 1000/T.

66

Electrical conductivity

It has been suggested that the difference in conductivity for the polycrystalline specimens from different starting powders could be due to a different level of impurities in the powders. The powder contained a small amount of residual carbonates. However, it is unlikely that carbonates are still present after sintering, particularly by using the SPS technique. At 900°C, σ varies from 75 to 83 mS/cm and in comparison a total ionic conductivity in the range of 40 – 60 mS/cm at 900°C is reported for YSZ (8% Y2O3) by Chen [5]. For YSZ (8.7% Y2O3) bulk materials with grain size varying from 6.5 down to 1.3 µm, a small variation in electrical conductivity (3 – 6 mS/cm at 600°C) has been reported by Tuller [6]. This is comparable to the measured electrical conductivity of 3 – 5 mS/cm at 600°C, however in this case no systematic variation with grain size is observed. Calculated activation energies varies from 94 to 102 kJ/mol which is somewhat lower than what is reported by Tuller (111 kJ/mol) [6] and Anselmi-Tamburini (109 kJ/mol) [24]. The colouring of SPS and HP specimens after sintering is an indication of reduced material as previously reported [24]. Increased electronic conductivity would be expected in YSZ specimen containing reduced species acting as charge carriers. Such an effect is not observed as the specimens are oxidized by heating to ~1000°C prior to the electrical conductivity measurement. A possible explanation for the higher EA values for SPS and HP specimens could be carbon impurities from the graphite pressing dies used during sintering. It should be pointed out that no significant weight gain or loss due to oxidation of reduced species or removal of carbon was observed upon heating.

Mechanical properties

Measured Vickers hardness are in agreement with reported values of 13-15 GPa [15,17,25] and no significant change is observed for different grain sizes. The fracture toughness is somewhat decreasing with increasing grain size. As intergranular fracture is observed, an increase in KIC with increasing grain size should be expected [26]. The observed decrease might be due to the segregation of pores along the grain boundaries, as seen for the SPS specimen in Fig. 6 a). Intergranular fracture in conventionally sintered YSZ (Fig. 6 b)) can be explained by the trapped pores inside individual grains. The intergranular fracture results in lower KIC value compared to the SPS sample sintered at 1100°C. The obtained KIC values are in agreement with literature values [15, 23].

67

Conclusions

In a comparative study of the sintering techniques SPS, HP and conventional sintering, dense YSZ materials (> 96% of theoretical), with average grain size of 210 nm, 370 nm and 12 µm, respectively, were obtained by all methods. SPS enables preparation of dense materials with limited grain growth provided appropriate sintering conditions are applied, if not very fast grain growth might occur. No significant differences in electrical conductivity were observed for YSZ materials with grain size varying from 210 nm to 12 µm, however the activation energy for spark plasma sintered material (99-102 kJ/mol) was slightly higher than for HP (96 kJ/mol) and conventional sintering (94 kJ/mol). No clear variation in Vickers hardness was found with varying grain size. The fracture toughness showed a small increase with decreasing grain size.

References

[1] I. Kosacki, T. Suzuki, V. Petrovsky, H.U. Anderson, Electrical

conductivity of nanocrystalline ceria and zirconia thin films, Solid State Ionics, 136-137, 1225-1233 (2000)

[2] I. Kosacki, H.U. Anderson, Microstructure – Property relations in

nanocrystalline oxide thin films, Ionics, 6, 294-311 (2000) [3] M.C. Martin, M.L. Mecartney, Grain boundary ionic conductivity of

yttrium stabilized zirconia as a function of silica content and grain

size, Solid State Ionics, 161, 67-79 (2003) [4] X. Guo, Z. Zhang, Grain size dependent grain boundary defect

structure: case of doped zirconia , Acta Materialia, 51, 2539-2547 (2003)

[5] X.J. Chen, K.A. Khor, S.H. Chan, L.G. Yu, Influence of

microstructure on the ionic conductivity of yttria-stabilized zirconia

electrolyte, Mater. Sci. Eng., A335, 246-252 (2002) [6] H.L. Tuller, Ionic conduction in nanocrystalline materials, Solid

State Ionics, 131, 143-157 (2000) [7] P. Mondal, A. Klein, W. Jaegermann, H. Hahn, Enhanced specific

grain boundary conductivity in nanocrystalline Y2O3-stabilized

zirconia, Solid State Ionics, 118, 331-339 (1999) [8] C. Petot, M. Filal, A.D. Rizea, K.H. Westmacott, J.Y. Laval, C.

Lacour, R Ollitrault, Microstructure and ionic conductivity of freeze-

dried yttria-doped zirconia, J. Eur. Ceram. Soc., 18, 1419-28 (1998) [9] Z. Shen, M. Johnsson, Z. Zhao, M. Nygren, Spark plasma sintering

of alumina, J. Am. Ceram. Soc., 85 (8), 1921-27 (2002)

68

[10] I. Kaus, P. Dahl, J. Mastin, T. Grande, M.-A. Einarsrud, Synthesis

and characterization of nanocrystalline YSZ powder by smoldering

combustion synthesis, Journal of Nanomaterials, 2006, 1-7 (2006) [11] Z. Shen, Z. Zhao, H. Peng, M. Nygren, Formation of tough

interlocking microstructures in silicon nitride ceramics by dynamic

ripening, Nature, 417, 266-269 (2002) [12] M.I. Mendelson, Average grain size in polycrystalline ceramics, J.

Am. Ceram. Soc., 52 (8), 443-446 (1969) [13] L.J. Van der Pauw, A method for measuring specific resistivity and

hall effect of discs of arbitrary shape, Philips Res. Repts., 13, 1-9 (1958)

[14] G.R. Anstis, P. Chantikul, B.R. Lawn, D.B. Marshall, A critical-

evaluation of indentation techniques for measuring fracture

toughness: 1. Direct crack measurements, J. Am. Ceram. Soc., 64, 533-538 (1981)

[15] L. Donzel, S.G. Roberts, Microstructure and mechanical properties

of cubic zirconia (8YSZ)/SiC nanocomposites, J. Eur. Ceram. Soc., 20, 2457-2462 (2000)

[16] U. Anselmi-Tamburini, J.E. Garay, Z.A. Munir, Spark plasma

sintering and characterization of bulk nanostructured fully stabilized

zirconia: Part I. Densification studies, J. Mater. Res., 19 (11), 3255-3262 (2004)

[17] H. Jiang, Y. Lu, W. Huang, X Li, M. Li, Microstructural evolution

and mechanical properties of the semisolid Al-4Cu-Mg alloy, Mater. Char., 51, 1-10 (2003)

[18] I.M. Lifshitz, V.V. Slyozov, The kinetics of precipitations from

supersaturated solid solutions, J. Phys. Chem. Solids, 19 (1/2), 35-40 (1961)

[19] C. Wagner, Theorie der alterung von niederschlagen durch umlosen

(Ostwald-Reifung), Zeitschrift für Elektrochemie, 65, 581-591 (1961)

[20] S. Tekeli, Influence of alumina addition on grain growth and room

temperature mechanical properties of 8YSCZ/Al2O3 composites, Comp. Sci. Tech., 65, 967-972 (2005)

[21] L. Yang, J.S. Wu, L.T. Zhang, Microstructure evolvements of a rare-

earth filled skutterudite compound during annealing and spark

plasma sintering, Materials and Design, 25, 97-102 (2004) [22] T.S. Zhang, J. Ma, Y.J. Leng, Z.M. He, Sintering, microstructure

and grain growth of Fe-doped Ce0.9Gd0.1O2−δ ceramics derived from

oxalate coprecipitation, J. Crystal Growth, 274, 603-611 (2005) [23] S. Tekili, Fracture toughness (K-IC), hardness, sintering and grain

growth behaviour of 8YSCZ/Al2O3 composites produced by colloidal

processing, J. All. Comp., 391, 217-224 (2005)

69

[24] U. Anselmi-Tamburini, Spark plasma sintering and characterization

of bulk nanostructured fully stabilized zirconia: Part II.

Characterization studies, J.E. Garay, Z.A. Munir, J. Mater. Res., 19

(11), 3263-69 (2004) [25] G.A. Gogotsi, S.N. Dub, E.E. Lomonova, B.I. Ozersky, Vickers and

Knoop indentation behaviour of cubic and partially-stabilized

zirconia crystals, J. Eur. Ceram. Soc., 15, 405-413 (1995) [26] P.F. Becher, Microstructural design of toughened ceramics, J. Am.

Ceram. Soc., 74 (2), 255-269 (1991)

70

PAPER III

72

73

Synthesis, densification and electrical properties of Strontium Cerate ceramics

Paul Inge Dahl, Hilde Lea Lein, Tor Grande, Mari-Ann Einarsrud †

Department of Materials Science and Engineering Norwegian University of Science and Technology

NO-7491 Trondheim, Norway

Truls Norby, Reidar Haugsrud

Centre for Materials Science and Nanotechnology Department of Chemistry, University of Oslo

NO-0349 Oslo, Norway

Abstract

Powders of pure and 5% ytterbium substituted strontium cerate (SrCeO3 / SrCe0.95Yb0.05O3-δ) were prepared by spray pyrolysis of nitrate salt solutions. The powders were single phase after calcination in nitrogen atmosphere at 1100°C (SrCeO3) and 1200°C (SrCe0.95Yb0.05O3-δ). Hollow spherical particles formed during spray pyrolysis were effectively broken down by ball milling giving particle size down to 0.06 µm. Dense SrCeO3 and SrCe0.95Yb0.05O3-δ materials were obtained by sintering at 1350 - 1400°C in air, but heat treatment at 850 and 1000°C, respectively, was necessary prior to sintering. The dense materials had homogenous microstructures with grain size in the range 6 - 10 µm for SrCeO3 and 1 - 2 µm for SrCe0.95Yb0.05O3-δ, after sintering up to 1400°C. The electrical conductivity of SrCe0.95Yb0.05O3-δ was in good agreement with reported data, showing mixed ionic-electronic conduction. The ionic contribution was dominated by protons in below 1000°C and the proton conductivity reached a maximum of ~0.005 S/cm above 900°C. In oxidizing atmosphere the p-type electronic conduction was dominating above ~700°C, while the contribution from n-type electronic conduction only was significant above ~1000°C in reducing atmosphere.

† Corresponding author. E-mail: mari-ann.einarsrud@material.ntnu.no

74

Introduction

Perovskite-type oxide materials (ABO3 structures), such as orthorhombic strontium cerate, SrCeO3, have been widely studied due to the electrical properties of these materials. When properly substituted with rare earth oxides on the cerium site, SrCeO3, exhibit high proton conductivity as first demonstrated by Iwahara et al. [1]. This property makes SrCeO3-based ceramics suitable for electrolyte materials in solid oxide fuel cells (SOFCs), hydrogen pumps and sensors [2-4]. In particular SrCeO3 substituted with 5% Yb (SrCe2.95Yb0.05O3-δ) has been classified as one of the perovskite-type oxides with the highest proton conductivity (~0.004 S/cm at 900°C) [5]. However, SrCeO3 based materials have often been rejected for use as proton conductors in commercial devices due to poor mechanical properties and chemical stability [2].

High stability of strontium carbonate (SrCO3) makes preparation of single phase SrCeO3 powders and bulk materials challenging. This is described in the literature [6,7], by the reaction between SrCeO3 and CO2 according to Eq. 1, forming SrCO3 which may decompose to strontium oxide (SrO) as described by Eq. 2.

SrCeO3(s) + CO2(g) = SrCO3(s) + CeO2(s) (1)

SrCO3(s) = SrO(s) + CO2(g) (2)

The onset temperature for synthesis of SrCeO3 from a mixture of SrCO3 and CeO2 (according to Eq. 1) has been reported to 800°C and between 1168 and 1190°C, for reaction in N2 and CO2 atmosphere, respectively. The corresponding onset temperature for decomposition of SrCO3 (according to Eq. 2) in N2 and CO2 atmosphere have been reported to 845°C and between 1220 and 1275°C, respectively [6,8]. Hence, at lower temperatures (<800°C), carbonate formation may be critical, with respect to phase purity and mechanical properties of SrCeO3 materials. Reactions with water, equivalent to Eq. 1 and 2, should also be considered when operating in atmospheres with high p(H2O) [7].

Preparation of SrCeO3 powders have been reported by methods such as complexation with EDTA [9] or citric acid [10] and combustion methods [11], but the dominating synthesis route for these materials is by the solid state ceramic method [1,6,8,12-20]. The main objectives of work on SrCeO3 ceramics have been to investigate the electrical proton conductivity. Hence, little focus has been put into densification and microstructure of sintered specimens. Generally sintering temperatures in the range 1500 - 1650°C (in air) are needed for preparing dense (>95%) SrCeO3 and SrCe2.95Yb0.05O3-δ

75

materials by the solid state ceramic method [13,15,16,20]. Higher densities (>97%) have been obtained from commercial powders by sintering at 1450°C in air for 2 hrs [21], and from powder prepared by a complexation route by sintering at 1300°C for 12 hrs in nitrogen atmosphere [9].

The limited literature data on synthesis and densification of SrCeO3 based materials leads to the motivation for the present work, where the overall aim have been to give descriptive routes for synthesis of SrCeO3 powders and preparation of dense materials with homogenous microstructures. Preparation routes for pure and 5% Yb-substituted SrCeO3 powders will be presented and the influence of carbonates (and other secondary phases) will be discussed. The sinterability of the materials has been studied with focus on how the microstructure in these materials evolves during densification. Furthermore, characterization of the electrical conductivity will be presented and thoroughly discussed, and thermodynamic data extracted from these.

Experimental

Powder synthesis and sample preparation

Powders of SrCeO3 and SrCe0.95Yb0.05O3-δ (noted SC and SC5Yb in the following) were synthesized by spray pyrolysis of nitrate salt solutions. Precursor solutions of cerium nitrate (Ce(NO3)2·xH2O, Alfa Aesar, 99.5%) and ytterbium nitrate (Yb(NO3)3·5H2O, Aldrich, 99.9%) dissolved in distilled water were thermogravimetrically standardized and mixed in stoichiometric proportions with dried (250°C, 24 hrs) strontium nitrate (Sr(NO3)2, Merck, >99%), giving a total cation concentration of 0.9 M. The mixed solutions were atomized into a furnace at a rate of 1 L/h. The temperature by the nozzle was 840-845°C and the output temperature was 540-550°C. The as-synthesized SC and SC5Yb powders were calcined at 900°C in ambient air for 48 hrs. Additional calcination at 1000 - 1200°C

for 6 hrs in nitrogen flow (2Op < 10-4 atm,

2COp < 2·10-7 atm) were performed

in order to remove carbonates and obtain phase pure powders. The calcined powders were ball milled with yttria stabilized zirconia (YSZ) balls in iso-propanol for 6 - 24 hrs. The milled powders were dried at 400°C for 12 hrs in ambient air and sieved (150 µm).

Green bodies were made by uniaxial pressing at 64 MPa. Generally powders calcined at higher temperatures (>1000°C) needed addition of 2 wt% binder (ethyl cellulose, Sigma) in order to increase the green strength and green density (~55%). No binder was needed for compaction of powders calcined at 900°C giving green bodies with green density of 50-

76

53%. Green bodies were sintered in ambient air at temperatures varying from 1250°C to 1450°C in a muffle furnace (Entech, SF-4/17). In order to remove carbonate species from the specimens prior to sintering, heat treatment at 850°C for 24 hrs and 1000°C for 1 h was used for sintering of SC and SC5Yb, respectively.

Characterization of sintered materials

The prepared powders were analyzed and crystal structure confirmed by X-ray diffraction (XRD) (Cu Kα, Phillips PW1730/10). Powders from crushed samples sintered at 1400°C and added 30 weight% Si standard were used to obtain XRD data from which the lattice parameters and theoretical densities were calculated by the Rietveld method. Infrared spectroscopy (IR) was performed on potassium bromide (Merck, KBr for IR spectroscopy Uvasol®) discs containing 1 wt% powder using a Bruker IFS 66v spectrophotometer. Thermogravimetric analysis (TGA) (Netzsch STA 449 C Jupiter) on powders was carried out in air with a heating rate of 3°C/min to 1450°C. Nitrogen adsorption on de-gassed powders (24 hrs, 250°C) was measured (Micromeritics ASAP 2000) using the 5 point BET equation to obtain specific surface area, from which the particle size was calculated assuming spherical particles.

Dilatometry (Netzch, DIL 402 C) on green body cylinders was performed in air with a heating rate of 3°C/min up to 1450°C. Density of sintered specimens was measured by the Archimedean method (ISO 5017) in iso-propanol. The prepared powders as well as sintered specimens were studied by scanning electron microscopy (SEM) (Hitachi S-3500N).

Conductivity measurements

A circular Pt electrode (~10 mm diameter) was attached to each side of sintered SC5Yb specimens, before mounting on top of the support tube in a ProboStat measurement cell for electrochemical characterization (two-point measurements) [22]. When the purpose was to measure the open-circuit voltage (OCV) from gradients in the pressures of oxygen, hydrogen and water vapor, a gold gasket was placed between the alumina support tube and the specimen. Consequently, the specimen served as a membrane between the inner and the outer gas compartment in the cell. Electrode leads (current supplies and voltage probes) from the cell base were contacted to the electrodes on the specimen. Relatively strong spring loads with alumina

77

parts in the hot zone held the assembly together; to maintain contact between the Pt leads and the electrodes and to facilitate the gold sealing.

The total conductivity was measured in the temperature range 300 to 1050°C by means of impedance spectroscopy and at a constant frequency of 10 kHz (Solartron 1260 FRA, oscillation voltages between 0.1 and 1 V) as a function of the oxygen partial pressure; in wet (0.025 atm H2O) O2-Ar mixtures for oxidizing and wet (0.025 atm H2O) H2-Ar mixtures for reducing conditions, as well as a function of the water vapor partial pressure. The partial conductivities, measured under wet conditions, were calculated based on transport numbers from EMF measurements (Solartron 7150+ high impedance voltmeter) as a function of temperature. Details regarding the EMF-method and the set-up of the gas mixer are described in the literature [23,24].

Impedance spectroscopy in the frequency range 1 MHz to 0.1 Hz was applied to determine the different contributions of bulk (b), grain boundaries (gb) and electrodes (e) to the overall impedance. Different circuits were assigned to fit the impedance data depending on the conditions (T, p(O2) and p(H2O)). Up to 650°C it was possible to separate the impedance of bulk and grain boundaries by fitting the impedance data to circuits of parallel resistors (R) and constant phase capacitive elements (Q) in series: (RbQb)(RgbQgb)(ReQe) at temperatures below 450°C, and Rb(RgbQgb)(ReQe) from 450 to 650°C. Above 650°C the bulk and grain boundary impedance could not be separated and the volume resistance, Rvol = Rb + Rgb, was determined. Under conditions where electronic conduction plays a significant role, it has been assumed that Relectronic, representing the electronic conductance, connects effectively directly between electrodes, yielding circuits which in a simplified version may look like Relectronic[Rvol(ReQe)]. At temperatures >500°C the impedance spectra were corrected for an inductive element resulting from different types of parasitic contribution. The (ReQe)-elements represent complex processes at the electrodes and can be further separated into several different circuit elements.

78

Results

Characterization of powders

The morphology of prepared powders is presented by SEM micrographs in Fig. 1. During spray pyrolysis characteristic agglomerates in the form of hollow spheres were formed as seen for the SC powder calcined at 900°C. The spherical shape was maintained upon heating to 1100°C, however, some coarsening of the powder was observed. The agglomerates in the SC and SC5Yb powders calcined at 900°C were effectively broken down to smaller particles by ball milling in iso-propanol. Due to coarsening of the powders during heat treatment at 1100°C, the particle size after milling seemed somewhat larger in the powders calcined at 1100°C. Ball milling in water resulted in decomposition of the calcined powders, confirming the chemical instability of SrCeO3.

Surface area and calculated particle size for a selection of powders is presented in Table 1. Ball milling of SrCeO3 powder calcined at 900°C resulted in an increase in surface area, from 1.7 to 5.4 m2/g. Calcination at 1100°C gave a slight reduction in surface area to 3.9 m2/g (after ball milling). The surface area of ball milled SC5Yb powder calcined at 900°C was 17.8 m2/g, corresponding to particles 3 - 8 times smaller compared to that found for the SC powders.

Table 1 Surface area (SA) of milled SrCeO3 and SrCe0.95Yb0.05O3-δ

powders. Spherical particles were assumed for calculation of

the particle size (D).

Compound Tcalcination

(°C)

tcalcination

(hrs) SA (m

2/g) D (µm)

SrCeO3* 900 48 1.7 0.62

SrCeO3 900 48 5.4 0.19

SrCeO3 1100 6 3.9 0.27

SrCe0.95Yb0.05O3-δ 900 48 17.8 0.06

SrCe0.95Yb0.05O3-δ 1100 6 9.2 0.11

* Not milled

79

Fig. 1 SEM micrographs of not milled SrCeO3 and SrCe0.95Yb0.05O3-δ

powders calcined at 900 and 1100°C before and after milling in

iso-propanol.

80

The orthorhombic phase was confirmed for SC and SC5Yb by XRD. The calculated lattice parameters and theoretical densities for both compositions are listed in Table 2. X-ray diffractograms of SC and SC5Yb powders are shown in Fig. 2 a) and b), respectively. Secondary phases (Sr2CeO4 and CeO2) were present in the SC powder after calcination at 900°C. After calcination at 1000°C in nitrogen flow these phases were still detected by XRD, however, calcination under these conditions at 1100°C resulted in phase pure SC powder. Larger amounts of secondary phases were detected in the SC5Yb powder calcined at 900°C in air. In addition to the secondary phases found in the SC powders, SrCO3 was detected in SC5Yb powder calcined at 900°C. SrCO3 was not detected in the SC5Yb powder calcined at 1000°C in nitrogen flow, but both Sr2CeO4 and CeO2 were present. Traces of Sr2CeO4 were also detected after calcination at 1100°C but the powder was apparently phase pure after calcination at 1200°C. XRD on the surface of sintered specimen exposed to ambient atmosphere for 2 weeks gave no indications of formation of SrCO3.

Table 2 Lattice parameters and theoretical density (TD) for SrCeO3 and

SrCe0.95Yb0.05O3-δ powders.

Compound a (Å) b (Å) c (Å) TD (g/cm3)

SrCeO3 6.16 8.60 6.01 5.76

SrCe0.95Yb0.05O3-δ 12.28 8.58 6.00 5.82

81

a)

2 θθθθ (°)

20 25 30 35 40

Re

lati

ve in

ten

sit

y

�

�

�

��

� �

�

��

�

��

�

�

�

�

�

SrCeO3

Sr2CeO4

CeO2

900°C

1000°C

1100°C

b)

2 θ θ θ θ (°)

20 25 30 35 40

Re

lati

ve

in

ten

sit

y

�

�

�

�

�

�

�� �

�

� �

�

�

�

��

�

�

�

SrCe0.95Yb0.05O2.975

SrCO3

Sr2CeO4

CeO2

�

�

�

900°C

1000°C

1100°C

1200°C

Fig. 2 X-ray diffractograms of a) SrCeO3 powders and b)

SrCe0.95Yb0.05O3-δ powders calcined in ambient air at 900°C and

in nitrogen flow from 1000 to 1200°C.

82

Fig. 3 shows TG data for SC and SC5Yb powders. Significant weight loss was observed for both SC and SC5Yb powder calcined at 900°C and ball milled in iso-propanol. The observed weight loss below 400°C (mainly seen for SC powder calcined at 900°C) is most likely due to adsorption of CO2 and/or water from the atmosphere as all powders were dried at 400°C after milling. As seen from Fig. 3, the main weight loss of milled powders was found between 700 and 800°C for SC and around 900°C for SC5Yb. This weight loss is associated with decomposition of SrCO3. The higher loss (4.0%) for SC5Yb compared to SC (1.6%) is in agreement with a larger amount of SrCO3 in the SC5Yb powder, as confirmed by XRD. A total weight loss of ~0.3% was observed for SC powder calcined at 1100°C. This indicates little or no presence of strontium carbonate after calcination at 1100°C.

Temperature (°C)

100 300 500 700 900 1100 1300 1500

Weig

ht

ch

an

ge

(%

)

96

97

98

99

100

SC - 900°C, milled

SC5Yb - 900°C, milled

SC - 1100°C, not milled

Fig. 3 TG curves obtained from milled SrCeO3 (SC) and

SrCe0.95Yb0.05O3-δ (SC5Yb) powders calcined at 900°C as well as

SC powder calcined at 1100°C before milling.

83

IR spectra of SC / SC5Yb powders calcined at different temperatures are presented in Fig. 4. A reference spectrum of SrCO3 is also included. The first broad band in the range of 500 to 800 cm-1, observed for all prepared powders, is due to the stretching of the metal-oxygen bonds [24]. The characteristic frequencies for SrCO3 are also found in the spectra of SC and SC5Yb powders calcined at 900°C, respectively. Higher intensity of the carbonate bands, indicate larger amount of SrCO3 in the SC5Yb powder compared to SC. SC powders calcined at 1000°C and 1100°C were carbonate free according to the IR spectra. SC5Yb powders calcined at higher temperatures (1000 - 1100°C) show similar behaviour as observed for SC. IR spectra of ball milled powders indicated that SrCO3 were reintroduced by CO2 adsorption from the air or from reaction with residual milling liquid in the powder.

Wave number (cm-1)

400 1000 1600 2200 2800 3400 4000

Re

lati

ve t

ran

sm

itta

nc

e

SrCO3SrCeO

3

SrCe0.95Yb0.05O2..975

900°C

900°C

1000°C

1100°C

Fig. 4 IR spectra of SrCO3 powder, SrCe0.95Yb0.05O3-δ and SrCeO3

powders calcined 900°C in air as well as SrCeO3 powders

calcined at 1000°C and 1100°C in nitrogen atmosphere.

84

Densification

The linear shrinkage versus temperature presented in Fig. 5 a) show an onset for densification at ~800°C for SC and ~1000°C for SC5Yb, both calcined at 900°C. The presintering feature above 900°C for SC5Yb is most likely due to decomposition of residual SrCO3, resulting in evolving CO2 gas that could cause a small expansion in the material. Calcination at 1100°C lead to a small shift in the onset temperature for densification. Though the onset temperature for densification is lower for SC, the sintering rate for this compound is lower than for SC5Yb, as displayed in Fig. 5 b). The highest sintering rate (Fig. 5 b)) for SC calcined at 900°C was found at ~1150°C and as high as ~1375°C for SC calcined at 1100°C. In comparison the highest sintering rate for SC5Yb calcined at 900°C and 1100°C was ~1330°C and ~1370°C, respectively.

Temperature (°C)

200 400 600 800 1000 1200 1400

d(d

L/L

0) /

dT

(%

/°C

)

-0.3

-0.2

-0.1

0.0

dL

/L0 (

%)

-20

-15

-10

-5

0

SC - 900°C

SC - 1100°C

SC5Yb - 900°C

SC5Yb - 1100°C

b)

a)

Fig. 5 a) Linear shrinkage and b) sintering rate from associated

derivative curves for SrCeO3 (SC) and SrCe0.95Yb0.05O3-δ

(SC5Yb) after calcination at 900°C and 1100°C.

85

Fig. 6 displays the variation in density and estimated grain size as a function of temperature during isothermal sintering of SC and SC5Yb. The presintering heat treatment at 850°C for 24 hrs for SC and 1000°C for 1 h for SC5Yb resulted in 1 - 4% reduction of closed porosity. As seen in Fig. 6 a), the density of SC and SC5Yb increases when the temperature is increased from 1250 to 1350°C, however further increase in temperature gave little change in density. After sintering at 1350°C the density was 96.6 and 97.6% of theoretical density for SC and SC5Yb, respectively.

Den

sit

y (

%)

85

90

95

100

SCSC5Yb

a)

Temperature (°C)

1250 1300 1350 1400 1450

Gra

in s

ize (

µm

)

0

5

10

15

b)

Fig. 6 a) density and b) estimated grain size of SrCeO3 (SC) and

SrCe0.95Yb0.05O3-δ (SC5Yb) specimens sintered at indicated

temperatures for 2 hrs in ambient air.

86

Microstructure and grain growth

Estimated grain size in sintered SC and SC5Yb specimens are presented in Fig. 6 b). The grain size in SC increased from ~6 to ~14 µm when the sintering temperature was increased from 1250 to 1450°C. For SC5Yb the grain growth was less severe giving average grain size in the range of 1 to 2 µm after sintering at 1250 - 1400°C. Sintering at 1450°C, however, resulted in grain size of ~8 µm for this composition as well. The homogenous microstructures of sintered SC and SC5Yb specimens are displayed by SEM micrographs in Fig. 7. The microstructure evolution is shown in the left and right row of micrographs for SC and SC5Yb, respectively, with increasing sintering temperature from 1250 to 1450°C downwards in the figure. While the grain size of the SC specimens were rather large, even after sintering at 1250°C, SC5Yb specimen sintered at this temperature exhibit grains of ~1 µm, however this specimen appears not fully dense. The inevitable amount of closed porosity observed in SC specimens, due to higher mobility of grain boundaries compared to pores, was not observed in the SC5Yb specimens sintered at temperatures up to 1400°C. Sintering at 1450°C produced larger grains resulting in some enclosed pores for SC5Yb as well.

The severe grain growth in SC resulted in microcracking reducing the mechanical properties of this material. This may be due to the anisotropic properties of the orthorhombic SrCeO3. Microcracking was not observed in the SC5Yb specimens, which may indicate critical grain size for microcracking in in the range 2 - 6 µm, explaining why this was observed for the larger grained SC materials.

87

Fig. 7 SEM micrographs of polished and etched (1200°C, 12 min)

surfaces of SrCeO3 and SrCe0.95Yb0.05O3-δ specimens sintered for

2 hrs 1250°C - 1450°C.

88

Electrical properties

The partial conductivities as a function of the inverse temperature for SC5Yb in wet oxygen (0.025 atm H2O + 0.975 atm O2) and wet hydrogen (0.025 atm H2O + 0.975 atm H2) are presented in Fig. 8 a) and b), respectively. These figures include the total AC (10 kHz) conductivity measured in continuous temperature ramps (12°C/h), along with partial conductivities deconvoluted from impedance spectroscopy at selected temperatures in wet hydrogen (Fig. 8 b)). The proton and oxygen ion conductivities are similar under oxidizing and reducing conditions, whereas the electronic contribution under oxidizing conditions (p-type) is higher than under reducing conditions (n-type). One should note that the OCV of concentration cells has only been measured in the temperature region where the proton transport number is less than 1 and below the melting point of gold (700 to 1050°C). The total conductivities measured at temperatures lower than ~700°C are predominantly protonic, although this is not indicated in Fig. 8, since it has not been directly measured. The grain boundary resistance in this material was low and, as mentioned, the bulk and grain boundary impedance could not be separated above 650°C.

The water vapor partial pressure dependence of the total conductivity from 500 to 1000°C, as measured in hydrogen, is shown in Fig. 9 a). The total conductivity increases as a function of increasing water vapor partial pressure. The typical conductivity behavior observed for a mixed ionic-electronic conductor upon large variations in the oxygen partial pressure, is shown in Fig. 9 b), for temperatures in the range 800 - 1000°C. The total conductivity increases with decreasing oxygen partial pressures under reducing conditions, and increases with increasing oxygen partial pressures under oxidizing conditions. These two behaviors reflect an increasing n-type and p-type contribution to the total conductivity, respectively. Between these two regimes the total conductivity is essentially independent of the oxygen partial pressure reflecting ionic conduction. Below 800°C under reducing conditions, the conductivity is predominantly ionic showing no significant n-type electronic conduction. One should note that the total conductivity degrades with time, and that this seems to depend particularly on the time measured under wet conditions and at low temperature.

89

a)

1000/T (K-1)

0.8 1.0 1.2 1.4 1.6 1.8

Co

nd

uc

tivit

y (

S/c

m)

10-5

10-4

10-3

10-2

Temperature (°C)

1000 800 600 400

σTotal(10 kHz)

σElectron

σProton

σOxygen ion

b)

1000/T (K-1)

0.8 1.0 1.2 1.4 1.6 1.8

Co

nd

uc

tivit

y (

S/c

m)

10-5

10-4

10-3

10-2

σTotal(10 kHz)

σTotal(IS)

σElectron

σProton

σOxygen ion

Temperature (°C)

1000 800 600 400

Fig. 8 Total conductivity and partial conductivities as a function of the

inverse absolute temperature for SrCe0.95Yb0.05O3-δ in a) wet

oxygen and b) wet hydrogen. σTotal(10 kHz) and σTotal(IS) are

obtained from a continuous ramp at 10 kHz and impedance

spectroscopy with 50°C steps, respectively.

90

a)

Water partial pressure (atm)

10-3 10-2

To

tal co

nd

uc

tivit

y (

S/c

m)

10-4

10-3

10-2

500°C

600°C

700°C

800°C

900°C

1000°C

b)

Oxygen partial pressure (atm)

10-20 10-15 10-10 10-5 100

To

tal c

on

du

cti

vit

y (

S/c

m)

0.010

1000°C

900°C

800°C

0.050

0.005

Fig. 9 Total conductivity of SrCe0.95Yb0.05O3-δ specimens with a)

varying water vapor partial pressure, measured in hydrogen at

500 - 1000°C, and b) varying oxygen partial pressure in the

range of 800 - 1000°C.

91

Discussion

Powder characteristics

As demonstrated in this work, phase pure SrCeO3 powders can be prepared by spray pyrolysis followed by calcination at 1100 - 1200°C in nitrogen atmosphere. Fine powders with particle size down to ~60 nm have been obtained by ball milling of calcined powders, however some coarsening was observed with increasing calcination temperature. The TG analysis (Fig. 3) of SC powders in air indicates that SrCO3 introduced during ball milling is removed below 800°C, in agreement with reported onset temperature for synthesis of SC from a mix of SrCO3 and CeO2 in nitrogen [6]. The IR spectrum of SC calcined at 900°C (Fig. 4), however, indicate that small amount of SrCO3 is present in the powder even before milling. A larger amount of SrCO3 in the SC5Yb powder calcined at 900°C, confirmed by XRD (Fig. 2 b)) and IR (Fig. 4)) and assisted by the larger weight loss from TG analysis (Fig. 3), indicate a stabilization of this secondary phase by substitution with Yb. Strontium rich secondary phase (Sr2CeO4) as well as ceria (CeO2) were also more dominating in the synthesized SC5Yb powder, the former being detected by XRD in the SC5Yb powder after calcination at 1100°C, but not in the SC powder after calcination at 1000°C. These features may be explained by the more basic nature of Yb-substituted material, as reported for ceria [25], which in turn e.g. will make the material more susceptible to CO2.

Densification and microstructure

By introducing a heat treatment at 850 and 1000°C prior to sintering, dense SC and SC5Yb materials, with well defined microstructures, were obtained by sintering for 2 hrs in air at 1350 - 1400°C. SC was found to sinter at a lower temperature compared to SC5Yb and onset temperatures for sintering being around ~200°C higher for the latter. The higher sintering temperature in the Yb-substituted material can not be assigned to any ionic size effects, as the ionic radius of Yb3+ equals that of Ce4+ [27]. The assisted introduction of vacancies in the material is also unlikely to increase the sintering temperature. Increased mobility of oxygen ions would rather aid the sintering, however, the cations are more likely the rate limiting species. The shift in onset temperature for sintering may be explained by the presence of secondary phase, inhibiting the sintering in the SC5Yb material. Sr2CeO4 was detected in the SC5Yb powder, even after calcination at 1100°C. All secondary phases seem to vanish at higher temperatures (1200°C), however, at such temperatures coarsening of the powder will reduce the sinterability as well. Despite lower onset temperature, the sintering rate was higher in

92

the SC5Yb material compared to SC, as demonstrated in Fig. 5. This effect may be explained by larger surface area in starting powders compared to that of SC. Densities of 96-98% in sintered SC and SC5Yb materials sintered at 1350 - 1450°C in air are comparable to reported densities >98% obtained by solid state ceramic method and sintering at 1550 - 1600°C in air for 10 - 12 hrs [16,17], and from powders prepared by complexation route, by sintering at 1300°C for 12 hrs in nitrogen flow [9].

Homogenous, fine grained microstructures were obtained for the SC5Yb specimens sintered up to 1400°C. The more severe grain growth in the SC materials, with resulting trapped pores indicates a high mobility of grain boundaries compared to pores. The reduced growth in SC5Yb may, as the sinterability, be explained by the presence of secondary phases in the material. Segregation of secondary phases on the grain boundaries will pin these and inhibit the growth. At elevated sintering temperatures the secondary phases will disappear (presuming stoichiometric cation ratio in the starting powder), alleviating the grain boundary pinning. This may be the effect observed in SC5Yb when increasing the sintering temperature from 1400 to 1450°C, resulting in remarkable grain growth.

Due to the anisotropic properties of SrCeO3, microcracking was observed in large grained SC materials (> 6 µm), while the finer grained (1 - 2 µm) SC5Yb materials were resistant towards microcracking. Based on these observations the critical grain size may be estimated between 2 and 6 µm. In a comparable work, grain size of ~7 µm have been reported for dense SC5Yb materials [9], however studies of microstructure and grain size in these materials are rare. The commonly used solid state ceramic route produces coarse powders, hence, high sintering temperatures (>1500°C) are needed to obtain dense materials. Consequently, the sintered materials will consist of rather large grains (10 - 20 µm) [13]. This exceeds the suggested critical size for microcracking, which in turn may help explain the poor mechanical properties observed for SrCeO3 materials [7]. SrCeO3 based materials with improved mechanical properties have been prepared by introducing SrZrO3 (SrCe1-xZrxO3), however, the lower conductivity of SrZrO3 based materials compared to SrCeO3 affects the final composite as well [18]. Observations made in the present work, points out the importance of high quality starting powders and controlled densification, with respect to the mechanical properties of the sintered SrCeO3-based materials.

93

Conductivity

The conductivity results from this investigation generally reflect the behaviour reported in the literature on acceptor substituted SrCeO3

[1,2,5,12-

18,28-33]. SrCe0.95Yb0.05O3-δ exhibit mixed ionic-electronic conduction where the ionic contribution is predominated by protons below ~1000°C. The proton conductivity reaches a maximum of ~0.005 S/cm above 900°C. Oxygen ion conductivity becomes the major ionic charge carrier above 1000°C. The p-type conductivity is considerably higher under oxidizing conditions than the corresponding n-type conductivity under reducing conditions. This conductivity behavior can be interpreted with basis in a few elementary point defect reactions. Substitution of trivalent Yb for fourvalent Ce yields an effectively negative defect (acceptor) that must be charge-compensated either by consumption of other effectively negative or formation of effectively positive defects. In SrCeO3 charge-compensation occurs by formation of oxygen vacancies according to Eq. 3.

xOO

xSr

'Ce32 O5vSr2Yb2SrO2OYb +++=+

•• (3)

Under wet conditions oxygen vacancies may be hydrated through interaction with water vapor. The protonic defect that forms is assumed to be associated with a structural oxygen ion through the reaction in Eq. 4.

•••=++ O

xOO2 OH2Ov)g(OH (4)

Native defects are furthermore in equilibrium with the surrounding atmospheres, as shown in Eq. 5, for oxygen vacancies in equilibrium with electron holes predominating under high-temperature oxidizing conditions.

)g(O2

1vh2O 2O

xO +=+

••• (5)

By taking the intrinsic ionization between electron holes and electrons into account all the point defects necessary to model the behavior encountered during this investigation are in place. However, in order to obtain a full mathematical description of the defect structure, and to determine physicochemical parameters reflecting the conductivity characteristics, the site balance and the electroneutrality must be included [34]. Within the experimental window of the present investigation, acceptor doping in SrCeO3 has, generally, been concluded to be charge compensated predominantly by oxygen vacancies and protons. Hence, the electroneutrality condition may be expressed by Eq. 6.

94

[ ] [ ] [ ]•••+= OO

'Ce OHv2Yb (6)

By combining this electroneutrality and the site balance with the equilibrium expression in Eq. 4, the concentration of protons and oxygen vacancies may be resolved. Moreover, the conductivity of the different charge carriers (indexed i) is proportional to their concentration (ci) and mobility (µi), σi = zieciµi. On these bases the dependencies of the partial and total conductivities with variations in the conditions can be modeled. Thermodynamic parameters extracted from this modeling include the standard entropy and enthalpy of the hydration reaction in Eq. 4, ∆S0 = -125

± 5 J/molK, ∆H0 = -145 ± 10 kJ/mol and, moreover, the preexponential mobility and the activation enthalpy of defect mobility for the protons and oxygen vacancies: µ0(H

+) = 20 ± 5 cm2K/Vs, ∆Hmob(H+) = 55 ± 5 kJ/mol,

µ0(••

Ov ) = 20 ± 10 cm2K/Vs, ∆Hmob(••

Ov ) = 60 ± 5 kJ/mol. Since it is not

possible to resolve independent expressions for the concentration of electrons and electron holes, only apparent activation energies for p- and n-type conductivity can be listed: EA(h·) = 65 kJ/mol and EA(e) = 350 kJ/mol. As a first approximation in the modeling the entropy of the hydration reaction was assumed to be -120 J/mol and, furthermore, the concentrations of protons and oxygen vacancies were assumed to be small relative to the concentration of oxygen sites [34-36].

Fig. 10 a) demonstrates, by way of example, the fit of the parameter-set to the water vapor partial pressure dependence in hydrogen. Moreover, Fig. 10 b) illustrates how the different partial conductivities individually influence the overall functional water vapor partial pressure dependence at 500 and 1000°C. One may note here how protons predominate at the low temperature as a consequence of the exothermic nature (negative enthalpy) of the hydration of oxygen vacancies. Oxygen ion conductivity barely influences the total conductivity even under dry conditions at 500°C, and n-type electronic conduction is orders of magnitude lower than the ionic conductivity. By increasing the temperature the influence of protons decreases, as indicated by reduced proton concentration relative to the oxygen vacancy concentration. At high temperature, under these rather reducing conditions, also n-type electronic conduction comes into play, particularly at high water vapor partial pressures. The shift in the relative influence of the three charge-carriers within the 500°C temperature interval explains the observed change in curvature of the functional water vapor partial pressure dependence (Fig. 10 a) and b)).

95

a)

Water partial pressure (atm)

10-5 10-4 10-3 10-2 10-1 100

To

tal co

nd

uc

tivit

y (

S/c

m)

10-4

10-3

10-2

500°C

600°C

700°C

800°C

900°C

1000°C

b)

Water partial pressure (atm)

10-5 10-4 10-3 10-2 10-1 100

Co

nd

ucti

vit

y (

S/c

m)

10-8

10-7

10-6

10-5

10-4

10-3

10-2σ

Total(1000°C)

σProton(1000°C)

σOxygen ion(1000°C)

σElectron(1000°C)

σTotal(500°C)

σProton(500°C)

σOxygen ion(500°C)

σElectron(500°C)

Fig. 10 a) Total conductivity curves fit to the experimental data for

SrCe0.95Yb0.05O3-δ, for varying water vapor partial pressure, in

hydrogen from 500 to 1000°C. b) Total and partial

conductivities with varying partial pressure of water vapor at

500°C (bold curves) and 1000°C (slim curves). Circles and

lines experimental results data and modeled conductivity curves,

respectively.

96

Despite the extensive literature on the functional properties of the acceptor substituted SrCeO3, only a few investigations have so far derived similar physicochemical properties as reported here. The thermodynamics of the dissolution of water according to the reaction in Eq. 4 have been measured by thermogravimetry and also estimated from conductivity measurements [17,37,38]. Hydration enthalpies are reported in the range from -115 to -160 kJ/mol, which correspond nicely with the values obtained here.

The degradation in the total conductivity encountered during these experiments may have several reasons. The basic high-temperature proton conducting perovskites, in particular BaCeO3 and SrCeO3, are renowned for their reactivity towards CO2. However, since gas mixtures with relatively low levels of CO2 have been applied during this investigation, this is not believed to cause the decline observed in the total conductivity with time. More likely this is a result of SrCeO3 decomposeing to Sr(OH)2 and CeO2

[7], due to long-term exposure in wet atmospheres at relatively low temperature (slow temperature ramps down to 300°C). It should also be recognized that SrCeO3 seems to react with Pt and that this may change both the properties of the material and affect the processes at the electrode.

Conclusions

Spray pyrolysis has been proven a good method for preparing powders of orthorombic SrCeO3 and SrCe0.95Yb0.05O3-δ. Effective breaking of agglomerate spheres formed during spray pyrolysis was done by ball milling in iso-propanol resulting in particle size down to 0.06 µm. All secondary phases present in the as-synthesized SrCeO3 and SrCe0.95Yb0.05O3-δ powders can be removed by calcination in nitrogen atmosphere at 1100 and 1200°C, respectively. Dense (>96%) materials with homogenous microstructures have successfully been prepared. Due to lower mobility of grain boundaries in the material the grain size in the SrCe0.95Yb0.05O3-δ (1 - 2 µm) materials were somewhat lower than those found in SrCeO3 (>6 µm). Some closed porosity and microcracking was observed in the SrCeO3 materials, while the fully dense (~98%) SrCe0.95Yb0.05O3-δ materials seemed were free of cracks and closed pores. The electrical properties of the Yb-substituted materials were in very good agreement with literature data, showing p-type conductivity in oxidizing and n-type conductivity in reducing atmosphere. The ionic contribution to the total conductivity is predominately from protons below 1000°C and the protonic conductivity reaches a maximum of 0.005 S/cm at 1000°C.

97

Acknowledgement

Work was supported by the Research Council of Norway, Grant No. 1585171431 (NANOMAT). Øystein Andersen is acknowledged for operating the spray pyrolysis.

References

[1] H. Iwahara, T. Esaka, H. Uchida, N. Maeda, Proton conduction in

sintered oxides and it’s application to steam electrolysis for

hydrogen production, Solid State Ionics, 3/4, 359-363 (1981) [2] H. Iwahara, Technological challenges in the application of proton

conducting ceramics, Solid State Ionics, 77, 289-298 (1995) [3] T. Norby, Solid-state protonic conductors: principles, properties,

progress and prospects, Solid State Ionics, 125, 1-10 (1999) [4] T. Schober, Applications of oxidic high-temperature proton

conductors, Solid State Ionics, 162-163, 277-281 (2003) [5] T. Yajima, H. Iwahara, Studies on proton behavior in doped

perovskite-type oxides: (II) Dependence of equilibrium hydrogen

concentration and mobility on dopant content in Yb-doped SrCeO3, Solid State Ionics, 53-56, 983-988 (1992)

[6] M.J. Scholten, J. Schoonman, Synthesis of strontium and barium

cerate and their reaction with carbon dioxide, Solid State Ionics, 61, 83-91 (1993)

[7] K.-D. Kreuer, On the development of proton conducting materials

for technological applications, Solid State Ionics, 97, 1-15 (1997) [8] A.N. Shirsat, K.N.G. Kaimal, S.R. Bharadwaj, D. Das,

Thermodynamic stability of SrCeO3, J. Solid State Chemistry, 177, 2007-2013 (2004)

[9] K.J. de Vires, Electrical and mechanical properties of proton

conducting SrCe0.95Yb0.05O3-α, Solid State Ionics, 100, 193-200 (1997)

[10] S. Cheng, V.K. Gupta, J.Y.S. Lin, Synthesis and hydrogen

permeation properties of asymmetric proton-conducting ceramic

membranes, Solid State Ionics, 176, 2653-2662 (2005) [11] S.V. Chavan, A.K. Tyagi, Preparation of Sr0.09Ce0.91O1.91, SrCeO3,

and Sr2CeO4 by glycine-nitrate combustion: Crucial role of oxidant-

to-fuel ratio, J. Mater. Res., 19 (11), 3181-3188 (2004) [12] T. Scherban, A.S. Nowick, Bulk protonic conduction in Yb-doped

SrCeO3, Solid State Ionics, 35, 189-194 (1989)

98

[13] N. Matsunami, T. Yajima, H. Iwahara, Permeation of implanted

deuterium through SrCeO3 (5% Yb), Nuclear Instruments and Methods in Physics Research, B65, 278-281 (1992)

[14] I. Kosacki, J. Schoonman, M. Balanski, Raman scattering and ionic

transport in SrCe1-xYbxO3, Solid State Ionics, 57, 345-351 (1992) [15] H. Iwahara, T. Yajima, T. Hibino, K. Ozaki, H. Suzuki, Protonic

conduction in calcium, strontium and barium zirconates, Solid State Ionics, 61, 65-69 (1993)

[16] I. Kosacki, H.L. Tuller, Mixed conductivity in SrCe0.95Yb0.05O3

protonic conductors, Solid State Ionics, 80, 223-229 (1995) [17] U. Reichel, R.R. Arons, W. Schilling, Investigation of n-type

electronic defects in the protonic conductor SrCe1-xYxO3-α,, Solid State Ionics, 86-88, 639-645 (1996)

[18] T. Matzeke, M. Cappadonia, Proton conductive perovskite solid

solutions with enhanced mechanical stability, Solid State Ionics, 86-88, 659-663 (1996)

[19] S.V. Chavan, A.K. Tyagi, Sub-solidus phase equilibria in CeO2-SrO

system, Thermochimica Acta, 390, 79-82 (2002) [20] H. Matsumoto, T. Shimura, H. Iwahara, T. Higuchi, K. Yashiro, A.

Kaimai, T. Kawada, J. Mizusaki, Hydrogen separation using proton-

conducting perovskites, Journal of Alloys and Compounds, 408-412, 456-462 (2006)

[21] H. Taherparvar, J.A. Kilner, R. Baker, M. Sahibzada, Effect of

humidification at anode and cathode in proton-conducting SOFCs, Solid State Ionics, 162-163, 297-303 (2003)

[22] T. Norby, "Electronic Book", Norwegian Electro Ceramics AS, www.norecs.com

[23] T. Norby, EMF method determination of conductivity contributions

from protons and other ions in oxides, Solid State Ionics, 28-30, 1586-1591 (1988)

[24] D.P. Sutija, T. Norby, P. Björnbom, Transport number

determination by the concentration-cell / open-circiut voltage

method for oxides with mixed electronic, ionic and protonic

conductivity, Solid State Ionics, 77, 167-174 (1995) [25] K. Nakamoto, Infrared and Raman Spectra of Inorganic and

Coordination Compounds, 2nd ed., John Wiley & Sons Inc., New York (1997)

[26] T. Mokkelbost, I. Kaus, T. Grande, M.-A. Einarsrud, Combustion

synthesis and characterization of nanocrystalline CeO2-based

powders, Chem. Mater., 16, 5489-5494 (2004) [27] R.D. Shannon, Revised effective ionic radii and systematic studies of

interatomic distances in halides and chalcogenides, Acta Cryst., A32, 751-767 (1976)

99

[28] H. Uchida, N. Maeda, H. Iwahara, Relation between proton and hole

conduction in SrCeO3-based solid electrolytes under water-

containing atmospheres at high temperatures, Solid State Ionics, 11, 117-124 (1983)

[29] H. Uchida, H. Yoshikawa, T Eseka, S. Ohtsu, H. Iwahara, Formation

of protons in SrCeO3-based proton conducting oxides. Part II.

Evaluation of proton concentration and mobility in Yb-doped

SrCeO3, Solid State Ionics, 36, 89-95 (1989) [30] T. Yajima, H. Iwahara, H. Uchida, K. Koide, Relation between

proton conduction and concentration of oxide ion vacancy in SrCeO3

based sintered oxides, Solid State Ionics, 40/41, 914-917 (1990) [31] T. Hibino, K. Mizutani, T. Yajima, H. Iwahara, Evaluation of proton

conductivity in SrCeO3, BaCeO3, CaZrO3 and SrZrO3 by

temperature programmed desorption method, Solid State Ionics, 57, 303-306 (1992)

[32] H. Iwahara, Proton conducting ceramics and their applications, Solid State Ionics, 86-88, 9-15 (1996)

[33] T. Schober, F. Krug, W. Schilling, Criteria for the application of

high temperature proton conductors in SOFCs, Solid State Ionics, 97, 369-373 (1997)

[34] R. Haugsrud, T. Norby, Proton conduction in rare-earth ortho-

niobates and ortho-tantalates, Nature Materials, 5, 193196 (2006) [35] K.D. Kreuer, Aspects of the formation and mobility of protonic

charge carriers and the stability of perovskite-type oxides, Solid State Ionics, 125, 285-302 (1999)

[36] Y. Larring, Protons and oxygen vacancies in acceptor-substituted

rare earth oxides, PhD. Thesis, University of Oslo (1998) [37] T. Yajima, H. Iwahara, Studies on behavior and mobility of protons

in doped perovskite-type oxides (I) In situ measurement of hydrogen

concentration in SrCe0.95Yb0.05O3-α at high temperatures, Solid State Ionics, 50, 281-286 (1992)

[38] F. Krug, T. Schober, T. Springer, In situ measurements of the water

uptake in Yb doped SrCeO3, Solid State Ionics, 81, 111-118 (1995)

100

PAPER IV

102

103

Preparation and characterization of Barium Zirconate ceramics

Paul Inge Dahl, Hilde Lea Lein, Tor Grande, Mari-Ann Einarsrud †

Department of Materials Science and Engineering Norwegian University of Science and Technology

NO-7491 Trondheim, Norway

Christian Kjølseth, Truls Norby, Reidar Haugsrud

Centre for Materials Science and Nanotechnology Department of Chemistry, University of Oslo

NO-0349 Oslo, Norway

Abstract

Powders of pure and 10% yttrium substituted barium zirconate (BaZrO3 / BaZr0.9Y0.1O2.95) were prepared by spray pyrolysis of nitrate salt solutions. The crystalline powders were calcined at 1000°C to remove secondary phases and agglomerates were effectively broken down by ball milling giving particle size in the range 0.09 – 0.17 µm. Despite of similar characteristics of the powders, the densification properties were poorer for the yttrium substituted material. Severe grain growth was observed during conventional sintering (1600°C) of BaZrO3 resulting in average grain size up to 18 µm and only < 92% relative density. Dense BaZrO3 and BaZr0.9Y0.1O2.95 materials (~98%) were prepared by hot pressing. The mobility of the grain boundaries was efficiently suppressed by application of uniaxial pressure during sintering, resulting in homogenous microstructures and average grain size down to 0.42 µm. The electrical properties of BaZr0.9Y0.1O2.95 are in agreement with the literature, and high grain boundary resistance was observed for both materials. Slightly lower conductivity was observed for BaZrO3 in wet compared to dry atmosphere.

† Corresponding author. E-mail: mari-ann.einarsrud@material.ntnu.no

104

Introduction

Perovskite oxides (ABO3) such as barium and strontium based cerates and zirconates (A = Ba, Sr and B = Ce, Zr) are well known for their electrical properties. Substituted with rare earth oxides these materials exhibit high proton conductivity. Potential application areas for such ceramic proton conductors include gas sensors, fuel cells and membrane technology [1]. Barium zirconate (BaZrO3) based materials exhibit lower proton conductivity compared to the isostructural barium cerates (BaCeO3). This is due to grain boundaries being highly resistive to proton transport, and consequently decreasing the conductivity by several orders of magnitude which is detrimental to applications like electrolytes in an SOFC. The reason for the high grain boundary resistance is not yet understood. Despite higher conductivity reported for the cerate based ceramics, zirconate materials seem superior with respect to chemical and mechanical properties [2-5]. The idea of combining the properties of BaCeO3 and BaZrO3 has been studied in mixed systems (BaCe0.9-xZrx Y0.1O3-δ) where increasing chemical stability against CO2 was reported with increasing Zr content. However, the increasing Zr content resulted in decreased proton conductivity [6,7]. Consequently, BaZrO3 substituted with trivalent rare earths such as Y or Yb has become state of the art materials among ceramic high temperature proton conductors.

The conductivity of yttrium substituted BaZrO3 is well documented in the literature [8-17]. For all these studies ceramic powders have been prepared by solid state reaction of oxides, carbonates and/or acetates at temperatures in the range 1200 – 1450°C. This method generally produces coarse powders. Fine powders with more defined morphology have been prepared by co-precipitation [18,19], sol-gel techniques [20,21] and hydrothermal synthesis [21].

A challenge in preparation and use of Ba/Sr-based zirconates (and cerates) is possible reactions with CO2 as indicated for BaZrO3 in Eq. (1).

BaZrO3(s) + CO2(g) = BaCO3(s) + ZrO2(s) (1)

The presence of BaCO3 has not been reported in BaZrO3 powders or dense specimens heat treated at high temperatures (>1000°C). However, as the equilibrium constant for Eq. (1) equals 1 around 320°C in air [4], formation of BaCO3 may occur during cooling. The stability of carbonates can narrow the temperature region of use for these ceramic proton conductors. It is not clear how the presence of BaCO3 affects the conductivity.

At high temperatures, BaZrO3 can decompose according to Eq. (2) as documented by thermodynamic studies [22,23].

105

BaZrO3(s) = BaO(g) + ZrO2(s) (2)

Evaporation of BaO(g) has been observed on the surface of Y-substituted BaZrO3 specimens heat treated above 1200°C and above 1500°C, bulk is affected as well [20]. These observations indicate that preparation of dense and phase pure BaZrO3 ceramics might be challenging.

No thorough sintering studies of BaZrO3 ceramics seem to be available. Dense (> 95%) BaZrO3 from powders made by solid state method can only be obtained by conventional sintering above 1700°C [7,11,12,15,16]. Using spark plasma sintering, dense (>95%) ceramics of pure and Y-substituted BaZrO3 can be obtained at 1500 and 1600°C, respectively [24]. Besides the latter work there are no consistent densification routes for controlled microstructure and grain size available for these materials unless undesirable sintering aids such as ZnO is added [25].

Based on the above information the preset study aims to establish reproducible procedures for preparation of dense BaZrO3 materials with designed microstructure. The quality of the starting powders is important for this matter. Hence, the first part of this deals with the preparation of pure and 10% Y-substituted BaZrO3 powders with special focus on the mentioned problem with formation of BaCO3. Two different sintering techniques have been used for densification of BaZrO3 specimens (from prepared powders) in order to design different microstructures. The materials have been thoroughly characterized, and being the most central property of the BaZrO3 materials, the electrical conductivity has been investigated. The conductivity measurements reported herein are, as such, preliminary results from a large matrix of measurements on Y-substituted BaZrO3 where the overall purpose is to contribute to the understanding of the grain boundary effects. The different electrical behaviour of the nominally BaZrO3 compared to the Y-substituted material are discussed in the present work, however, for a deeper discussion on the conductivity characteristics of acceptor substituted BaZrO3 and in particular of the effects of grain boundary resistance, forth-coming papers should be consulted [26].

Experimental

Powder synthesis

Powders of BaZrO3 (BZ) and BaZr0.90Y0.10O2.95 (two similar batches: YBZ-1, YBZ-2) were synthesized by spray pyrolysis of nitrate salt solutions. Aqueous solutions of zirconyl nitrate (ZrO(NO3)2·xH2O, Acros, 99.5%) and

106

yttrium nitrate (Y(NO3)3·6H2O, Merck, 99.9%) were standardized by thermogravimetry, mixed in stoichiometric proportions with dried (250°C, 24 hrs) barium nitrate powder (Ba(NO3)2, Merck, 99%). Due to the limited solubility of Ba(NO3)2 (10.1 g / 100 g H2O), the total cation concentration in the mixed solution was 0.2 M. The spray pyrolysis was performed with a feed rate of 2 L/h. The temperature by the nozzle was 840-845°C and the output temperature was ~560°C. The as-synthesized powders were calcined at 1000°C for 48 hrs in ambient air and 1100°C for 6 hrs in nitrogen flow. Calcined powders were ball milled with yttria stabilized zirconia (YSZ) balls in isopropanol for 24 hrs, dried at 400°C for 24 hrs in ambient air and sieved (150 µm).

Sample preparation

Green bodies were made by uniaxial pressing at 64 – 128 MPa giving green densities in the range 44 – 51% of theoretical value. Conventional sintering of BZ and YBZ-1 was performed in a muffle furnace (Entech, SF-4/17) at 1600°C in ambient air with a presintering step at 1000°C for 24 hrs. For BaZrO3, the sintering time was varied from 1 to 18 hrs while the sintering time for BaZr0.9Y0.1O2.95 was 6 hrs. Two different hot presses were used, depending on desired temperature region and applied pressure. Hot pressing (HP) up to 1500°C was performed in a clam furnace (Entech, VSTF 40/15) under flowing nitrogen gas, with an applied pressure of 25 MPa. A commercial hot press (Thermal Technology Inc. HP50-7010G) was used for hot pressing at 1450 – 1750°C in flowing argon gas, with an applied pressure of 50 MPa. Cylindrical graphite pressing dies (15 – 25 mm) was used. The uniaxial pressure was applied after heating to sintering temperature with a rate of 600°C/h. The sintering time was varied from 1 to 6 hrs and the pressure was released before cooling to room temperature at 600°C/h. As the HP was performed in reducing atmosphere the specimens were reoxidized at 1400°C in flowing O2 for 24 hrs after sintering.

Characterization

Thermogravimetric analysis (TGA) (Netzsch STA 449 C Jupiter) was performed on as-synthesized powders (BZ and YBZ-1) by heating to 1500°C in air with a heating rate of 3°C/min. As-synthesized and calcined powders as well as surface of sintered specimens were analyzed by X-ray diffraction (XRD) (Cu Kα, Phillips PW1730/10). Powders from crushed samples sintered at 1500°C for 6 hrs and added 30 weight% Si standard were used to obtain XRD data from which the lattice parameters and

107

theoretical densities were calculated by the Rietveld method. Infrared spectroscopy (IR) was performed on potassium bromide (Merck, KBr for IR spectroscopy Uvasol®) pellets containing 1 wt% BZ or YBZ-1 powder using a Bruker IFS 66v spectrophotometer. Nitrogen adsorption on de-gassed powders (250°C, 24 hrs) was measured (Micromeritics Tristar) using the 5 point BET equation to give the specific surface area. The particle size was calculated from these results assuming spherical particles.

Dilatometer studies (Netzch, DIL 402 C / 402 E) were performed in air using a heating rate of 3°C/min up to 1500°C and 1600°C for BaZrO3 (BZ) and BaZr0.9Y0.1O2.95 (YBZ-1 and YBZ-2), respectively. Densities of sintered specimens were measured by the Archimedean method (ISO 5017) in isopropanol. The microstructure of synthesized powders and dense specimens (oxidized/etched at 1400°C) was studied by scanning electron microscopy (SEM) (Hitachi S-3500N). Average grain size of sintered specimens was calculated by the linear intercept method [27] over a minimum of 50 grains.

Electrical properties

Electrical characterization was performed on (BZ), with > 95 % density, and (YBZ-1), with > 88% density. Circular Pt electrodes (~10 mm diameter) were attached to each side of the specimens and mounted in ProbostatTM measurement cells [28]. The electrical characterization was carried out in wet (p(H2O) ~ 0.025 atm) and dry (p(H2O) ~ 3·10-5 atm) oxygen by means of impedance spectroscopy (Hewlett Packard (HP) 4192A) in the frequency range 5 MHz to 5 Hz, with an oscillation voltage of 0.5 V. Dependencies of the oxygen partial pressure at different temperatures were obtained by varying the ratio of argon and oxygen at constant water partial pressure. Impedance spectroscopy was performed every 50°C between 200 and 1000°C for BZ and between 150 and 600°C for YBZ-1. However, due to instrument limitations it was only possible to get reasonable values from deconvolution of the spectra in the temperature range 450-750°C (wet) and 450-600°C (dry) for BZ and 200-400°C for YBZ-1 (wet).

The impedance spectra were deconvoluted using the Equivalent Circuit for Windows program [29] and fitted to equivalent circuits consisting of two parallel RQ elements in series, where R is the impedance and Q is a constant

phase capacitive element. Q has the impedance [ ] 1nQ )j(YZ

−

ω= , where

1j −= , =ω frequency, Y and n are constants, and n ranges between 0 and

1. Experimentally, the exponent n characterizing the subcircuit element

108

(RQ) is rather close to 1, and thus these elements behave much like

capacitors. The capacitance is given by( ) ( )1

n1

n1

RYC−

= . The equivalent circuit (RbCb)(RgbCgb) was used in the deconvolution and had capacitances for bulk and grain boundaries in the typical order of 10 pF and 10 nF, respectively. From these data specific conductivities for both bulk and grain boundaries were, furthermore, calculated applying the Brick-Layer model [30]. The presented specific conductivity data have not been corrected for porosity. This can be done using the measured density ρ and the theoretical density, ρth, following the empirical relationship: σ = σmeasured / (ρ/ρth)

2 [31].

Results

Characterization of synthesized powders

The TG curves in Fig. 1 show weight loss observed for as-synthesized powder of both BaZrO3 (BZ) and BaZr0.9Y0.1O2.95 (YBZ-1) upon heating to 1500°C. Though there was a continuous weight loss all the way to 1500°C, less than 0.5wt% weight loss was observed above ~950°C.

Temperature (°C)

0 300 600 900 1200 1500

Weig

ht

ch

an

ge

(%

)

94

95

96

97

98

99

100

BZYBZ-1

Fig. 1 TG analysis of as-synthesized BaZrO3 (BZ) and BaZr0.9Y0.1O2.95

(YBZ-1) powders.

109

From XRD of the powder, the lattice parameter for cubic BaZrO3 was calculated to a = 4.185 Å giving a theoretical density of 6.26 g/cm3. Lattice parameter and theoretical density for BaZr0.9Y0.1O2.95 (YBZ-1) was calculated to a = 4.193 Å and 6.21 g/cm3, respectively. Calculated lattice parameters are slightly lower, but still in accordance with reported values [11-12].

X-ray diffractograms of BaZrO3 and BaZr0.9Y0.1O2.95 powders are displayed in Fig. 2 a) and b), respectively. For the as-synthesized BZ powder (Fig. 2 a)) presence of secondary phases (BaCO3 and BaO) was found, however, these phases were apparently removed by calcination at 1000°C. During hot pressing at 1500°C, a secondary phase of ZrO2, indicated by the additional reflection near 2θ = 29°, was formed on the specimen surface. The reflection, though weak, was also observed for bulk material. No traces of BaCO3 could be detected by XRD in the hot pressed BZ specimens. The X-ray diffractograms in Fig. 2 b) demonstrates that BaCO3 is present in the as-synthesized YBZ-1 powder, however, a lower intensity of the BaCO3 reflections indicates a smaller amount compared to the as-synthesized BaZrO3 powder (Fig. 2 a)). After calcination at 1000°C and hot pressing at 1500°C no secondary phases were detected by XRD for YBZ-1. For YBZ-2, treated the same way as YBZ-1, BaCO3 was still observed by XRD after calcination at 1000°C. As for BZ, a reflection around 2θ = 29° was detected for the hot pressed YBZ-2 specimens, even in the bulk. The intensity of this reflection was not significantly different for YBZ-2 specimens hot pressed at 1450 and 1750°C, indicating that the composition of bulk is constant in this temperature range.

110

a)

2 θ (θ (θ (θ (°))))

20 30 40 50 60

Rela

tive

in

ten

sit

y

BaZrO3: �

BaCO3: �

BaO: � ZrO2: �

�

�� �

� �

�

��� �

�

� � ��

BZ as-synth.

BZ 1000°C

BZ HP 1500°Csurface

BZ HP 1500°Cbulk

b)

2 θ (θ (θ (θ (°))))

20 30 40 50 60

Re

lati

ve

in

ten

sit

y

BaZr0.9Y0.1O2.95: �

BaCO3: �

BaO:�

ZrO2:�

�� �

� �

�� � ��

YBZ-1 as-synth.

YBZ-1 1000°C

YBZ-1 HP 1500°C

YBZ-2 1000°C

YBZ-2 HP 1450°C

YBZ-2 HP 1750°C

� � ��

� �

�� �

Fig. 2 a) XRD of as-synthesized and calcined (1000°C) BZ powder, as

well as surface and bulk of BZ specimen hot pressed (25 MPa,

1 h) at 1500°C.

b) XRD of as-synthesized and calcined (1000°C) YBZ-1 and YBZ-

2 powders and bulk of YBZ-1 specimen hot pressed (25 MPa, 1

h) at 1500°C and YBZ-2 specimens hot pressed (50 MPa, 1 h)

at 1450 and 1750°C.

111

Fig. 3 presents SEM micrographs of calcined BaZrO3 (BZ) and BaZr0.9Y0.1O2.95 (YBZ-1 and YBZ-2) powders, before and after milling. Spray pyrolysis produced homogenous powders with particle size <5 µm, as seen for the BZ powder shown in Fig. 3 a). The particles are typically shaped like distorted hollow spheres as displayed for BZ and YBZ-1 in Fig. 3 b) and c), respectively. The particle size was reduced significantly by ball milling, as demonstrated by the SEM micrographs of ball milled BZ, YBZ-1 and YBZ-2 in Fig. 3 d), e) and f), respectively. Table 1 demonstrates how the surface area and calculated particle size of BZ and YBZ-1 powders vary with calcination temperature. Only minor decrease in surface area was observed for BZ and YBZ-1 with increasing calcination temperature showing that these powders are resistant to coarsening with increasing calcination temperature. The YBZ-1 powder calcined at 1000°C for 48 hrs had a higher surface area (10.3 m2/g) compared to the YBZ-2 powder treated the same way (5.8 m2/g).

Table 1 Surface area of BaZrO3 (BZ) and BaZr0.9Y0.1O2.95 (YBZ-1 and

YBZ-2) powders calculated from nitrogen adsorption using the

BET equation. Spherical particles were assumed for estimation

of the particle size.

Compound Tcalcination

(°C)

tcalcination

(hrs)

Surface area

(m2/g)

Particle size

(µm)

BZ 850* 0 9.0 0.11

BZ 1000 48 7.5 0.13

BZ 1100 6 5.8 0.16

YBZ-1 1000 48 10.3 0.09

YBZ-1 1100 6 7.0 0.14

YBZ-2 1000 48 5.8 0.17

* As-synthesized powder.

112

Fig. 3 SEM micrographs of BZ, YBZ-1 and YBZ-2 powders calcined at

1000°C. a) and b) BZ before ball milling. c)YBZ-1 before ball

milling. d) BZ after ball milling. e)YBZ-1 after ball milling and

f) YBZ-2 after ball milling.

113

IR spectra of BZ and YBZ-1 powders are shown in Fig. 4. The BaCO3 spectrum is included as a reference. The broad band in the range of 500 to 800 cm-1, observed for all prepared powders, is due to the stretching of the metal-oxygen bonds [32]. The intensity of the BaCO3 bands in the YBZ-1 powders is reduced with increasing calcination temperature. No bands assigned to BaCO3 were detected for YBZ-1 powder from specimen sintered at 1600°C. In addition to the wide carbonate band in the range of 1200 to 1700 cm-1, a narrow sharp peak was detected for the as-synthesized YBZ-1 powder, at frequency 1384 cm-1, probably due to physically adsorbed CO2

[33]. The wide band in the frequency region from 3000 to 3600 cm-1 is assigned to O-H stretching of physiosorbed water or from surface adsorbed hydroxyl groups. IR spectra of BZ powders calcined at different temperatures show the same trends as the YBZ-1 powders.

Wave number (cm-1)

400 1000 1600 2200 2800 3400 4000

Rela

tive t

ran

sm

itta

nce

BaCO3

YBZ-1BZ

as-synthesized

as-synthesized

1000°C

1100°C

1600°C

Fig. 4 IR spectra of BaCO3 powder, as-synthesized BZ and YBZ-1

powder, YBZ-1 powder calcined at 1000°C and 1100°C as well

as YBZ-1 powder from crushed YBZ-1 specimen sintered at

1600°C.

114

Densification

The linear shrinkage curve for BZ in Fig. 5 shows onset for sintering around 1200°C. The curve for YBZ-1 shows the same on-set temperature, however, the curve is remarkably wider due to lower sintering rate. The shrinkage for YBZ-2 starts around 950°C and flattens out in a plateau before the main shrinkage appears above 1300°C. The presence of a secondary phase (BaCO3) as confirmed by XRD (Fig. 2 b)), may explain the first shrinkage step for YBZ-2. The density and porosity of BZ conventionally sintered at 1600°C are shown in Fig. 6 a). An increase in sintering time from 1 to 6 hrs resulted in an increase in density from 82.5% to 91.1%. No significant change in density was observed with increased hold time beyond 6 hrs at 1600°C, and the porosity was mainly closed (7.5%). In comparison, the density of YBZ-1 specimen sintered at 1600°C for 6 hrs was 74%. Density and grain size of sintered BaZrO3 and BaZr0.9Y0.1O2.95 specimens are presented in Table 2.

Temperature (°C)

800 1000 1200 1400 1600

dL

/Lo

(%

)

-12

-10

-8

-6

-4

-2

0

BZ

YBZ-1

YBZ-2

Fig. 5 Linear shrinkage curves for BaZrO3 (BZ) and BaZr0.90Y0.10O2.95

(YBZ-1 and YBZ-2). All powders were calcined at 1000°C for 48

hrs, ball milled with YSZ balls in iso-propanol for 24 hrs and

dried for 24 hrs at 400°C. Green body densities were in the

range 49 – 51% of theoretical density.

115

a)

Time (hrs)

0 2 4 6 8 10 12 14 16 18

De

ns

ity (

%)

80

85

90

95

100

Po

ros

ity (

%)

0

2

4

6

8

10

Density

Closed porosity

Open porosity

b)

Time (hrs)

0 1 2 3 4 5 6

De

ns

ity (

%)

80

85

90

95

100

Po

ros

ity (

%)

0

2

4

6

8

10

Density

Closed porosity

Open porosity

c)

Temperature (°C)

1450 1500 1550 1600 1650 1700 1750

De

ns

ity (

%)

80

85

90

95

100P

oro

sit

y (

%)

0

2

4

6

8

10

Density

Closed porosity

Open porosity

Fig. 6 Density, closed and open porosity for a) BaZrO3 conventionally

sintered at 1600°C for 1 – 18 hrs in ambient air, b) BaZrO3 hot

pressed at 1500°C (25 MPa) for 1 – 6 hrs and c) YBZ-2

specimens hot pressed for 1 h (50 MPa) at indicated

temperatures.

116

Density and porosity as a function of HP time at 1500°C (25 MPa) are shown for BaZrO3 (BZ) in Fig. 6 b). As for the conventionally sintered BZ (Fig. 6 a)) some closed porosity seems inevitable for HP at 1500°C, however, the amount is lower (~3%). Density, closed and open porosity for YBZ-2 specimens obtained by HP at different temperatures are shown in Fig. 6 c). A significant increase in the density, from 85.6 to 96.7%, was observed for YBZ-2 when the HP (1 h, 50 MPa) temperature was increased from 1450 to 1550°C. 97.6% density was obtained for both YBZ-2 after HP (1h, 50 MPa) at 1750°C and BZ after HP (1 h, 50 MPa) at 1650°C. The results of HP are summarized in Table 2. The BZ and YBZ-1 specimens appeared black due to reduction after hot pressing. The colour changed to light brown for BaZrO3 and green/white for BaZr0.9Y0.1O2.95 after reheat treatment at 1400°C for 24 hrs in flowing oxygen.

Table 2 Density and grain size of conventionally sintered and hot

pressed BaZrO3 (BZ) and BaZr0.90Y0.10O2.95 (YBZ-1 / YBZ-2)

specimens.

Compound Tsintering

(°C)

tsintering

(hrs)

Pressure

(MPa)

Density

(%)

Gran size

(µm)

BZ 1600 1 - 82.5

BZ 1600 18 - 91.9 18

BZ 1400 1 25 87.5

BZ 1500 1 25 93.4 0.46

BZ 1500 3 25 95.8 0.77

BZ 1500 6 25 96.1 1.0

BZ 1650 1 50 97.6 0.71

YBZ-1 1600 6 - 74.2

YBZ-1 1500 1 25 88.5 0.4*

YBZ-2 1450 1 50 85.6 0.4*

YBZ-2 1550 1 50 96.7 0.46

YBZ-2 1650 1 50 96.3 0.49

YBZ-2 1750 1 50 97.8 0.60

*Porous material – exact value not attainable.

117

Microstructure and grain growth

As revealed by the optical micrograph in Fig. 7 a), BZ conventionally sintered at 1600°C for 12 hrs contained large grains (17 µm). In Fig. 7 b) the SEM micrograph of BZ conventionally sintered at 1600°C for 18 hrs confirms the presence of pores enclosed inside grains. Only a minor increase in average grain size (16 - 18 µm, as given in Table 2, was observed for BZ when the sintering time at 1600°C was increased from 6 to 18 hrs. The SEM micrographs in Fig. 7 c), d) and e) demonstrates how the grain size in BZ increases when the HP time is increased from 1 to 6 hrs at 1500°C. HP at 1650°C for 1 h, with an applied pressure of 50 MPa also resulted in grains (Fig. 7 f)) similar to, or smaller than those obtained by HP at 1500°C for 3 and 6 hrs (Fig. 7 d) and e)). Fig. 7 g) and h) reveals the homogenous fine grained microstructures of YBZ-2 specimens HP (1 h, 50 MPa) at 1550 and 1750°C, respectively. Table 2 summarizes the average grain size of hot pressed specimens of BZ and YBZ-2.

118

Fig. 7 a) Optical micrograph of BZ conventionally sintered at 1600°C

for 12 hrs. b) through h) SEM micrographs of: b) BZ

conventionally sintered at 1600°C for 18 hrs, c) BZ specimens

hot pressed at 1500°C (1 h, 25 MPa) for 1 h, d) BZ hot pressed

119

at 1500°C (3 hrs, 25 MPa), e) BZ hot pressed at 1500°C (6 hrs,

25 MPa), f) BZ hot pressed at 1650°C (1 h, 50 MPa), g) YBZ-2

hot pressed at 1550°C (1 h, 50 MPa) and h) YBZ-2 hot pressed

at 1750°C (1 h, 50 MPa)

Electrical properties

Specific bulk and grain boundary conductivities as a function of the inverse absolute temperature in wet and dry oxygen, are shown for hot pressed BZ (1500°C, 3h, 25 MPa), with >95 % density and average grain size ~0.8 µm, in Fig. 8 a). The conductivity in BZ was slightly higher under dry as compared to wet conditions and the bulk conductivity was considerably higher than the specific grain boundary conductivity both under wet and dry conditions. The grain boundary conductivity under wet conditions exhibits a straight-line Arrhenius behavior with an apparent activation energy of 170 kJ/mol. The bulk data exhibits lower activation energy and the temperature dependence decreases with increasing temperature from 600 to 800°C. Fig. 8 b) shows the p(O2)-dependecies for the BZ specimen under relatively oxidizing conditions at 500, 600 and 700°C. The conductivity increases with increasing oxygen partial pressure.

As demonstrated in Fig. 9 a), the impedance spectra of hot pressed YBZ-1 (1500°C, 1 h, 25 MPa), with >88% density and average grain size ~0.4 µm, consisted of two distinguishable semicircles, with values of the capacitance typical representing grain boundaries. Hence, two sets of grain boundary conductivities have been reported, in Fig. 9 b), where the temperature dependence (from 200 to 400°C) for bulk and grain boundary conductivities of YBZ-1 is presented. The reason for this behaviour is not yet clear and needs further investigation. The bulk semicircle was not observed in the YBZ-1 spectra presumably due to instrument limitations. The spectra were, consequently, fitted to the equivalent circuit Rb(Rgb1Qgb1)(Rgb2Qgb2) where Rb corresponds to the apex of the first grain boundary semi-circle in the Nyquist plot. The specific grain boundary conductivities in Fig. 9 b) have been calculated as for the BZ sample using the Brick Layer Model. However, in the absence of bulk semi-circles for the YBZ-1 spectra, we have assumed an average empirical value for the bulk capacitance based on data from the low temperature spectra of the YBZ-1 specimen where only the bulk semi-circle is visible. This value was used as input to calculate specific grain boundary conductivities for YBZ-1.

120

a)

1000/T (K-1

)

1.0 1.1 1.2 1.3 1.4

Co

nd

ucti

vit

y (

S/c

m)

10-9

10-8

10-7

10-6

10-5

10-4

10-3

σBulk(wet O2)

σGb

(wet O2)

σBulk(dry O2)

σGb(dry O2)

Temperature (°C)

800 700 500600

b)

Oxygen partial pressure (atm)

10-5 10-4 10-3 10-2 10-1 100

Bu

lk c

on

du

cti

vit

y (

s/c

m)

1e-5

1e-4

500°C600°C

700°C

Fig. 8 a) Conductivity as a function of inverse temperature for BZ in

wet and dry O2. The specific conductivity of the grain boundary

was calculated applying the Brick Layer Model. b) pO2-dependence for BZ, bulk, at 500, 600 and 700°C.

121

a)

R (Ohm)

0 20000 40000 60000 80000

-Z (

Oh

m)

0

20000

40000

60000

80000YBZ-1, 400°C, wet O2

b)

1000/T (K-1)

1.4 1.6 1.8 2.0 2.2

Ap

pa

ren

t c

on

du

cti

vit

y (

S/c

m)

10-9

10-8

10-7

10-6

10-5

10-4

10-3

10-2

σBulk

σGb-1

σGb-2

Temperature (°C)

400 300 200

Fig. 9 a) Impedance spectrum for YBZ-1 at 400°C in wet O2. The

spectrum shows two grain boundary semicircles.

b) Apparent conductivities as a function of inverse temperature

for YBZ-1 in wet O2. The apparent specific grain boundary

conductivities were calculated applying the Brick Layer Model.

122

Discussion

Powder characteristics and densification

The TG curves in Fig. 1 demonstrated that BaCO3 detected in as-synthesized powders by XRD (Fig. 2) was hard to remove completely, as weight loss was observed even above 1000°C. Calcination at 1000°C for 48 hrs resulted in apparently phase pure BZ (Fig. 2 a)) and YBZ-1 (Fig. 2 b)) powders, however, carbonates were detected by IR even after calcination at 1100°C in nitrogen (as seen for YBZ-1 in Fig. 4). For YBZ-2 powder, BaCO3 along with ZrO2, was detected by XRD after calcination at 1000°C for 48 hrs (Fig. 2 b)). The presintering step observed for YBZ-2 (Fig. 5) can be explained by the presence and decomposition of BaCO3. YBZ-1 and YBZ-2 showed considerably poorer sinterability compared to BaZrO3 (BZ). This could indicate a slight deviation in the stoichiometric composition of the BaZr0.9Y0.1O2.95 powders. This is the case for YBZ-2, where the secondary phase of ZrO2 becomes more apparent after hot pressing. A general explanation for the reduced sinterability can also be the substitution of yttrium for zirconium. As Y3+ is larger than Zr4+ (0.90 versus 0.72 Å [34]) reduced sinterability will be observed if the Zr/Y-site cation diffusion is rate limiting.

Conventional sintering of BaZrO3 at 1600°C resulted in density <92% independent of sintering time. This is in agreement with the literature [7,11,12,15,16] stating that T >1700°C is needed in order to obtain >95% dense BaZrO3 specimens. Sintering at 1600°C also involved severe grain growth giving average grain size ≥16 µm. Due to the high mobility of grain boundaries pores became entrapment in grains, as seen in Fig. 7 b). Entrapped pores have also been reported for BaZrO3 prepared from fine grained powders sintered at 1450°C [19]. Decomposition of traces of BaCO3 may also contribute to pore growth and increased porosity. In the presented work, HP has been shown to effectively suppress the grain growth observed for conventional sintering. The reduced sintering temperature in addition to applied uniaxial pressure reduces the mobility of grain boundaries, and dense (>96%) materials with uniform microstructures and average grain size in the range of 0.42 to 1.0 µm have been obtained by HP. Close to 98% density was obtained for both BaZrO3 and BaZr0.9Y0.1O2.95 after HP (1h, 50 MPa) at 1650 and 1750°C, respectively. Based on these results, HP is clearly superior to conventional sintering when it comes to densification of BaZrO3-based materials.

123

Chemical stability

It should be remarked that the high temperatures used in particular during HP (up to 1750°C) have been reported to give evaporation of BaO(g), as shown by Eq. 2. In the HP experiments the system is fairly closed and the atmosphere highly reducing. The reaction between BaZrO3 and the carbon pressing dies resulting in evaporation of Ba(g), according to Eq. 3, is therefore more significant than the evaporation of BaO(g).

BaZrO3(s) + C(s) = ZrO2(s) + CO(g) + Ba(g) (3)

In Fig. 10 the equilibrium constants for the reactions in Eq. 2 and 3 are plotted as a function of temperature. Fig. 9 clearly demonstrates how the reaction described by Eq. 3 will dominate at temperatures above ~1250°C when carbon is present. The equilibrium constant for the reaction in Eq. 3 equals 10-6 below 1600°C. This implies that the partial pressure of Ba(g) may exceed 10-3 bar when the HP temperature is increased beyond 1600°C, giving significant loss of Ba.

If Ba(g) has diffused out of the pressing die one would expect to observe ZrO2(s) in the material, particularly close to the surface and at grain boundaries. The latter is possibly what was observed for hot pressed BZ and specially YBZ-2 specimens. As all specimens were polished after densification one would expect possible ZrO2(s) on the surface to disappear. In the case of YBZ-2 hot pressed at T > 1450°C, ZrO2(s) was detected even in bulk. For YBZ-2 there are indications that the starting powder differ slightly from optimal stoichiometry, resulting in ZrO2 rich materials after hot pressing. No obvious observation of secondary phases were observed by XRD of polished pellets of both BZ and YBZ-1 hot pressed at 1500°C, however the surface of the BZ specimen contained ZrO2(s) according to XRD. Density larger than 96% can be obtained for both BaZrO3 and BaZr0.9Y0.1O2.95 by HP at 1500 and 1550°C, respectively and these temperatures might be sufficiently low to obtain stoichiometric, phase pure materials. For full densification (~98% density) higher temperatures are needed (1650 -1750°C) and evaporation of BaO(g) and Ba(g) is expected to take place. Loss of Ba(g) or BaO(g) during sintering may result in ZrO2 at grain boundaries, which will explain the grain boundary resistance discussed further below.

124

Temperature (K)

1000 1200 1400 1600 1800 2000

log

K

-20

-15

-10

-5

Temperature (°C)

900 1100 1300 1500 1700

BaZrO3(s) + C(s) = ZrO2(s) + CO(g) + Ba(g)

BaZrO3(s) = ZrO

2(s) + BaO(g)

Fig. 10 Equilibrium constants, K, for decomposition of BaZrO3 plotted

as a function of temperature. For the reaction in normal

atmosphere the equilibrium constant equals the BaO partial

pressure (K = p(BaO)) while for the reaction in reducing

atmosphere the constant is a product of the Ba- and the CO

partial pressure (K = p(Ba)·p(CO)).

Electrical properties

The oxygen partial pressure dependence of the bulk conductivity for the BZ specimen can be interpreted as consisting of two different contributions. At high temperatures and under oxidizing conditions the conductivity increases with increasing oxygen partial pressure, generally concluded to reflect p-type electronic conductivity. Towards lower oxygen partial pressures, the conductivity becomes independent of the oxygen partial pressure which either reflects that an aliovalent impurity determines the concentration of the charge carrier or that intrinsic disorder dominates the defect structure. Since the conductivity is higher under dry than under wet conditions and that the difference increases with increasing temperature, one may conclude that electron holes are the predominating charge carrier under high-temperature oxidizing conditions. It is however not straight-forward to decide which defect is responsible for the ionic contribution. On basis of the protonic dominance reported in the literature for acceptor substituted BZ and the fact that water decreases the p-type contribution, as shown in Fig 8 a), protons

125

clearly influence the defect structure. Whether these protons are charge-compensated by an effectively negative impurity (acceptor), frozen in metal vacancies or in thermal equilibrium with metal vacancies cannot be established. One may note, though, that the apparent activation energy for the bulk conductivity in the low temperature region is relatively high (~100 kJ/mol) as compared to the YBZ-1 (~30 kJ/mol), where the acceptor substitution clearly is charge compensated by protons.

From the conductivity characteristics of the BZ material one should also recognize that there is a large difference between the bulk and the grain boundary conductivity and, furthermore, that essentially the same effects are observed for the bulk conductivity under wet versus dry conditions. One may speculate whether this indicates that the same defect situation prevails in the grain boundaries as in the bulk. The apparent activation energy of the grain boundary conductivity is, however, considerably higher than for bulk.

The conductivities of BZ and YBZ-1 are compared to the state-of-the-art

values reported for proton conductivity in BaZr0.9Y0.1O3-δ by Bohn and Schober [15] and Kreuer [17] in Fig. 11. Acceptor substitution clearly has a tremendous impact increasing the conductivity several orders of magnitude. The values for the bulk conductivity of YBZ-1 are essentially the same as those observed by Bohn and Schober [15]. Although it has not been directly measured we may assume, based on the similar temperature dependence of the present data and the literature, that the conductivity is predominantly protonic in the present temperature region (200 to 400°C).

There are different hypotheses in the literature so as to explain the high grain boundary resistance of acceptor substituted BaZrO3. It has been suggested that a non-uniform distribution of charge carriers across grain boundaries may result in high resistance towards the ionic transport [35].

Through-out the discussion above, one must bear in mind that interpretation of nominally BaZrO3 materials always is associated with high uncertainties due to the unknown level and type of impurities. In our case, this is of particular importance since grain boundaries are sensitive in this respect. Difference in the chemical composition at the grain boundaries due to loss of Ba(g) / BaO(g) during sintering, may be of importance for understanding the grain boundary resistance in BaZrO3 ceramics.

126

1000/T (K-1

)

1.0 1.5 2.0 2.5 3.0

Co

nd

uc

tivit

y (

S/c

m)

10-6

10-5

10-4

10-3

10-2 BZ

YBZ-1

YBZ Bohn

YBZ Kreuer

Temperature (°C)

600 400 200800

Fig. 11 Bulk conductivity as a function of inverse temperature for BZ

and YBZ-1 in wet O2, obtained in the present investigation,

compared with bulk data for YBZ in wet O2 presented by Bohn

and Schober [35]

and data for single crystal YBZ by Kreuer [36]

.

Conclusions

High quality powders of BaZrO3 and BaZr0.9Y0.1O2.95 with particle size of ~0.1 µm were prepared by spray pyrolysis of nitrate salt solutions followed by calcination at 1000°C and ball milling in iso-propanol. By conventional sintering of BaZrO3, a maximum density of 92% was obtained and the average grain size for these specimens was in the range of 16 to 18 µm. Both for BaZrO3 and BaZr0.9Y0.1O2.95, a density >96% was obtained by hot pressing at 1500 and 1550°C, respectively. The HP specimens showed homogenous microstructures with average grain size down to 0.42 µm. Density close to 98% was obtained for BaZrO3 and BaZr0.9Y0.1O2.95 after HP (1h, 50 MPa) at 1650 and 1750°C, respectively. However, at these temperatures evaporation of BaO(g) and Ba(g) is an issue. The prepared specimens exhibit electrical properties comparable to what has been reported in the literature. The conductivity in BaZrO3 is slightly higher in dry atmosphere compared to wet, and for both compositions the proton conductivity is clearly limited by high grain boundary resistance.

127

Acknowledgement

Work was supported by the Research Council of Norway, Grant No. 1585171431 (NANOMAT). Øystein Andersen is acknowledged for operating the spray pyrolysis.

References

[1] H. Iwahara, Y. Asakura, K. Katahira, M. Tanaka, Prospect of

hydrogen technology using proton-conducting ceramics, Solid State Ionics, 168, 299-310 (2004)

[2] N. Taniguchi, K. Hatoh, J. Niikura, T. Gamo, H. Iwahara, Proton

conductive properties of gadolinium doped barium cerates at high

temperatures, Solid State Ionics, 53-56, 998-1003 (1992) [3] M.J. Scholten, J. Schoonman, J.C. van Miltenburg, H.A.J. Oonk,

Synthesis of strontium and barium cerate and their reaction with

carbon dioxide, Solid State Ionics, 61, 83-91 (1993) [4] K.D. Kreuer, On the development of proton conducting materials for

technological applications, Solid State Ionics, 97, 1-15 (1997) [5] T. Norby, Solid-state protonic conductors: principles, properties,

progress and prospects, Solid State Ionics, 125, 1-10 (1999) [6] K.H. Ryu, S.M. Haile, Chemical stability and proton conductivity of

doped BaCeO3 – BaZrO3 solid solutions, Solid State Ionics, 125, 355-367 (1999)

[7] K. Katahira, Y. Kohchi, T. Shimura, H. Iwahara, Protonic

conduction in Zr-substituted BaCeO3, Solid State Ionics, 138, 91-98 (2000)

[8] K.D. Kreuer, Proton-conducting oxides, Annu. Rev. Mater. Res., 33, 333-359 (2003)

[9] T. Norby, Y. Larring, Concentration and transport of protons in

oxides, Current Opinion in Solid State & Materials Science, 2, 593-599 (1997)

[10] R.C.T. Slade, S.D. Flint, N. Singh, Investigation of protonic

conduction in Yb- and Y-doped barium zirconates, Solid State Ionics, 82, 135-141 (1995)

[11] K.D. Kreuer, St. Adams, W. Münch, A. Fuchs, U. Klock, J. Maier, Proton conducting alkaline eart zirconates and titanates for high

drain electrochemical applications, Solid State Ionics, 145, 295-306 (2001)

[12] T. Schober, H.G. Bohn, Water vapor solubility and electrochemical

characterization of the high temperature proton conductor

BaZr0.9Y0.1O2.95, Solid State Ionics, 127, 351-360 (2000)

128

[13] H. Iwahara, T. Yajima, T. Hibino, K. Ozaki, H. Suzuki, Protonic

conduction in calcium, strontium and barium zirconates, Solid State Ionics, 61, 65-69 (1993)

[14] W. Wang, A.V. Virkar, Ionic and electron-hole conduction in

BaZr0.93Y0.07O3-δ by 4-probe dc measurements, J. Power Sources, 142, 1-9 (2005)

[15] H.G. Bohn, T. Schober, Electrical conductivity of the high-

temperature proton conductor BaZr0.9Y0.1O2.95, J. Am. Ceram. Soc., 83 [4], 768-772 (2000)

[16] F.M.M. Snijkers, A. Nuekenhoudt, J. Cooymans, J.J. Luyten, Proton

conductivity and phase composition in BaZr0.9Y0.1O3-δ, Scripta Materialia, 50, 655-659 (2004)

[17] K.D. Kreuer, Aspects of the formation and mobility of protonic

charge carriers and the stability of perovskite-type oxides, Solid State Ionics, 125 (1-4), 285-302 (1999)

[18] J. Brzezinska-Miecznik. K. Haberko, M.M. Bucko, Barium zirconate

ceramic powder synthesis by the coprecipitation – calcination

technique, Materials Letters, 56, 273-278 (2002) [19] F. Boschini, B. Robertz, A. Rulmont, C. Cloots, Preparation of

nanosized barium zirconate powder by thermal decomposition of

urea in an aqueous solution containing barium and zirconium, and

by calcination of the precipitate, J. Eur. Ceram. Soc., 23, 3035-3042 (2003)

[20] A. Magrez, T. Schober, Preparation, sintering and water

incorporation of proton conducting BaZr0.9Y0.1O3-δ: comparison

between three different synthesis techniques, Solid State Ionics, 175, 585-588 (2004)

[21] P.P. Phule, D.C. Grundy, Pathways for the low temperature

synthesis of nano-sized crystalline barium zirconate, Materials Science and Engineering, B23, 29-35 (1994)

[22] T. Tsuneo, S. Stølen, H. Yokoi, Thermodynamic study of barium

zirconates by mass-spectrometry, J. Nuclear Materials, 209, 174-179 (1994)

[23] T. Tsuneo, Thermodynamic properties of ternary barium oxides, Thermochimica Acta, 253, 155-165 (1995)

[24] U. Anselmi-Tamburini, M.T. Buscaglia, M. Viviani, M. Bassoli, C. Bottini, V. Buscaglia, P. Nanni, Z.A. Munis, Solid-state synthesis

and spark plasma sintering of submicron BaYxZr1-xO3-x/2 (x = 0, 0.08

and 0.16) ceramics, J. Eur. Ceram. Soc., 26, 2313-2318 (2006) [25] Babilo, S.M. Haile, Enhanced sintering of yttrium doped barium

zirconate by addition of ZnO, J. Am. Ceram. Soc., 88 [9], 2362-2368 (2005)

129

[26] C. Kjølseth, R. Haugsrud, T. Norby, Electrical characterization of

undoped and acceptor doped BaZrO3 - effects of grain boundaries, To be published (2007)

[27] M.I. Mendelson, Average grain size in polycrystalline ceramics, J. Am. Ceram. Soc., 52 (8), 443-446 (1969)

[28] T. Norby, "Electronic Book", Norwegian Electro Ceramics AS, www.norecs.com

[29] B.A. Boukamp, Enschede. p. AC-immittance analysis system, EqC. 2003, University of Twente/WisseQ

[30] T. Norby, Electrical measurements, KJM-MEF 4010, Oslo: University of Oslo (2005)

[31] S. Marion, A.I. Becerro, T. Norby, Ionic and electronic conductivity

in CaTi1-xFexO3-δ (x = 0.1-0.3), Ionics, 5 (5-6), 385-392 (1999) [32] K. Nakamoto, Infrared and Raman Spectra of Inorganic and

Coordination Compounds, 2nd ed., John Wiley & Sons, Inc., New York, (1997)

[33] T. Mokkelbost, I. Kaus, T. Grande, M.-A. Einarsrud, Combustion

synthesis and characterization of nanocrystalline CeO2-based

powders, Chem. Mater., 16, 5489-5494 (2004) [34] R.D. Shannon, Revised effective ionic radii and systematic studies of

interatomic distances in halides and chalcogenides, Acta Cryst., A32, 751-767 (1976)

[35] X. Guo. R. Waser, Electrical properties of the grain boundaries of

oxygen ion conductors: Acceptor-doped zirconia and ceria, Progress in Materials Science, 51 (2), 151-210 (2006)

130

PAPER V

132

133

Oriented LaFeO3 thin films grown on NdGaO3 by spin-coating

Paul Inge Dahl, Tor Grande, Mari-Ann Einarsrud †

Department of Materials Science and Engineering

Norwegian University of Science and Technology

N-7491 Trondheim, Norway

Abstract

LaFeO3 precursor solution was prepared using nitrate salts dissolved in methanol. Addition of acetic acid and acetyl acetone as chelating agent resulted in a solution with good spinning properties. Spin-coating was performed at 2000 rpm for 1 min on (100)- and (110)-oriented single crystalline NdGaO3 substrates as well as (0001)-oriented single crystal Al2O3 substrates. Controlled heat treatment up to 400°C with heating rate of 0.1°C/min resulted in homogenous continuous amorphous films on all substrates. Annealing for 1 h at 500 – 1000°C caused the formation of oriented polycrystalline thin films on NdGaO3 with both (100)- and (110)-orientation. Films prepared on sapphire were polycrystalline, randomly oriented and inhomogenous after annealing at 700°C. The LaFeO3 films grown on the NdGaO3 substrates crystallize between 400 and 500°C and the average grain size increased from 40 to 250 nm as the temperature was increased from 500 to 700°C.

† Corresponding author. E-mail: Mari-Ann.Einarsrud@material.ntnu.no

134

Introduction

Many perovskite-type oxide materials, such as lanthanum ferrite (LaFeO3) exhibit mixed ionic-electronic conductivity [1]. LaFeO3, with orthorhombic structure, is an antiferromagnetic insulator at room temperature, making the material feasible for use in magnetic sensors and as read heads in computer hard drives [2]. At elevated temperatures the material shows significant conductivity. At 1000°C a linear response for log σ versus log p(O2) have been reported for LaFeO3 thin films, making the material promising for oxygen sensor applications [3]. Furthermore, LaFeO3 show excellent sensitivity towards non flammable gases such as CO [4] and NOx

[4-8] and exhibit selective sensitivity towards flammable gases such as ethanol [9], methane [4] and volatile sulfides such as CH3SH [10]. LaFeO3 films have also been investigated for γ-radiation dosemetry purposes [11]. To optimize response time, thin film sensors are advantageous, and these sensors have the benefit of needing no reference cell. Additionally the gas sensitivity has been reported to increase with decreasing grain size [12], making the morphology an important feature in gas sensing films. In the literature, the preparation of LaFeO3 films is reported by use of sputtering techniques [3,5,13-16], screen printing of slurries from nanosized powder [6,7,11,17], and electrochemical reduction [18].

Among the above mentioned techniques, only sputtering has resulted in epitaxial LaFeO3 thin films. (110)-oriented LaFeO3 films have been grown on rhombohedral (001)-oriented LaAlO3 substrates [3,13], and (001)-oriented LaFeO3 films have been prepared on cubic (001)-oriented MgO substrates [14]. The crystal structure and lattice parameters of LaFeO3 and a selection of substrate materials are given in Table 1. NdGaO3 only give a minor mismatch in the orthorhombic lattice parameters compared to LaFeO3. Hence, NdGaO3 substrate is a suitable candidate for studying epitaxial growth of LaFeO3 films. The aim of the present fundamental study has been to develop a simple sol-gel preparation route for precursor solution of LaFeO3, and to obtain uniform thin films with defined microstructure via spin-coating. Further we have investigated the crystallization behaviour of LaFeO3 films on single crystal NdGaO3 substrates ((100)- and (110)-oriented). Sapphire substrates, with a noncompatible hexagonal crystal structure, were chosen for comparable LaFeO3 film growth of random orientation.

135

Table 1 Crystal structure and lattice parameters of LaFeO3 [19]

and

selected substrate materials [20]

.

Lattice parameters (Å) Composition Crystal structure

a b c

LaFeO3

Orthorhombic

5.555

5.566

7.855

NdGaO3 Orthorhombic 5.43 5.50 7.71

Al2O3 (0001) Hexagonal 4.758 - 12.99

Experimental

LaFeO3 precursor solution was prepared by a sol-gel synthesis route comparable to well accepted routes with metal alkoxides/acetates [21,22], however modified with nitrate salts as starting agents. Stoichiometric quantities of La and Fe nitrate salts (La(NO3)3·6H2O and Fe(NO3)2·xH2O, Merck or Fluka, >99%) were dissolved in methanol (CH3OH, Normapur, >99.8%) before addition of acetic acid (CH3COOH, Merck, 99.5%) followed by stirring for 1 h. Finally, acetyl acetone (CH2(COCH3)2, Merck, >99%) was added as a chelating agent. The addition of acetyl acetone resulted in an immediate reaction, observed in form of a rapid colour change in the solution from transparent red-brownish to dark opaque. A molar cation : acetic acid : acetyl acetone ratio of 1 : 6 : 6 was used and the total

cation concentration in the solution was ∼0.20 M. The solution was

refluxed at 65°C for 1 h. The as-prepared precursor solution was stable for 1 – 2 months after which the wetting properties onto the substrates were found to deteriorate.

The LaFeO3 precursor solution was deposited onto single crystal neodymium gallate (NdGaO3, (100)-oriented, MTI Corp. and (110)-oriented, Crystal & Coating Tech.) and sapphire (Al2O3, (0001)-oriented, MTI Corp.) substrates by spin-coating (Laurell, WS-400A-6NPP/C-1). The precursor solution was dispersed on the substrate surface and spun at 2000 rpm for 1 min. The films were dried at 100°C for 5 min or heat treated at 400°C for 1 h with heating and cooling rates of 0.1 and 1°C/min, respectively. A total of 1 - 5 film layers were deposited onto the substrates,

after which they were annealed (Entech, MF 1/12) at 500 – 1000°C for 1 h with heating/cooling rate of 1 – 3°C/min. Additionally, a part of the precursor solution was evaporated forming a thick gel and after drying at 100°C (1 h) a hard gel was obtained, which was easily crushed down to a

136

black powder. This powder self ignited at 135°C in a volatile reaction, to give a fluffy product which was calcined at 1000°C. The prepared LaFeO3 powder was single phase with orthorhombic crystal structure, indicating stoichiometric La:Fe ratio in the precursor solution. The calculated lattice parameters were a = 5.553 Å, b = 5.566 Å and c = 7.859 Å, in agreement with literature [19].

The crystallisation of the films was studied by X-ray diffraction (XRD) (Siemens D5000 and Bruker AXS, D8 Advance) with Cu Kα radiation. The alignment of the substrate and film was performed using rocking curve scans over the main reflections to record the offset angle. Crystallite size of crystalline films was estimated by Scherrer’s equation using the full width at half maximum (FWHM) of the dominating film reflections, and FWHM of comparable reflections from LaB6 (NIST standard no. 660A [23]) to indicate the broadening contribution from the instrument. The morphology of prepared films was studied by atomic force microscopy (AFM) (Digital Instruments, MMAFM-1) in tapping mode using doped silicon tips. The average grain size of polycrystalline films with defined grain structure, was calculated by use of the linear intercept method over at least 50 grains.

Results

X-ray diffractograms of LaFeO3 films grown on NdGaO3 substrates are shown in Fig. 1, and compared to the diffractogram of powder obtained from the same precursor solution as used for film preparation. LaFeO3 films grown on (100)-oriented NdGaO3 (Fig. 1 a) were oriented in the substrate direction, as only reflections corresponding to (200)/(112) and (400)/(224) were detected. Assuming the d-values of film reflections correspond to the (100)-plane distances (d200 and d400), the lattice parameter a, was calculated to 5.543 Å for the film (annealed at 700°C) compared to a = 5.553 Å for bulk (powder calcined at 1000°C).

Oriented LaFeO3 films were also grown on NdGaO3 with (110)-orientation (Fig. 1 b)) and only reflections corresponding to LaFeO3 (110)/(002) and (220)/(004) were detected. These LaFeO3 films (annealed at 700°C) issued d-values from which the lattice parameter c = 7.825 Å was calculated, assuming (hkl) to be (002) and (004) for this set of d-values. In comparison the corresponding lattice parameter for bulk was c = 7.858 Å. Prepared LaFeO3 film on (110)-oriented NdGaO3 was amorphous after annealing at 400°C. LaFeO3 film deposited on (100)-oriented NdGaO3 substrate showed crystallinity after annealing at 500°C. The crystallization temperature can

137

therefore be estimated to be between 400 and 500°C, assuming it is independent of substrate orientation.

a) 2 θ θ θ θ (°)

25 30 35 40 45 50 55 60 65 70 75

Re

lati

ve

in

ten

sit

y

LaFeO3 powder

LaFeO3 film on (100)-oriented NdGaO3

(11

2)/

(20

0)

(22

4)/

(40

0)

(20

2)/

(02

2)

(22

0)/

(00

4)

(20

4)/

(31

2)

(02

3)/

(22

1)

Nd

Ga

O3 (

10

0)

Nd

Ga

O3 (

20

0)

La

Fe

O3 (

20

0)

La

Fe

O3 (

40

0)

(11

4)/

(22

2)

(13

1)

b) 2 θ θ θ θ (°)

20 25 30 35 40 45 50

Re

lati

ve in

ten

sit

y

LaFeO3 powder

LaFeO3 film on (110)-oriented NdGaO3

(11

0)/

(00

2)

(22

0)/

(00

4)

(20

2)/

(02

2)

(11

2)/

(20

0)

Nd

Ga

O3 (

11

0)

Nd

Ga

O3 (

22

0)

La

FeO

3 (

11

0)

La

Fe

O3 (

22

0)

Fig. 1 X-ray diffractograms of LaFeO3 films grown on NdGaO3

substrate with a) (100)-orientation and b) (110)-orientation.

(200

)

(40

0)

138

Both films have been annealed at 700°C. Diffractogram of

powder calcined at 1000°C is shown for comparison.

The evolution of LaFeO3 film crystallization on (100)-oriented NdGaO3 substrates, with increasing annealing temperature, is demonstrated in Fig. 2. Increased intensity of the reflection at 2θ = 32.3°, corresponding to LaFeO3 (200), was observed with increasing temperature up to 1000°C. The full width at half maximum (FWHM) decreased from 0.234° to 0.178° over the temperature range 500 to 1000°C, corresponding to a small increase in crystallite size from 37 to 51 nm.

2 θ θ θ θ (°)

32.0 32.2 32.4 32.6

Rela

tive

in

ten

sit

y

Substrate

500°C

700°C

1000°C

Fig. 2 X-ray diffractograms of LaFeO3 films grown on (100)-oriented

NdGaO3 substrates annealed at various temperatures for 1 h.

The offset angles for the NdGaO3 substrates and corresponding films were in the range of 0.3 – 0.5°. The difference in offset angle of LaFeO3 film and NdGaO3 substrate was generally <0.1°, confirming epitaxial growth of the film. No offset angle was recorded for films grown on sapphire, indicating random orientation of these films. The random orientation of LaFeO3 films deposited on sapphire (annealed at 700°C) was confirmed by XRD, as demonstrated in Fig. 3. All high intensity reflections of LaFeO3 are present

139

in the diffractogram, however, the intensities are weak compared to those obtained for the oriented films (Fig. 1). Characteristic features from the sapphire substrate dominate the diffractogram.

2 θ θ θ θ (°)

20 25 30 35 40 45 50

Re

lati

ve

in

ten

sit

y

LaFeO3 on sapphire 700°C

Sapphire substrate(1

11

)

(11

2)/

(20

0)

(22

0)/

(00

4)

(20

2)/

(02

2)

Fig. 3 X-ray diffratogram of LaFeO3 film on sapphire substrate, after

annealing at 700°C for 1 h. Diffraction profile for uncoated

substrate is added for comparison. The non-linear background

is due to the properties of the detector used.

AFM micrographs of LaFeO3 films grown on (100)-oriented NdGaO3, are presented in Fig. 4, where Fig. 4 a) shows that fast heating (3°C/min) to relatively high temperature (1000°C) resulted in films with a rather rough, porous microstructure. Slow heating (0.1°C/min up to 400°C) resulted in more continuous fine grained films after annealing at 500°C as shown by the AFM micrograph in Fig. 4 b). As the slow heating (0.1°C) improved the homogeneity, it was used for all LaFeO3 films grown on (110)-oriented NdGaO3 substrates. The morphology of these films are presented in Fig. 5. The homogenous film morphology obtained by slow heating to 400°C (Fig. 5 a)) was maintained upon further annealing at 700°C (1°C/min) for 1 h (Fig. 5 b)). A fine grained LaFeO3 film structure (~40 nm) was obtained after heating to 400°C, as seen in Fig. 5 c), although the film was

140

amorphous. Further annealing at 700°C resulted in a crystalline film with a larger grain structure (~250 nm) (Fig. 5 d)).

Fig. 4 AFM micrographs of LaFeO3 films on (100)-oriented NdGaO3

substrates after annealing at a) 1000°C with a heating rate of

3°C/min and b) 500°C with a heating rate of 0.1°C/min up to

400°C, then 1°C/min.

The morphology of LaFeO3 films grown on sapphire substrate is presented in Fig. 6. The film is homogenous after heat treatment at 400°C (Fig. 6 a)), however, Fig. 6 b) clearly shows that this is not the case after 700°C. The formation of islands, giving a rough noncontinuous film, confirms that sapphire was not a suitable substrate for LaFeO3 films. Although the surface roughness was high, the average grain size obtained for LaFeO3 film on sapphire annealed at 700°C was the same (~250 nm) as for the film on (110)-oriented NdGaO3, heat treated the same way. The crystallite size obtained from XRD was lower than the grain size obtained from AFM. The trend of increasing crystallite/grain size with increasing temperature was the same.

141

Fig. 5 AFM micrographs of LaFeO3 film grown on (110)-oriented

NdGaO3 substrate after heat treatment with a heating rate of

0.1°C/min up to 400°C (a) and c)) followed by annealing at

700°C (1°C/min) (b) and d)). The scalebar on the bottom right

indicates height differences of the surfaces in Fig. c) and d).

142

Fig. 6 AFM micrographs of LaFeO3 film grown on (0001)-oriented

sapphire substrate after slow heat treatment up to 400°C

(0.1°C/min) (Fig. a)) and annealing at 700°C for 1 h (1°C/min)

(Fig. b)). Corresponding height scalebars are placed right of

micrographs.

Discussion

The importance of low heating rate, during first heat treatment of the film layers have clearly been demonstrated by the AFM micrographs in Fig. 4. The volatile self ignition reaction observed for the dried LaFeO3 sol at 135°C can explain why the fast heat treatment resulted in a rough, nonhomogenous film surface (Fig. 4 a)). When using slow heating rate the solvent (with boiling point at 65°C) will be completely evaporated by the time the ignition temperature is reached. Hence, the reduction of nitrates may be more controlled as the additional fuel (methanol) no longer is present.

Epitaxial growth of LaFeO3 thin films, as demonstrated in the present work, have to our knowledge not been reported by similar methods (only by sputtering techniques). (110)-oriented LaFeO3 films have been grown on rhombohedral (001)-oriented LaAlO3 substrates [3,13] and (001)-oriented LaFeO3 films have been prepared on cubic (001)-oriented MgO substrates [14]. Both LaFeO3 and NdGaO3 exhibit an orthorhombic crystal structure with < 2% mismatch in lattice parameters. LaFeO3 film on NdGaO3 substrate is therefore expected to orient in the same direction as the substrate. The observed LaFeO3 film reflections can be assigned to two sets of hkl values. As an example, distance between two corresponding (200)-planes is close to identical to that for (112)-type planes, hence the d-values for these two reflections (d200 and d112) will be the same. Although, not

143

evidenced, the LaFeO3 films on the NdGaO3 substrates are likely to orient in the substrate direction

The lattice parameters for the NdGaO3 substrates are 1-2% smaller, giving ~5% lower unit cell volume, compared to LaFeO3 bulk material. Consequently it is expected that a LaFeO3 film epitaxially grown on NdGaO3 will be under compression, which was confirmed by the smaller lattice parameters (a and c for films on (100)- and (110)-oriented substrates, respectively) for LaFeO3 films on NdGaO3 substrates compared to those obtained from bulk. The prepared films were polycrystalline, indicating growth starting by nucleation at several sites on the substrate surface. The compressive stress in the film may be alleviated at the grain boundaries in the polycrystalline film while maintaining the film continuity. If the stress gets too large, however, cracking of the film will occur. This was not the case for LaFeO3 grown on NdGaO3 as the prepared polycrystalline films were all crack free.

In order to be epitaxial, the growth must start at the interface between the film and the substrate surface. If the growth starts at the film surface moving down towards the substrate, the growth is unlikely to be epitaxial. In addition, the surface structure of the substrate must be compatible with a corresponding surface of the film. As demonstrated, the prepared LaFeO3 films on (0001)-oriented Al2O3 substrates were randomly oriented and polycrystalline. As the crystal structure of substrate and film did not match (hexagonal versus orthorhombic) this was expected, however not obvious. Packing of the atoms of one crystal structure (film) on top of a different structure (substrate) is possible if the interatomic distances of the corresponding surfaces match. Furthermore, epitaxial growth is possible if a repetitive order can be found between the interatomic distances of the film and the substrate surface. This could not be obtained for LaFeO3 film on (0001)-oriented sapphire, hence the random orientation of these films. A more critical aspect of films prepared on sapphire was the formation of islands, observed when heating from 400 to 700°C. The inhomogenous and rough film surfaces may be a result of high strain effects in the films due to non compatibility with the substrate surface. The strain will be introduced when crystallization starts and this may explain why the formation of islands was not observed in the amorphous films on sapphire heat treated at 400°C.

144

Conclusion

A simple preparation route for LaFeO3 precursor solution from nitrate salts dissolved in methanol was obtained using acetyl acetone as a chelating agent. Single crystal NdGaO3 was proved to be a suitable substrate for growth of oriented polycrystalline LaFeO3 thin films by spin-coating of the LaFeO3 precursor solution. By gentle heat treatment to 400°C (0.1°C/min heating rate) amorphous films with well defined smooth microstructure were obtained. Annealing of these films at up to 700°C resulted in continuous, highly oriented polycrystalline films due to epitaxial growth and average grain size up to 250 nm. Hexagonal sapphire was excluded as a suitable substrate due to formation of LaFeO3 islands upon heating to 700°C.

Ackowledgement

Dr. Julian Tolchard is acknowledged for assisting the interpretation of XRD data.

Literature

[1] O. Yamamoto, Y. Takeda, R. Kanno, M. Noda, Perovskite-type

oxides as oxygen electrodes for high temperature oxide fuel cells,

Solid State Ionics, 23, 241-246 (1987) [2] J.B. Kortright, D.D. Awschalom, J. Stöhr, S.D. Bader, Y.U, Idzerda,

S.S.P. Parkin, I.K. Schuller, H.-C. Siegmann, Journal of Magnetism

and Magnetic Materials, 207, 7-44 (1999) [3] I. Hole, T. Tybell, J.K. Grepstad, I. Wærnhus, T. Grande, K. Wiik,

High temperature transport kinetics in heteroepitaxial LaFeO3 thin

films, Solid State Electronics, 47, 2279-2282 (2003) [4] N.N. Toan, S. Saukko, V. Lantto, Gas sensing with semiconducting

perovskite oxide LaFeO3, Physica B: Condensed Matter, 327, 279-282 (2003)

[5] E. Traversa, S. Matsushima, G. Okada, Y. Sadaoka, Y. Sakai, K. Watanabe, NO2 sensitive LaFe3 thin films prepared by r.f. sputtering, Sensors and Actuators B, 25, 661-664 (1995)

[6] M. Carotta, M. Butturi, G. Martinelli, Y. Sadaoka, P. Nunziante, E. Traversa, Microstructural evolution of nanosized LaFeO3 powders from the thermal decomposition of a cyano-complex for thick film gas sensors, Sensors and Actuators B, 44, 590-594 (1997)

145

[7] J. Yoon, M. Grilli, E. Di Bartolomeo, R. Polini, E. Traversa, The

NO2 response of solid electrolyte sensors made using nano-sized

LaFeO3 electrodes, Sensors and Actuators B, 76, 483-488 (2001) [8] H. Aono, E. Traversa, M. Sakamoto, Y. Sadaoka, Crystallographic

characterization and NO2 gas sensing property of LnFeO3 prepared by thermal decomposition of Ln---Fe hexacyanocomplexes, Ln[Fe(CN)6]·nH2O, Ln = La, Nd, Sm, Gd, and Dy, Sensors and Actuators B, 94, 132-139 (2003)

[9] S. Zhao, J. Sin, B. Xu, M. Zhao, Z. Peng, H. Chai, A high performance ethanol sensor based on field-effect transistor using a LaFeO3 nano-crystalline thin-film as a gate electrode, Sensors and Actuators B, 64, 83-87 (2000)

[10] C. Xiangfeng, P. Siciliano, CH3SH-sensing characteristics of LaFeO3 thick-film prepared by co-precipitation method, Sensors and Actuators B, 94, 197-200 (2003)

[11] K. Arshak, O. Korostynska, S. Clifford, Screen printed thick films of

NiO and LaFeO3 as gamma radiation sensors, Sensors and Actuators A, 110, 354-360 (2004)

[12] C. Xu, J. Tamaki, N. Miura, N. Yamazoe, Grain size effects on gas

sensitivity of porous SnO2-based elements, Sensors and Actuators B, 3, 147-155 (1991)

[13] J.K. Grepstad, Y. Takamura, A. Scholl, I. Hole, Y. Suzuki, T. Tybell, Effects of thermal annealing in oxygen on the antiferromagnetic order and domain structure of epitaxial LaFeO3 thin films, Thin Solid Films, 486, 108-112 (2005)

[14] Y.-H. Lee, J.-M. Wu, Epitaxial growth of LaFeO3 thin films by RF

magnetron sputtering, J. Crystal Growth, 263, 436-441 (2004) [15] J.P. Locquet, J. Perret, J. Fompeyrine, E. Mächler, J.W. Seo, G. Van

Tendeloo, Doubling the critical temperature of La1.9Sr0.1CuO4

using epitaxial strain, Nature, 394, 453-456 (1998) [16] A. Scholl, J. Stöhr, J. Lüning, J.W. Seo, J. Fompeyrine,H. Siegwart,

J.P. Locquet, F. Nolting, S. Anders, E.E. Fullerton, M.R. Scheinfein, H.A. Padmore, Observation of antiferromagnetic domains in

epitaxial thin films, Science, 287, 1014-1016 (2000) [17] E. Traversa, Y. Sadaoka, M. Carotta, G. Martinelli, Environmental

monitoring field tests using screen-printed thick-film sensors based on semiconducting oxides, Sensors and Actuators B, 65, 181-185 (2000)

[18] Y. Mastumoto, J. Hombo, Preparation of LaFeO3 perovskite film

using electrochemical reduction, J. Electroanal. Chem., 348, 441-445 (1993)

[19] S.E Dann, D.B. Currie, M.T. Weller, M.F. Thomas, A.D. Al-Rawwas, The Effect of Oxygen Stoichiometry on Phase Relations

146

and Structure in the System La1-xSrxFeO3-δ (0 ≤ x ≤ 1, 0 ≤ δ ≤ 0.5), J. Solid State Chem., 109, 134-144 (1994)

[20] http://www.mticrystal.com/products_crystal.html [21] N. Tohge, S. Takahashi, T. Minami, Preparation of PbZrO3-PbTiO3

ferroelectric thin-films by the sol-gel process, J. Am. Ceram. Soc., 74 [1], 67-71 (1991)

[22] D.M. Tahan, A. Safari, L.C. Klein, Preparation and characterization

of BaxSr1-xTiO3 thin films by a sol-gel technique, J. Am. Ceram. Soc., 79 [6], 1593-1598 (1996)

[23] http://ts.nist.gov/ts/htdocs/230/232/232.htm

PAPER VI

148

149

Crystallization and surface morphology of oriented LaCoO3 films prepared by three

different sol-gel routes

Paul Inge Dahl, Hasan Okuyucu*, Julian Tolchard Tor Grande, Mari-Ann Einarsrud †

Department of Materials Science and Engineering Norwegian University of Science and Technology

N-7491 Trondheim, Norway

* Present address: Gazi University, Faculty of Technical Education Besevler, 06500 Ankara, Turkey

Abstract

Precursor solutions for preparation of LaCoO3 and La0.8Ca0.2CoO3 films were prepared by three different sol-gel routes. The precursor solutions were synthesized from acetates, alcoxides or nitrate salts of the respective cations (La, Co and Ca), using methanol or 2-methoxyethanol as solvent and suitable chelating agents. The solutions were deposited onto single crystal yttria stabilized zirconia (YSZ) and sapphire substrates by spin- or dip-coating. Oriented polycrystalline films of both LaCoO3 and La0.8Ca0.2CoO3 were grown on the cubic (100)-oriented YSZ substrates. These films were (104)-oriented when indexing according to the hexagonal structure, corresponding to (110)-orientation in the cubic structure, and the orientation was independent of both type of precursor solution and deposition technique. Crystallization of the films started between 800 and 900°C. The most critical influence on surface roughness was the heating rate. Controlled heating, with heating rate of 0.1°C/min to 400°C resulted in smooth homogenous surface morphology (Rq = 0.6 nm) of the prepared films, and smooth surfaces were maintained by heating of these films up to 900°C. Heating to 1000°C was assisted with a severe anisotropic grain growth giving increased surface roughness. Films prepared on hexagonal (0001)-oriented sapphire were polycrystalline and randomly oriented.

† Corresponding author. E-mail: Mari-Ann.Einarsrud@material.ntnu.no

150

Introduction

Lanthanum cobaltite, LaCoO3, with perovskite-type structure (ABO3), exhibit interesting magnetic and electronic behaviour with change in temperature. Depending on the spin-state of the trivalent Co-ions, LaCoO3 goes from being an insulator at low temperatures (T < 35 K), to showing semiconducting behaviour at first, and then at elevated temperatures (T > 650 K), close to metallic conductivity [1]. The rhombohedral crystal structure of LaCoO3 is described by hexagonal lattice parameters as listed in Table 1. At elevated temperatures (above ~1400°C) the material undergoes a phase transition from rhombohedral to cubic structure [2]. Orientation of ferroelastic domains, arising from the cubic to rhombohedral phase transition, gives the LaCoO3-based materials interesting ferroelastic features [3-5]. The mentioned properties may be adjusted by substitution with e.g. calcium, strontium or barium on the A-site [3-6]. Such substitution will lower the rhombohedral to cubic phase transition temperature, and the materials become cubic at ambient temperature for ~50 mol% substitution [2]. Furthermore, mixed ionic/electronic conductivity at elevated temperatures, makes LaCoO3-materials applicable in membrane technology and as cathode materials in solid oxide fuel cells (SOFCs) [7]. LaCoO3 materials are also investigated with respect to catalytic properties [8] and for use in gas sensors [9,10].

The literature reveals thin films of LaCoO3 (pure and Ca-/Sr-substituted) prepared by different techniques, such as pulsed laser deposition (PLD) [10], chemical vapour deposition (CVD) [11], atomic layer epitaxy (ALE) [12], electrochemical oxidation [13], ion beam sputtering [14] and by spray pyrolysis in inductively coupled plasma (spray-ICP) [15]. LaCoO3 films have also been prepared from various sol-gel precursor solutions by spin- or dip-coating [16-25]. Sol-gel precursor solutions have been synthesized by traditionally accepted routes using metal alkoxides dissolved in alcohol with proper chelating agents, such as 2-ethylacetoacetate or polyethylene glycol (PEG) [16-18]. Less expensive variants using nitrate salts or acetates as starting materials for alcohol-based precursor solutions, with e.g. butylacetate or polyvinyl alcohol (PVA) as chelating agents, have also been reported [19-23]. LaCoO3 films have also been prepared from water based routes using ethylendiaminetetraacetic acid (EDTA) or diethylenetriamineoentaacetic acid (DTPA) as complexing agents [23-26]. None of these so-called chemical solution deposition (CSD) techniques have resulted in epitaxial growth of the films. Due to the often superior properties of crystalline epitaxial films, these are more attractive for applications and fundamental studies, compared to randomly oriented polycrystalline films [27].

151

Among the variety of film preparation techniques mentioned above, the present work puts focus on CSD methods. The aim of the work has been to obtain knowledge and experience on film preparation, by fundamental studies of LaCoO3 films prepared by several CSD techniques. Different sol-gel precursor routes have been investigated and films prepared from the different routes, by both spin- and dip-coating, have been compared. Furthermore, the heat treatment of prepared films has been studied, with the goal of obtaining crystalline films with smooth, homogenous surfaces.

Experimental

Synthesis routes

Precursor solutions for preparation of pure and calcium substituted lanthanum cobaltite (LaCoO3 and La0.8Ca0.2CoO3) films, noted LC and LCC in the following, were prepared by three different sol-gel routes. Each synthesis route presented in the following has been named based on the starting precursors used: acetate route, alkoxide route and nitrate route. Literature references to comparable sol-gel routes are given for each of the presented three. However, all routes have been modified from those reported in the literature.

Acetate route [21,22,28]

Stoichiometric amounts of acetates of La, Co and Ca [La(C2H2O2)3·1.5H2O, Co(C2H2O2)2·4H2O and Ca(C2H2O2)2·xH2O, Acros or Alfa Aesar, >98%] were dissolved in methanol [CH3OH, Normapur, >99.8%]. Trifloroacetic acid [CF3COOH, Fluka, >98%] and triethanolamine [N(C2H5O)3, Fluka, 99%) were added as chelating agents. The molar ratio of trifluoroacetic acid/triethanolamine/cations was 6.4/0.30/1.0 and the total cation concentration was 0.07 M. The prepared solutions were stirred for 12 hrs and appeared transparent with a bright violet color.

Alkoxide route [16-18]

Lanthanum isopropoxide [La(C3H7O)3, Gelest, >95%], cobalt acetate and calcium acetate [Co(C2H3O2)2·4H2O, Ca(C2H3O2)2·xH2O, Acros, >99%] were used as starting materials. Precursor solutions were prepared in N2 atmosphere in order to obtain control over hydrolysis/condensation reactions of lanthanum isopropoxide. The solvent, 2-methoxyethanol [C3H9O2, Acros, >99%), was dried over molecular sieves [4 Å, Merck] for

24 hrs and distilled (122 - 124°C) in N2 flow. The alkoxide was dissolved

152

in 2-methoxyethanol and the solution transferred to flasks containing cobalt/calcium acetates, dried to constant weight at 100°C in vacuum. The acetates immediately dissolved giving clear violet solutions. After heating and stirring for 1 h at 60°C, ethylacetoacetate was added as a chelating agent. The molar ratio of chelating agent/cations was 1 and the total cation concentration in the prepared solution was ~0.1 M. The solution was refluxed at 75°C for 2 hrs in nitrogen atmosphere giving a slight colour change from violet to dark pink at first, and then, to intense dark red (LC-sol) or brown (LCC-sol). To start a partial hydrolysis and condensation reaction a small amount of water was added (water/alkoxide molar ratio equal to 1).

Nitrate route [19-22]

Stoichiometric amounts of nitrate salts of La, Co and Ca [La(NO3)3·6H2O, Co(NO3)2·6H2O, Ca(NO3)2·4H2O, Fluka, KeboLab or Merck, >99%] were dissolved in methanol or 2-methoxyethanol. Acetic acid [CH3COOH, Merck, 99.5%] was added and the solutions stirred for 1 h. Acetyl acetone [CH2(COCH3)2, Merck, >99%] was added as a chelating agent. A molar cation/acetic acid/acetyl acetone ratio of 1/6/6 was used and the total cation concentrations in the solutions were 0.07 or 0.20 M. The solutions were

refluxed at 65°C or 124°C for 1 h.

Film processing

Films were prepared by spin- and dip-coating of the precursor solutions onto single crystal substrates of (100)-oriented cubic YSZ [ZrO2 with 9.5 mol% Y2O3, Crystal & Coating Inc.] and (0001)-oriented hexagonal sapphire [Al2O3, MTI Corp.]. The crystal structure, lattice parameters and thermal expansion coefficients (TEC) for substrate and film materials are compared in Table 1. In the dip-coating method, ultrasonically cleaned substrates were dipped into the precursor solution and manually pulled up into a vertical furnace preheated to 550°C. In the spin-coating method, the precursor solution was dispensed at substrates cleaned with methanol and spun at 2000 rpm for 30 sec [Laurell, WS-400A-6NPP/C-1]. Each film layer was dried at 100 – 150°C for 5 – 10 min or heat treated at 400°C for 1 h, with a heating rate of 0.1°C/min. A total of 1 – 5 film layers were deposited. Prepared films were annealed [Entech, MF 1/12] at 800 –

1000°C in ambient air for 20 min – 3 hrs using heating/cooling rates from 0.1 to 3°C/min. Parts of the precursor solutions were evaporated forming hard gels that could be crushed down to powders. The evaporated nitrate precursor solution self ignited, in a volatile combustion reaction around

153

135°C, while decomposition of the two other precursor solutions were more controlled. The synthesized powders were calcined at 1000°C in ambient air for 1 – 12 hrs, giving single phase powders of rhombohedral crystal structure (confirmed by X-ray diffraction).

Table 1 Crystal structure, lattice parameters and thermal expansion

coefficients (TEC) for film [2]

and substrate materials [29,30]

.

Compound Structure Lattice

parameters (Å)

TEC

(10-6

K-1

)

YSZ

Cubic a = 5.140 10.3

Al2O3 Hexagonal a = 4.758 c = 12.99

7.50

LaCoO3 Rhombohedral (hexagonal)

ah = 5.437 ch = 13.09

22.4

La0.80Ca0.20CoO3 Rhombohedral (hexagonal)

ah = 5.427 ch = 13.09

19.3

Characterization

The prepared films and powders were characterized by X-ray diffraction (XRD) [Siemens D5000] with Cu Kα radiation through a primary monochromator. The alignment of the substrate and film was performed using rocking curve scans over the main reflections to record the offset angle. Additional studies of the film crystallisation was performed in situ by high temperature X-ray diffraction (HTXRD) [Siemens D5005] using heating/cooling rates of 2°C/h and 1 h dwell time at each temperature. The morphology of prepared films was studied by atomic force microscopy (AFM) [Digital Instruments, MMAFM-1] in tapping mode using doped silicon tips. The roughness of film surfaces was estimated by calculation of root mean square roughness (Rms, Rq) from scanned 5 x 5 µm2 areas. Average grain size of the crystalline films were estimated from the AFM images using the linear intercept method over 30 – 60 grains.

154

Results

Dip-coated films from acetate solution

LC/LCC films completely covered the YSZ substrate surfaces after dip-

coating and heat treatment at 550°C. The morphology of dip-coated LC films on YSZ is presented by AFM micrographs in Fig. 1. The as-prepared films after heat treatment at 550°C showed uniform surfaces (Rq = 0.3 nm), as indicated by Fig. 1 a). After annealing at 900°C for 20 min, with a heating rate of 3°C/min (Fig. 1 b)), the surface was still smooth (Rq = 2.7 nm) and the formation of grains had taken place. Annealing at 1000°C for 20 min led to a significant increase in surface roughness (Rq = 22 nm), as demonstrated by Fig. 1 c). The increase in annealing temperature from 900 to 1000°C was accompanied by grain growth, average grain size increasing from ~300 nm to ~580 nm.

X-ray diffractograms of the dip-coated films prepared on YSZ substrates from the acetate solutions are shown in Fig. 2 an X-ray diffractogram of synthesized powder is included for comparison. Fig. 2 a) shows that both LC and LCC films annealed at 1000°C are (104)-oriented, as only reflections corresponding to the (104)- and (208)-planes in the diffractogram for LC powder, can be observed. The HTXRD diffractograms in Fig. 1 b) shows that crystallization of the LC film takes place between 850 and 900°C.

Spin-coated films from alkoxide solution

Homogenous films of LC and LCC were obtained by spin-coating of the alkoxide solutions onto YSZ substrates. AFM micrographs of the films are shown in Fig. 3. As indicated by Fig. 3 a), fast heating (3°C/min) to 800°C (3 hrs) results in formation of islands on top of a finer grained underlying film surface (Fig. 3 b)). Annealing at 1000°C for 3 hrs resulted in complete domination of islands on the surface, hence the rough morphology shown in Fig. 3 c). The surface roughness after fast heating to 800 and 1000°C, was calculated to Rq = 4.4 and 28 nm, respectively. The corresponding increase in grain/particle size was from ~75 to ~400 nm. Slow heating (0.1°C/min) to 400°C produced smooth films (Rq = 0.6 nm) as demonstrated by the AFM micrograph in Fig. 3 d). Annealing of such film at 800 for 1 h, with a heating rate of 1°C, resulted in a homogenous morphology (Fig. 3 e)) and surface roughness of 4.0 nm. Further annealing at 1000°C for 1 h, resulted in island formation and a rather rough surface (Rq = 9.8 nm), as indicated by Fig. 3 f).

155

Fig. 1 AFM micrographs of LC films prepared by dipcoating in acetate

precursor solution; a) as prepared film (heat treated at 550°C),

b) and c) films annealed at 900°C and 1000°C, respectively, for

20 min with heating rate of 3°C/min.

156

a)

2 θ θ θ θ (°)

25 30 35 40 45 50 55 60 65 70 75

Re

lati

ve

in

ten

sit

y

LaCoO3-δ powder

YS

Z (

10

0)

La0.8

Ca0.2

CoO3-δ

film

YS

Z (

20

0)

LaCoO3-δ

film

(110

)

(10

4)

(20

2)

(006

)

(12

2)

(11

6)

(300

)

(01

8) (2

14

)

(22

0)

(02

4)

(20

8)

b)

2 θ θ θ θ (°)

32.5 33.5 34.5

Re

lati

ve

in

ten

sit

y

850°C

900°C

950°C

LaCoO3-δ film

Fig. 2 a) X-ray diffractograms of LC and LCC films, prepared on YSZ

substrates by dipcoating with acetate-sol and annealed at

1000°C for 20 min. The diffractogram of powder obtained by

evaporation of sol and calcination at 1000°C for 12 hrs is added

for comparison. Indexing (hkl) based on hexagonal structure [31]

. b) X-ray diffractograms of LC film prepared on YSZ

substrates by dipcoating with acetate-sol, obtained by in situ

HTXRD.

157

Fig. 3 AFM micrographs of LCC films obtained by spincoating of

alkoxide precursor solution onto YSZ substrates. a) and b) 5-

layered film Annealed at 800°C for 3 hrs with a heating rate of

158

3°C/min, c) 5-layered film annealed at 1000°C for 3 hrs with a

heating rate of 3°C/min, d) 1-layered film heat treated at 400°C

for 1 h with a heating rate of 0.1°C, e) and f) 1-layered film heat

treated as in d) followed annealing at 800 and 1000°C,

respectively, for 3 h.

The X-ray diffractograms of LC and LCC films in Fig. 4 demonstrate the evolution of crystallization with increasing temperature from 800 to

1000°C. For both compositions, only one reflection, corresponding to the (104)-reflection in the powder diffractogram, can be assigned to the films. This shows that the films of both LC and LCC were textured. Rocking-curve scans confirmed this by showing similar offset angles for the films compared to the substrate. As an example the offset angle for a LCC film annealed at 1000°C for 3 hrs was measured to 0.41°, compared to 0.44° for the substrate. Some smaller reflections (that may be assigned to Co3O4) could be detected, indicating a slight deviation from stoichiometric composition of the precursor solutions.

Spin- and dip-coated films from nitrate solutions

Due to poor wetting properties, films on YSZ substrates where not obtained by spin-coating of the nitrate-methanol precursor solution. This effect may be due to the different acidity of the substrate materials (zirconia being acidic, alumina having amphoteric properties). Solutions prepared without acetic acid or with 2-methoxyethanol as solvent did not improve the wetting on the YSZ substrates. Films were however successfully prepared on YSZ substrates by dip-coating and rapid heat treatment. The AFM micrographs in Fig. 5 a) and b) shows the surface morphology of dip-coated films from nitrate precursor solution after annealing at 800 and 1000°C, respectively, for 1 h with a heating rate of 3°C/min. The corresponding surface roughness (Rq) actually decreased from 17 to 11 nm, however the homogenous grain structure (~140 nm) obtained after annealing at 800°C was lost when increasing the temperature to 1000°C. The X-ray diffractograms of dip-coated film annealed at 800 and 1000°C, given in Fig. 5 c), are equivalent to those in Fig. 2 and 4, showing (112)-orientation of the films.

Homogenous LC films were obtained by spin-coating of methanol based precursor solution onto (0001) sapphire substrates, as demonstrated by the AFM micrograph in Fig. 6 a). Surface roughness and average grain size for such film, annealed at 900°C, was 2.8 nm and ~220 nm, respectively. XRD

159

of this film (Fig. 6 b)) reveals reflections assigned to several crystal planes in LaCoO3, indicating random orientation. The intensity of these reflections was low, compared to those obtained for textured films, which is reasonable as all crystal planes are not oriented in the same direction.

Estimated d-values for (104)-reflections selected of LC / LCC powders and films on (100) YSZ substrates, prepared by the three different techniques and annealed at 1000°C, are listed in Table 2. As seen from Table 2, all film d-values match well with those for powders, indicating no significant lattice distortion perpendicular to the (104)-planes.

2 θ θ θ θ (°)

20 25 30 35 40 45 50

Re

lati

ve in

ten

sit

y

LaCoO3 powder

YS

Z (

10

0)

(01

2)

(10

4)

LaCoO3 films

(11

0)

(20

2)

(006

)

(02

4)

La0.8

Ca0.2

CoO3 films

800°C900°C1000°C

800°C1000°C

1000°C

Fig. 4 X-ray diffractograms of 5-layered LC and LCC films, prepared

by spincoating of alkoxide-sol onto YSZ substrates and annealed

at indicated temperatures for 3 hrs. The diffractograms of

powder obtained by evaporation of sol and calcination at

1000°C for 12 hrs is added for comparison. Hexagonal

indexing (hkl) [31]

.

160

c)

2 θ θ θ θ (°)

25 30 35 40 45 50 55 60 65 70 75

Rela

tive in

ten

sit

y

Powder 1000°C

YS

Z (

100)

Film 1000°C

YS

Z (

20

0)

Film 800°C

(110)

(104)

(202)

(006)

(122)

(116)

(30

0)

(018) (2

14)

(220)

(024)

(208)

Fig. 5 AFM micrographs of LC films on YSZ substrate dipcoated from

nitrate precursor solution and annealed for 1 h at a) 800 and b)

1000°C, with a heating rate of 3°C/min. c) Corresponding X-

ray diffractograms with hexagonal indexing (hkl) [31]

.

161

a)

b)

2 θ θ θ θ (°)

20 30 40 50 60 70

Re

lati

ve

lo

ga

rith

mic

in

ten

sit

y

(110

)

(104

)

(20

2)

(00

6)

(12

2)

(11

6)

(300

)

(018

) (2

14)

(22

0)

(02

4)

(01

2)

Substrate

LaCoO3 film

LaCoO3 powder

Fig. 6 a) AFM micrograph of 5-layered LCfilm on (0001)-oriented

sapphire substrate annealed at 900°C for 1 h. b) X-ray

diffractograms of (0001)-oriented sapphire substrate with and

without LC film annealed at 900°C for 1 h. Arrows indicate

reflections assigned to the film. Powder diffractograms, with

hexagonal indexing (hkl) [31]

, is added for comparison.

162

Table 2 Comparison of d-values obtained from XRD on LC/LCC

powders and films, all heat treated at 1000°C.

Compound Precursor

solution Technique d104 (Å) d208 (Å)

LaCoO3 Acetate Powder 2.6868

LaCoO3 Acetate Dip coating 2.6849 1.3438

La0.80Ca0.20CoO3 Acetate Powder 2.6802

La0.80Ca0.20CoO3 Acetate Dip coating 2.6801 1.3409

LaCoO3 Alkoxide Powder 2.6862

LaCoO3 Alkoxide Spin coating 2.6892

La0.80Ca0.20CoO3 Alkoxide Powder 2.6804 1.3424

La0.80Ca0.20CoO3 Alkoxide Spin coating 2.6869 1.3447

LaCoO3 Nitrate Powder 2.6858 1.3466

LaCoO3 Nitrate Dip coating 2.6821 1.3453

Discussion

Film orientation

Films of LC and LCC grown on (100)-oriented YSZ showed preferential growth, independent of precursor solution and deposition technique. The observed (104)-orientation (hexagonal indexing, equivalent to rhombohedral (112)) have also been reported for La0.8Sr0.2CoO3 films grown on (100)-oriented YSZ by sputtering techniques [32], however no explanation for this growth orientation was given.

The (104)-planes in the hexagonal structure of the LaCoO3 correspond to the cubic (110)-planes, as obtained from the transformation matrices presented in Eq. (1). Subscript indexes c and h notes cubic and (h k l) values, respectively.

163

( ) ( )

=

61

61

61

31

31

31

31

31

32

hhhccc lkhlkh (1)

Considering the correlation between the hexagonal (104) and the cubic (110)-planes may assist the understanding of how LaCoO3 films may orient in this direction on top of cubic (100)-oriented YSZ, despite the different crystal structures. However, in order to explain such orientation of the film the interatomic distances in the (104)-plane of LaCoO3 must be compared with those of the YSZ (100)-plane. In Fig. 7, this is done by comparing the (100)-surface of cubic YSZ with the (104)-surface of LaCoO3 (Fig. 7 a) and b), respectively). The interatomic O – O / La – La distance in the LaCoO3 ccp in the horizontal direction of the illustrated (104)-plane is 3.87 Å (Fig. 7 b)). Compared to the interatomic O – O distance of 2.57 Å in YSZ (Fig. 7 a)), this makes for close to perfect repetitive stacking in this direction, as 3 O- or La-atoms in LaCoO3 fits on top of 4 octahedral holes on the YSZ surface, with a mismatch of < 0.4%. In the vertical direction of the illustrated crystal planes (Fig. 7) the interatomic O – La distances in LaCoO3 are 2.46 and 3.04 Å, the latter giving a more significant mismatch (~18%) with the O – O distance in YSZ. This mismatch may be reduced by a local compression of the film, in this direction, at the interface between the film and the substrate.

As the illustrated vertical interatomic distances in the film are larger than those of the underlying substrate, an ordered the film is expected to be under compressive stress. Some of the stress introduced from lattice mismatch may have been relieved at the grain boundaries, and in the case of fast heated films assisted by severe grain growth/island formation. Additionally, the compressive stress from lattice mismatch may to some extent be extinguished by tensile stress introduced during cooling of the films, due to the higher TEC of LC/LCC compared to YSZ (Table 1). In Fig. 7 c), the two planes, illustrated in Fig. 7 a) and b), are overlaid, illustrating a possible way for oriented growth to occur. A similar match to the (0001)-surface in Al2O3 was not feasible, hence orientated growth did not occur in films grown on sapphire substrates.

164

Fig. 7 Surface structures of a) (001)-plane YSZ [29]

, b) (104)-plane

LaCoO3 [2]

and c) the latter structure on top of the first one,

suggesting a possible way of oriented growth.

Surface morphology

The surface morphology of the films has been a central part in this study. The prepared films were polycrystalline, indicating that growth has started by nucleation at several sites on the substrate surface. Both the heating rate and film preparation technique are important for obtaining uniform films with a smooth surface (here defined by Rq < 5 nm). It was demonstrated in Fig.1 how smooth films on YSZ were obtained from the acetate precursor solution by dip-coating and annealing up to 900°C. The heating rate was not critical for these films as the dip-coated films were directly pulled up

165

into a furnace preheated at 550°C. An increase in the annealing temperature, from 900 to 1000°C (Fig. 1 b) and c), respectively) resulted in a significant roughening of the surface (Rq = 2.7 and 23 nm, respectively). A possible explanation for this phenomenon is anisotropic growth of the grains in a preferred direction (perpendicular to the (104)-plane). The film crystallization was shown to start between 850 and 900°C, so after annealing at 900°C for only 20 min the film may not have been fully crystalline. Dip-coated films made from the nitrate precursor solution showed rougher surfaces compared the film obtained from acetate precursor annealed at 900°C (Fig. 1 b), Rq = 2.7 nm), even after annealing at 800°C (Fig. 5 a), Rq = 18 nm). As the temperature was 100°C lower, the longer annealing time (1 h versus 20 min) can not explain this difference. More likely the rough surface in the dip-coated films from nitrate precursor solution is due to the vigorous combustion reaction observed at 135°C for the precursor. By pulling the dip-coated films directly into a furnace at 550°C, this combustion reaction was allowed to happen uncontrolled, hence causing a rough surface morphology before crystallization.

Controlled heating (0.1°C/min up to 400°C followed by 1°C/min to 900°C) of a film prepared from the nitrate sol by spin-coating onto sapphire, resulted in a homogenous film with a smooth surface (Fig. 6 a)). For fast heating (3°C/min) of films spin-coated from the alkoxide precursor solution, the roughening of the surfaces were found to start around 800°C, and further increase of annealing temperature gave a severe anisotropic (104)-oriented grain growth. Again, controlled heating (0.1°C/min) gave smooth films after heating up to 400°C (Fig. 3 d)), followed by annealing for 1 h at 800°C, with a heating rate of 1°C/min (Fig. 3 e)). As for the dip-coated films, increase in the annealing temperature to 1000°C resulted in a significant grain growth and roughening of the surface.

Based on these results we conclude that controlled heating up to 400°C improved the film quality. This temperature range corresponds to where the organics (solvent, chelating agents) and nitrates decompose, hence the controlled heating in this range is critical. Uncontrolled decomposition of these species may give rough surface morphology.

Conclusion

Highly oriented LaCoO3-based films can be grown on (100)-oriented YSZ substrates by both spin- and dip-coating, independent of preparation route for precursor solution. The observed (104)-orientation of these films includes some lattice mismatch, resulting in compressive stress on the

166

interface between film and substrate. As the films are polycrystalline some of this stress may be relieved on the grain boundaries, and possibly to some extent be extinguished by tensile stress introduced during cooling of the LaCoO3 films, due to the higher TEC values compared to that of the YSZ substrates. Controlled decomposition of organics and nitrates in the precursor solutions, performed by slow heating (0.1°C/min) up to 400°C, followed by annealing at 800 – 900°C, seemed to be the key to obtaining homogenous films with smooth surfaces (Rq < 5 nm). These annealing temperatures should not be exceeded, as the surface roughness increased quite dramatically when the annealing temperature was increased to 1000°C (Rq up to 28 nm).

Acknowledgement

Laura Bertolo is acknowledged for experimental work done on LC films prepared by the alkoxide route and Dr. Mohan Menon for his contribution on heat treatment of these.

Literature

[1] M.A. Señarís-Rodríguez, J.B. Goodenough, Magnetic and transport

properties of the system La1-xSrxCoO3-δ (0 < x ≤ 0.50), J. Solid State Chem., 118, 323-336 (1995)

[2] J. Mastin, M.-A. Einarsrud, T. Grande, Crystal structure and thermal

properties of La1-xCaxCoO3-delta (0 <= x <= 0.4), Chem. Mater., 18, 1680-1687 (2006)

[3] K. Kleveland, N. Orlovskaya, T. Grande, A.M.M. Moe, M.-A. Einarsrud, Ferroelastic behaviour of LaCoO3-based ceramics, J. Am. Ceram. Soc., 84, 2029-2033 (2001)

[4] S. Faaland, P.E. Wullum, R. Holmestrand, T. Grande, M.-A. Einarsrud, Stress-strain behaviour during compression of

polycrystalline La1-xCaxCoO3-based ceramics, J. Am. Ceram. Soc., 88 [3], 726-730 (2005)

[5] J. Mastin, H.L. Lein, T. Grande, M.-A. Einarsrud, Mechanical

properties of LaCoO3-based materials, to be published [6] M.A. Señarís-Rodríguez, M.P. Breijo, S. Castro, C. Rey, M.

Sanchez, R.D. Sanchez, J. Mira, A. Fondado, J. Rivas, Peculiarities

in the electrical and magnetic properties of cobalt perovskites Ln1-

167

xMxCoO3 (Ln3+

: La3+

, M2+

: Ca2+ , Sr2+

, Ba2+

; Ln3+

: Nd3+

, M2+

:

Sr2+), Int. J. Inorg. Mater., 1, 281-287 (1999)

[7] V.V. Kharton, E.N. Naumovich, A.V. Kovalevsky, A.P. Viskup, F.M. Figueiredo, I.A. Bashmakov, F.M.B Marques, Mixed electronic

and ionic conductivity of LaCo(M)O3 (M=Ga, Cr, Fe or Ni), Solid State Ionics, 138, 135-148 (2000)

[8] R.N. Singh, B. Lal, High surface area lanthanum cobaltate and its A

and B sites substituted derivatives for electrocatalysis of O2

evolution in alkaline solution, Int. J. Hydrogen Energy, 27, 45-55 (2002)

[9] E. Brosha, R. Mukundan, D.R. Brown, F. H. Garzon, J.H. Visser, M. Zanini, Z. Zhou, E.M. Logothetis, CO/HC sensors based on thin

films of LaCoO3 and La0.8Sr0.2CoO3-8 metal oxides, Sensors and Actuators B, 69, 171-182 (2000)

[10] D.T.V. Anh, W. Olthuis, P. Bergveld, Sensing properties of

perovskite oxide La0.5Sr0.5CoO3-8 obtained by using pulsed laser

deposition, Sensors and Actuators B, 103, 165-168 (2004) [11] M. Losurdo, A. Sacchetti, P. Capezzuto, G. Bruno, L. Armelao, D.

Batteca, G. Bottaro, A. Gasparotto, C. Maragno, E. Tondello, Optical and electrical properties of nanostructure LaCoO3 thin

films, Applied Physics Letters, 87, 0601909 (2005) [12] H. Seim, M. Nieminen, L. Niinistö, H. Fjellvåg, L.-S. Johansson,

Growth of LaCoO3 thin films from β-diketonate precursors, Applied Surface Science, 112, 243-250 (1997)

[13] Y. Matsumoto, T. Sesaki, J. Hombo, A new preparation method of

LaCoO3 perovskite using electronic oxidation, Inorg. Chem., 31, 738-741 (1992)

[14] T. Hattori, T.Matsui, H. Tsuda, H. Mabuchi, K. Morii, Fabrication

and electric properties of LaCoO3 thin films by ion-beam sputtering, Thin Solid Films, 388, 183-188 (2001)

[15] H. Ichinose, H. Katsuki, M. Nagano, Deposition of LaMO3(M=Co,

Cr, Al) films by spray pyrolysis in inductively coupled plasma, J. Crystal Growth, 144, 59-64 (1994)

[16] H.J. Hwang, J. Moon, M. Awano, K. Maeda, Sol-Gel Route to

porous lanthanum cobaltite (LaCoO3) thin films, J. Am. Ceram. Soc., 83, 2852-2854 (2000)

[17] H.J. Hwang, M. Awano, Preparation of LaCoO3 catalytic thin film

by the sol-gel process and its NO decomposition characteristics, J. Europ. Ceram. Soc., 21, 2103-2107 (2001)

[18] H.J. Hwang, A. Towata, M. Awano, K. Maeda, Sol-Gel route to

perovskite-type Sr-substituted LaCoO3 thin films and effects of

polyethylene glycol on microstructure evolution, Scripta Mater., 44, 2173-2177 (2001)

168

[19] B. Trummer, O. Fruhwirth, K. Reichmann, M. Holzinger, W. Sitte, P. Pölt, Preparation and characterisation of LaNixCo1-xO3, J. Europ. Ceram. Soc., 19, 827-829 (1999)

[20] S. Javorič, G. Dražič, M. Kosec, J. Europ. Ceram. Soc., 21, 1543-1546 (2001)

[21] E. Bontempi, L. Armelao, D. Barreca, L. Bertolo, G. Bottaro, E. Pierangelo, L.E. Depero, Structural characterized of sol-gel

lanthanum, cobaltite thin films, Crystal Engineering, 5, 291-298 (2002)

[22] M. Gelfi, E. Bontempi, R. Roberti, L. Armelao, L.E. Depero, Residual stress analysis of thin films and coatings through XRD

2

experiments, Thin Solid Films, 450, 143-147 (2004) [23] Y. Zhang, Y. Zhu, R. Tan, W. Yao, L. Cao, Influence of PEG

additive and precursor concentration on the preparation of LaCoO3

film with peroskite structure, Thin Solid Films, 388, 160-164 (2001) [24] Y. Zhu, R. Tan, J. Feng, S. Ji, L. Cao, The reaction and poisoning

mechanism of SO2 and perovskite LaCoO3 film model catalysts,

Applied Catalysis A: General, 209, 71-77 (2001) [25] L. Hong, X. Chen, Z. Cao, Preparation of a perovskite

La0.2Sr0.8CoO3-x membrane on a porous MgO substrate, J. Europ. Ceram. Soc., 21, 2207-2215 (2001)

[26] Y. Zhu, R. Tan, T. Yi, S. Ji, X. Ye, L. Cao, Preparation of nanosized

LaCoO3 perovskite oxide using amorphous heteronuclear complex

as a precursor at low temperature, J. Mater. Sci., 35, 5415-5420 (2000)

[27] D.P. Norton, Synthesis and properties of epitaxial electronic oxide

thin-film materials, Mater. Sci. Engin. R, 43, 139-247 (2004) [28] T. Schuller, M.A. Aegerter, Optical, electrical and structural

properties of sol gel ZnO:Al coatings, Thin Solid Films, 351, 125-131 (1999)

[29] http://www.coatingandcrystal.com/html/super.html [30] http://www.mticrystal.com/products_crystal.html [31] V.V. Sikolenko, E.V. Pomjakushina, S.Y. Istomin, Neutron

diffraction studies of La1-xSrxCoO3 magnetic structure at x = 0.15

and x = 0.30, J. Magnetism and Magnetic Materials, 258-259, 300-301 (2003)

[32] E.L. Brosha, B.W. Chung, F.H. Garzon, I.D. Raistrick, R.J. Houlton, M.E. Hawley, In Situ Growth and Characterization of La0.8Sr0.2CoO3

Perovskite Mixed Conductor Films, J. Electrochem. Soc., 142 (5), 1702-1705