Effect of micro- and nano-structures on the properties of ionic conductors

Solid State Ionics 70/71 (1994) 83-95

North-Holland

SOLID STATE IOWlCS

Effect of micro- and nano-structures on the properties of ionic conductors

S.P.S. Badwal CSIRO, Division of Materials Science and Technology, Private Bag 33, Rosebank MDC, Clayton, Victoria, Australia 3169

and

S. Rajendran Ceramic Fuel Cells Limited, 710 Blackburn Road, Clayton, Victoria, Australia 3168

Many ceramics are commonly used in solid electrolyte devices. Micro- and nano-structural features, apart from intrinsic behav-

iour, play a critical role in defining their electrical and mechanical properties. An overview of the effect of powder and ceramic

processing techniques on the micro- or nano-structures and phase assemblage has been given for different types of ceramics with

main emphasis on zirconia-based materials. The effect of micro- and nano-structural features both within the grains and across

grain boundaries on both the electrical and mechanical properties has been discussed along with nano-phase processing of ce-

ramics and the actual micro- or nano-structural features in the sintered ceramics. The effect of grain size and grain boundary

phases on superplastic deformation and degradation of tetragonal zirconia in moist environments has been considered.

1. Introduction

A large number of ceramic materials, because of their favourable ion transport properties, are used in sensors, batteries, fuel cells, electrochemical reactors and thermoelectric converters [ l-3 1. The initial se-

lection of the electrolyte materials is usually made based on their ionic transport properties in device operating environments. However, in terms of device fabrication and operation, other properties such as mechanical strength, toughness and creep behaviour of the ceramic play an important role and cannot be ignored. The micro- or nano-structural features in ceramic materials such as grain size, porosity, grain shape, grain boundaries, phase assemblage, coherent growth of precipitates, inhomogeneous composition phase boundaries, deliberate inclusions and local ordering (formation of micro domains or modulated structures) affect both electrical and mechanical properties and therefore performance of an electrochemical cell.

Tailoring of microstructures, to control or decrease flaw size, in the past has resulted in consid-

erable improvements to mechanical properties (strength, toughness) of ceramics [ 4-71. Strengths in excess of 2 GPa have been reported for partially stabilised zirconia ceramics. Micro- and nano-structural features to a large extent depend upon powder preparation, powder processing, and ceramic fabrication techniques and processing procedures. Spe- cial micro- or nano-structural features in the sintered ceramic can be created by careful control of powder morphology, sintering and annealing conditions, and the resultant phase assemblage. Alternatively ther- modynamically inert inclusions are incorporated at

micro- or nano-scale to produce composites and to deliberately alter properties of materials. However, incorporation or growth of secondary phases can considerably alter ionic transport properties of ceramic electrolytes and often a compromise is nec- essary between mechanical and electrical properties. Powders from which ceramics are made often contain impurities, which are present due to the high cost of powder purification or sometimes are deliberately added to enhance sintering. These impurities tend to segregate at the grain boundaries and at the external

0167-2738/94/S 07.00 0 1994 Elsevier Science B.V. All rights reserved.

84 S.P.S. Badwal, S. Rajendran /Micro- and nano-structures

surfaces during sintering and subsequent heat treatment of the ceramic. Depending upon the conducting properties of these phases and their quantity and

location along grain boundaries, the effect of grain boundary segregation on the electrical properties of a ceramic electrolyte and electrochemical performance of the device may be quite significant [ 8,9].

In this paper an overview of the effect of powder and ceramic processing techniques on the micro- or

nano-structures and phase assemblage has been given for different types of ceramics with main emphasis

on zirconia-based materials. Zirconia-based ceramics have attracted considerable attention because of the interesting electrical and mechanical properties which can be obtained for different dopants, microstructures and phase assemblage. The effect of micro- and nano-structural features within the grains and across grain boundaries on both the electrical and mechanical properties has been discussed. The paper is not intended as a thorough review article on

micro- or nano-structures and their influence on material properties. The information reported in the lit- erature is rather extensive and cannot be covered in this short article.

bution at atomic level, high density and uniform grain size distribution) in the sintered ceramic and to lower

ceramic processing temperatures, and can produce materials with superior properties. However, the grain size in the sintered ceramics is often not in the nano-size range (i.e. less than 0.1 pm). Neverthe- less, nano-structural features developed in the sintered ceramic, for example by annealing in the two phase field, play an equally important role in defining electrical and mechanical properties of ceramic materials, as discussed below.

3. Powder preparation and processing

2. Micro- and nano-structures

Most materials of interest have micro- or nano- structural features of some kind which may have a significant effect on materials properties. It has been shown in the past that as the microstructural features vary from macro (>lOO pm) to micro (0.1-100 urn) to nano scale ( l-100 nm), material properties can be altered significantly [ lo]. These microstructural features in the micro- or nano-scale within the grains and/or across grain boundaries may be created by deliberate microstructural engineering e.g. by the control of powder morphology, powder or ceramic processing procedures, or may be the result of natural thermodynamic equilibrium constraints.

It is essential to control properties of precursor powders as powder characteristics such as size, shape, morphology and distribution of particles, and strength and structure of agglomerates and aggregates play an important role in defining the microstructure and properties of sintered ceramics. With the availability of appropriate powders, fabrication of sintered ceramics with submicron grain size is fea- sible. Extremely fine powders, with very low impurity content and free of aggregates, are preferred. Powder processing techniques to remove or break agglomerates, powder compaction and control over firing of green compacts are equally important in obtaining high performance ceramics. A number of methods have been developed over recent years for the production of single or multicomponent oxide powders with high surface area and high reactivity. These include precipitation or coprecipitation (hydroxide, oxalate, etc. ), sol-gel synthesis, spray or drip pyrolysis, polymeric methods, cryochemical techniques, emulsion precipitation and hydrothermal techniques [ 11,12 1. Depending upon the technique used and processing of powders, the quality of the precursor powder and ceramic microstructure can be quite different as illustrated below with some examples.

Often there is confusion between nano-phase pro- Fig. la and b shows micrographs of calcined alu-

cessing of ceramics and the actual micro- or nano- mina powders synthesised by the hydroxide precip- structural features in the sintered ceramic. There is itation technique (ammonium hydroxide and alu-

no doubt that nano-phase processing of powders, (e.g. minium nitrate solutions) [ 131. For the powder

nanocrystalline, high surface area, high reactivity shown in fig. 1 b, 5% a-Al,O, seed particles (crystals

powders free of hard agglomerates) can lead to much of about 100-200 nm in size) were dispersed in the

better microstructures (more homogeneous distri- aluminium nitrate solution before precipitation. The

S.P.S. Badwal, S. Rajendran /Micro- and nano-structures 85

j I”:..

Fig. 1. Powder and sintered ceramic microstructure of alpha-alumina. (a) Precipitated powder calcined at 12OO”C, (b) prepared from seeded hydrous alumina by calcining at 95O”C, (c): powder (a) sintered at 1600°C. (d): powder (b) sintered at 1450°C.

powder prepared without seed particles had to be

calcined at 1200°C to completely transform to a- A&O3 and required to be sintered at 1600°C to ob- tain a ceramic with density > 98% of the theoretical. The average grain size was 9 pm (fig. lc). The powder prepared with 5% a-Al,@ seed particles almost fully transformed to a-Alz03 at 950°C and could be sintered to >99% density by sintering at 1450°C. The average grain size in this material was less than 1 .O urn (fig. Id). The fracture strength was 460 MPa in the first case and it was about 800 MPa in the case

of powder prepared with a-A&O3 seed particles dispersed in aluminium nitrate solution.

Special care is required when preparing multicomponent phases. The homogeneity of the starting powders plays a dominant role in defining the ceramic

microstructure. Fig. 2 shows microstructures of two Y-TZP (2.5 mol% Yz03-ZrO*)-30 wt% alumina ceramics. The powders were prepared by the hydroxide precipitation method but with variation in the precipitation procedure. Two grossly different stirring speeds were used during the addition of nitrate

86 SAY. Badwal, S. Rajendran /Micro- and nano-structures

Fig. 2. Sintered (1600°C) and HIP’ed ( 1SOO’C) ceramic micro-

structures of 2.5 mol% Y20,-ZrO*-30 wt% Al,OJ powders prepared by (a) lower stirring rate (b) higher stirring rate during

co-precipitation.

solution of Y, Zr and Al in to the ammonium hydroxide solution. The phase distribution is significantly different in both cases being more uniform in the ceramic obtained from the powder prepared using the fast stirring speed [ 141.

Precursor oxide powders often contain both soft and hard agglomerates. A number of different hy- potheses have been put forward over the years to ex- plain the formation of aggregates of strongly bonded crystallites [ 15-l 71. The structure, strength and size of agglomerates depends on the calcination conditions and the powder morphology or gel structure obtained with a particular powder synthesis technique. In any case, green compacts prepared from powders containing both strongly and weakly bonded agglomerates will have dense and loosely packed regions and an uneven pore structure. Sintering these compacts will produce ceramics with inhomoge-

neous grain growth, low density, large pores, high density of flaws and consequently poor electrical and mechanical properties.

Powders with soft agglomerates can be sintered to near theoretical density at lower temperatures. Ni- cholson [ 18 ] showed that the removal of hard agglomerates improves sintering kinetics and produces ceramics with low flaw density and high strength. For yttria partially stabilised zirconia an improvement in the strength from 880 MPa to 1400 MPa was demonstrated with this secondary treatment of the powder.

Rajendran [ 141 has shown that for the hydroxide precipitation route, powders free of hard agglomerates can be produced by controlling gel polymeris- ation and condensation processes during the initial stages of heating. The process involved simple treatment of gels, produced by the normal precipitation process, with an organic solvent such as an alcohol. The treatment apparently reduces the amount of free water and induces surface alkoxide groups which re- sist cross linking and condensation process therefore reducing aggregation of clusters into strongly bonded crystallites. Fig. 3 shows the microstructure of 2.5 mol% Y203-ZrO,-5 wt% AllO, coprecipitated powder treated with alcohol and calcined at 1000’ C. The powder has a uniform microstructure with tine crystallites (20-40 nm range). On sintering at 1450°C. the ceramic densified to >99% of the theoretical density. The same powder calcined at the same conditions without any treatment with alcohol showed strongly bonded crystallites over 100 nm in size and produced ceramics with a maximum density of only 85% of the theoretical even after sintering at 1700°C (fig. 3). Further improvements to the hydroxide

S.P.S. Badwal, S. Rajendran /Mcro- and nano-structures 87

Fig. 3. Powder (calcined at 1000°C for two hours) and ceramic (sintered at 16OO”C, HIP’ed at ISOO°C) microstructures of 2.5 mol%

YzO,-ZxQ-5 wt% Alz03 specimens: (a) calcined powder and (b) specimen dispersed in iso-propanol first and then calcined, (c)

ceramic prepared from powder (a), and (d ) ceramic prepared from powder (b )

precipitation process has been made by better dis- persion of precipitated powder in an alcohol e.g. by a high speed shear mixing process. The 2-2.5% YQ- ZrOz powders produced with this procedure could

be sintered between 1150°C and 1300°C to near theoretical density [ 191. The calcined powder par- title size was less than 10 nm and grain size in the sintered ceramic was between 100-300 nm (fig. 4a


Fig. 4. Powder calcined at 550°C for one hour and then sintered (1150°C for two hours) ceramic microstructures of 2 mol% Y,O,-ZrO,. The precipitated specimen was dispersed in ethanol with high speed shear mixer before any heat treatment.

and b). The green compacts sintered at 1250°C ex- hibited maximum fracture strength of 1550 MPa.

Van de Graaf et al. [ 201 have developed a special precipitation and powder processing technique to produce weakly agglomerated fine tetragonal Yz03- Zr02 powders using metal chlorides and ammonium hydroxide. These authors report an average grain size of 100-500 nm depending on the conditions used and density > 95% of theoretical for 1250°C sintered ceramics. However, the fracture strength was some- what lower (480-550 MPa) [21] than usually ob-

served for such materials (in the vicinity of 800- 1000 MPa).

Advanced processing techniques such as colloidal processing and consolidation [ 22,231 and electro- phoretic deposition can help to produce green compacts with high density and uniform pore size. Sin- tering of green compacts prepared by these techniques have been reported to produce small grain size ceramics and high density at low sintering temperatures.

4. Micro- and nano-structures in solid electrolytes

In ceramic electrolyte materials, ionic transport can be divided into two main components: volume or lattice conductivity resulting from ion-transport through the grains, and grain boundary resistivity due to partial blocking of migrating species at the grain boundaries. The electrical, thermal and mechanical properties of ceramics depend on the microstructure within the grains and across grain boundaries. The ceramic grain size, pores, flaws, phase assemblage and composition of grain boundary phases all contribute to the material properties.

4.1. Intragrain features and their injluence on properties

Micro- and nano-structural features within the grains such as coherent growth of precipitates of an- other phase, second phase inclusions, compositional variations and formation of microdomains, modulated structures or ordered phases can all alter mechanical and electrical properties. Although in most instances the usual emphasis is on obtaining single phase materials with uniform microstructures, special micro- and nano-structural features within the ceramic can be used to a considerable advantage. For example, the transformation toughening mechanism used to enhance mechanical strength and toughness of partially stabilised zirconia takes advantage of the special nano-structural features and phase assemblage created deliberately (growth of tetragonal zirconia precipitates and their size by controlled annealing) [ 4-7, lo]. In the preparation of MgO- partially stabilised zirconia (MgO-PSZ) conven- tional powders and ceramic processing routes (pow-

S.P.S. Badwal. S. Rajendran /Micro- and nano-structures 89

der milling) are used and the grain size of the sintered ceramic is in the vicinity of 50 pm. However, superior strength of the material is attributed to the presence of small (of the order of 100 nm) metastable precipitates of tetragonal zirconia and an ordered &phase. By control of the ageing process at about 1100°C (below the eutectoid temperature of 1400°C) optimum mechanical properties have been obtained [ 51. When the ceramic is placed under stress, the tetragonal zirconia precipitates undergo martensitic transformation to monoclinic zirconia (with about 4% volume change) in the stress field of a propagating crack [ 7 1, thus imparting toughness to the ceramic. The transformation process absorbs en- ergy which is otherwise available to the propagating crack. Materials with high strength or high thermal

shock resistance can be prepared by controlling the morphology of tetragonal zirconia precipitates. These features are important for their use as structural ceramics. However, such materials have also found use as an electrolyte in oxygen sensors requiring high thermal shock resistance such as for direct immer- sion in molten steel. However, the electrical properties of these materials degrade strongly (fig. 5 ) with the increased precipitation of the monoclinic phase within the grains of MgO-PSZ as well as at grain boundaries and cannot be used in solid electrolyte devices requiring high ionic conductivity.

Zirconia ceramics containing 2-3 mol% Yz03 (Y- TZP) with high density and small grain size are nearly 1OOo/6 in tetragonal structure and are known to have high strength and fracture toughness [ lo]. Fracture strengths in Y-TZP materials of above 1000

3.4 wt% MgO - PSZ

I E -2.10

-0 I c -2.15

b x-2.20

s -2.25

-2.30 i ' ' 0 1000 2000 3000 4000 5000

Time, min

Fig. 5. Conductivity degradation with time for 3.4 wt% MgO- Zr02 at 1000°C.

MPa (compared with 250-300 MPa in large grained cubic phase) have been reported. These materials have been the subject of intense investigation re-

cently for possible use as structural ceramics and also in solid electrolyte cells. Some solid oxide fuel cell developers are contemplating the use of Y-TZP materials in the planar design fuel cell despite the fact that at 1 OOO”C, their conductivity is lower by about a factor of three compared with 8 mol% YzOs-ZrOz. However, at lower temperatures around 400-500°C their conductivity is slightly better [ 241.

The typical microstructures observed in Y-TZP ceramics are shown in fig. 6. The microstructure is a function of the sintering temperature and the impurity content. Normally sintering below 1400- 1450°C produces ceramics with uniform grain size distribution and high strength. With increase in the sintering temperature, grain size increases, inhomogeneous grain growth occurs and mechanical properties deteriorate. Inhomogeneous grain growth has been associated with phase partitioning as de- termined by the equilibrium phase diagram but is promoted by the presence of glassy impurities (fig. 6~)) in the starting powders. The role of the glassy phase in promoting grain growth is demonstrated more convincingly in fig. 6d where a disk of high purity Y-TZP ceramic was heated at 1400” C in contact with a glass phase.

Y-TZP materials show significant variation in the lattice conductivity as a function of sintering temperature, cooling rates and post-sinter heat treat- ments [ 251. They also show degradation in the lattice conductivity with time when annealed in the vicinity of 900-1200°C (fig. 7). Badwal et al. [26] have shown that in Y-TZP materials fine precipitates of t-ZrOz form within the grains during sintering and slow cooling of the ceramic from the sintering temperature. The intensity and size of t-ZrO, precipitates increase with subsequent annealing in the 900-1200°C temperature range. Compositional inhomogeneities are common and different variants of the tetragonal phase with different Y/Zr ratios have been observed within the same grain [ 26 1. If the size of these precipitates is below a critical size then they remain in the metastable form. However, if they are allowed to grow above a critical size they transform spontaneously on cooling. Fig. 8 shows microstructures of two Y-TZP ceramics sintered at

Fig. 6. Scanning electron micrographs of 3 mol% Yz03-ZrOz (Y-TZP) ceramics. Sintering temperature: (a) 1450°C; (b) 1600°C; (c)

1500°C; (d) 15OO”C, subsequently heated in contact with glass at 1400°C. Inhomogeneous grain growth in (c) is due to the large

impurity phase

/ 8 i

“,

07 i 27500

/

25000 L 0-t 1 o-o __A .__L_-L-- 1 ------~

300 500 700 900 1100 1300 1500

Tatlneal~ OC

Fig. 7. Conductivity (lattice) as a function of the annealing tem-

perature for high purity 3 mol% Y203-Zr02 measured at 350°C in air. Ceramic sintered at 1500°C for four hours and quenched

in air. Annealing time at each temperature was 50 h.

1700°C and cooled at different rates from the sintering temperature. The ceramics cooled at the slow rate of 50°C h- ’ showed massive twinning, about 60% m-ZrOz, and were severely cracked. The conductivity was relatively low. The twinning is asso-

ciated with the martensitic transformation of large t- ZrOz precipitates to m-Zr02 in the slow cooled specimens. The ceramic disks cooled at the fast rate of 300” C hh ’ showed minimum twinning, had only 5- 10% m-Zr02 and relatively high conductivity.

Similarly, thermodynamic equilibrium constraints on other systems have a detrimental effects on the electrical properties. In the Y203-Zr02 system, the highest conductivity has been reported for 8 mol% Y,03 (8-YSZ) content [24]. For this reason, this composition is commonly used as the electrolyte material in fuel cells. However, the composition is in the two phase field near the (c-t t)/c phase boundary [ 271 and conductivity degradation with time has been observed at 1000°C due to formation of tine precipitates of t-ZrO, as observed with transmission electron microscopy of specimens annealed at 1000°C [28].

In the scandia-zirconia system, homogeneously prepared compositions in the zirconia-rich end also show substantial change in the conductivity with


Fig. 8. Scanning electron micrographs oftwo 3 mol% Y203-Zr02

ceramics sintered at 1700°C and then cooled at: (a) 300°C h-l;

and (b) and (c) 50°C h-‘. Severe twinning is clearly obvious in

the slow cooled specimen.

time. Ciacchi et al. [28] and Badwal and Drennan [ 29 ] have investigated the cause of this in detail. The Scz03-ZrOz composition after sintering in the cubic phase field transforms to a scandia-rich t’-phase. The phase is characterised by the herring-bone structure shown in fig. 9. A similar phase has also been reported in rapidly cooled Y,OJ-Zr02 compositions (Yz03 content between 3-6 mol%). The t’-phase in the Sc203-Zr02 system has a high conductivity (0.32

Fig. 9. Transmission electron micrographs of 7.8 mol% Sc,03-

ZrOz electrolyte showing (a) herringbone structure characteris-

tic of t’-phase; and (b) formation of t-ZrOz precipitates on an-

nealing the specimen at 1000” C for 2000 h.

S cm-’ at 1000°C for 7.8 mol% Sc203) [29]. On annealing in the vicinity of 800-lOOO”C, the t’-phase decomposes into a cubic phase and t-ZrO, precipitates (low conductivity phase). In addition, diffuse scatter observed in the electron diffraction patterns is indicative of the existence of an ordered phase [ 291. Most of these features are observed in the nano-


scale range but nevertheless have a detrimental effect on the ionic conductivity.

Compositional inhomogeneities, microdomains, modulated structures and formation of ordered phases within grains are all examples of nano-structures created within the grains. The relationship be-

tween these nano-structural features and material properties is not well documented. However, the existence of short and long range ordering has been observed in cubic solid solutions in the YzOj-ZrOz and

CaO-ZrOz systems and the observed decrease in the ionic conductivity with time for higher dopant con-

centrations (around 1OOO’C) appears to be related to the formation of microdomains [ 9 1.

Addition of secondary phases can influence the electrical and mechanical properties. Improvements in the strengths of both cubic and tetragonal Yz03- ZrOz have been reported with the addition of alumina [ 30,3 11. Additions of small quantities of alumina to fully stabilised zirconia pins grain growth which is clearly obvious in the case of cubic zirconia (fig. 10). Although large increases in the strength have been reported for Y-TZP materials [ 3 1 ] with the addition of alumina, the effect on cubic stabilised zirconia is marginal. For example with the ad-

dition of 10 wt% alumina to 8 mol% Y203-Zr02, only a 10% increase in the room temperature strength was observed. In general, inert inclusions within the grains lead to a reduction in the lattice conductivity of solid electrolytes and the conductivity is very much a function of the distribution of phases in a multi-

0.25

‘; 0.20

‘; c

d 0.15

O.lC IL 0.0

1 OOO~C

L ,

2.5 5.0 7.5 10.0 12.5

wt% A1203

Fig. 10. Graph showing change in the ionic conductivity at 1000°C with alumina addition to 8 mol% Y203-210~. Scanning electron

micrographs (a) without and (b) with the addition of 10 wt%

AlzOl are shown as insets.

phase system. For example, the addition of 10 wt% alumina to 8-YSZ leads to a decrease in the ionic conductivity by about 60% at 1000°C. Minor vari-

ations to the processing procedures can produce significantly different phase distributions and consequently different electrical and mechanical properties.

4.2. Intergrain features and their influence on properties

The grain boundary regions are often characterised by impurity phase segregation, imperfect contact between grains, pores, presence of secondary inclusions, dislocations and lattice mismatching. Therefore as the grain size changes from micron size to nano size, the volume occupied by the grain boundary region and the ratio of grain boundary surface area/ceramic volume increases. Although there are advantages in terms of superior mechanical properties of decreasing the grain size (small flaw size) and forging of ceramics to desired shapes because of the high creep rates for small grained ceramics, the effect on electrical properties is not al- ways complementary.

All the features associated with grain boundary regions have a detrimental effect on the ion-transport properties in ceramic electrolytes. The glassy segre- gated phases have often higher resistivity than the lattice resistivity and impede the flow of migrating ions [ 8 1. Also the presence of glassy phases at grain boundaries reduces fracture strength and toughness of the ceramic as cracks can easily migrate through the weak glassy phase. In general, for high performance ceramics, grain boundaries should be clean to achieve good mechanical and electrical properties.

Silica is one of the most common impurities present in zirconia powders and is reasonably well dispersed. It enhances sintering kinetics and densili- cation of the ceramic via liquid phase sintering. However, this occurs at the expense of mechanical and electrical properties. SiOz is known to react with the stabiliser such as CaO, MgO and Y,03 leading to the removal or loss of stabiliser from grains and formation of grain boundary phases during sintering. The presence of other impurities such as oxides of alkali metals, Fe, Ti, Al substantially affects the wetting properties of the grain boundary phase(s). The

S.P.S. Badwal, S. Rajendran /Micro- and namstructures 93

composition of the grain boundary phase(s) and its location change with the sintering temperature and post-sinter heat treatment. The location of the grain boundary phase whether continuously wetting all grains or present in isolated pockets or triple grain junctions is likely to have a significant effect on the material properties. Fig. 11 shows the impedance diagram of relatively pure (20 ppm SiOz) 3 mol% Y203-Zr02 ceramic before and after the addition of an impurity phase. The increase in the grain bound-

ary resistance is dramatic and clearly demonstrates the low oxygen-ion transport through the glass phase.

Drennan and Hannink [ 321, in the case of MgO- partially stabilised zirconia, have shown that the del- eterious grain boundary glassy phase can be removed by appropriate processing of the ceramic. Small additions of SrO uniformly dispersed in the MgO-ZrOz powder forms a low melting glassy phase on reacting with silica preferentially and prevents loss of MgO to grain boundaries. This glassy phase is ejected from the bulk of the ceramic to the external surface during the sintering/densification process and the grain boundaries become relatively clean [ 321. Normally,

the grain boundary microstructure of MgO-PSZ aged at 1100°C (well below the eutectoid temperature of 1400’ C) consists of m-ZrO, and MgO rods [ 5 1. The decomposition products are much weaker than the MgO-PSZ matrix and leads to reduction in the strength. The MgO-PSZ ceramic prepared with no addition of SrO has a network of crystalline MgzSi04 present at grain boundaries which leads to the decomposition of the MgO-ZrOz solid solution and formation of m-ZrO,, the nucleation sites for further

Z’, ohm cm

Fig. 11. Impedance diagram of 3 mol% Y203-Zr02 ceramic. Small

right (grain boundary) arc: high purity material. Large grain

boundary arc: with addition of ( r0.2 wt%) silica based impurity.

growth of the phase. In the presence of SrO, when the grain boundaries are clean, the relative density of the nucleation sites is low. Drennan and Hannink [ 321 have clearly demonstrated that the amount of m-Zr02 was significantly higher in ceramics pre-

pared with no SrO ( SrC03) in the starting powders. Thus by controlling the grain boundary chemistry, a considerable enhancement in the fracture strength

can be obtained [ 32 1. Butler and Drennan [33] and Rajendran et al.

[34] have shown that alumina scavenges silica at grain boundaries. In the case of 2.5 mol% Yz03- Zr02, a reduction in the grain boundary resistivity by a factor of six was observed with the addition of 10 wt% alumina [ 341.

4.3. Tortuosity

In anisotropic polycrystalline conductors, the orientation of grains with respect to neighbouring grains also has a significant effect on the ionic conductivity. In beta-aluminas, ionic conduction is limited in the two-dimensional conduction planes. No conduction

takes place in the direction of the c-axis. The total conductivity of polycrystalline materials is considerably lower (up to an order of magnitude) than that in single crystals. The major reason for this is the tor-

tuous path conducting ions (random distribution of conduction planes in polycrystalline materials) have

to follow because of the anisotropic nature of the ceramic. The relative orientation of each grain with respect to neighbouring grains plays a major role and the conductivity is very much a function of the fabrication procedures as they affect the grain size and orientation of grains and the ceramic microstructure.

Increased path lengths for migrating ions can also occur in isotropic materials due to precipitation of low conducting phases within the grains of the ceramic electrolyte as discussed earlier.

The presence of SiOz and CaO as impurities in the powders or added as sintering aids in beta-alumina

form blocking phases at grain boundaries with detrimental consequences for ionic conduction espe- cially in view of the anisotropic nature of the conduction.

94 S.P.S. Badwal, S. Rajendran /Micro- and nano-strucm-es

4.4. Superplastic deformation

Ceramic materials in general do not undergo superplastic deformation. Wakai et al. [ 351 first reported superplastic deformation in Y-TZP materials. Since then a number of other similar materials (Y-TZP/alumina, ceria-TZP) have been shown to

undergo superplastic deformation [ 36 1. The creep rates are usually high above about 1400’ C and offer the mechanism for near net shape forming of ceramics. The superplastic behaviour is considered to be associated with the small grain size (0.2-0.5 urn) of Y-TZP or similar materials. The composition of

the intergranular layer between grains and the area it occupies play an important role. Grain boundary sliding appears to be the dominant mechanism for the superplastic deformation.

features demonstrates clearly that micro and nano- structural features affect materials properties and are critical to the performance of solid electrolytes and devices made from such materials.

Acknowledgement

The authors are thankful to Fabio Ciacchi and Ky- lie Crane for experimental assistance, Dr. J. Dren-

nan for TEM work and to Drs. K. Foger and J. Par- rott for reviewing this manuscript.

References

4.5. Moisture sensitivity qf Y-TZP

Y-TZP materials when exposed to moist environments in the 150-400°C temperature range or im- mersed in water in the 100-200’ C temperature range undergo a phase transformation to m-ZrOz with con- sequent degradation in the mechanical properties. In

extreme cases a complete breakdown of the ceramic microstructure has been observed [ 10 1. Similar behaviour is also observed in dry air at the surface of

Y-TZP electrolyte disks during current flow but only on the anodic side of the cell [ 371. The transformation starts at the surface and moves to the interior of the material. Several different mechanisms have been proposed for the transformation of tetragonal zirconia in moist environments [ 381. The microstructure of the material plays a major role in estab- lishing the stability of the material. Grain size, distribution of dopant within grains and nature and location of the grain boundary phase are some of the major factors which influence the rate of degradation. Materials with small grain size and more uniform distribution of yttrium within grains show better stability.

[ 1 ] T. Takahashi, ed., High Conductivity Solid Iomc Conductors

(World Scientific, Singapore, 1989).

[2] B.V.R. Chowdari and S. Radhakrishna, ed., Solid State Ionic

Devices (World Scientific, Singapore, 1988).

31 S.P.S. Badwal, M.J. Bannister and R.H.J. Hannink, ed.,

Proc. Fifth Internat. Conf. on the Science and Technology

of Zirconia (Technomic Publishing Co., Lancaster, USA,

1993).

4) R.C. Garvie, R.H.J. Hannink and R.T. Pascoe, Nature 258

(1975) 703.

I 5) R.H.J. Hannink, J. Mater. Sci, 18 (1983) 457.

61 A.H. Heuer, J. Am. Ceram. Sot. 70 ( 1987) 689.

[7] M.V. Swain, Mater. Forum 13 (1989) 237.

[ 8 ] S.P.S. Badwal, J. Drennan and A.E. Hughes, in: The Science

of Ceramic Interfaces, ed. J. Nowotny (Elsevier,

Amsterdam. 199 1) p. 227.

[9] S.P.S. Badwal, in: Materials Science and Technology, A

Comprehensive Treatment, ed. M.V. Swain, Vol. 11 (Verlag

Chemie, Weinheim, Germany, 1993), to be published.

[ 10) S. Somiya, N. Yamamoto and H. Yanagida, eds., Science

and Technology of Zirconia III, Advances in Ceramics Vol.

24 A and 24 B (The Am. Ceram. Sot., Westerville. OH.

1986).

[ll ] G.L. Messing, S. Hirano and H. Hausner. eds., Ceramic

Powder Science III, Ceramic Transactions, Vol. 12 (The Am.

Ceram. Sot., Westerville, OH, 1990).

[12 ] B.J.J. Zelinski, G.J. Brinker, D.E. Clark and D.R. Ulrich, eds., Better Ceramics Through Chemistry IV, Materials

Research Society Symp. Proc. Vol. 180 (Materials Research

Society, 1990).

[ 131 S. Rajendran, J. Mater. Sci. (1993), submitted for publication.

5. Conclusions [ 141 S. Rajendran, J. Mater. Sci. 27 (1992) 433.

[ 151 R.M. Dell and S. Weller, Trans. Faraday Sot. 55 (1959)

2203. This overview of the powder and ceramic pro- [ 161 W.D. Kingery, in: Ceramic Powders, ed. P. Vincenzini

cessing techniques on micro- and nano-structural (Elsevier, Amsterdam, 1983) p. 3.


[ 171 MS. Kaliszewski and A.H. Heuer, J. Am. Ceram. Sot. 73

(1990) 1504.

[ 18 ] P.S. Nicholson, in: Solid State Ionic Devices, eds., B.V.R.

Chowdari and S. Radhakrishna (World Scientific,

Singapore, 1988) p. 605.

[ 19 ] S. Rajendran, Mat. Forum ( 1993), submitted for

publication.

[20]M.A.C.G.VandeGraaf,J.H.H.TermaatandA.J.Burggraaf,

J. Mater. Sci. 20 (1985) 1407.

[ 2 1 ] K. Keizer, M. Van Hemert, M.A.C.G. Van de Graaf and

A.J. Burggraaf, Solid State Ionics 16 (1985) 67.

[22] J. Cesarano III and LA. Aksay, J. Am. Ceram. Sot. 71

(1988) 1062.

(231 I.A. Aksay, F.F. Lange and B.I. Davis, J. Am. Ceram. Sot.

66 (1983) C. 190.

[24] S.P.S. Badwal, Solid State Ionics 52 (1992) 23.

[25] S.P.S. Badwal, Appl. Phys. A 50 ( 1990) 449.

[26] S.P.S. Badwal, ET. Ciacchi and R.H.J. Hannink, Solid State

Ionics 40/41 (1990) 882.

[27] H.G. Scott, J. Mater. Sci. 10 (1975) 1527.

[ 281 F.T. Ciacchi, S.P.S. Badwal and J. Drennan, J. Europ.

Ceram. Sot. 7 (1991) 185.

[29] S.P.S. Badwal and J. Drennan, Solid State Ionics 53-56

(1992) 769.

[ 3010. Yamamoto, Y. Takeda, N. Imanishi, T. Kawahara, G.Q.

Shen, M. Mori and T. Abe, Proc. Second Internat. Conf.

Solid Oxide Fuel Cells, Athens, 199 1, eds. F. Grosz, P. Zeger,

S.C. Singhal and 0. Yamamoto (Commiss. of the European

Commun., Brussels, 199 1) p. 437.

[ 3 1 ] K. Tsukuma, T. Takahat and M. Shiomi, Science and

Technology of Zirconia III, Advances in Ceramics, eds. S.

Somiya, N. Yamamoto and H. Yanagida, Vol. 24 A and 24

B (The Am. Ceram. Sot., Westerville, OH, 1986) p. 721.

[32] J. Drennan and R.H.J. Hannink, J. Am. Ceram. Sot. 69

(1986) 541.

[ 331 E.P. Butler and J. Drennan, J. Am. Ceram. Sot. 65 (1982)

474.

[ 341 S. Rajendran, J. Drennan and S.P.S. Badwal, J. Mater. Sci.

Lett.6 (1987) 1431.

[ 351 F. Wakai, S. Sakaguchi and Y. Matsuno, Adv. Ceram. Mater.

1 (1986) 259.

[ 361 A.H. Chokshi, A.K. Mukherjee and T.G. Langdon, Mater.

Sci. Eng. R 10 (1993) 237.

[ 371 S.P.S. Badwal and N. Nardella, Appl. Phys. A 49 (1989)

13.

[38] A.E. Hughes, F.T. Ciacchi and S.P.S. Badwal, Proc. Fifth

Intemat. Conf. Science and Technology of Zirconia eds.,

S.P.S. Badwal, M.J. Bannister and R.H.J. Hannink

(Technomic Publishing Co., Lancaster, USA, 1993).

Documents

Effect of micro- and nano-structures on the properties of ionic conductors