《复合材料 Composites》课程教学资源（学习资料）第五章 陶瓷基复合材料_ZrO2-17

DENTAL MATERIALS 24(2008)289-298 availableatwww.sciencedirect.com materials ° Science Direct ElseVierjournalhomepagewww.intl.elsevierhealth.com/journals/dema Systematic review article Stabilized zirconia as a structural ceramic. an overview J. Robert Kelly a,, Isabelle Denry Department of Reconstructive Sciences, Center for Biomaterials, University of Connecticut Health Center, Farmington, CT USA b Department of Restorative and Prosthetic Dentistry, The Ohio State University, Columbus, OH USA ARTICLE INFO A BSTRACT This review introduces concepts and background from the ceramics engineering literature Received 24 April 2007 regarding metastable zirconia ceramics to establish a context for understanding current and Accepted 11 May 2007 emerging zirconia-based dental ceramics e 2007 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved. Zirconia ramiCs Dental ceramics ghening Stabilized tetragonal R-curve Dispersion-toughened zirconia Partially stabilized zirconia Toughened zirconia polycrystalline Y-TZP 1. Toward improved reliability: (1)flaw control and(2)flaw tolerance 2. Zirconia polymorphs: temperature dependence and transformation strains 3. Three distinct zirconia ceramics: terminology, processing and microstructures 3.1 rtially stabilized zi 3.3. Single-phase, polycrystalline t-zro2 4. Mechanism(s) and consequences of transformation 293 5. R-curve behavior: description, definition and implications 6. Strength versus toughness Cyclic fatigue of transformation-toughened ceramics Presentend nt the annua cessio hef th cecnatde ms f ental Materias, e Etobein 2t55 06 30 1615. sa Tll fax:+18606791370 E-mail address: Kelly@nsol. uchc. edu .R Kelly) 0109-5641/$-see front matter e 2007 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved. doi:10.1016/ denta2007.05005

dental materials 24 (2008) 289–298 available at www.sciencedirect.com journal homepage: www.intl.elsevierhealth.com/journals/dema Systematic review article Stabilized zirconia as a structural ceramic: An overview J. Robert Kelly a,∗, Isabelle Denry b a Department of Reconstructive Sciences, Center for Biomaterials, University of Connecticut Health Center, Farmington, CT USA b Department of Restorative and Prosthetic Dentistry, The Ohio State University, Columbus, OH USA article info Article history: Received 24 April 2007 Accepted 11 May 2007 Keywords: Zirconia Ceramics Dental ceramics Transformation toughening Stabilized tetragonal R-curve Dispersion-toughened zirconia Partially stabilized zirconia Toughened zirconia polycrystalline Y-TZP abstract This review introduces concepts and background from the ceramics engineering literature regarding metastable zirconia ceramics to establish a context for understanding current and emerging zirconia-based dental ceramics. © 2007 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved. Contents 1. Toward improved reliability: (1) flaw control and (2) flaw tolerance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290 2. Zirconia polymorphs: temperature dependence and transformation strains. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290 3. Three distinct zirconia ceramics: terminology, processing and microstructures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291 3.1. Dispersion-toughened ceramics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291 3.2. Partially stabilized zirconia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291 3.3. Single-phase, polycrystalline t-ZrO2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292 4. Mechanism(s) and consequences of transformation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293 5. R-curve behavior: description, definition and implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295 6. Strength versus toughness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295 7. Cyclic fatigue of transformation-toughened ceramics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296 Presented at the annual session of the Academy of Dental Materials, October 23–25, 2006, Sao Paulo, Brazil. ˜ ∗ Corresponding author at: UConn Health Center, 263 Farmington Avenue, Farmington, CT 06030-1615, USA. Tel.: +1 860 679 3747; fax: +1 860 679 1370. E-mail address: Kelly@nso1.uchc.edu (J.R. Kelly). 0109-5641/$ – see front matter © 2007 Academy of Dental Materials. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.dental.2007.05.005

DENTAL MATERIALS 24(2008)289-298 8. Low temperature degradation of 3Y-TZP 8.1. Cubic phase and accelerated aging 2.31%; t-m approximately 4.5%. Sintered structures trans 1. Toward improved reliability: ( 1)flaw control and (2) flaw tolerance forming from t to m on cooling from sintering temperatures (approximately 1300-1500C)undergo spallation with por- tions crumbling into multi-grained powders Two research paths aimed at increasing the structural relia Beginning about 1972, the ceramic engineering commu- bility of ceramics have been pursued over the past 30 years. nity was discovering that alloying with lower valance oxides The first involved efforts to minimize the number and size such as Cao, Mgo, La203, and Y203, disfavored the strained of critical flaws based on the well-accepted Griffith flaw- m phase at room temperature and favored more symmetric ize/strength relationship for linearly elastic brittle fracture c"and t lattice structures(with'indicating metastability)[6] These cand t' phases are analogous to those in pure zirco- (1) nia but have dopant ions substituted on Zrt sites and have a fraction of oxygen sites vacant to retain charge neutrality where of is the fracture strength, c the critical flaw radius, 14]. The amount of dopant required for full cubic stabiliza- Kic critical stress intensity in mode I opening and Y is the tion is substantial; 8 mo1% in the case of the dopant Y2O3 crack shape factor. This research, part of a discipline known with one oxygen vacancy created for every two yttrium ions as ceramics processing, continues to investigate a multitude [7]. Partial stabilization of tetragonal zirconia can occur at of steps including powder fabrication (to control chemi- dopant concentrations of 2-5mo1% depending on grain size, 1 homogeneity particle size, size uniformity, etc ) particle to be discussed below. These metastable c and I phases have dispersion in processing media, powder consolidation and prolonged stability at room temperature given that cation and ho mobility in zirconia is quite low and that the oxygen vacan sities), sintering control, and"flaw kind"finish machining cies are locally ordered 17]. Recent consistent, but apparently (or finishing to net-shape avoiding the need for machining) independent work, attributes tetragonal metastability solely [1]. Second were efforts controlling ceramic microstructures to the presence of the oxygen vacancies that allow both anion to increase their resistance toward crack propagation, and cation relaxations to occur dependent on their vacancy to increase toughness. Counter to the Griffith relationship, proximity 14, 7,8. Overall, three mechanisms are discussed for increased strength and increased toughness do not gener ally correlate in transformation-toughened ceramics as will Y203 and CeO2: ()dopants inducing oxygen vacancies that are beelaborated on later. However, suchhigh toughness ceramics generally trivalent (e.g. Gd", Fe3+, Ga3+, and Y3); (i)tetrava- of lower strength are appealing for structural use due to their lent dopants being undersized or oversized with respect to the damage tolerance. Table 1 lists five major ceramic toughen ing mechanisms along with engineering material examples 2 One of these mechanisms, transformation toughening along with microcracking and deflection mechanisms are the Table 1-Ceramic toughening mechanisms [2] toughening mechanisms now prominent in zirconia-based or Mechanism Highest toughnes Example zirconia-containing ceramics and are the topic of this paper. aterials Transformation 2. Zirconia polymorphs: temperature dependence and transformation strains Microcracking Al O3/ZrO? SinG/sic Pure zirconia is monoclinic (m) at room temperature and pressure. With increasing temperature the material trans Metal dispersion AlO/Al forms to tetragonal (t), by approximately 1170 C and then Al2O3/Ni to a cubic (c) fluorite structure starting about 2370C with melting by 2716 C[3, 4]. These lattice transformations are Whiskers/platelets SigNa/Sic martensitic, characterized by(1)being diffusionless (i.e. involv SigN4/SigN Al2 O3/Sic ing only coordinated shifts in lattice positions versus transport of atoms),(2) occurring thermally implying the need for a Fibers CAS/SiC temperature change over a range rather than at a specific temperature and,(3) involving a shape deformation [S]. This Al2O3/SiC C/SiC transformation range is bounded by the martensitic start(Ms) and martensitic finish temperatures. Volume changes on cool Al2O3/Al2O ing associated with these transformations are substantial enough to make the pure material unsuitable for applica- a Calcium aluminum silicate glass ceramic. tions requiring an intact solid structure: c- t approximately Lithium aluminum silicate glass ceramic

290 dental materials 24 (2008) 289–298 8. Low temperature degradation of 3Y-TZP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296 8.1. Cubic phase and accelerated aging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297 1. Toward improved reliability: (1) flaw control and (2) flaw tolerance Two research paths aimed at increasing the structural relia￾bility of ceramics have been pursued over the past 30 years. The first involved efforts to minimize the number and size of critical flaws based on the well-accepted Griffith flaw￾size/strength relationship for linearly elastic brittle fracture: f = KIC Y √c (1) where f is the fracture strength, c the critical flaw radius, KIC critical stress intensity in mode I opening and Y is the crack shape factor. This research, part of a discipline known as ceramics processing, continues to investigate a multitude of steps including powder fabrication (to control chemi￾cal homogeneity, particle size, size uniformity, etc.), particle dispersion in processing media, powder consolidation and packing (creating high and homogeneous greenware den￾sities), sintering control, and “flaw kind” finish machining (or finishing to net-shape avoiding the need for machining) [1]. Second were efforts controlling ceramic microstructures to increase their resistance toward crack propagation, i.e. to increase toughness. Counter to the Griffith relationship, increased strength and increased toughness do not gener￾ally correlate in transformation-toughened ceramics as will be elaborated on later. However, such high toughness ceramics of lower strength are appealing for structural use due to their damage tolerance. Table 1 lists five major ceramic toughen￾ing mechanisms along with engineering material examples [2]. One of these mechanisms, transformation toughening along with microcracking and deflection mechanisms are the toughening mechanisms now prominent in zirconia-based or zirconia-containing ceramics and are the topic of this paper. 2. Zirconia polymorphs: temperature dependence and transformation strains Pure zirconia is monoclinic (m) at room temperature and pressure. With increasing temperature the material trans￾forms to tetragonal (t), by approximately 1170 ◦C and then to a cubic (c) fluorite structure starting about 2370 ◦C with melting by 2716 ◦C [3,4]. These lattice transformations are martensitic, characterized by (1) being diffusionless (i.e. involv￾ing only coordinated shifts in lattice positions versus transport of atoms), (2) occurring athermally implying the need for a temperature change over a range rather than at a specific temperature and, (3) involving a shape deformation [5]. This transformation range is bounded by the martensitic start (Ms) and martensitic finish temperatures. Volume changes on cool￾ing associated with these transformations are substantial enough to make the pure material unsuitable for applica￾tions requiring an intact solid structure: c→t approximately 2.31%; t→m approximately 4.5%. Sintered structures trans￾forming from t to m on cooling from sintering temperatures (approximately 1300–1500 ◦C) undergo spallation with por￾tions crumbling into multi-grained powders. Beginning about 1972, the ceramic engineering commu￾nity was discovering that alloying with lower valance oxides, such as CaO, MgO, La2O3, and Y2O3, disfavored the strained m phase at room temperature and favored more symmetric c* and t* lattice structures (with * indicating metastability) [6]. These c* and t* phases are analogous to those in pure zirco￾nia but have dopant ions substituted on Zr4+ sites and have a fraction of oxygen sites vacant to retain charge neutrality [4]. The amount of dopant required for full cubic stabiliza￾tion is substantial; 8mol% in the case of the dopant Y2O3 with one oxygen vacancy created for every two yttrium ions [7]. Partial stabilization of tetragonal zirconia can occur at dopant concentrations of 2–5mol% depending on grain size, to be discussed below. These metastable c* and t* phases have prolonged stability at room temperature given that cation mobility in zirconia is quite low and that the oxygen vacan￾cies are locally ordered [7]. Recent consistent, but apparently independent work, attributes tetragonal metastability solely to the presence of the oxygen vacancies that allow both anion and cation relaxations to occur dependent on their vacancy proximity [4,7,8]. Overall, three mechanisms are discussed for stabilization of t-ZrO2 with the most common dopants being Y2O3 and CeO2: (i) dopants inducing oxygen vacancies that are generally trivalent (e.g. Gd3+, Fe3+, Ga3+, and Y3+); (ii) tetrava￾lent dopants being undersized or oversized with respect to the Table 1 – Ceramic toughening mechanisms [2] Mechanism Highest toughness (MPam1/2) Example materials Transformation ∼=20 ZrO2 (MgO) HfO2 Microcracking ∼=10 Al2O3/ZrO2 Si3N4/SiC SiC/TiB2 Metal dispersion ∼=25 Al2O3/Al Al2O3/Ni WC/Co Whiskers/platelets ∼=15 Si3N4/SiC Si3N4/Si3N4 Al2O3/SiC Fibers ≥30 CASa/SiC LASb/SiC Al2O3/SiC SiC/SiC SiC/C Al2O3/Al2O3 a Calcium aluminum silicate glass ceramic. b Lithium aluminum silicate glass ceramic.

DENTAL MATERIALS 24(2008)289-298 oxide cations(e.g. Ti++, Ge++, Ce4+); or, (ii) dopants resulting 3.1. Dispersion-toughened ceramics In minor, stabilization mechanism involves matrix constraint of The simplest material utilizes the dispersion of zirco- t-zrO2 grains held within non-transforming materials nia particles in another matrix and appears to be the The"absorption of energy"during the room temper- least widely published and commercially important. These ature t-m transformation in partially stabilized zirconia dispersion-toughened materials, such as Zro2-toughened alu (Cao-ZrO2; described in more detail below) was recognized mina(Al2O3)or ZrO2-toughend mullite(3Al2O3 2SiO2)have as a strengthening mechanism in 1975 [10]. In the 1975 been termed ZTA and ZTM [11]. In contrast with the other two publication reference was made to similarities between trans- classes, stability of the t phase to room temperature does not forming zirconia and austenite-martensite strengthening primarily involve the use of dopants but is controlled instead of TRIP steels(transformation-induced plasticity)[10]. TRIP by particle size, particle morphology and location (intra-or steels contain retained austenite when they have carbon con- intergranular, see Fig. 1a). In ZTA, for example, particles above centrations in excess of 1 wt % Some mechanical properties, a critical size will transform to monoclinic symmetry upon in particular toughness and ductility, rely on the diffusionless cooling to room temperature [12]. Since this t-m transfor- transformation of this austenite into high-carbon martensite mation is known to be martensitic, a useful way to describe nduced by stress and strain. Three features, in particular, particle size effects has been to examine their influence on vere seen shared with strengthened steels leading Garvie et the martensitic start(Ms) temperature; essentially all t-phase al. [10] to term these new zirconia compositions"ceramic stabilization can be viewed as decreasing the Ms to below steel":(1)three allotropes,(2)martensitic transformations, room temperature. Such investigation has suggested that the and ()metastable phases. Stabilized zirconia and steels also particle size effect is likely due to difficulties in nucleating share similarities in elastic moduli and coefficients of thermal the transformation, although considerations have also been expansion Garvie et al. further predicted that a vast range of given to the possible effects of surface and strain energy and ceramic materials with different properties would be devel- chemical free energy driving forces[12. Within dental mate oped analogously to iron systems [10] rials the sole commercial example of a dispersion-toughened ceramic is In-Ceram Zirconia (Vita Zahnfabrik) which is an interpenetrating composite of 30% glass and 70% polycrys 3. Three distinct zirconia ceramics: talline ceramic consisting of Al2O3 ZrO2 in a vol. ratio of terminology, processing and microstructures As foreseen by Garvie wide latitude was found in the applica- 3. 2. Partially stabilized zirconia tion of the zirconia t-m transformation in ceramics, leading to development of three different materials each having an These materials are the most widely studied,commer- associated terminology [10]. These three classes are listed cially important, microstructurally complex, and in the case in Table 2 along with some dental examples; the first two of Mg-doped some of the toughest of the transformation of these are at least two-phase materials with t-ZrO2 as the toughened ceramics. In these ceramics t-ZrO2 intra-granular minor phase(dispersed and precipitated respectively) and precipitates exist within a matrix of stabilized c-zrO2. Sta- the last is essentially a single-phase t-zrO2. The origin and bilization involves dopant addition, such as with Cao, Mgo, details of stabilization of the t phase differs among these three La2O3, and Y203, in concentrations lower than that required toughened microstructures; photomicrographs of representa- for full c-Zro2 stabilization. Full stabilization is purpose- tive materials are presented in Fig. 1. The three materials share fully not achieved in these materials, hence the historical that stabilization of t occurs and that toughness involves the derivation for the term "partially stabilized zirconia"or martensitic t- m transformation PSZ, to which the relevant dopant is often appended: Ca PSZ, Mg-PSZ, Y-PSZ, etc. [11]. Precipitates are fully coherent with the cubic lattice, forming on a nanometer scale with lenticular morphology(approximately 200nm diameter and Table 2-Forms of transformation-toughened zirconia 75nm thick) parallel to the three cubic axes (refer 1. Zirconia(dispersed phase)toughened ceramics; e.g, ZTA Fig. 1b)[12, 13]. Following sintering or solution annealing in (alumina), ZTM(mullite) the cubic solid solution single-phase field(approximately Dental example: >1850C), precipitates are nucleated and grown at lower In-Ceram ita Zahnfabrik) temperatures(approximately 1100C)within the two-phase 2. Partially stabilized zirconia(PSZ; e.g. Ca-PSZ, Mg-PSZ, Y-psz) tetragonal solid solution plus cubic solid solution phase fiel Lenticular (ens shaped) tetragonal precipitates in a cubic matrix a process termed"aging"[12, 13]. Aging optimization(time- Denzir-MDentror 3. Tetragonal zirconia polycrystals (TZP; e.g. Y-TZP, Ce-TZP and phase stability [14]. Metastability can be lost when tetrag Nominally 98% tetragonal, fine grain size onal precipitates are too small(they will not transform) and Dental examples. when precipitates grow too large spontaneous transformation DC Zirkon(DCS Precident, Schreuder Co can occur to m with twinning and microcracking [14] Complex decomposition and tertiary precipitation pro- Lava(3M ESPE cesses have also been reported to occur with aging of Mg-PS In-Ceram YZ(Vita Zahnfabrik) [15] along with the development of quite some range of phys

dental materials 24 (2008) 289–298 291 oxide cations (e.g. Ti4+, Ge4+, Ce4+); or, (iii) dopants resulting in charge-compensations (YNbO4, YTaO4) [9]. Another, more minor, stabilization mechanism involves matrix constraint of t-ZrO2 grains held within non-transforming materials. The “absorption of energy” during the room temper￾ature t→m transformation in partially stabilized zirconia (CaO–ZrO2; described in more detail below) was recognized as a strengthening mechanism in 1975 [10]. In the 1975 publication reference was made to similarities between trans￾forming zirconia and austenite–martensite strengthening of TRIP steels (transformation-induced plasticity) [10]. TRIP steels contain retained austenite when they have carbon con￾centrations in excess of 1 wt.%. Some mechanical properties, in particular toughness and ductility, rely on the diffusionless transformation of this austenite into high-carbon martensite induced by stress and strain. Three features, in particular, were seen shared with strengthened steels leading Garvie et al. [10] to term these new zirconia compositions “ceramic steel”: (1) three allotropes, (2) martensitic transformations, and (3) metastable phases. Stabilized zirconia and steels also share similarities in elastic moduli and coefficients of thermal expansion. Garvie et al. further predicted that a vast range of ceramic materials with different properties would be devel￾oped analogously to iron systems [10]. 3. Three distinct zirconia ceramics: terminology, processing and microstructures As foreseen by Garvie wide latitude was found in the applica￾tion of the zirconia t→m transformation in ceramics, leading to development of three different materials each having an associated terminology [10]. These three classes are listed in Table 2 along with some dental examples; the first two of these are at least two-phase materials with t-ZrO2 as the minor phase (dispersed and precipitated respectively) and the last is essentially a single-phase t-ZrO2. The origin and details of stabilization of the t phase differs among these three toughened microstructures; photomicrographs of representa￾tive materials are presented in Fig. 1. The three materials share that stabilization of t occurs and that toughness involves the martensitic t→m transformation. Table 2 – Forms of transformation-toughened zirconia 1. Zirconia (dispersed phase) toughened ceramics; e.g., ZTA (alumina), ZTM (mullite) • Dental example: In-Ceram zirconia (Vita Zahnfabrik) 2. Partially stabilized zirconia (PSZ; e.g. Ca-PSZ, Mg-PSZ, Y-PSZ) • Lenticular (lens shaped) tetragonal precipitates in a cubic matrix • Dental example: Denzir-M (Dentronic AB) 3. Tetragonal zirconia polycrystals (TZP; e.g. Y-TZP, Ce-TZP • Nominally 98% tetragonal, fine grain size • Dental examples: DC Zirkon (DCS Precident, Schreuder & Co) Cercon (Dentsply Prosthetics) Lava (3M ESPE) In-Ceram YZ (Vita Zahnfabrik) 3.1. Dispersion-toughened ceramics The simplest material utilizes the dispersion of zirco￾nia particles in another matrix and appears to be the least widely published and commercially important. These dispersion-toughened materials, such as ZrO2-toughened alu￾mina (Al2O3) or ZrO2-toughend mullite (3Al2O3·2SiO2) have been termed ZTA and ZTM [11]. In contrast with the other two classes, stability of the t* phase to room temperature does not primarily involve the use of dopants but is controlled instead by particle size, particle morphology and location (intra- or intergranular; see Fig. 1a). In ZTA, for example, particles above a critical size will transform to monoclinic symmetry upon cooling to room temperature [12]. Since this t→m transfor￾mation is known to be martensitic, a useful way to describe particle size effects has been to examine their influence on the martensitic start (Ms) temperature; essentially all t-phase stabilization can be viewed as decreasing the Ms to below room temperature. Such investigation has suggested that the particle size effect is likely due to difficulties in nucleating the transformation, although considerations have also been given to the possible effects of surface and strain energy and chemical free energy driving forces [12]. Within dental mate￾rials the sole commercial example of a dispersion-toughened ceramic is In-Ceram Zirconia (Vita Zahnfabrik) which is an interpenetrating composite of 30% glass and 70% polycrys￾talline ceramic consisting of Al2O3:ZrO2 in a vol.% ratio of approximately 70:30. 3.2. Partially stabilized zirconia These materials are the most widely studied, commer￾cially important, microstructurally complex, and in the case of Mg-doped some of the toughest of the transformation￾toughened ceramics. In these ceramics t-ZrO2 intra-granular precipitates exist within a matrix of stabilized c-ZrO2. Sta￾bilization involves dopant addition, such as with CaO, MgO, La2O3, and Y2O3, in concentrations lower than that required for full c-ZrO2 stabilization. Full stabilization is purpose￾fully not achieved in these materials, hence the historical derivation for the term “partially stabilized zirconia” or PSZ, to which the relevant dopant is often appended: Ca￾PSZ, Mg-PSZ, Y-PSZ, etc. [11]. Precipitates are fully coherent with the cubic lattice, forming on a nanometer scale with lenticular morphology (approximately 200 nm diameter and 75 nm thick) parallel to the three cubic axes (refer to Fig. 1b) [12,13]. Following sintering or solution annealing in the cubic solid solution single-phase field (approximately >1850 ◦C), precipitates are nucleated and grown at lower temperatures (approximately 1100 ◦C) within the two-phase tetragonal solid solution plus cubic solid solution phase field; a process termed “aging” [12,13]. Aging optimization (time￾temperature-transformation) involves both precipitate size and phase stability [14]. Metastability can be lost when tetrag￾onal precipitates are too small (they will not transform) and when precipitates grow too large spontaneous transformation can occur to m with twinning and microcracking [14]. Complex decomposition and tertiary precipitation pro￾cesses have also been reported to occur with aging of Mg-PSZ [15] along with the development of quite some range of phys-

292 DENTAL MATERIALS 24(2008)289-298 dispersed t-Zro2 phase lenticular t-zro, precipitates on cubic faces a. Ziconia-toughenend alumina(ZTA b Mg partially-stabilized zirconia(Mg-PSz) Yttria-stabilized tetragonal zirconia polycrystalline(y-tZP Fig. 1- Microstructural features of the three major categories of transformation-toughened zirconia Compiled with permission from Heuer[11] and Matsui et al.22 ical properties, for example, KIc ranging from 3 to 16MPa m12 8 phase is thought to directly explain the improved mechan with lower values occurring in both un-aged and over-aged ical properties of Mg-PSZ during aging[18]. One commercial ceramic [16]. Specific insights into bulk phase developments Mg-PSZ appears to be available as a dental ceramic( Denzir-M and microstructural control over toughening in Mg-PSZ came Dentronic AB, Sweden) and has received attention during with the advent and application of neutron diffraction; X-ray vitro testing of fixed partial dentures [191 penetration in these materials is extremely limited compared to centimeter depths possible with neutrons. Aging in Mg-PSZ 3.3. Single-phase, polycrystalline t-Zro2 involves the development of a complex microstructure form- ing from essentially a two-phase c+t starting system. These In 1977 it was reported that fine grain ZrO2 (gener changes include (1)tertiary t-phase precipitation,(2)some ally <0.5um) with small concentrations of stabilizing Y203 C-m transformation,(3)limited orthorhombic (o) phase for- could contain up to 98% of the metastable t phase fol- mation, and most critically,(4)growth of an anion-ordered lowing sintering [20]. High strengths coincided with high acancy phase termed delta( a )having the composition tetragonal phase content and low strengths coincided with Mg2ZrsO12[17, 18]. This 8 compound nucleates on the broad high monoclinic phase content [20). Subsequent investigation lenticular tetragonal-cubic phase boundary and grows with revealed that the highest strengths(700 MPa)and toughness onsumption of c-ZrO2. Although the t-8-phase boundary is KIc 6-9 MPa m1)were only found below a critical average coherent, some lattice parameter mismatch exists leaving the grain size(<0.3 um)[21]. This critical grain-size phenomenon t phase increasingly susceptible to transformation as the 8 indicated a strength/toughness mechanism beyond the sim- layer thickens [18].Growth of this Mg-rich 8 phase also appears ple flaw-related grain-size effects generally recognized for to occur with Mg depletion of the t precipitates. It has been polycrystalline ceramics. However, many attributes are still calculated roughly that the stress required for the t-m trans- shared with other polycrystalline materials, including the formation decreases from 470 to 70 MPa with aging, in a linear simplicity of processing; norequirement for"aging"heattreat fashion with &-phase formation [18].Thus, the precipitation of ments, and reciprocally, relative insensitivity to follow-on heat

292 dental materials 24 (2008) 289–298 Fig. 1 – Microstructural features of the three major categories of transformation-toughened zirconia. Compiled with permission from Heuer [11] and Matsui et al. [22]. ical properties, for example, KIC ranging from 3 to 16 MPam1/2 with lower values occurring in both un-aged and over-aged ceramic [16]. Specific insights into bulk phase developments and microstructural control over toughening in Mg-PSZ came with the advent and application of neutron diffraction; X-ray penetration in these materials is extremely limited compared to centimeter depths possible with neutrons. Aging in Mg-PSZ involves the development of a complex microstructure form￾ing from essentially a two-phase c + t starting system. These changes include (1) tertiary t-phase precipitation, (2) some c→m transformation, (3) limited orthorhombic (o) phase for￾mation, and most critically, (4) growth of an anion-ordered vacancy phase termed delta (ı) having the composition Mg2Zr5O12 [17,18]. This ı compound nucleates on the broad lenticular tetragonal-cubic phase boundary and grows with consumption of c-ZrO2. Although the t-ı-phase boundary is coherent, some lattice parameter mismatch exists leaving the t phase increasingly susceptible to transformation as the ı layer thickens [18]. Growth of this Mg-rich ı phase also appears to occur with Mg depletion of the t precipitates. It has been calculated roughly that the stress required for the t→m trans￾formation decreases from 470 to 70 MPa with aging, in a linear fashion with ı-phase formation [18]. Thus, the precipitation of ı phase is thought to directly explain the improved mechan￾ical properties of Mg-PSZ during aging [18]. One commercial Mg-PSZ appears to be available as a dental ceramic (Denzir-M, Dentronic AB, Sweden) and has received attention during in vitro testing of fixed partial dentures [19]. 3.3. Single-phase, polycrystalline t-ZrO2 In 1977 it was reported that fine grain ZrO2 (gener￾ally < 0.5m) with small concentrations of stabilizing Y2O3 could contain up to 98% of the metastable t phase fol￾lowing sintering [20]. High strengths coincided with high tetragonal phase content and low strengths coincided with high monoclinic phase content [20]. Subsequent investigation revealed that the highest strengths (∼=700 MPa) and toughness (KIC ∼= 6–9 MPam1/2) were only found below a critical average grain size (<0.3m) [21]. This critical grain-size phenomenon indicated a strength/toughness mechanism beyond the sim￾ple flaw-related grain-size effects generally recognized for polycrystalline ceramics. However, many attributes are still shared with other polycrystalline materials, including the simplicity of processing; no requirement for “aging” heat treat￾ments, and reciprocally, relative insensitivity to follow-on heat

DENTAL MATERIALS 24(2008)289-298 293 treatments (e.g. porcelain firings in dental usage)and the much broader grain size range, approximately 8-0.25 um[26] opportunity to explore chemistry-based powder fabrication Both grain size and the test temperature in relationship to the and nano-scale microstructures Ms temperature control the size of the transformation zone (t These materials consist primarily of equiaxed grains of or h)associated with a growing crack(to be discussed momen t-zro2 sintered (and sometimes hot isostatically pressed or tarily); with rr directly influencing toughness mechanisms HIP'ed)to 96-99.5% of theoretical density. It has recently For 2 mol% Y-TZP, AK was clearly shown to decrease as tem- been reported that some grain-boundary segregation of Y3+ peratures moved away(higher from the Ms temperature[26] ions occurs during late-stage sintering[22]. It is energetically Transformation zone size of a crack extended at two different favorable for the cubic phase to form at yttria-rich triple junc- temperatures indicated an rr of slum at 295K and =10 um ons and grain boundaries at temperatures between 1300 and at 80K(with Ms presumed to be below 80K)(26] 1500 C[22]. The implications for hydrolytic stability(low tem- Nano-scale transformation-toughened Zro2 is alread erature degradation) by having minor volume fractions of appearing in the literature and in limited commercializa Zro2 segregated at grain boundaries along with the depletion tion. It was reported in 2002 that the trend toward increased of Y3+ from t-Zro2 will be discussed later. phase stability with decreasing particle-size of t-ZrO2 could Properties of these essentially homogeneous ceramics are be overcome by adjusting the yttria dopant concentration[23] primarily a function of grain size, in that grain size controls the Whereas 3 mol% Y2O3 has been found to optimize toughness Ms temperature and the ease of transformation(and hence the in micrometer and sub-micrometer t-ZrO2, dopant concen- toughness effect). The closer the test or service temperature trations and critical grain size for nano-scale material were is to the Ms temperature(but still above it avoiding sponta- identified as 1.0 mol% for 90nm and 1.5 mol% for 110 nm; neous transformation) the less stable are the t grains and the both combinations reaching around 16 MPa m[23. As with higher the available toughness increment. For a given dopant micro-scale zirconias, strong grain size effects(decreasing concentration the toughness increment decreases as grains toughness with decreasing grain size) were exhibited in these become much smaller than the critical size, presumably due to nano-scale ceramics as well[23]. At least one commercial over-stabilization"of the grains, eliminating the t-m trans- nano-scale Ce-TZP containing 20%Al2 O3(Nanozir, Matsushita formation upon introduction of a crack[23] Electrical Works, Ltd. is being examined as a dental ceramic This grain size effect itself is controlled by the type [28 having a reported fracture toughness of approximately dopant and the dopant concentration that are essentially 20 MPa m12(E. Rothbrust, Ivoclar Vivadent, Inc, personal com- determining(1)the degree of tetragonality (i.e. crystal lattice munication). It may be that nano-scale t-zro2 will primarily length ratio of c/a being>1.0)and(2) the thermal expan- appear in a polycrystalline form due to the difficulties for sion anisotropy (c versus a directions) of the unit cells. The intra-granular precipitation and tertiary phase development compositional effect of the dopant can be represented by required in PSZ. materials the resultant unit cell tetragonality (c/a[24, 25]. In general higher tetragonality contributes to a less stable material char- cterized by an increased Ms temperature [24]. Based on4. Mechanism(s)and consequences of tragonality, at similar grain size and dopant concentra- transformation tions yttrium appears to be a stronger stabilizer than cerium and especially titanium [25]. Anisotropic thermal expansie Numerous factors are discussed in the literature as (1)nucle- for the c and a axes, can influence residual strains in t ating and driving the transformation and (2)controlling grains; higher residual stresses can lower the nucleation stress the consequences of transformation. Two main phenomena threshold for the t-m in the presence of crack-tip strain resulting from transformation include (1)increased resistance energy[26]. Linear thermal expansion coefficients()are avail- for growth of both short(0.3mm)cracks obtained by direct measure oflattice parameters from 800 to with crack length (termed R-curve behavior) until generally room temperature. These measures indicate that anisotropy reaching a toughness plateau. These transforming is minimal (aa and ac crossover or, equivalently, aa/ac =1) at step away from the simple Griffith dependence on flaw size 4mol%Y2O3[27] and many have strengths that depend on the stress needed In its most basic form, a transformation-toughening con to trigger transformation rather than being flaw-size ser tribution(ak)has been defined as[26 tive. Quite non-linear behavior is exhibited by the toughest materials bordering on quasi-plasticity with measurable pre (2) failure deformation. Therefore, as will be discussed below, high strength and high toughness do not present in the same where ko is the fracture resistance inherent to the matrix with- material out transformation toughening. In general, the toughness con Driving forces and the role of temperature particularly for tribution from transformation(15 MPam1/2)exceeds that for the t-m transformation can be simply considered within microcracking (2-6 MPam 2) or deflection(-2-4MPamv2) a thermodynamic framework, as reproduced here following the work of Becher and Swain [26]. The total unit volume At a given temperat ransformation-toughening free energy change AFo for the transformation, including an ontribution(△K)for2 TZP decreased from approx applied stress is ain size decreased from 0.5 um; for 12 mol% Ce-TzP this same range occurred △Fo=△FcH+△Ue+△Us-△U1

dental materials 24 (2008) 289–298 293 treatments (e.g. porcelain firings in dental usage) and the opportunity to explore chemistry-based powder fabrication and nano-scale microstructures. These materials consist primarily of equiaxed grains of t-ZrO2 sintered (and sometimes hot isostatically pressed or HIP’ed) to 96–99.5% of theoretical density. It has recently been reported that some grain-boundary segregation of Y3+ ions occurs during late-stage sintering [22]. It is energetically favorable for the cubic phase to form at yttria-rich triple junc￾tions and grain boundaries at temperatures between 1300 and 1500 ◦C [22]. The implications for hydrolytic stability (low tem￾perature degradation) by having minor volume fractions of c-ZrO2 segregated at grain boundaries along with the depletion of Y3+ from t-ZrO2 will be discussed later. Properties of these essentially homogeneous ceramics are primarily a function of grain size, in that grain size controls the Ms temperature and the ease of transformation (and hence the toughness effect). The closer the test or service temperature is to the Ms temperature (but still above it avoiding sponta￾neous transformation) the less stable are the t grains and the higher the available toughness increment. For a given dopant concentration the toughness increment decreases as grains become much smaller than the critical size, presumably due to “over-stabilization” of the grains, eliminating the t→m trans￾formation upon introduction of a crack [23]. This grain size effect itself is controlled by the type of dopant and the dopant concentration that are essentially determining (1) the degree of tetragonality (i.e. crystal lattice length ratio of c/a being > 1.0) and (2) the thermal expan￾sion anisotropy (c versus a directions) of the unit cells. The compositional effect of the dopant can be represented by the resultant unit cell tetragonality (c/a) [24,25]. In general higher tetragonality contributes to a less stable material char￾acterized by an increased Ms temperature [24]. Based on tetragonality, at similar grain size and dopant concentra￾tions yttrium appears to be a stronger stabilizer than cerium and especially titanium [25]. Anisotropic thermal expansion, for the c and a axes, can influence residual strains in t grains; higher residual stresses can lower the nucleation stress threshold for the t→m in the presence of crack-tip strain energy [26]. Linear thermal expansion coefficients (˛) are avail￾able for yttria-doped ZrO2 over a limited concentration range, obtained by direct measure of lattice parameters from 800 ◦C to room temperature. These measures indicate that anisotropy is minimal (˛a and ˛c crossover or, equivalently, ˛a/˛c ∼= 1) at 4.5mol% Y2O3 [27]. In its most basic form, a transformation-toughening con￾tribution (KT) has been defined as [26]: Kc = Ko + KT (2) where Ko is the fracture resistance inherent to the matrix with￾out transformation toughening. In general, the toughness con￾tribution from transformation (≈15 MPa m1/2) exceeds that for microcracking (≈2–6 MPam1/2) or deflection (≈2–4 MPam1/2) mechanisms [9]. At a given temperature, the transformation-toughening contribution (KT) for 2mol% Y-TZP decreased from approx￾imately 12–2.5 MPam1/2 as grain size decreased from 2 to 0.5m; for 12mol% Ce-TZP this same range occurred over a much broader grain size range, approximately 8–0.25m [26]. Both grain size and the test temperature in relationship to the Ms temperature control the size of the transformation zone (rT or h) associated with a growing crack (to be discussed momen￾tarily); with rT directly influencing toughness mechanisms. For 2mol% Y-TZP, KT was clearly shown to decrease as tem￾peratures moved away (higher) from the Ms temperature [26]. Transformation zone size of a crack extended at two different temperatures indicated an rT of ≤1m at 295 ◦K and ∼=10m at 80 ◦K (with Ms presumed to be below 80 ◦K) [26]. Nano-scale transformation-toughened ZrO2 is already appearing in the literature and in limited commercializa￾tion. It was reported in 2002 that the trend toward increased phase stability with decreasing particle-size of t-ZrO2 could be overcome by adjusting the yttria dopant concentration [23]. Whereas 3mol% Y2O3 has been found to optimize toughness in micrometer and sub-micrometer t-ZrO2, dopant concen￾trations and critical grain size for nano-scale material were identified as 1.0mol% for 90 nm and 1.5mol% for 110 nm; both combinations reaching around 16 MPam1/2 [23]. As with micro-scale zirconias, strong grain size effects (decreasing toughness with decreasing grain size) were exhibited in these nano-scale ceramics as well [23]. At least one commercial nano-scale Ce-TZP containing 20% Al2O3 (Nanozir, Matsushita Electrical Works, Ltd.) is being examined as a dental ceramic [28]; having a reported fracture toughness of approximately 20 MPam1/2 (F. Rothbrust, Ivoclar Vivadent, Inc., personal com￾munication). It may be that nano-scale t-ZrO2 will primarily appear in a polycrystalline form due to the difficulties for intra-granular precipitation and tertiary phase development required in PSZ materials. 4. Mechanism(s) and consequences of transformation Numerous factors are discussed in the literature as (1) nucle￾ating and driving the transformation and (2) controlling the consequences of transformation. Two main phenomena resulting from transformation include (1) increased resistance for growth of both short (≤100m) and long (≥0.3mm) cracks and, for many ceramics, (2) toughness continuing to increase with crack length (termed R-curve behavior) until generally reaching a toughness plateau. These transforming ceramics step away from the simple Griffith dependence on flaw size and many have strengths that depend on the stress needed to trigger transformation rather than being flaw-size sensi￾tive. Quite non-linear behavior is exhibited by the toughest materials bordering on quasi-plasticity with measurable pre￾failure deformation. Therefore, as will be discussed below, high strength and high toughness do not present in the same material. Driving forces and the role of temperature particularly for the t→m transformation can be simply considered within a thermodynamic framework, as reproduced here following the work of Becher and Swain [26]. The total unit volume free energy change FO for the transformation, including an applied stress is: FO = FCH + Ue + US − UI (3)

DENTAL MATERIALS 24(2008)289-298 where AFcH is the chemical free energy change, AUe the strain increase in Ms temperatures with increasing grain size [26] energy change, AUs the surface energy change, and AU the Local residual stresses will also scale with grain size, provid- interaction term due to stress application. The surface free ing an additional or altemative explanation for the Ms tracking energy change is considered negligible and the chemical free with grain size [26] energy term can be stated as a function of transformation Transformation involves the development of a transforma entropy change Asm and temperature [26] tion zone first associated with the crack tip and later becoming a crack wake feature The size of this zone and features of the △FcH=△stm(To-T) (4) zone microstructure(grain size and microcracking, in particu- lar) control toughening. Following the analysis of Evans 2, the where T is the test temperature and To the transformation toughening increment due to transformation is a function of temperature for an unconstrained t-phase particle. Tetrago- the transformation zone width h, the local elastic modulus of nal particles embedded within a matrix will be stabilized by transformed material e, the volume fraction of transformable Constrained particles, that are metastable, will transform at is Poisson's ratio e ational strain e as given below(where v both matrix constraint as well as the strain energy term AUe. phase f, and the dilat a temperature Ms<To (with Ms being the martensitic start temperature discussed earlier). Dopants such as Y203 act to √h decrease both the Ms and To temperatures. This decrease in △Kc=0.2 22a-可 Ms below To creates an additional free energy change term essentially capturing matrix and dopant stabilization, that is The simplest, most commonly"understood"toughening equal to As -m To-Ms)[26]. Thus, without an applied stress, mechanism concept involves crack-tip shielding(from the the free energy change for transformation can be written applied stress)by the compressive dilatational stress asso- terms of the test(m)and Ms temperatures as ciated with transformation. In fact, it is now thought that purely dilatant transformation within an included angle AFo=ASmI(To-T)-(To-Ms)]=Ast-m(Ms-T)(5) 120 ahead of the crack tip(hydrostatic component in Fig. 2) leads to a decrease in toughness[9. Other toughening factors reflecting that AFo=0 as T=Ms, initiating transformation have been advanced related to the microstructure of the trans- At test(or service)temperatures above Ms, an applied stress formed material that act both locally near the crack tip and Un can reduce AFo to zero and trigger thet+m transformation in the crack wake. For example, the crack tip is embedded in a microcracked material having an elastic modulus that dif- fers from the bulk ceramic, creating an additional crack tip (6) shielding term. As cracks grow to lengths of approximately h-5 to h.10 the toughness reaches a plateau [16] determined Returning to more familiar and useful concepts involving by crack wake and crack tip toughening mechanisms (29,30) tensile stress and strain is achieved by defining the interactie to be elucidated below in the discussion of R-curve behavior term UI in terms of applied stress aa and the transformation dilatational strain g[26] 3 Formed transformation zone △U1=aE Now the critical applied stress for transformation o becomes[26] 3. Asymptote =0.22.. a-V I+ISVD=ED (8)△K Two practical implications come from this analys △a born-out by experimental evidence. First the stress required for transformation increases with temperature above Ms and 2. Partially developed zone second at a given test or service temperature(above Ms)this critical applied stress will increase as Ms is decreased [26]. The kinetics of transformation are governed by nucleation, with the probability of nucleation enhanced by local resid ual stresses as well as applied stress[ 29. Sources of local hydrostatic component residual stresses have been described above(e.g 8-phase/t phase lattice mismatch in PSZ and t-phase a anisotropy in TZP). Recall from above that for PSZ the stress required for △a/h the t- m transformation decreases from 470 to 70 MPa with Fig. 2- Schematic views of transformation zone and 8-phase formation during aging [18]. Since the probability of a toughness increment(AKe)development with crack potent nuclei existing within a grain should scale with grain extension, recreated with permission per description of volume, nucleation considerations also partially explain the Evans [2] and Evans and Heuer

294 dental materials 24 (2008) 289–298 where FCH is the chemical free energy change, Ue the strain energy change, US the surface energy change, and UI the interaction term due to stress application. The surface free energy change is considered negligible and the chemical free energy term can be stated as a function of transformation entropy change St→m and temperature [26]: FCH = St→m(TO − T) (4) where T is the test temperature and TO the transformation temperature for an unconstrained t-phase particle. Tetrago￾nal particles embedded within a matrix will be stabilized by both matrix constraint as well as the strain energy term Ue. Constrained particles, that are metastable, will transform at a temperature Ms < TO (with Ms being the martensitic start temperature discussed earlier). Dopants such as Y2O3 act to decrease both the Ms and TO temperatures. This decrease in Ms below TO creates an additional free energy change term essentially capturing matrix and dopant stabilization, that is equal to St→m (TO − Ms) [26]. Thus, without an applied stress, the free energy change for transformation can be written in terms of the test (T) and Ms temperatures as: FO = St→m[(TO − T) − (TO − Ms)] = St→m(Ms − T) (5) reflecting that FO = 0 as T = Ms, initiating transformation. At test (or service) temperatures above Ms, an applied stress UI can reduce FO to zero and trigger the t→m transformation [26]: FO = St→m(Ms − T) − UI (6) Returning to more familiar and useful concepts involving tensile stress and strain is achieved by defining the interaction term UI in terms of applied stress a and the transformation dilatational strain εT [26]: UI = aεT (7) Now the critical applied stress for transformation T c becomes [26]: a = T c = St→m(Ms − T) εT (8) Two practical implications come from this analysis, both born-out by experimental evidence. First the stress required for transformation increases with temperature above Ms and second at a given test or service temperature (above Ms) this critical applied stress will increase as Ms is decreased [26]. The kinetics of transformation are governed by nucleation, with the probability of nucleation enhanced by local resid￾ual stresses as well as applied stress [29]. Sources of local residual stresses have been described above (e.g. ı-phase/t￾phase lattice mismatch in PSZ and t-phase anisoptropy in TZP). Recall from above that for PSZ the stress required for the t→m transformation decreases from 470 to 70 MPa with ı-phase formation during aging [18]. Since the probability of a potent nuclei existing within a grain should scale with grain volume, nucleation considerations also partially explain the increase in Ms temperatures with increasing grain size [26]. Local residual stresses will also scale with grain size, provid￾ing an additional or alternative explanation for the Ms tracking with grain size [26]. Transformation involves the development of a transforma￾tion zone first associated with the crack tip and later becoming a crack wake feature. The size of this zone and features of the zone microstructure (grain size and microcracking, in particu￾lar) control toughening. Following the analysis of Evans [2], the toughening increment due to transformation is a function of the transformation zone width h, the local elastic modulus of transformed material E, the volume fraction of transformable phase f, and the dilatational strain εT ij , as given below (where  is Poisson’s ratio) [2]: KC = 0.22EεT ijf √ h (1 − ) (9) The simplest, most commonly “understood” toughening mechanism concept involves crack-tip shielding (from the applied stress) by the compressive dilatational stress asso￾ciated with transformation. In fact, it is now thought that purely dilatant transformation within an included angle of 120◦ ahead of the crack tip (hydrostatic component in Fig. 2) leads to a decrease in toughness [9]. Other toughening factors have been advanced related to the microstructure of the trans￾formed material that act both locally near the crack tip and in the crack wake. For example, the crack tip is embedded in a microcracked material having an elastic modulus that dif￾fers from the bulk ceramic, creating an additional crack tip shielding term. As cracks grow to lengths of approximately h·5 to h·10 the toughness reaches a plateau [16] determined by crack wake and crack tip toughening mechanisms [29,30] to be elucidated below in the discussion of R-curve behavior; Fig. 2 – Schematic views of transformation zone and toughness increment (Kc) development with crack extension, recreated with permission per description of Evans [2] and Evans and Heuer [5].

DENTAL MATERIALS 24(2008)289-298 elements of this concept are presented schematically in Fig. 2. ximated as[11] More recent theoretical modeling places much more empha sis on shear versus dilatational stresses for both activating GRiffith+△K*(△C the transformation and leading to an increase up to fourfold Y(co+△c)05 in transformation zone height[9). Perhaps the most compre- hensive review of the current state of theoretical work is found where Acf is the stable crack extension prior to failure, co the in Hannink et al. [91 initial flaw size, and Y a flaw geometry constant. Acf is further Experimentally the size of the transformation zone is found defined as [11] to be a function of test temperature and grain size; increasing [25]. Zone size as a function of temperature(relative to M. Ac=2 as t drops toward Ms and increasing as grain size increases has been visually demonstrated by Becher and Swain for a (r)05/(coag)os Y-TZP[26]. Zone size plays a critical role in the toughness with d being the transformation zone size described above increment achieved as a result of the microstructural changes Adding to this complexity is the finding that many occurring with transformation Microcracking transformation the highest toughness ceramics(especially high Mg-PSz zones have been visualized by Lutz et al. for a 20% Al203 Y- demonstrate non-linear, non-elastic yielding prior to fail- TZP in a dramatic series of photomicrographs [31]. Zone sizes ure [30]. Deformation mechanisms include both reversible have been measured by several techniques including opti- and irreversible components, with the reversible associated cal and transmission electron microscopy, X-ray diffraction with transformation and the irreversible with microcracking and Raman spectrometry [32]. Reported zone sizes include: [30, 35]. Marshall and James [35] provided both photographic 2um to 70 um in Mg-PSZ having Kic values between 4 and (Nomarski interference)and x-ray diffraction data to convinc- ues varying from 5 to 10MPam2[32 e gly establish the presence of a reversible transformation around 200 MPa in a 9mol% Mg-PSZ. This reversibility raises an issue regarding the concept of nucleation being a R-curve behavior: description, definition key feature of the transformation in this material and implications same material was found to exhibit a non-linear stress-strain curve with irreversible strain at failure being as large as the elastic strain [30] as do many transformation-toughened An increase in the resistance to crack growth (i.e. toughness) zirconia ceramics having high Kic values(36). Both the stress- during crack extension has been termed R-curve behavior. R- induced transformation along with microcracking prior to curve behavior was originally noted for large-graine failure account for the nonlinear yielding Two major implica having thermal expansion anisotropy and the increasing tions arise for highly toughened ceramics: (1)actual stresses resistance to crack extension with crack length was mainly at failure can be much lower(up to 40%)than calculated by attributed to mechanisms active in the crack wake that linear elastic analysis and(2) these materials are remarkably shielded the crack tip from the applied stress intensity 33, 34. damage tolerant, e.g. strengths of Vickers-indented speci In transforming ceramics additional toughening mechanisms mens (up to 1000N) can equal polished-surface specimens contribute to R-curve behavior, with the total available being (301 subdivided into three basic groups [31]: (1)crack deflection and crack branching: (2) contact shielding by wedging and bridg ing involving broken-out grains or rough crack-wake surfaces 6. Strength versus toughness and (3)stress-induced zone shielding involving transforma tion, microcracking and residual stress fields For linearly elastic brittle materials the highest strength and The KI given in Eq(1)as the Griffith failure criteria is no highest toughness occur in the same material; this is not longer"critical. This applied k is suffcient for initial flaw the case for transforming ceramics. As cracks grow from an growth but insufficient for catastrophic crack propagation initial size, transformation events create an incrementally ow growing with increasing length increasing toughness, AK, with toughness increments scal (Fig. 2).One new description of the critical stress intensity for ing proportional to h, where h is the transformation zone fracture, Kf, has been given by size in the nomenclature of Swain and Rose 36]. This scal- ing yields R-curve behavior discussed above, until h is fully K= GRiffith+△K*(△c) (10) developed and the toughness reaches a plateau. However, incremental crack extension, Ac, scales directly with h while mental crack extension Ac AK(Ac)(essentially the R-curve)is in scaling properties, vh versus h, yields a strength maxi- a function of the depth and location of the initial crack, speci- mum with increasing steady-state toughness 36]. In practical men dimensions, test conditions and methods; hence there is terms the ceramic begins to weaken(crack growth o h)even no unique R-curve for a material [31]. This implies that Kr is not though the toughness is still increasing(AK o h). Accor specifying a unique material failure criterion, or is a"material ing to Swain and Rose [36] strengths beyond this maximum property"in a sense analogous with KIc become bounded by the critical stress at the crack tip nec Keeping with the formulation of Heuer the fracture essary for the t-m transformation. This analytical concept trength of transformation-toughened ceramics can be appro- was illustrated using measured strength and toughness data

dental materials 24 (2008) 289–298 295 elements of this concept are presented schematically in Fig. 2. More recent theoretical modeling places much more empha￾sis on shear versus dilatational stresses for both activating the transformation and leading to an increase up to fourfold in transformation zone height [9]. Perhaps the most compre￾hensive review of the current state of theoretical work is found in Hannink et al. [9]. Experimentally the size of the transformation zone is found to be a function of test temperature and grain size; increasing as T drops toward Ms and increasing as grain size increases [25]. Zone size as a function of temperature (relative to Ms) has been visually demonstrated by Becher and Swain for a Y-TZP [26]. Zone size plays a critical role in the toughness increment achieved as a result of the microstructural changes occurring with transformation. Microcracking transformation zones have been visualized by Lutz et al. for a 20% Al2O3 Y￾TZP in a dramatic series of photomicrographs [31]. Zone sizes have been measured by several techniques including opti￾cal and transmission electron microscopy, X-ray diffraction and Raman spectrometry [32]. Reported zone sizes include: 0.2m to 70m in Mg-PSZ having KIC values between 4 and 14 MPa m1/2 and 0.8–4.6m for Y-TZP (2–4mol%) with KIC val￾ues varying from 5 to 10 MPam1/2 [32]. 5. R-curve behavior: description, definition and implications An increase in the resistance to crack growth (i.e. toughness) during crack extension has been termed R-curve behavior. R￾curve behavior was originally noted for large-grained ceramics having thermal expansion anisotropy and the increasing resistance to crack extension with crack length was mainly attributed to mechanisms active in the crack wake that shielded the crack tip from the applied stress intensity [33,34]. In transforming ceramics additional toughening mechanisms contribute to R-curve behavior, with the total available being subdivided into three basic groups [31]: (1) crack deflection and crack branching; (2) contact shielding by wedging and bridg￾ing involving broken-out grains or rough crack-wake surfaces and (3) stress-induced zone shielding involving transforma￾tion, microcracking and residual stress fields. The KI given in Eq. (1) as the Griffith failure criteria is no longer “critical”. This applied K is sufficient for initial flaw growth but insufficient for catastrophic crack propagation since crack resistance is now growing with increasing length (Fig. 2). One new description of the critical stress intensity for fracture, Kf, has been given by Heuer as [11]: Kf = KGriffith + K ∗ (c) (10) where K* is the toughening increment as a function of incre￾mental crack extension c. K*(c) (essentially the R-curve) is a function of the depth and location of the initial crack, speci￾men dimensions, test conditions and methods; hence there is no unique R-curve for a material [31]. This implies that Kf is not specifying a unique material failure criterion, or is a “material property” in a sense analogous with KIC. Keeping with the formulation of Heuer the fracture strength of transformation-toughened ceramics can be appro￾ximated as [11]: f = [KGriffith + K ∗ (cf)] Y(cO + cf) 0.5 (11) where cf is the stable crack extension prior to failure, cO the initial flaw size, and Y a flaw geometry constant. cf is further defined as [11]: cf = 2 () 0.5 (cOd) 0.5 (12) with d being the transformation zone size described above. Adding to this complexity is the finding that many of the highest toughness ceramics (especially high Mg-PSZ) demonstrate non-linear, non-elastic yielding prior to fail￾ure [30]. Deformation mechanisms include both reversible and irreversible components, with the reversible associated with transformation and the irreversible with microcracking [30,35]. Marshall and James [35] provided both photographic (Nomarski interference) and X-ray diffraction data to convinc￾ingly establish the presence of a reversible transformation at around 200 MPa in a 9mol% Mg-PSZ. This reversibility raises an issue regarding the concept of nucleation being a key feature of the transformation in this material [35]. This same material was found to exhibit a non-linear stress-strain curve with irreversible strain at failure being as large as the elastic strain [30] as do many transformation-toughened zirconia ceramics having high KIC values [36]. Both the stress￾induced transformation along with microcracking prior to failure account for the nonlinear yielding. Two major implica￾tions arise for highly toughened ceramics: (1) actual stresses at failure can be much lower (up to 40%) than calculated by linear elastic analysis and (2) these materials are remarkably damage tolerant, e.g. strengths of Vickers-indented speci￾mens (up to 1000 N) can equal polished-surface specimens [30]. 6. Strength versus toughness For linearly elastic brittle materials the highest strength and highest toughness occur in the same material; this is not the case for transforming ceramics. As cracks grow from an initial size, transformation events create an incrementally increasing toughness, K, with toughness increments scal￾ing proportional to √h, where h is the transformation zone size in the nomenclature of Swain and Rose [36]. This scal￾ing yields R-curve behavior discussed above, until h is fully developed and the toughness reaches a plateau. However, incremental crack extension, c, scales directly with h while R-curve toughness becomes developed [36]. This difference in scaling properties, √h versus h, yields a strength maxi￾mum with increasing steady-state toughness [36]. In practical terms the ceramic begins to weaken (crack growth ∝ h) even though the toughness is still increasing (K ∝ √h). Accord￾ing to Swain and Rose [36] strengths beyond this maximum become bounded by the critical stress at the crack tip nec￾essary for the t→m transformation. This analytical concept was illustrated using measured strength and toughness data

DENTAL MATERIALS 24(2008)289-298 Strength(GPa) 7 orders of magnitude higher than for chemically assisted (water-enhanced)crack growth at equivalent crack-tip stress intensities 37 In fact, a number of authors have demon Y-TZP+ Al2O4 strated crack growth under cyclic conditions that arrests when the same specimens are then held statically with the same Kmax and then resume growth under resumed cycling] Crack growth rates, da/dN, are found to be a function of the cyclic stress intensity range, AK, in a power law relationship 37: where Cis a constant. The exponent, m, is in the range of 21-42 (for metals the same relationship holds with m ranging from gness(MPavm) 2to4)B37] It is also found that for small cracks($100 um; i.e. natural g 3- Strength vs toughness curves for four types of flaw size range)(1)growth rates are far in excess of that for transformation-toughened zirconia recreated with long cracks(> mm) at equivalent applied stress intensities permission from work of Swain and Rose[36]. Solid lines and (2) that crack extension occurs at stress intensities below are analytical fits to experimental data (not shown). Dashed the threshold for long-crack growth[39]. Both of these differ- line represents the critical stress for the t-m ences in short crack behavior are likely due to the incomplete transformation. Origin of the strength maximum and the R-curve toughness development related to the incomplete non-coincidence of maximum strength and maximum establishment of a steady-state transformation zone for the toughness are discussed in the text. short crack Cyclic crack growth rates can also be influenced by cycling history under variable amplitude loading; demonstrating for four different transformation-toughened ceramics, as is accelerated growth following an increase in maximum load reproduced in Fig 3[ 29, 36 (increasing Ak) and growth retardation following a reduction Another feature of this analysis by Swain and Rose is the in peak cyclic load(decreasing AK)[37. This history effect prediction that a lower toughness is achieved at mum is thought to be related to(1) the transformation zone size trength for ceramics having smaller initial flaw sizes [36]. being sensitive to the crack-tip Kmax and(2)the steady-state This prediction comes from the feature that transformation shielding becoming maximal after crack extensions of approx zone size, h, at maximum strength depends on the initial mately five times the zone width [ 37] flaw size, Ci, with ho(Gi)[2, 36]. Recall from above that this Many characteristics discussed above regarding premature zone-size/filaw-size relationship is one reason why there is failure of zirconia materials under cyclic conditions compli no characteristic R-curve for any given material. This insight cate lifetime predictions: of Swain is also consistent with experimental data with flaw sizes known to be the smallest in y-tzP alumina materials and esign data from static tests have little value when the largest in Mg-PSZ. apparent threshold for fatigue crack growth can be approx- This strength-toughness"disconnect"also creates two imately 50% of the fracture toughness, K classes of materials at either end of the toughness continuum. Purely elastic failure criteria cannot be applied(even for The very high-strength and lower-toughness materials remain non-cyclic loading sensitive to processing flaws. In contrast the high-toughness Very high crack growth exponents(m)indicate extreme sen- and lower-strength ceramics are flaw and damage tolerant. sitivity to applied stres Swain and Rose cite the example that a 20 kg Vickers indenta- Transient crack acceleration/retardation effects imply that tion of peak-strength Y-TZP 20% alumina results in a ten-fold data from constant AK cyclic loading may be inadequate rength degradation whereas a peak-strength Mg-PSZ can The role of crack size in influencing cyclic behavior is incom- withstand twice that indentation load with no strength loss pletely understood [36]. Such considerations also lead to the generalization that for materials having a Kic less than a8 MPa"the strength is These factors imply that lifetime predictions may be enor limited by flaw size while above this value strength is limited mously sensitive to assumptions made regarding the initial by stress-activated transformations 91 defect size and in-service stresses 7. Cyclic fatigue of Low temperature degradation of 3Y-TZP transformation-toughened ceramics If most aspects of the transformation toughening ability Lifetimes of transformation-toughened ceramics are found to zirconia are positive, resulting in high strength and toughne be lower under cyclic loading than under equivalent static and reduced brittleness compared to alumina, considerable loading. Crack growth rates under cyclic conditions can amount of work has been devoted to the characterization of a

296 dental materials 24 (2008) 289–298 Fig. 3 – Strength vs. toughness curves for four types of transformation-toughened zirconia recreated with permission from work of Swain and Rose [36]. Solid lines are analytical fits to experimental data (not shown). Dashed line represents the critical stress for the t→m transformation. Origin of the strength maximum and the non-coincidence of maximum strength and maximum toughness are discussed in the text. for four different transformation-toughened ceramics, as is reproduced in Fig. 3 [29,36]. Another feature of this analysis by Swain and Rose is the prediction that a lower toughness is achieved at maximum strength for ceramics having smaller initial flaw sizes [36]. This prediction comes from the feature that transformation zone size, h, at maximum strength depends on the initial flaw size, ci, with h ∝ (ci) [2,36]. Recall from above that this zone-size/flaw-size relationship is one reason why there is no characteristic R-curve for any given material. This insight of Swain is also consistent with experimental data with flaw sizes known to be the smallest in Y-TZP alumina materials and largest in Mg-PSZ. This strength-toughness “disconnect” also creates two classes of materials at either end of the toughness continuum. The very high-strength and lower-toughness materials remain sensitive to processing flaws. In contrast the high-toughness and lower-strength ceramics are flaw and damage tolerant. Swain and Rose cite the example that a 20 kg Vickers indenta￾tion of peak-strength Y-TZP 20% alumina results in a ten-fold strength degradation whereas a peak-strength Mg-PSZ can withstand twice that indentation load with no strength loss [36]. Such considerations also lead to the generalization that for materials having a KIC less than ≈8 MPam1/2 the strength is limited by flaw size while above this value strength is limited by stress-activated transformations [9]. 7. Cyclic fatigue of transformation-toughened ceramics Lifetimes of transformation-toughened ceramics are found to be lower under cyclic loading than under equivalent static loading. Crack growth rates under cyclic conditions can be 7 orders of magnitude higher than for chemically assisted (water-enhanced) crack growth at equivalent crack-tip stress intensities [37]. In fact, a number of authors have demon￾strated crack growth under cyclic conditions that arrests when the same specimens are then held statically with the same Kmax and then resume growth under resumed cycling [37–39]. Crack growth rates, da/dN, are found to be a function of the cyclic stress intensity range, K, in a power law relationship [37]: da dN = C(K) m (13) where C is a constant. The exponent, m, is in the range of 21–42 (for metals the same relationship holds with m ranging from 2 to 4) [37]. It is also found that for small cracks (≤100m; i.e. natural flaw size range) (1) growth rates are far in excess of that for long cracks (≥3mm) at equivalent applied stress intensities and (2) that crack extension occurs at stress intensities below the threshold for long-crack growth [39]. Both of these differ￾ences in short crack behavior are likely due to the incomplete R-curve toughness development related to the incomplete establishment of a steady-state transformation zone for the short crack. Cyclic crack growth rates can also be influenced by cycling history under variable amplitude loading; demonstrating accelerated growth following an increase in maximum load (increasing K) and growth retardation following a reduction in peak cyclic load (decreasing K) [37]. This history effect is thought to be related to (1) the transformation zone size being sensitive to the crack-tip Kmax and (2) the steady-state shielding becoming maximal after crack extensions of approx￾imately five times the zone width [37]. Many characteristics discussed above regarding premature failure of zirconia materials under cyclic conditions compli￾cate lifetime predictions: • Design data from static tests have little value when the apparent threshold for fatigue crack growth can be approx￾imately 50% of the fracture toughness, Kc; • Purely elastic failure criteria cannot be applied (even for non-cyclic loading); • Very high crack growth exponents (m) indicate extreme sen￾sitivity to applied stress; • Transient crack acceleration/retardation effects imply that data from constant K cyclic loading may be inadequate; • The role of crack size in influencing cyclic behavior is incom￾pletely understood. These factors imply that lifetime predictions may be enor￾mously sensitive to assumptions made regarding the initial defect size and in-service stresses. 8. Low temperature degradation of 3Y-TZP If most aspects of the transformation toughening ability of zirconia are positive, resulting in high strength and toughness and reduced brittleness compared to alumina, considerable amount of work has been devoted to the characterization of a

DENTAL MATERIALS 24(2008)289-298 297 less appealing characteristic of zirconia: its susceptibility to further establishing that the enrichment of the cubic grains in low temperature degradation(LTD). This phenomenon was Y+ ions corresponded to a depletion in yttrium in the neigh first reported by Kobayashi in Ref. [40]. It was shown that boring tetragonal grains which became less stable and acted a slow t-m transformation from the metastable tetragonal as nucleation sites for the(t-m)transformation. phase to the more stable monoclinic phase occurs in surface n excellent review by Chevalier, evaluating the future grains in a humid environment at relatively low temperatures of zirconia as a biomaterial has recently been published (150-400C) The zirconia studied in this work was a polycrys- [51]. With zirconia becoming increasingly popular as a den talline zirconia stabilized by 4.5-6 mol% yttria. These results tal restorative material, under various forms and with a wide prompted a series of investigations on low temperature degra- range of processing conditions(Table 2), it seems wise to keep dation of zirconia stabilized with various amounts of Y2O3 in mind that some forms of zirconia are susceptible to aging that helped define the characteristics of the transformation and that processing conditions can play a critical role in the 41-47] low temperature degradation of zirconia. Classically, LTD initiates at the surface of polycrystalline zirconia and later progresses toward the bulk of the mate- rEFERENCES rial. The transformation of one grain is accompanied by an increase in volume that causes stresses on the surrounding grains and microcracking Water penetration then exacerbates the process of surface degradation and the transf [1] Messing GL, Hirano S, Gauckler L Ceramic processing science. J Am Ceram Soc 2006: 89 (6): 1769-70 progresses from neighbor to neighbor. The growth of the trans- [2] Evans AG. Perspectives on the development of formation zone results in severe microcracking, grain pullout high-toughness ceramics. J Am Ceram Soc and finally surface roughening, which ultimately leads to 1990;73(2):187-206 strength degradation. Any factor that would be detriment 3 Subbarao EC. Zirconia-an overview In: Heuer AH, Hobbs LW, to the stability of tetragonal zirconia is susceptible to promot editors. Science and technology of zirconia. Columbus, OH: low temperature degradation. Amongst these factors are the The American Ceramic Society: 1981. p. 1-24 grain size[48, the amount of stabilizer [9] and the presence of 4 Goff JP, Hayes W, Hull S, Hutchings MT, Clausen KN Defect structure of yttria-stabilized zirconia and its influence on residual stresses[49). he ionic conductivity at elevated temperatures. Phys Rev B One consequence of the above mentioned observations 1999:59(22):14202-19 was a 1997 posting by the Food and Drug Administration [5]Evans AG, Heuer AH. Review-transformation toughening in (http://www.fda.gov/cdrh/steamst.htmlcautioningagain ceramics:martensitic transformations in crack-tip stress fields. J Am Ceram Soc 1980; 63(5-6): 241-8 steam sterilization of zirconia femoral heads for total hip pros-(6] Garvie RG, Nicholson PS Structure and thermomechanical theses and specifying thatit could cause phase transformation and roughening of the material, later leading to increased roperties of partially stabilized zirconia in the Cao-zr02 Fabris S, Paxton A, Finnis MW. A stabilization mechanism of zirconia based on oxygen vacancies only. Acta Mater 8.1. Cubic phase and accelerated aging 8 Foschini CR, Filho T, Juiz SA, Souza AG, oliveira More recently, in 2001, some series of Y-TZP femoral heads t al. On the stabilizing behavior of zirconia: a co experimental and tical study. J Mater Sci were recalled due to spontaneous fractures. These inci- 200439:1935-41 dents were traced back to a manufacturing issue that led to 19) Hannink RH], Kelly PM, Muddle BC.Transformation accelerated tetragonal to monoclinic transformation in the toughening in zirconia-containing ceramics. J Am Ceram thecatastrophicfailures(http://www.prozyr.com/pages.Uk/[10GarvieRc,hAnninkRh,PascoeRt.Ceramicsteel?nature Biomedical/Committee. htm). These events had a noticeable 1975;258:703-4 negative impact on the use of zirconia as an implant bioma- [11 Heuer AH. Transformation toughening in ZrOz-containing ceramics.JAm Ceram Soc 1987: 70(10) 689-98 terial, triggering a considerable amount of research work in [12 Heuer AH, Claussen N, Riven WM, Ruhle M Stability of order to elucidate the possible origin of the failures (50, 51] tetragonal Zro2 particles in ceramic matrices. J Am Ceram Although anticipated from the phase diagram established Soc1982:65(12):642-50 earlier by Scott[52], recent work has demonstrated that both [13] Porter DL, Heuer AH. Mechanisms of toughening parti bic and tetragonal phases co-exist in 3Y-TZP [22, 50, 53 .Mat sui et al. reported that the amount of cubic zirconia in 3Y-TZP [4 Montross CS. Comparison of bulk properties of Mg-PSZ with sintered at 1300oC was 12.7 mass% and reached 18.6 mass% 1993;76(8):1993-7 when the sintering temperature was 1500C. These values [15] Hughan RR, Hannink RH). Precipitation during controlled were determined by X-ray diffraction and Rietveld calcula- ions [22]. In addition, the distribution of the Y3+ ions inside Ceo ing of magnesia-partially stabilized zirconia.JAm geneous while r fter sintering at 1300 C was nearly homo- [16] Steffen AA, Dauskardt RH, Ritchie RO Cyclic fatigue life and the zirconia ns appeared to concentrate within the rack-growth behavior of microstructurally small cracks in larger cubic grains after sintering at 1500 C. It was also shown magnesia-partially stabilized zirconia ceramics. JAm Ceram Soc1991;74(6):125968 that cubic phase regions started to form from grain bound- [171 Hannink RHJ Microstructural development of the aries and triple junctions where the y=+ ions had segregated sub-eutectoid aged Mgo-zro2 alloys. J Mater Sci 22 These results were confirmed by Chevalier et al. 50 1983;18:457-70

dental materials 24 (2008) 289–298 297 less appealing characteristic of zirconia: its susceptibility to low temperature degradation (LTD). This phenomenon was first reported by Kobayashi in Ref. [40]. It was shown that a slow t→m transformation from the metastable tetragonal phase to the more stable monoclinic phase occurs in surface grains in a humid environment at relatively low temperatures (150–400 ◦C). The zirconia studied in this work was a polycrys￾talline zirconia stabilized by 4.5–6mol% yttria. These results prompted a series of investigations on low temperature degra￾dation of zirconia stabilized with various amounts of Y2O3 that helped define the characteristics of the transformation [41–47]. Classically, LTD initiates at the surface of polycrystalline zirconia and later progresses toward the bulk of the mate￾rial. The transformation of one grain is accompanied by an increase in volume that causes stresses on the surrounding grains and microcracking.Water penetration then exacerbates the process of surface degradation and the transformation progresses from neighbor to neighbor. The growth of the trans￾formation zone results in severe microcracking, grain pullout and finally surface roughening, which ultimately leads to strength degradation. Any factor that would be detrimental to the stability of tetragonal zirconia is susceptible to promote low temperature degradation. Amongst these factors are the grain size [48], the amount of stabilizer [9] and the presence of residual stresses [49]. One consequence of the above mentioned observations was a 1997 posting by the Food and Drug Administration (http://www.fda.gov/cdrh/steamst.html) cautioning against steam sterilization of zirconia femoral heads for total hip pros￾theses and specifying that it could cause phase transformation and roughening of the material, later leading to increased wear on the acetabular component. 8.1. Cubic phase and accelerated aging More recently, in 2001, some series of Y-TZP femoral heads were recalled due to spontaneous fractures. These inci￾dents were traced back to a manufacturing issue that led to accelerated tetragonal to monoclinic transformation in the central area of the femoral heads that likely played a role in the catastrophic failures (http://www.prozyr.com/PAGES UK/ Biomedical/Committee.htm). These events had a noticeable negative impact on the use of zirconia as an implant bioma￾terial, triggering a considerable amount of research work in order to elucidate the possible origin of the failures [50,51]. Although anticipated from the phase diagram established earlier by Scott [52], recent work has demonstrated that both cubic and tetragonal phases co-exist in 3Y-TZP [22,50,53]. Mat￾sui et al. reported that the amount of cubic zirconia in 3Y-TZP sintered at 1300 ◦C was 12.7 mass% and reached 18.6 mass% when the sintering temperature was 1500 ◦C. These values were determined by X-ray diffraction and Rietveld calcula￾tions [22]. In addition, the distribution of the Y3+ ions inside the zirconia grains after sintering at 1300 ◦C was nearly homo￾geneous while Y3+ ions appeared to concentrate within the larger cubic grains after sintering at 1500 ◦C. It was also shown that cubic phase regions started to form from grain bound￾aries and triple junctions where the Y3+ ions had segregated [22]. These results were confirmed by Chevalier et al. [50], further establishing that the enrichment of the cubic grains in Y3+ ions corresponded to a depletion in yttrium in the neigh￾boring tetragonal grains which became less stable and acted as nucleation sites for the (t→m) transformation. An excellent review by Chevalier, evaluating the future of zirconia as a biomaterial has recently been published [51]. With zirconia becoming increasingly popular as a den￾tal restorative material, under various forms and with a wide range of processing conditions (Table 2), it seems wise to keep in mind that some forms of zirconia are susceptible to aging and that processing conditions can play a critical role in the low temperature degradation of zirconia. references [1] Messing GL, Hirano S, Gauckler L. Ceramic processing science. J Am Ceram Soc 2006;89(6):1769–70. [2] Evans AG. Perspectives on the development of high-toughness ceramics. J Am Ceram Soc 1990;73(2):187–206. [3] Subbarao EC. Zirconia-an overview. In: Heuer AH, Hobbs LW, editors. Science and technology of zirconia. Columbus, OH: The American Ceramic Society; 1981. p. 1–24. [4] Goff JP, Hayes W, Hull S, Hutchings MT, Clausen KN. Defect structure of yttria-stabilized zirconia and its influence on the ionic conductivity at elevated temperatures. Phys Rev B 1999;59(22):14202–19. [5] Evans AG, Heuer AH. Review—transformation toughening in ceramics: martensitic transformations in crack-tip stress fields. J Am Ceram Soc 1980;63(5–6):241–8. [6] Garvie RG, Nicholson PS. Structure and thermomechanical properties of partially stabilized zirconia in the CaO–ZrO2 system. J Am Ceram Soc 1972;55(3):152–7. [7] Fabris S, Paxton A, Finnis MW. A stabilization mechanism of zirconia based on oxygen vacancies only. Acta Mater 2002;50:5171–8. [8] Foschini CR, Filho T, Juiz SA, Souza AG, Oliveira JBL, Longo E, et al. On the stabilizing behavior of zirconia: a combined experimental and theoretical study. J Mater Sci 2004;39:1935–41. [9] Hannink RHJ, Kelly PM, Muddle BC. Transformation toughening in zirconia-containing ceramics. J Am Ceram Soc 2000;83(3):461–87. [10] Garvie RC, Hannink RH, Pascoe RT. Ceramic steel? Nature 1975;258:703–4. [11] Heuer AH. Transformation toughening in ZrO2-containing ceramics. J Am Ceram Soc 1987;70(10):689–98. [12] Heuer AH, Claussen N, Kriven WM, Ruhle M. Stability of ¨ tetragonal ZrO2 particles in ceramic matrices. J Am Ceram Soc 1982;65(12):642–50. [13] Porter DL, Heuer AH. Mechanisms of toughening partially stabilized zirconia (PSZ). J Am Ceram Soc 1977;60(3–4):183–4. [14] Montross CS. Comparison of bulk properties of Mg-PSZ with temperature-time contour diagrams. J Am Ceram Soc 1993;76(8):1993–7. [15] Hughan RR, Hannink RHJ. Precipitation during controlled cooling of magnesia-partially stabilized zirconia. J Am Ceram Soc 1986;69(7):556–63. [16] Steffen AA, Dauskardt RH, Ritchie RO. Cyclic fatigue life and crack-growth behavior of microstructurally small cracks in magnesia-partially stabilized zirconia ceramics. J Am Ceram Soc 1991;74(6):1259–68. [17] Hannink RHJ. Microstructural development of the sub-eutectoid aged MgO-ZrO2 alloys. J Mater Sci 1983;18:457–70

DENTAL MATERIALS 24(2008)289-298 [18] Hannink RH], Howar rd C), Kisi EH, Swain MV. Relationship [36 Swain MV, Rose LRF Strength limitations of between fracture toughness and phase assemblage in transformation-toughened zirconia alloys. J Am Ceram Soc Mg-PSZ. J Am Ceram Soc 1994; 77(2: 571-9 1986;69(7):511-8 [19 Sundh A, Sjoegren G Fracture resistance of all-ceramic [37 Dauskardt RH, Marshall DB, Ritchie RO Cyclic fatigue-crack zirconia bridges with differing phase stabilizers and quality propagation in magnesia-partially stabilized zirconia of sintering. Dent Mater 2006; 22: 778-84. ceramics.J Am Ceram Soc 1990; 73(4): 893-903 [20 Gupta TK, Bechtold JH, Kuznicki RC, CadoffLH, Rossing BR. 38 Liu S-Y, Chen I-W Fatigue of yttria-stabilized zirconia: Il, Stabilization of tetragonal phase in polycrystalline zirconia. J crack propagation, fatigue striations, and short-crack Mater Sci1977;12(12):2421 behavior. J Am Ceram Soc 1991; 74(6: 1206-16 [21] Gupta TK, Lange FF, Bechtold JH. Effect of stress-induced 39 Hoffman M, Dauskardt RH, Mai Y-W, Ritchie RO. A review phase transformation on the properties of polycrystalline the mechanics and mechanisms of cyclic fatigue-crack zirconia containing metastable tetragonal phase. J Mater Sci propagation in transformation-toughened zirconia 1978;13:1464-70 ceramics. In: Badwal SPS, Bannister M, Hannink RHJ [22 Matsui K, Horikoshi H, Ohmichi N, Ohgai M, Yoshida H, editors. Science and technology of zirconia V. Lancaster, PA Ikuhara Y. Cubic-formation and grain-growth mechanisms Technomic Publ. Co; 1993. p 321 tetragonal zirconia polycrystal. J Am Ceram So [40 Kobayashi K, Kuwajima H, Masaki T Phase change and 00386(8):1401-8 mechanical properties of ZrO2-Y2O3 solid electrolyte after [23] Bravo-Leon A, Morikawa Y, Kawahara M, Mayo M]. Fracture aging. Solid State Ionics 1981: 3/4: 489-95 toughness of nanocrystalline tetragonal zirconia with low [41 Sato TMS Transformation of yttria-doped tetragonal Zro2 yttria content. Acta Mater 2002: 50: 4555-62. lycrystals by annealing in water. J Am Ceram Soc [24 Kim D-J. Effect of Ta2O5, Nb2 O5, and HfO2 alloying on the 198568(6):356-9 transformability of Y2O3-stabilized tetragonal ZrO2JAm [42] Guo X On the degradation of zirconia ceramics during Ceram soc199073(1):115-20 low-temperature annealing in water or water vapor. J Phys en GSAM, Bouma Js, Winnubst AJA, Burggraff AJ Chem solids199960(4):53946 Mechanical properties of ultra-fine grained zirconia [43 Sato T, Ohtaki S, Shimada M. Transformation of yttria ceramics. J Mater Sci 1992: 27: 4429-38 /-temperature annealing in [26 Becher PF, Swain MV Grain-size-dependent transformation air. J Mater Sci 1985; 20(4): 1466-70 behavior in polycrystalline tetragonal zirconia. J Am Ceram [44 Sato T, Shimada M. Crystalline phase-change in 1992:75(3)493-502 yttria-partially stabilized zirconia by low-temperature [27 Schubert H Anisotropic thermal expansion coefficients of annealing. J Am Ceram Soc 1984; 67(10): C212-3 Y2O3-stabilized tetragonal zirconia. J Am Ceram Soc [45 Chevalier J, Cales B, Drouin JM. Low-temperature aging of 98669(3)270-1 Y-TZP ceramics. J Am Ceram Soc 1999: 82(8): 2150-4 28] Fischer J, Stawarczyk B, Haemmerle CHE Bonding of a [46 Lange FF, Dunlop GL, Davis BI Degradation during aging of Veneering Ceramic to Ce-TZP/Al2O3 Nanocomposite IADR transformation-toughened ZrO2-Y2O3 materials at 250C. J Abstracthttp://iadr.confex.com/iadr/2006brisb/ m Ceram Soc1986;69(3):237 techprogram/abstract.79542htm [47 Cales B, Stefani Y, Lilley E. Long-term in vivo and in vitro 9 Heuer AH, Lange FE, Swain MV, Evans AG. Transformation aging of a zirconia ceramic used in orthopaedy. J Biomed an overview. J Am Ceram Soc 1986 69 Mater Res1994;28:619-24. [30 Marshall DB Strength characteristics of 48 Lawson S Environmental degradation of zirconia ceramics transformation-toughened zirconia. J Am Ceram So Eur Ceram Soc 1995: 15: 485-502. 173-80 [49 Deville S, Chevalier J, Gremillard L Influence of surface finish [31] Lutz EH, Claussen N, Swain MV K-curve behavior of duplex dual stresses on the ageing ceramics.J Am Ceram Soc 1991: 74(1): 11-8. grade zirconia. Biomaterials 2006; 27(10): 2186-5192 e) Swain MV Limitation of maximum strength of 50 Chevalier J, Deville S, Munch E, Jullian R, Lair F Critical effect rconia-toughened ceramics by transf of cubic phase on aging in 3mol% yttria-stabilized zirconia increment. J Am Ceram Soc 1985; 68(4): C-97-9 [33] Huebner H, Jillek W. Sub-critical crack extension and crack 2004;25(24):5539-45 resistance in polycrystalline alumina. J Mater Sci [51 Chevalier J. What future for zirconia as a biomaterial? 1977;12:117-25 Biomaterials 2006: 27(4): 535-43. [34 KnehansR, Steinbrech R, Schaaraechter W Increase of crack [52 Scott HG Phase relationships in the zirconia-yttria system J esistance during slow crack growth in Al2O3 bend Mater Sci1975;10:1527-35 pecimens. J Mater Sci 1983 18: 265-70 [53 Ruiz L, Readey M]. Effect of heat-treatment on grain size, [35] Marshall DB, James MR Reversible stress-induced phase assemblage, and mechanical properties of 3 mol% martensitic transformation in Zro2. J Am Ceram Soc Y-TZP. J Am Ceram Soc 1996: 79(9): 2331-40 1986;69(3)215-7

298 dental materials 24 (2008) 289–298 [18] Hannink RHJ, Howard CJ, Kisi EH, Swain MV. Relationship between fracture toughness and phase assemblage in Mg-PSZ. J Am Ceram Soc 1994;77(2):571–9. [19] Sundh A, Sjoegren G. Fracture resistance of all-ceramic zirconia bridges with differing phase stabilizers and quality of sintering. Dent Mater 2006;22:778–84. [20] Gupta TK, Bechtold JH, Kuznicki RC, Cadoff LH, Rossing BR. Stabilization of tetragonal phase in polycrystalline zirconia. J Mater Sci 1977;12(12):2421–6. [21] Gupta TK, Lange FF, Bechtold JH. Effect of stress-induced phase transformation on the properties of polycrystalline zirconia containing metastable tetragonal phase. J Mater Sci 1978;13:1464–70. [22] Matsui K, Horikoshi H, Ohmichi N, Ohgai M, Yoshida H, Ikuhara Y. Cubic-formation and grain-growth mechanisms in tetragonal zirconia polycrystal. J Am Ceram Soc 2003;86(8):1401–8. [23] Bravo-Leon A, Morikawa Y, Kawahara M, Mayo MJ. Fracture toughness of nanocrystalline tetragonal zirconia with low yttria content. Acta Mater 2002;50:4555–62. [24] Kim D-J. Effect of Ta2O5, Nb2O5, and HfO2 alloying on the transformability of Y2O3-stabilized tetragonal ZrO2. J Am Ceram Soc 1990;73(1):115–20. [25] Theunissen GSAM, Bouma JS, Winnubst AJA, Burggraff AJ. Mechanical properties of ultra-fine grained zirconia ceramics. J Mater Sci 1992;27:4429–38. [26] Becher PF, Swain MV. Grain-size-dependent transformation behavior in polycrystalline tetragonal zirconia. J Am Ceram Soc 1992;75(3):493–502. [27] Schubert H. Anisotropic thermal expansion coefficients of Y2O3-stabilized tetragonal zirconia. J Am Ceram Soc 1986;69(3):270–1. [28] Fischer J, Stawarczyk B, Haemmerle CHF. Bonding of a Veneering Ceramic to Ce-TZP/Al2O3. Nanocomposite. IADR Abstract. http://iadr.confex.com/iadr/2006Brisb/ techprogram/abstract 79542.htm. [29] Heuer AH, Lange FF, Swain MV, Evans AG. Transformation toughening: an overview. J Am Ceram Soc 1986;69(3):i–iv. [30] Marshall DB. Strength characteristics of transformation-toughened zirconia. J Am Ceram Soc 1986;69(3):173–80. [31] Lutz EH, Claussen N, Swain MV. KR-curve behavior of duplex ceramics. J Am Ceram Soc 1991;74(1):11–8. [32] Swain MV. Limitation of maximum strength of zirconia-toughened ceramics by transformation toughening increment. J Am Ceram Soc 1985;68(4):C-97–9. [33] Huebner H, Jillek W. Sub-critical crack extension and crack resistance in polycrystalline alumina. J Mater Sci 1977;12:117–25. [34] Knehans R, Steinbrech R, Schaaraechter W. Increase of crack resistance during slow crack growth in Al2O3 bend specimens. J Mater Sci 1983;18:265–70. [35] Marshall DB, James MR. Reversible stress-induced martensitic transformation in ZrO2. J Am Ceram Soc 1986;69(3):215–7. [36] Swain MV, Rose LRF. Strength limitations of transformation-toughened zirconia alloys. J Am Ceram Soc 1986;69(7):511–8. [37] Dauskardt RH, Marshall DB, Ritchie RO. Cyclic fatigue-crack propagation in magnesia-partially stabilized zirconia ceramics. J Am Ceram Soc 1990;73(4):893–903. [38] Liu S-Y, Chen I-W. Fatigue of yttria-stabilized zirconia: II, crack propagation, fatigue striations, and short-crack behavior. J Am Ceram Soc 1991;74(6):1206–16. [39] Hoffman MJ, Dauskardt RH, Mai Y-W, Ritchie RO. A review of the mechanics and mechanisms of cyclic fatigue-crack propagation in transformation-toughened zirconia ceramics. In: Badwal SPS, Bannister MJ, Hannink RHJ, editors. Science and technology of zirconia V. Lancaster, PA: Technomic Publ. Co; 1993. p. 321–38. [40] Kobayashi K, Kuwajima H, Masaki T. Phase change and mechanical properties of ZrO2-Y2O3 solid electrolyte after aging. Solid State Ionics 1981;3/4:489–95. [41] Sato TMS. Transformation of yttria-doped tetragonal ZrO2 polycrystals by annealing in water. J Am Ceram Soc 1985;68(6):356–9. [42] Guo X. On the degradation of zirconia ceramics during low-temperature annealing in water or water vapor. J Phys Chem Solids 1999;60(4):539–46. [43] Sato T, Ohtaki S, Shimada M. Transformation of yttria partially stabilized zirconia by low-temperature annealing in air. J Mater Sci 1985;20(4):1466–70. [44] Sato T, Shimada M. Crystalline phase-change in yttria-partially stabilized zirconia by low-temperature annealing. J Am Ceram Soc 1984;67(10):C212–3. [45] Chevalier J, Cales B, Drouin JM. Low-temperature aging of ` Y-TZP ceramics. J Am Ceram Soc 1999;82(8):2150–4. [46] Lange FF, Dunlop GL, Davis BI. Degradation during aging of transformation-toughened ZrO2-Y2O3 materials at 250 ◦C. J Am Ceram Soc 1986;69(3):237. [47] Cales B, Stefani Y, Lilley E. Long-term in vivo and in vitro aging of a zirconia ceramic used in orthopaedy. J Biomed Mater Res 1994;28:619–24. [48] Lawson S. Environmental degradation of zirconia ceramics. J Eur Ceram Soc 1995;15:485–502. [49] Deville S, Chevalier J, Gremillard L. Influence of surface finish and residual stresses on the ageing sensitivity of biomedical grade zirconia. Biomaterials 2006;27(10):2186–5192. [50] Chevalier J, Deville S, Munch E, Jullian R, Lair F. Critical effect ¨ of cubic phase on aging in 3mol% yttria-stabilized zirconia ceramics for hip replacement prosthesis. Biomaterials 2004;25(24):5539–45. [51] Chevalier J. What future for zirconia as a biomaterial? Biomaterials 2006;27(4):535–43. [52] Scott HG. Phase relationships in the zirconia-yttria system. J Mater Sci 1975;10:1527–35. [53] Ruiz L, Readey MJ. Effect of heat-treatment on grain size, phase assemblage, and mechanical properties of 3 mol% Y-TZP. J Am Ceram Soc 1996;79(9):2331–40