Availableonlineatwww.sciencedirect.com Science Direct E噩≈RS ELSEVIER Joumal of the European Ceramic Society 28(2008)2301-2308 www.elsevier.comlocate/jeurceramsoc Microstructures, crystallography of interfaces, and creep behavior of melt-growth composites L Mazerolles a,*,L. Perriere a, b,S. Lartigue-Korineka, N. Piquet a, b, M. Parlier b CECM, UPR 2801, CNRS, 15 rue Georges Urbain, F-94407 Vitry sur Seine, France b ONERADMSC 29 avenue de la Division Leclerc. F-92322 chatillon Cedex franc Available online 4 March 2008 Oxide eutectic ceramics were prepared from Al2O3 and Ln]O3-based systems by unidirectional solidification from the melt. The microstructure consists of two single-crystal phases continuously entangled in a three-dimensional interpenetrating network without grain boundaries, pores or colonies. The outstanding stability of these microstructures gives rise to a high strength and creep resistance at high temperature. Preferred growth directions, orientation relationships between phases and single-crystal homogeneity of specimens were revealed. Creep behavior at high temperatur has been studied, mechanisms of deformation by dislocation motion and twinning were revealed from Transmission Electron Microscopy (TEM) observations. Extension to ternary eutectics with a three-dimensional microstructure consisting in the addition of a toughening phase(zrO2)to the previous binary eutectics has been investigated By using this method, significant improvement of fracture toughness was obtained. C 2008 Elsevier Ltd. all rights reserved. Keywords: Oxides; Ceramic eutectics; Microstructure; Creep; Dislocations Introduction their melting point, as compared with conventional composites and monolithic ceramics. , Furthermore oxide-based materi In the field of structural materials, eutectic ceramic oxides als are very attractive because of their inherent thermochemical prepared by solidification from the melt appear as potential can- stability in oxidizing environments at high temperature. More didates in the future for thermomechanical applications at very recently, Waku et al. have developed binary eutectics, called high temperature. Indeed, challenges related to future energy melt-growth composites(MGC), with novel microstructures in equirements impose the need to develop novel ultra-high- which continuous networks of single-crystal Al2O3 phases and temperature structural materials which display good mechanical single-crystal oxide compounds interpenetrate without grain properties(tensile strength, creep resistance, fracture tough- boundaries. These composites present a flexural strength con- ness)at temperatures above 1500C. For example in aircraft stant from room temperature up to high temperatures and a ge engines, the use of Ni-based single-crystal cast superalloys creep resistance allowing to consider applications in gas turbine for turbine blades is only possible at temperatures lower than and power generation systems with non-cooled turbine blades 100-1150oC. Silicon carbide-based composites are not sta- at very high temperatures. -S In this paper, we will present ble enough in an oxidizing atmosphere when temperature is results concerning similar microstructures obtained by direc- higher than 1300C and, finally, ceramic oxides, usually pre- tional solidification in various Al2O3 and Ln2O3-based system pared by sintering, have a too high brittleness due to the grain Morphology of microstructures, crystallography of constituent boundaries and the amorphous phases observed at grain bound- phases and interfaces, and single-crystal homogeneity of grown aries. Early studies on some oxide-oxide systems(such as samples will be reported. Creep behavior at high tempera- Al2O3-ZrO2)demonstrate the outstanding mechanical proper- ture has been studied. Factors controlling the deformation ties and the thermal and microstructural stability of directionally mechanisms will be analyzed taking into account microstruc- solidified eutectic ceramic oxides up to temperatures close to tural characteristics and Transmission Electron Microscopy (TEM)observations performed on deformed specimen. Finally, first results relative to the extension to ternary systems that display a significant increase of fracture toughness will be pre- E-mail address: mazerolles@glvt-cnrs fr(L Mazerolles) sented 0955-2219/S-see front matter o 2008 Elsevier Ltd. All rights reserved. doi: 10.1016/j-jeurceramsoc 2008.01.014
Available online at www.sciencedirect.com Journal of the European Ceramic Society 28 (2008) 2301–2308 Microstructures, crystallography of interfaces, and creep behavior of melt-growth composites L. Mazerolles a,∗, L. Perriere a,b, S. Lartigue-Korinek a, N. Piquet a,b, M. Parlier b a CECM, UPR 2801, CNRS, 15 rue Georges Urbain, F-94407 Vitry sur Seine, France b ONERA/DMSC, 29 avenue de la Division Leclerc, F-92322 Ch ˆatillon Cedex, France Available online 4 March 2008 Abstract Oxide eutectic ceramics were prepared from Al2O3 and Ln2O3-based systems by unidirectional solidification from the melt. The microstructure consists of two single-crystal phases continuously entangled in a three-dimensional interpenetrating network without grain boundaries, pores or colonies. The outstanding stability of these microstructures gives rise to a high strength and creep resistance at high temperature. Preferred growth directions, orientation relationships between phases and single-crystal homogeneity of specimens were revealed. Creep behavior at high temperature has been studied, mechanisms of deformation by dislocation motion and twinning were revealed from Transmission Electron Microscopy (TEM) observations. Extension to ternary eutectics with a three-dimensional microstructure consisting in the addition of a toughening phase (ZrO2) to the previous binary eutectics has been investigated. By using this method, significant improvement of fracture toughness was obtained. © 2008 Elsevier Ltd. All rights reserved. Keywords: Oxides; Ceramic eutectics; Microstructure; Creep; Dislocations 1. Introduction In the field of structural materials, eutectic ceramic oxides prepared by solidification from the melt appear as potential candidates in the future for thermomechanical applications at very high temperature. Indeed, challenges related to future energy requirements impose the need to develop novel ultra-hightemperature structural materials which display good mechanical properties (tensile strength, creep resistance, fracture toughness) at temperatures above 1500 ◦C. For example in aircraft engines, the use of Ni-based single-crystal cast superalloys for turbine blades is only possible at temperatures lower than 1100–1150 ◦C. Silicon carbide-based composites are not stable enough in an oxidizing atmosphere when temperature is higher than 1300 ◦C and, finally, ceramic oxides, usually prepared by sintering, have a too high brittleness due to the grain boundaries and the amorphous phases observed at grain boundaries. Early studies on some oxide–oxide systems (such as Al2O3–ZrO2) demonstrate the outstanding mechanical properties and the thermal and microstructural stability of directionally solidified eutectic ceramic oxides up to temperatures close to ∗ Corresponding author. E-mail address: mazerolles@glvt-cnrs.fr (L. Mazerolles). their melting point, as compared with conventional composites and monolithic ceramics.1,2 Furthermore, oxide-based materials are very attractive because of their inherent thermochemical stability in oxidizing environments at high temperature. More recently, Waku et al. have developed binary eutectics, called melt-growth composites (MGC), with novel microstructures in which continuous networks of single-crystal Al2O3 phases and single-crystal oxide compounds interpenetrate without grain boundaries. These composites present a flexural strength constant from room temperature up to high temperatures and a good creep resistance allowing to consider applications in gas turbine and power generation systems with non-cooled turbine blades at very high temperatures.3–5 In this paper, we will present results concerning similar microstructures obtained by directional solidification in various Al2O3 and Ln2O3-based systems. Morphology of microstructures, crystallography of constituent phases and interfaces, and single-crystal homogeneity of grown samples will be reported. Creep behavior at high temperature has been studied. Factors controlling the deformation mechanisms will be analyzed taking into account microstructural characteristics and Transmission Electron Microscopy (TEM) observations performed on deformed specimen. Finally, first results relative to the extension to ternary systems that display a significant increase of fracture toughness will be presented. 0955-2219/$ – see front matter © 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jeurceramsoc.2008.01.014
L Mazerolles et al. /Joumal of the European Ceramic Society 28(2008)2301-2308 Table I Chemical compositions and constituent phases of studied eutectics Composition( mol%) Eutectic phases Structure of phases 77Al2O3-23Gd2O3 2O3-19Er2O3 Al2O3-Yb3 AlsAlO12 Corundum(hexagonal) 77Al2O3-23Y2O3 Al2O3-Y3Als AlO12 +Garmet(cubic) 2. Experimental procedures Fabrication from the melt microstructures requires planar growth front conditions in order to keep flat solid-liquid interfaces during growth at the macro- scopic and microscopic level. These process conditions are obtained with equipments displaying high-thermal gradients in the solidification direction. Our experiments were carried out with a floating-zone translation device consisting in an arc image furnace operating with a 6-kW xenon lamp as a radiation source cation was driven in air at a constant speed ranging from 2 to 20 mmh. Cylindrical specimens of about 8 mm in diame ter,and 50 mm in length were grown by this method. hism elt- 10 m powders(99.99%)were used for starting materials. Before melt ing, these powders were mixed and molded in green cylinder Fig. 2. Al2O3-Eu2O3 eutectic: SEM micrograph of longitudinal section cold isostatic pressure then, consolidated by sintering at 1300°Cfor12h. major axis being parallel to the solidification direction Microstructural observations and chemical analyses of eutec- Constant-load compressive creep tests were conducted in air tic phases were performed by Scanning Electron Microscopy environment at 1450 and 1600C within the stress range from using a Leo 1530 ( Leo, Germany)equipped with a Princeton 70 to 200 MPa. Compression rams were made from sintered Gamma Tech(USA) EDX spectrometer. They were carried out alumina with sapphire rods for applying loads on the specimen on sections of rods parallel and perpendicular to the growth direction. The crystalline homogeneity of specimen was con- 3. Results and discussion trolled at the macroscopic scale(1-2 mm) by using the Electron ack Scattering Diffraction(EBSD)technique on a Zeiss DSM 3.1. Microstructure 950 microscope equipped with a TSL detector. Maps of the crys- tallographic orientations were obtained by this method. Growth The studied eutectic composites were prepared in directions, orientation relationships between eutectic phases and Al2O3-Ln2O3 systems. Phase diagrams of these systems tron Microscopy. These observations were performed on thinned at temperatures close to 1800C. Eutectic phases conss e structure of interfaces were investigated by Transmission Elec- all display an eutectic composition on the rich alumina si foils of transverse sections using either a conventional micro- an a-Al2O phase(corundum structure)associated to either scope (EOL 2000EX)or a high-resolution electron microscope a perovskite-type phase LnAIO3(Ln=Sm, Eu, and Gd)or (Topcon 002B) both operating at 200k a garnet-type phase Ln AlsO12 for other elements of the From cylinder bars, compressive creep specimens were lanthanides series Data about the studied eutectic compositions machined with dimensions of 3 mm x mm x 5.5 mm, their are reported in Table 1 10 um μm Fig. 1. SEM micrographs of directionally solidified eutectic cross-sections showing the interlocking microstructure(a)Al2O3-Gd2O3: (b)Al2O3-Eu2O3;(c) Al2O3-Er2O3; (d)Al2O3-Y2O3). The white regions are perovskite (a and b)or garnet(c and d) and the black regions are alumina
2302 L. Mazerolles et al. / Journal of the European Ceramic Society 28 (2008) 2301–2308 Table 1 Chemical compositions and constituent phases of studied eutectics Composition (mol%) Eutectic phases Structure of phases 77Al2O3–23Gd2O3 Al2O3–GdAlO3 Corundum (hexagonal) 76Al2O3–24Eu2O3 Al2O3–EuAlO3 +Perovskite (orthorhombic) 81Al2O3–19Er2O3 Al2O3–Yb3Al5AlO12 Corundum (hexagonal) 77Al2O3–23Y2O3 Al2O3–Y3Al5AlO12 +Garnet (cubic) 2. Experimental procedures Fabrication from the melt of homogeneous eutectic microstructures requires planar growth front conditions in order to keep flat solid–liquid interfaces during growth at the macroscopic and microscopic level. These process conditions are obtained with equipments displaying high-thermal gradients in the solidification direction. Our experiments were carried out with a floating-zone translation device consisting in an arc image furnace operating with a 6-kW xenon lamp as a radiation source. Solidification was driven in air at a constant speed ranging from 2 to 20 mm h−1. Cylindrical specimens of about 8 mm in diameter, and 50 mm in length were grown by this method. High-purity powders (99.99%) were used for starting materials. Before melting, these powders were mixed and molded in green cylinder bars by cold isostatic pressure then, consolidated by sintering at 1300 ◦C for 12 h. Microstructural observations and chemical analyses of eutectic phases were performed by Scanning Electron Microscopy using a Leo 1530 (Leo, Germany) equipped with a Princeton Gamma Tech (USA) EDX spectrometer. They were carried out on sections of rods parallel and perpendicular to the growth direction. The crystalline homogeneity of specimen was controlled at the macroscopic scale (1–2 mm2) by using the Electron Back Scattering Diffraction (EBSD) technique on a Zeiss DSM 950 microscope equipped with a TSL detector. Maps of the crystallographic orientations were obtained by this method. Growth directions, orientation relationships between eutectic phases and structure of interfaces were investigated by Transmission Electron Microscopy. These observations were performed on thinned foils of transverse sections using either a conventional microscope (JEOL 2000EX) or a high-resolution electron microscope (Topcon 002B) both operating at 200 kV. From cylinder bars, compressive creep specimens were machined with dimensions of 3 mm × 3 mm × 5.5 mm, their Fig. 2. Al2O3–Eu2O3 eutectic: SEM micrograph of longitudinal section. major axis being parallel to the solidification direction. Constant-load compressive creep tests were conducted in air environment at 1450 and 1600 ◦C within the stress range from 70 to 200 MPa. Compression rams were made from sintered alumina with sapphire rods for applying loads on the specimen. 3. Results and discussion 3.1. Microstructure The studied eutectic composites were prepared in Al2O3–Ln2O3 systems. Phase diagrams of these systems all display an eutectic composition on the rich alumina side at temperatures close to 1800 ◦C. Eutectic phases consist in an -Al2O3 phase (corundum structure) associated to either a perovskite-type phase LnAlO3 (Ln = Sm, Eu, and Gd) or a garnet-type phase Ln3Al5O12 for other elements of the lanthanides series. Data about the studied eutectic compositions are reported in Table 1. Fig. 1. SEM micrographs of directionally solidified eutectic cross-sections showing the interlocking microstructure ((a) Al2O3–Gd2O3; (b) Al2O3–Eu2O3; (c) Al2O3–Er2O3; (d) Al2O3–Y2O3). The white regions are perovskite (a and b) or garnet (c and d) and the black regions are alumina.
L Mazerolles et al. Journal of the European Ceramic Society 28(2008)2301-2308 S mm/h 30mm熙8 S20 Fig. 3. Al2O3-Y2O3 eutectic: Morphology of microstructure vs. the solidification rate The eutectic microstructures grown in various composites are by using an arc image furnace). When this rate increases the shown in Fig. 1. These SEM micrographs of sections perpen- eutectic growth undergoes a transition from the planar to the dicular to the growth direction, reveal, in every case, continuous cellular regime that does not correspond anymore to a couple networks of an alumina phase(dark contrast)and lanthanide growth. Fig. 3 reveals that the three-dimensional microstruc and aluminum oxide compounds ((a) GAP=GdAlO3;(b) ture of the Al2O3-YAG eutectic, is not immediately modified AP=EuAIO3;(c)EAG= Er3AlsO12: (d) YAG=Y3Al5O12). when the growth rate increases and persists up to rates close to The observations performed on sections parallel (Fig. 2)to 30 mmh- with similar solidification conditions(thermal gra- the growth direction show similar morphologies indicating the dient, diameter of specimen) before entering into the cellular three-dimensional configuration of the microstructure. The two growth regime(Fig. 3). Similar results were obtained with the phases are continuously entangled in a three-dimensional inter- other eutectic compositions penetrating network without grain boundaries, pores or colonies The average size of each phase does not vary with the added 3. 2. Crystallography and interfaces rare-earth oxide except for the Al2O3-YAG composite( Fig. ld) ndeed, at this eutectic composition, prepared at a growth rate Aligned eutectic microstructures(lamellae, fibers or disper (5 mmh- )similar to other composites, a microstructure with soids), grown by unidirectional solidification, usually consist of harp angle facets and dimensions about five times larger than single-crystal phases growing preferentially along well-defined that of other eutectics has been observed. This difference could crystallographic directions. These directions are not necessarily be related to the diffusion coefficient values of yttrium higher the directions of easy-growth of the components but often cor- than that of lanthanide elements. However, for the Al2O3-YAG respond to minimum interfacial energy configurations between eutectic, a microstructure displaying dimensions very similar phases. These perfectly aligned lattices are related by orienta much higher solidification ales o d ut modifying the Chinese tion relationships which are unique in most systems and produce to other AlO3-Ln2 O3 eutectics with script morphology has been prod for growth conditions with well-defined interface planes corresponding to dense atomic arrangements in the component phases. 8- Contrarily to these In most directionally solidified oxide eutectics, coupled aligned microstructures, interconnected microstructures, shown growth is mainly controlled by the growth rate. For example, in in the previous figures, display a very isotropic morphology the case of the Al2O3-ZrO2(Y2O3)eutectic, the planar growth However, electron diffraction studies performed on thin plates matrix,only exists at very low solidification rates(<5mmh-I cut perpendicularly to the rod axes reveal growth directions also regime, leading to zirconia fibers embedded in an alumina corresponding to well-defined crystallographic directions 121 Fig 4. Electron diffraction patterns performed at the Al2O3-YAG (a)and Al2O3-EuAlO(b)interfaces
L. Mazerolles et al. / Journal of the European Ceramic Society 28 (2008) 2301–2308 2303 Fig. 3. Al2O3–Y2O3 eutectic: Morphology of microstructure vs. the solidification rate. The eutectic microstructures grown in various composites are shown in Fig. 1. These SEM micrographs of sections perpendicular to the growth direction, reveal, in every case, continuous networks of an alumina phase (dark contrast) and lanthanide and aluminum oxide compounds ((a) GAP = GdAlO3; (b) EAP = EuAlO3; (c) EAG = Er3Al5O12; (d) YAG = Y3Al5O12). The observations performed on sections parallel (Fig. 2) to the growth direction show similar morphologies indicating the three-dimensional configuration of the microstructure. The two phases are continuously entangled in a three-dimensional interpenetrating network without grain boundaries, pores or colonies. The average size of each phase does not vary with the added rare-earth oxide except for the Al2O3–YAG composite (Fig. 1d). Indeed, at this eutectic composition, prepared at a growth rate (5 mm h−1) similar to other composites, a microstructure with sharp angle facets and dimensions about five times larger than that of other eutectics has been observed. This difference could be related to the diffusion coefficient values of yttrium higher than that of lanthanide elements. However, for the Al2O3–YAG eutectic, a microstructure displaying dimensions very similar to other Al2O3–Ln2O3 eutectics without modifying the Chinese script morphology has been produced for growth conditions with much higher solidification rates.6 In most directionally solidified oxide eutectics, coupled growth is mainly controlled by the growth rate. For example, in the case of the Al2O3–ZrO2(Y2O3) eutectic, the planar growth regime, leading to zirconia fibers embedded in an alumina matrix, only exists at very low solidification rates (<5 mm h−1 by using an arc image furnace). When this rate increases the eutectic growth undergoes a transition from the planar to the cellular regime that does not correspond anymore to a coupled growth.7 Fig. 3 reveals that the three-dimensional microstructure of the Al2O3–YAG eutectic, is not immediately modified when the growth rate increases and persists up to rates close to 30 mm h−1 with similar solidification conditions (thermal gradient, diameter of specimen) before entering into the cellular growth regime (Fig. 3). Similar results were obtained with the other eutectic compositions. 3.2. Crystallography and interfaces Aligned eutectic microstructures (lamellae, fibers or dispersoids), grown by unidirectional solidification, usually consist of single-crystal phases growing preferentially along well-defined crystallographic directions. These directions are not necessarily the directions of easy-growth of the components but often correspond to minimum interfacial energy configurations between phases. These perfectly aligned lattices are related by orientation relationships which are unique in most systems and produce well-defined interface planes corresponding to dense atomic arrangements in the component phases.8–11 Contrarily to these aligned microstructures, interconnected microstructures, shown in the previous figures, display a very isotropic morphology. However, electron diffraction studies performed on thin plates cut perpendicularly to the rod axes reveal growth directions also corresponding to well-defined crystallographic directions. Fig. 4. Electron diffraction patterns performed at the Al2O3–YAG (a) and Al2O3–EuAlO3 (b) interfaces
L Mazerolles et al. /Joumal of the European Ceramic Society 28(2008)2301-2308 th directions and orientation relationships of eutectic phases in the AlzO3-Ln2O3 eutectics Eutectic phases Growth directions Orientation relationships [1010]A1203//[1 10]perovskite (1 120)Al203//(00 1)perovskit (Ln= Gd, Eu) or[11 20Al203//100 1]perovskite (0001)Al2O3/(1 00)perovskit Al2O3-Ln3 Als O12 [10I0Al2O3 (0001)Al2O3//(12 1)garnet _n=Er, Dy, Yb, Y) (2 10)or(1 10)gamet In these eutectic composites, the most frequently observed The minimal energy configuration at the interfaces is well growth direction for Al2O3 was [1010]. Sometimes, the illustrated by the HRTEM image(Fig. 5a)of the inter- [1 120] direction was also observed but, in all cases, the basal face between the corundum and perovskite structures for the plane of the corundum structure is always parallel to the solid- Al2O3-EuAlO3 eutectic. This image corresponds to a eutectic ification axis whatever the considered eutectic system. The structure grown along the [1 120]alumina and [00 l]perovskite electron diffraction patterns shown in Fig. 4 were performed directions. No intermediary phase is detected at the interface and on platelets cut perpendicularly to the growth directions of the transition on both sides of the interface operates on one or A12O3-YAG and Al2O3-GAP eutectics. The selected area aper- two atomic planes. Bragg filtering in the reciprocal space, from ture is centered on the interface, and consequently diffraction the numerical Fourier Transform of the digitized image, reveals spots of both phases are superimposed on the same pattern Crys- that accommodation between the two structures is restored by a llographic principal directions are strictly aligned according to periodic array of dislocations along the interface(Fig 5b). Sim- the following relations ilar results were also obtained with eutectics consisting of garnet and alumina phases. These TEM observations did not reveal any (0001)A2O3//(121)Y3Al5O12 stress fields at the interfaces, and are in good agreement with low and:(1120)Al2O3//(001GdAO3 values of residual stresses measured at room temperature from X-ray diffraction ex riments or from spectroscopic studies through the shift of fluorescence lines of Cr in sapphire. 4 Other growth directions of the garnet phase were also observed((11 0)in major cases) but the orientation relationship 3.3. Compressive creep deformation between the two phases persisted In the case of eutectics associ- ating perovskite and corundum structures, two sets of orientation Fig. 6 shows a typical strain vs. time relationship for the relations were determined. These results are summarized in Al2O3-YAG eutectic in a compressive test at 1450C with Table 2 stresses ranging from 70 to 200 MPa. The creep deformation EBSD studies have shown that these preferred growth direc- curves show a primary creep regime where the deformation rate tions and orientation relationships are retained on a large central decreases continuously. After this short primary stage(. g strain part of the specimenand confirmed the single-crystal quality of about 0.5%), when the secondary creep rate is reached, the applied load is modified and primary and secondary creep rates 102104 b 5.(a)HRTEM image of the AlzO3-EuAlO3 interface(b)Inverse Fourier Transform from the digitized image of the (a) interface built with 0003A10, and
2304 L. Mazerolles et al. / Journal of the European Ceramic Society 28 (2008) 2301–2308 Table 2 Growth directions and orientation relationships of eutectic phases in the Al2O3–Ln2O3 eutectics Eutectic phases Growth directions Orientation relationships Al2O3–LnAlO3 [1 0 1 0]Al ¯ 2O3//[1 1 0]perovskite (1 1 2 0) Al ¯ 2O3//(0 0 1) perovskite (Ln = Gd, Eu) or [1 1 2 0]Al ¯ 2O3//[0 0 1]perovskite (0 0 0 1) Al2O3//(1 0 0) perovskite Al2O3–Ln3Al5O12 [1 0 1 0]Al ¯ 2O3 (0 0 0 1)Al2O3//(1 2 1)garnet (Ln = Er, Dy, Yb, Y)) 2 1 0¯ or110garnet In these eutectic composites, the most frequently observed growth direction for Al2O3 was [1 0 1 0]. Sometimes, the ¯ [1 1 2 0] direction was also observed but, in all cases, the basal ¯ plane of the corundum structure is always parallel to the solidification axis whatever the considered eutectic system. The electron diffraction patterns shown in Fig. 4 were performed on platelets cut perpendicularly to the growth directions of Al2O3–YAG and Al2O3–GAP eutectics. The selected area aperture is centered on the interface, and consequently diffraction spots of both phases are superimposed on the same pattern. Crystallographic principal directions are strictly aligned according to the following relations: (0 0 0 1)Al2O3//(1 2 1)Y3Al5O12 and : (1 1 2 0)Al ¯ 2O3//(0 0 1)GdAlO3 Other growth directions of the garnet phase were also observed (110 in major cases) but the orientation relationship between the two phases persisted. In the case of eutectics associating perovskite and corundum structures, two sets of orientation relations were determined. These results are summarized in Table 2. EBSD studies have shown that these preferred growth directions and orientation relationships are retained on a large central part of the specimen12 and confirmed the single-crystal quality of grown samples. The minimal energy configuration at the interfaces is well illustrated by the HRTEM image (Fig. 5a) of the interface between the corundum and perovskite structures for the Al2O3–EuAlO3 eutectic. This image corresponds to a eutectic structure grown along the [1 1 2 0] alumina and [0 0 1] perovskite ¯ directions. No intermediary phase is detected at the interface and the transition on both sides of the interface operates on one or two atomic planes. Bragg filtering in the reciprocal space, from the numerical Fourier Transform of the digitized image, reveals that accommodation between the two structures is restored by a periodic array of dislocations along the interface (Fig. 5b). Similar results were also obtained with eutectics consisting of garnet and alumina phases. These TEM observations did not reveal any stress fields at the interfaces, and are in good agreement with low values of residual stresses measured at room temperature from X-ray diffraction experiments13 or from spectroscopic studies through the shift of fluorescence lines of Cr in sapphire.14 3.3. Compressive creep deformation Fig. 6 shows a typical strain vs. time relationship for the Al2O3–YAG eutectic in a compressive test at 1450 ◦C with stresses ranging from 70 to 200 MPa. The creep deformation curves show a primary creep regime where the deformation rate decreases continuously. After this short primary stage (e.g. strain of about 0.5%), when the secondary creep rate is reached, the applied load is modified and primary and secondary creep rates Fig. 5. (a) HRTEM image of the Al2O3–EuAlO3 interface. (b) Inverse Fourier Transform from the digitized image of the (a) interface built with 0003Al2O3 and 200EuAlO3 Bragg spots
L Mazerolles et al. Journal of the European Ceramic Society 28(2008)2301-2308 100 MPa 8E 06 △Al2O3 A2O3·GAF 0 104 Applied Stress(MPa) Time(s) Fig. 8. Plot of the creep rates vs. the applied stress for various eutectics. Fig. 6. Al2O3-YAG eutectic: Creep curve at 1450C with load increments and decrements(70, 100, 140, 200 then 140 and 100 MPa) low stresses(<140 MPa) and even higher at high stresses. Thi behavior is especially noteworthy as the Al2O3-ZrO2 eutec- e is very sensitive to the applied stress o following a power-law Al2O3-Ln2O3 eutectics (1710<Tm<1827C. Consequently, relationship: for the same T/Tm, these former display a better creep resistance. The incremental application of the load during one single E=Ao exp RT periment allows to determine readily the n exponent at each stress step by extrapolation of the minimum creep rate values where A is a material constant, n is the stress exponent, Q is for a given strain value. Eventual changes of the n value for var- the activation energy for creep, R is the gas constant and T ious applied stresses reveal different creep mechanisms. From the absolute temperature. The quasi-steady-state regime in thi the results summarized in Table 3 we can see that the work was determined by plots of the strain rate as a function of is higher than 2 from a 100-140 MPa step suggesting a defor- the true strain(Fig. 7). Fig. 8 shows the quasi-steady-state rates mation mechanism controlled by a dislocation motion. Values for the eutectics and regime stress tested. These plots reveal the close to 4 have also been measured for creep tests performed at high creep resistance of these interlocked microstructures with 1600 C. These high stress components are incompatible with strain rates very similar to that of Al2O3-ZrO2 eutectics5, 16 the interpretation of plasticity controlled by pure diffusion. In the case of deformation due to pure lattice diffusion or grain boundary diffusion the resulting creep rate is linearly prop tional to the stress and n=1. Moreover, hrtEM images of interfaces( Fig. 5)indicate a considerable coherence and strong bonding between phases at the interfaces and consequently boundary sliding mechanisms are impossible. TEM studies were performed on these specimens deformed at 1600C. Images presented in Fig. 9 corresponding to the AlzO3-GAP eutectic, clearly show dislocations in the two eutectic phases. In alumina, the dislocations observed in Fig. 9a are basal type dislocations(b= 1 /3(21)aligned in parallel basal slip planes(0001). The basal plane is seen edge-on in this orienta tion. The basal slip system has the lowest critical resolved shear Table 3 Values of the stress exponent as a function of the stress increment for vari AlO3-Lng O3 eutectics deformed at 1450oC 02040608 70→100MPa 100→140MPa 140→200MPa True strain(%) Al2O3-EAG AlO3-YAG 1.13 2.06 -state regime In various eutectics
L. Mazerolles et al. / Journal of the European Ceramic Society 28 (2008) 2301–2308 2305 Fig. 6. Al2O3–YAG eutectic: Creep curve at 1450 ◦C with load increments and decrements (70, 100, 140, 200 then 140 and 100 MPa). are again measured at a higher stress. The secondary creep rate ε˙ is very sensitive to the applied stress σ following a power–law relationship: ε˙ = Aσn exp − Q RT where A is a material constant, n is the stress exponent, Q is the activation energy for creep, R is the gas constant and T is the absolute temperature. The quasi-steady-state regime in this work was determined by plots of the strain rate as a function of the true strain (Fig. 7). Fig. 8 shows the quasi-steady-state rates for the eutectics and regime stress tested. These plots reveal the high creep resistance of these interlocked microstructures with strain rates very similar to that of Al2O3–ZrO2 eutectics15,16 at Fig. 7. Strain rates as a function of true strain allowing to determine the minimum creep rates corresponding to the quasi-steady-state regime in various eutectics. Fig. 8. Plot of the creep rates vs. the applied stress for various eutectics. low stresses (<140 MPa) and even higher at high stresses. This behavior is especially noteworthy as the Al2O3–ZrO2 eutectic has melting temperature (Tm = 1910 ◦C) higher than that of Al2O3–Ln2O3 eutectics (1710 < Tm < 1827 ◦C). Consequently, for the same T/Tm, these former display a better creep resistance. The incremental application of the load during one single experiment allows to determine readily the n exponent at each stress step by extrapolation of the minimum creep rate values for a given strain value. Eventual changes of the n value for various applied stresses reveal different creep mechanisms. From the results summarized in Table 3 we can see that the n value is higher than 2 from a 100–140 MPa step suggesting a deformation mechanism controlled by a dislocation motion. Values close to 4 have also been measured for creep tests performed at 1600 ◦C. These high stress components are incompatible with the interpretation of plasticity controlled by pure diffusion. In the case of deformation due to pure lattice diffusion or grain boundary diffusion the resulting creep rate is linearly proportional to the stress and n = 1.17 Moreover, HRTEM images of interfaces (Fig. 5) indicate a considerable coherence and strong bonding between phases at the interfaces and consequently boundary sliding mechanisms are impossible. TEM studies were performed on these specimens deformed at 1600 ◦C. Images presented in Fig. 9 corresponding to the Al2O3–GAP eutectic, clearly show dislocations in the two eutectic phases. In alumina, the dislocations observed in Fig. 9a are basaltype dislocations (b = 1/3[2 1¯ 1 0]) aligned in parallel basal slip ¯ planes (0 0 0 1). The basal plane is seen edge-on in this orientation. The basal slip system has the lowest critical resolved shear Table 3 Values of the stress exponent as a function of the stress increment for various Al2O3–Ln2O3 eutectics deformed at 1450 ◦C 70→100 MPa 100→140 MPa 140→200 MPa Al2O3–GAP 1.10 2.08 2.60 Al2O3–EAG 1.20 2.10 2.72 Al2O3–YAG 1.13 2.06 2.99