Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.issp.ac.ru/lsc/files/strength_1999.pdf Äàòà èçìåíåíèÿ: Thu Feb 9 10:14:58 2012 Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 07:23:49 2012 Êîäèðîâêà:

Journal of Crystal Growth 198/199 (1999) 227 --231

Effect of growth conditions on the strength of shaped sapphire
P.A. Gurjiyants , M.Yu. Starostin , V.N. Kurlov *, F. Theodore , J. Delepine
Institute of Solid State Physics of RAS, Chernogolovka, Moscow District 142432, Russia CEA Grenoble, DTA/DEM/SPCM, 17 rue des Martyrs, 38054 Grenoble Cedex 9, France

Abstract The strengths of sapphire crystals grown both by Stepanov/EFG and GES (growth from an element of shape) methods were compared. Tests were carried out in four-point loading between ambient and 1500°C. Crystals of various shape and of different orientations were investigated. 1999 Published by Elsevier Science B.V. All rights reserved. PACS: 62.20.Mk; 81.05.!t; 81.10.!h; 81.10.Fq Keywords: Shaped crystal growth; Sapphire; Single crystals; Strength; Stepanov/EFG; GES method

1. Introduction Recently sapphire single crystals gained a wider application domain due to some growth methods allowing preparation of practically finished products during the growth. The potential of sapphire crystal as structure material is due to its high elastic modulus and tensile strength, chemical stability, oxidation resistance, and optical transparency. These properties are valid over a wide range of temperatures. The strength of sapphire fibers, disks, and single crystals of various orientations [1--5] have been investigated previously as a function of temperature, but adequate data for shaped crystals is lacking.

In this paper we report the strength of sapphire shaped crystals grown by EFG and GES (growth from an element of shape) methods as a function of temperature, orientation and annealing.

2. Experimental procedure The sapphire shaped crystals for mechanical tests were grown in a 8 kHz induction heated graphite susceptor/molybdenum crucible setup held within a growth chamber. The dies were molybdenum in all experiments. The feed material was crushed Verneuil crystals. The atmosphere was high purity argon. Sapphire rods of 2.5;2.5 mm cross section and 250--300 mm length were grown from the melt by the EFG technique. This technique was described in detail in Ref. [6]. Pulling rate was 1 mm/min. Experimental runs used c-axis (called a 0°) and

* Corresponding author. Fax: #33 7 096 576 4111; e-mail: kurlov@issp.ac.ru.

0022-0248/99/$ -- see front matter 1999 Published by Elsevier Science B.V. All rights reserved. PII: S 0 0 2 2 - 0 2 4 8 ( 9 8 ) 0 1 1 3 1 - 2


228

P.A. Gurjiyants et al. / Journal of Crystal Growth 198/199 (1999) 227 ­ 231

perpendicular to c-axis (called a 90°) sapphire seeds to initiate growth. Tubes of 30 mm length with external diameter 23 mm and wall thickness 3.5 mm were grown by the GES method. The approach of the GES method consists of pulling a shaped crystal from a melt meniscus which is only a small element of the whole transverse cross section of the growing crystal [7]. The die is placed at some distance from the rotation axis, and it predetermines only the thickness of the tube wall. The tubular crystal grows layer by layer. The thickness of one layer was determined by »/ ratio, where » is the pulling rate and is the frequency of rotation. The pulling rate was varied from 0.07 to 0.1 mm/min, the frequency of rotation was varied from 0.75 to 20 rpm. Seed crystals were tubes grown in the [0 0 0 1] direction. The EFG rods were cut into 25 mm lengths, the GES tubes were cut and polished to bars with dimensions 2.5;2.5;25 mm. The crystals were annealed in a vacuum (4 h at 1900°C). In order to compare the mechanical properties of annealed and as-grown samples, some EFG rods were not annealed. The strengths of sapphire EFG crystals of two different crystallographic orientations and GES bars were measured in four-point loading between ambient and 1550°C. Mechanical testing was conducted with an Instron Model 1195 (at 20°C; 600°C; 900°C) and a Nikimp (at 1550°C) testing machine at a crosshead speed of 0.05--2 mm/min. Elastic modulus was measured by composite oscillator method (Marx) at 10 kHz. Microhardness was measured for basal and 90° planes of annealed and as grown crystals by Vickers indentation with 100 g loading. Crystal surfaces were free of any treatment. Optical examination of indentations was done with a "Neophot-21" light microscope. Observations of crystal defects were carried out under a Docuval light microscope.

boundaries cannot exceed 2° at the end of EFG rods. These grain boundaries can not cause the sharp change in strength measurement [8]. The EFG crystals present a quite traditional peripheral distribution of gaseous inclusions. Just below (50--100 m) the surface is a layer of micro-voids of 1--5 m in size. Regular striations of voids [9] are observed in the longitudinal sections of sapphire GES crystals with height of crystallization layers »/ "20 and 120 m. Those periodic layers revealed by optical methods are due to gas bubbles, solid inclusions and inhomogeneously doped impurities. The regular striations were absent for the tubes with height of crystallization layers 3.5 m. The misorientation between grain boundaries in the GES tubes did not exceed 2°. 3.2. Mechanical properties Elastic modulus is found as E"455 GPa for 0° specimens, and E "422 GPa and E "426 GPa for two types of 90° specimens. These values are near the measurements for crystals of perfect structure, cut from boules. Therefore the voids do not significantly affect the elastic properties of our crystals. Vickers microhardness measurements are shown in Fig. 1. For as grown crystals the average microhardness values are 1424 and 2157 kg/mm for 90° and basal planes, respectively. For annealed crystals the average microhardness values are 2212 and 2878 kg/mm for 90° and basal planes, respectively. The variation coefficient is estimated as 7--11%. It is obvious that high-temperature annealing leads to the fixation of dislocation. Therefore the microhardness of annealed crystals is higher. During bending test three types of sample failures were observed. The first is the failure with shattering when the initiation of fracture is not identified. The second is the failure in two (or three) pieces between inner support points which is typical for a majority of specimens. And the third is the failure originated at the outer support point which is observed for some 90° specimens. The presence of second and third types of failure for different crystal orientations is in accordance with failure modes described in Ref. [5].

3. Results and discussion 3.1. Structure of grown crystals The second-phase precipitations in the EFG rods were absent, and misorientation between grain


P.A. Gurjiyants et al. / Journal of Crystal Growth 198/199 (1999) 227 ­ 231

229

Fig. 1. Vickers microhardness measurements.

Fig. 2a presents the variation of strength with temperature for EFG specimens. The strong difference between values of strength for as grown and annealed crystals is in the temperature range 600--900°C for both crystal orientations. It can be explained by a higher amount of mobile dislocation in as grown crystals at these temperatures. This fact is in good agreement with the microhardness investigation. At room temperature there is no difference in strength for 0° as grown and annealed specimens. The slight increasing in room temperature strength of 90° as grown specimens is likely connected with the possibility of a weak dislocation motion in the rhombohedral planes. 90° orientation of crystals is more suitable for motion than 0°. Fig. 2b presents the variation of strength with temperature for the GES specimens. One can see the significant differences in the strength of the crystals of small height of layer (3.5--20 m) and of large height of layer (70--120 m). At 1550°C the strength of a GES crystal of 120 m layer is close to that of annealed 0° specimens whereas the strength of crystals of small height of layer is higher. It can be assumed that at 1550°C the improved value of strength for GES crystals of small height of layer such as for 90° specimens is connect-

ed with a greater mobile dislocation density in comparison to GES crystals of large height of layer and 0° specimens. The stress--strain curve's deviation from straight line behaviour before failure confirms the significant plastic strain in the crystals. Unal and Lagerlof [10] reported the absence of plastic deformation for 0° specimens at 1500°C. The deformation of the crystals can be via slip on rhombohedral planes but the brittle fracture is achieved at lower stress than needed for the slip occurrence. The fibers having a orientation near 6° off of the c-axis demonstrate the noticeable plastic strain at 1450°C [10]. Existence of plastic strain is the likely explanation for the improved strength of 90° specimens in comparison with 0° ones at 1550°C. A higher level of strength was observed for 90° specimens than the strength of 0° specimens [4]. Plastic strain occurs in the layered crystals near the layer boundaries, which results in more homogeneous strain and consequently increased plasticity [11]. Perhaps, in our case, the great number of such boundaries in GES crystals of a small height of layer contributes to the plastic deformation of these crystals. It is the likely reason for increased strength in the crystals.


230

P.A. Gurjiyants et al. / Journal of Crystal Growth 198/199 (1999) 227 ­ 231

Fig. 2. Variation of flexural strength with temperature for (a) the EFG specimens (as grown and annealed crystals); (b) the GES crystals of different height of layer.

Fig. 3. Fracture surface of GES crystal, SEM: (a) slip lines and stairs; (b) origin of failure (showed by arrow); (c) zone of high concentration of voids (magnified Fig. 3b).


P.A. Gurjiyants et al. / Journal of Crystal Growth 198/199 (1999) 227 ­ 231

231

3.3. Fractography Whatever the crystal orientation (0° or 90° specimens) and growth method (EFG or GES), the fracture surfaces are observed to be roughly 45° tilted from the sample's axis. Room temperature failure is purely brittle, with a strong tendency to rhombohedral cleavage. The evidence of plastic strain is in the crystals at temperatures above 600°C. Rhombohedral slip is activated before the sample failure. In GES crystals the stairs were observed on the fracture surface (Fig. 3a). The height of the stair is correlated with the height of layer. Therefore, the crack propagates in some jumps from one level to another. The origin of fracture is situated near the surface for EFG crystals. Thin EFG fibers demonstrate two possible origins of fracture [10]: surface flaws and internal voids located rather far from surface [12]. Since in our EFG crystals voids are situated near the surface it is difficult to discriminate the origin of the critical flaw. In GES crystals failure initiation is ductile but complete failure is achieved in a brittle mode (Fig. 3b). Failure origin is a zone of high concentration of voids (Fig. 3c). The location of the zones coincides with the bubbles distributed along each new crystallized layer.

2. Failure of the GES sapphire crystals initiates in a zone of voids between crystallized layers, and then crack propagates in jumps from one crystallized level to another. Acknowledgements This work was carried out partially under the financial support of the French Ministry of Defense (DRET) and Copernicus network (Grant PL 978078). References
[1] G.H. Hurley, Appl. Polym. Symp. 21 (1973) 121. [2] J.T.A. Pollock, G. Hurley, J. Mater. Sci. 8 (1973) 1595. [3] R.L. Gentilman, E.A. Maguire, H.S. Starret, T.M. Hartnett, H.P. Kirchner, Commun. Am. Ceram. Soc., September 1981, p. C-116. [4] J.W. Fisher, W.R. Compton, N.A. Jaeger, D.C. Harris, in: Window and Dome Technologies and Materials II, SPIE Proc. 1326 (1990) 11. [5] D.C. Harris, F. Schmid, J.J. Mecholsky Jr., Y.L. Tsai, SPIE Proc. 2286 (1994) 16. [6] H.E. LaBelle, J. Crystal Growth 50 (1980) 8. [7] P.I. Antonov, Yu.G. Nosov, S.P. Nikanorov, Bull. Acad. Sci. USSR, Phys. Ser. 49 (1985) 2295. [8] M.Yu. Starostin, T.N. Yalovets, A.Zh. Rozenflants, V.A. Borodin, Bull. Rus. Acad. Sci. Phys. 58 (1994) 1461 (in Russian). [9] V.A. Borodin, V.V. Sidorov, T.A. Steriopolo, V.A. Tatarchenko, T.N. Yalovets, Bull. Acad. Sci. USSR, Phys. Ser. 52 (1988) 118. [10] O. Unal, K.P. Lagerlof, J. Amer. Ceram. Soc. 77 (1994) 2609. [11] A.V. Nikiforov, Yu.G. Nosov, O.V. Kljavin, P.I. Antonov, M.B. Muchamedzhanova, Bull. Acad. Sci. USSR, Phys. Ser. 52 (1988) 131. [12] S.A. Newcomb, R.E. Tressler, J. Amer. Ceram. Soc. 77 (1994) 3030.

4. Conclusions 1. The crystal orientation and height of crystallization layer influence the strength of sapphire crystals. Since the temperature as low as 600°C it can be connected with dislocation motion.