Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.issp.ac.ru/lsc/files/ICM_modif_1999.pdf Дата изменения: Thu Feb 9 10:14:55 2012 Дата индексирования: Tue Oct 2 07:41:48 2012 Кодировка:

Journal of Crystal Growth 204 (1999) 499 } 504

Sapphire "bres grown by a modi"ed internal crystallisation method
V.N. Kurlov*, V.M. Kiiko, A.A. Kolchin, S.T. Mileiko
Solid State Physics Institute, Russian Academy of Sciences, Chernogolovka, Moscow District 142432, Russia Received 24 December 1998; accepted 7 April 1999 Communicated by D.T.J. Hurle

Abstract Single-crystal c-axis Al O "bres were grown by the modi"ed internal crystallisation method, and were evaluated as reinforcements for high-temperature composites. Fibres were characterised by orientation, shape and mechanical properties. Growth procedures and "bre properties are reported. 1999 Elsevier Science B.V. All rights reserved.
PACS: 81.10.!h; 81.10.Fq; 62.20.!x; 77.84.Lf Keywords: Oxide "bre; Sapphire; Crystallization; Internal crystallization; Strength

1. Introduction The use of single-crystal "bres as a reinforcement for metal and ceramic matrix composites promises a way to obtain materials to serve in structural elements at temperatures in excess of 12003C in oxidising conditions. Among oxides, sapphire is the number one candidate for "bre material because of its combination of mechanical, thermal, and chemical properties. The following principal melt growth techniques have been successfully used to produce single-crystal oxide "bres: (1) the edge-de"ned "lm-fed growth (EFG) [1] or Stepanov [2] method, (2) the laser heated #oating zone directional solidi"cation process (LHFZ) [3], (3) the laser heated pedestal

* Corresponding author. Fax: #7-096-5764111. E-mail address: kurlov@issp.ac.ru (V.N. Kurlov)

growth (LHPG) method [4], and (4) the micropulling-down ( -PD) method [5]. However, the cost of the "bres obtained by using the above-mentioned methods is too high to use them for structural applications. Therefore, to stimulate the usage of oxide-"bre composites, it is necessary to overcome the problem of high cost of the "bres which perhaps can be achieved by developing new methods to crystallise oxide "bres from the melt. The internal crystallisation method (ICM) was originally developed to produce refractorymetal matrix composites. It has been successfully used to obtain various oxide "bres (sapphire, complex oxides, some oxide eutectics, and so on) [6,7]. The spontaneous crystallisation of the "bres in the channels takes place after the in"ltration of a melt into channels in a matrix and subsequently moving the matrix to a cold zone of the furnace. Using this technique crystallisation yields both polycrystalline and single-crystal "bres with nonhomogeneous

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


500

V.N. Kurlov et al. / Journal of Crystal Growth 204 (1999) 499 } 504

crystallographic orientation. The histograms of angle between the c-axis of the sapphire crystal and the "bre axis show that the values of lie between 453 and 903 [6,8]. This yields a decreasing creep strength of composites containing such "bres since "bres of unfavourable orientation start to creep at rather low stresses and the stress redistribution leads to overloading "bres of a favoured orientation that yields the failure of the latter [8]. Sapphire exhibits very high yield stress and creep resistance at high temperatures if it is oriented so that shear stresses on the basal slip system are zero, as is the case when the tensile axis is parallel to the c-axis [9]. The aim of the present study is to develop single-crystal sapphire "bres of homogeneous crystallographic orientation by the modi"ed internal crystallisation method. The main feature of this particular technique is that the crystallisation of the "bres within a volume of an auxiliary matrix begins on an oriented seed located at the top of a molybdenum block with a system of internal channels. Hence, we call this modi"ed technique as the top seeded internal crystallisation method.

Fig. 1. A cross-section of the molybdenum-matrix block. The open parts to be "lled by an oxide melt and, hence, form the "bre cross-section.

2. Fabrication of sapphire 5bres by the internal crystallisation method The experiments were conducted in an 8 kHz induction heated graphite susceptor/molybdenum crucible set-up. The raw material was crushed Verneuil boules. An argon atmosphere under a pressure of 1.1}1.2 atm was used as ambient. Experimental runs used c-axis (100012) seeds to initiate growth. The pulling rates are between 1 and 30 mm/min. The continuous cylindrical channels in auxiliary molybdenum matrix are shaped by di!usion bonding under special conditions of an assembly of foil and wire as described elsewhere [6,7]. The shape of a transverse section of the matrix is shown in Fig. 1. Normally, the size of the molybdenum block used for "bre crystallisation is about 5;40 mm in a cross-section, and 65 mm in length. Nearly 50% of the block volume is occupied by the channels. A sequence of the "bre production process using the modi"ed ICM is illustrated schematically in Fig. 2. The initial arrangement includes a molyb-

denum block containing the channels, a seed of 5;40 mm in a cross-section contacting the blocktop surface and a crucible "lled with crushed Verneuil boules in the lower position (Fig. 2a). Then a susceptor/crucible system is heated up to the beginning of melting of the sapphire seed at the zone of seed/matrix-block contact. The temperature necessary to reach this condition is adjusted by trial. The seed is lowered at the melting point. Immediately the crucible containing the melt is moved up to contact the matrix block which becomes "lled with the melt (Fig. 2b). The in"ltration of the melt into channels of the matrix takes place as a result of capillary forces. The next stage of the process is the directional crystallisation of the "bres in the channels. Actual growth starts with pulling up the matrix/seed system. Crystallographic growth axes are determined simply by pre-oriented seed crystals. Since the seed crystal is attached to the melt into channels in a matrix, the "bre axis follows the orientation of the seed provided the pulling-up is initiated after the melt zone has stabilised. With the small crosssections of the "bres, preferred growth axes do not manifest themselves as they do for larger crystals. Moving the matrix/seed system to a cold zone, a solid/liquid interface in each channel moves from


V.N. Kurlov et al. / Journal of Crystal Growth 204 (1999) 499 } 504

501

Fig. 3. A bundle of the sapphire "bres.

Fig. 2. Schematic of the experimental arrangement for crystallising homogeneously oriented sapphire "bres by the modi"ed ICM.

the top to the bottom. The pulling-up is continued to obtain homogeneous oriented "bres along the whole channel length. Finally, pulling-up is stopped as the sapphire/molybdenum block exits from the molybdenum ring (Fig. 2c). The block is then cooled at a rate of about 303C/min. Extraction of the "bres after crystallisation is achieved by dissolving the molybdenum matrix in acid mixtures. Sapphire "bres obtained by this method are shown in Fig. 3. It should be noted that molybdenum could be regenerated fully to appear as a powder. While dealing with "ve and eight molybdenum blocks in one process, 120}150 g of the sapphire "bres are produced.

the "rst requirement for them is a su$ciently high e!ective strength. This means that they should have high strength at a length of an order of the critical length that a "bre length which can be e!ectively loaded in a particular matrix [7]. Normally this is about 10 diameters or average diameter. Actually, the "bre strength should be measured inside an appropriate matrix since a "bre/matrix interaction can change the strength characteristics drastically. This is especially true regarding sapphire "bre in a nickel-based matrix [10], which is normally considered as an appropriate matrix for heat-resistant composites. Hence, statistical strength characteristics and a scale dependence of the "bre shall be determined here. These characteristics can be used for both comparing various "bres and studying an e!ect of microstructural parameters on "bre properties. 3.1. Crystallography orientation Observing the "bres in polarised light (Fig. 4) shows that they have a single-crystalline structure along the whole length for all the above-mentioned pulling rates. Total extinction was observed under changes in the angle from 03 (Fig. 4a) to 453 (Fig. 4b). The orientations of grown "bres and seed coincide closely. 3.2. Room temperature strength

3. Fibre characteristics ICM sapphire "bres are mainly to be used as a reinforcement for "brous composites. Therefore,

The scale dependence of the "bres strength was determined following the procedure outlined in detail elsewhere [11] that consists of looping a "bre over a series of rigid cylinders of decreasing


502

V.N. Kurlov et al. / Journal of Crystal Growth 204 (1999) 499 } 504

It can be seen that changing the pulling rate by more than an order of magnitude does not yield a noticeable change in the "bre strength. Coating the "bres with a thin layer of several microns, of either silicon carbide or silicon oxy-carbide by using a CVD-process changes the "bre strength characteristics drastically (Fig. 6). The main conclusion that can be drawn from this observation is that defects located on the "bre surface are rougher than those distributed within the "bre volume. This contradicts partially the conclusion made as a result of testing such "bres in the molybdenum matrix [6,8], although the latter was drawn after high-temperature testing, at 13003C. At the same time, the only visible defect, which can e!ect the "bre failure, is a pore or a chain of the pores located in the vicinity of the "bre surface (Fig. 7). The e!ect of the "bre surface pro"le that replicates the pro"le of the molybdenum foil should not be discounted. This surface feature is obviously revealed in Fig. 7. Let us now write the well-known Weibull distribution function to describe the "bre strength in the form @ g P( , g)"1!exp ! , (1) g where g is a geometrical parameter equal to either "bre surface S or "bre volume < depending on a particular pattern of the characteristic defects location, g is a corresponding Weibull parameter, being either S or < , and are two other Weibull parameters. Obviously, one of the three parameters introduced can be chosen arbitrarily. Eq. (1) yields the average "bre strength for surface S or volume < g@ 1 1# , (2) g where (z) denotes the gamma function. Eqs. (1) and (2) are used to interpret experimental data of a type presented in Fig. 5 to obtain the Weibull parameters, , , S or < , describing the bending strength as given in Table 1. Technical details of the procedure are outlined in Ref. [11]. If one assumes either surface or volume location of the defects that determine the "bre strength then one can calculate tensile strength characteristics by 1 (g)2" dP( )"

Fig. 4. The sapphire "bres photographed in polarised light.

diameter d , d , 2, d , 2, d , measuring the aver G I age distance between "bre breaks, calculating the ultimate bending stress versus the length (the average inter-break distance), calculating the tensilestrength/"bre-length dependence. Suppose we are looping a "bre of a length ё the cylinder of diameter d and the number of "bre breaks is n . Then G G we repeat the same with cylinder of diameter d and get n of the "bre breaks. Because the G> G> maximum "bre stress is "Eh/d , where h is the G G "bre thickness, the average "bre bending strength corresponding to "bre length l "ё/n is G> G> H "( # )/2. G> G G> Experimental dependencies H(lH) for "bres in GG Table 1 are presented in Fig. 5.






V.N. Kurlov et al. / Journal of Crystal Growth 204 (1999) 499 } 504 Table 1 The geometrical characteristics, fabrication parameter, and Weibull parameters of sapphire "bres Fibre block Pulling rate (mm/min) Circumference (mm) Average area (mm) Volume hypothesis bending (MPa) 3.096 3.378 2.795 3.875 3.107 2.496 2.128 733.2 1973.4 1480.1 789.6 969.6 1131.9 1180.6 tension (MPa) 483.6 1334.9 965.7 550.9 665.0 713.7 741.9 < (mm) Surface hypothesis bending (MPa) 3.141 3.299 3.154 4.170 3.154 2.511 2.490 432.4 1217.6 593.6 552.9 597.9 414.3 587.4 tension (MPa) 172.4 516.0 244.2 271.6 260.6 138.0 212.5

503

S (mm)

V086 V086c V118 V020 V145 V121 V123

1.3 1.3 1.3 1.3 1.3 3.3 30.0

0.6499 0.7262 0.6752 0.6515 0.4791 0.6810 0.7526

0.01157 0.01374 0.01186 0.01017 0.00493 0.01138 0.01409

0.05784 0.06871 0.05931 0.05087 0.02465 0.05693 0.07048

17.04 24.93 60.64 16.38 12.45 41.91 23.00

Fig. 5. Bending strength of "bres extracted from sapphire/molybdenum blocks crystallised at two pulling rates versus "bre length.

Fig. 6. Bending strength of as-extracted and coated "bres versus "bre length. See Table 1 for block characteristics. The "bres are coated by a layer of SiC O of thickness 4}6 m.

using original bending strength characteristics. The procedure is based on comparison of the average strength values obtained for a homogeneous stress distribution for which in the integral, Eq. (2), does not depend of the co-ordinates of a point within a tensile specimen and for a nonhomogeneous stress distribution in which case depends on the co-ordinates of a point in a specimen subjected to bending. Again, technical details of the procedure are published elsewhere [11,12]. The Weibull parameters, , , S or < , to describe tensile charac teristics of the "bres are presented in Table 1. Labels surface and volume relate the characteristics to one of the hypothesises accepted. Certainly, a real tensile behaviour is somewhere in between the two patterns mentioned.

Fig. 7. A sapphire "bre after failure. The arrow indicates the pore located in the vicinity of the "bre surface.


504

V.N. Kurlov et al. / Journal of Crystal Growth 204 (1999) 499 } 504

One can see that strength of the "bres under investigation at lengths of about the critical length is comparable with that of other types of aluminabased "bres analysed in Ref. [7].

nicus network (PL 97-8078). The authors appreciate the help by Dr. G.K. Strukova, Mrs. T.I. Boromychenko and Mrs. S.I. Gvozdeva.

References 4. Conclusions E The single-crystal sapphire "bres of homogeneous crystallographic orientation have been grown by the top seeded internal crystallisation method. E The strength properties of the sapphire "bres in terms of those that are important for the design of composites are comparable with those of other types of alumina-based "bres. Acknowledgements The work was supported by the International Science and Technology Center (ProjectC507-97), INTAS-RFBR (Project C 95-0599) and Coper[1] H.E. LaBelle Jr., A.I. Mlavsky, Mater. Res. Bull. 6 (1971) 571. [2] L.N. Dmitruk, V.I. Shelyubskii, Sov. Phys. Crystallogr. 24 (1979) 506. [3] J.S. Haggerty, W.P. Menashi, NASA, Final Report Contract No. NAS 3-13479, February, 1971. [4] R.S. Feigelson, J. Crystal Growth 79 (1986) 669. [5] D.H. Yoon, I. Yonenaga, N. Ohnishi, T. Fukuda, J. Crystal Growth 142 (1994) 339. [6] S.T. Mileiko, V.I. Kazmin, J. Mater. Sci. 27 (1992) 2165. [7] S.T. Mileiko, Metal and Ceramic Based Composites, Elsevier, Amsterdam, 1997. [8] S.T. Mileiko, V.I. Kazmin, Compos. Sci. Technol. 45 (1992) 209. [9] J.S. Haggerty, K.C. Wills, J.E. Sheehan, Ceram. Eng. Sci. Proc. 12 (1991) 1785. [10] R. Asthana, S.N. Tewari, S.L. Draper, Metall. Mater. Trans. 29 A (1998) 1527. [11] V.M. Kiiko, S.T. Mileiko, Compos. Sci. Technol., in press. [12] S.T. Mileiko, Compos. Sci. Technol., in press.