- Email: [email protected]
aluminium titanate composite using grazing-incidence synchrotron radiation diffraction
August 2002
Materials Letters 55 (2002) 344 – 349 www.elsevier.com/locate/matlet
Depth profiling of near-surface information in a functionally graded alumina/aluminium titanate composite using grazing-incidence synchrotron radiation diffraction M. Singh, P. Manurung, I.M. Low * Materials Research Group, Department of Applied Physics, Curtin University of Technology, P.O. Box U1987 Perth 6845 W.A., Australia Received 24 August 2001; accepted 25 October 2001
Abstract Grazing-incidence synchrotron radiation diffraction (GISRD) has been successfully used for near-surface depth profiling of phase composition in a functionally graded alumina (A)/aluminium titanate (AT) composite prepared by an infiltration process. Depth profiling of near-surface layers both within the nanometer and micrometer range could be done by using angles of incidence below and above the critical angle (ac) for total external reflection. The penetration depth increased to several hundred angstroms as a approached ac. However, above ac there was a rapid increase in penetration depth to several thousand angstroms or more. As the penetration depth increased, the intensity of aluminium titanate peaks relative to those of alumina became less intense, indicating a graded change in the phase abundance. The presence of graded residual strains in the composite due to the thermal expansion mismatch between the phases has been computed and verified from the display of line shifts. The unique but powerful capability of GISRD as a tool for depth-profiling the near-surface information of graded materials has been demonstrated in this work. D 2002 Elsevier Science B.V. All rights reserved. Keywords: Depth profiling; Phase abundance; Grazing-incidence synchrotron radiation; Functionally graded; Alumina; Aluminium titanate
1. Introduction Functionally graded alumina-based composites with improved damage tolerance properties have been recently produced with success by an infiltration technique [1 –7]. These layered composites, comprising of a homogenous alumina layer and a heterogenous layer of alumina/aluminium titanate (AT) or alumina/calcium-hexaluminate (CA6), are formed by controlled
*
Corresponding author. Tel.: +61-8-92667544; fax: +61-892662377. E-mail address: [email protected] (I.M. Low).
infiltration of liquid precursors into a porous preform and subsequent firing to form the desired phases. The heterogeneous layer in the composite is designed to give functional gradation of hardness and fracture toughness so as to result in improved damage-tolerance properties. Depth profiling of phase composition in the composite is particularly of importance as it is expected to have a profound influence on both physical and mechanical properties of the composite as a whole. X-ray diffraction has been previously used for depth profiling of phase composition and abundance in various ceramic systems such as alumina/AT, alumina/mullite, alumina/mullite/zirconia, and alumina/ CA6 [1 –10]. This is usually done by gradual polishing
0167-577X/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 5 7 7 X ( 0 2 ) 0 0 3 9 0 - 7
M. Singh et al. / Materials Letters 55 (2002) 344–349
with emery paper or cutting the sample into thin slices of f 1 mm thick. However, the information obtained in these cases is from depths of few micrometers from the surface. In addition, the use of polishing or cutting may introduce undesirable attributes such as surface damage, grain pull-outs and residual strains. Hence, it is impossible to depth profile the compositions in these materials in a nanometer scale using conventional XRD. In order to allow in situ measurement of nearsurface information on a nanometer scale, grazing angles of incidence are used, which allow the penetration of the beam to be varied by variation of the angle of incidence [11– 13]. Marra et al. [14] has demonstrated that surfaces of solids can be studied to advantage by using a highly collimated beam of X-rays striking the surface at a grazing angle near or within the range at which total external reflection occurs. In this situation, only the near-surface layer is illuminated and a diffraction pattern is produced which reveals the structure of the surface region in preference to that of the interior. Synchrotron sources are particularly well suited for providing the highly collimated radiation with necessary intensity [15]. The wavelength can also be selected to avoid fluorescent background and to obtain the highest possible peak to background. The beam broadening due to asymmetric geometry will not lead to peak broadening if the diffraction angle is determined by a set of parallel foils or an analyser crystal in front of the detector [15]. Going from symmetric diffraction to lower angles of incidence reduces the penetration depth and thus contribution of the near-surface layers. At grazing incidence, not only the intensities but also the peak positions (due to refraction) and peak widths (due to total reflection) are changed. Furthermore, the drastic reduction of the penetration depth as the angle of total reflection leads to a steep decrease in intensity from the layers below this interface. In the present paper, we describe the use of grazingincidence synchrotron radiation diffraction (GISRD) as a powerful tool for depth profiling the near-surface composition and residual strains of a functionally graded alumina/AT composite both at the nanometer and micrometer range through the use of angles of incidence below and above the critical angle for total external reflection. The residual strains at different depths from the surface were computed from line shifts
345
due to thermal expansion mismatch between the phases in the graded composite.
2. Experimental procedure 2.1. Preparation of graded alumina/aluminium titanate composite Reactive alumina powder of 0.4 mm average particle size and 99.9% purity from Mandoval, UK was used to fabricate alumina preforms for infiltration processing. The as-received powder was uni-axially pressed into discs of 19-mm diameter and 2-mm thickness at a pressure of 75 MPa, and then presintered at a temperature of 1100 C for 2 h. The resultant preforms had a porosity of f 46%. The preforms were degassed in a vacuum chamber, and then immersed in TiCl4 solution for 30 min. Multiple infiltrations were performed for increased retention of the infiltrant. After each infiltration, the samples were fired to 600 C for 2 h for better adherence of the previous infiltrant. The green infiltrated discs were subsequently fired at a rate of 3 C/min to 1600 C for 2 h with 2 h of soaking at 1310 C in a high temperature furnace (Ceramic Engineering, Model HT 04/ 17), to facilitate the formation of aluminium titanate. The cooling rate was 1 C/min till room temperature. In order to observe the graded microstructure, the cross-section of sintered sample was cut, mounted in epoxy resin and polished with diamond paste to a surface finish of 1 mm. The polished was then thermally etched at 1400 C for 15 min, gold-coated and examined with back-scattered electrons using a JOEL scanning-electron microscope. 2.2. Depth profiling using grazing-incidence synchrotron radiation diffraction Near-surface depth profiling by GISRD was performed using the BIGDIFF instrument located at the Photon Factory in Tsukuba, Japan. Thin slices of f 1 mm thick were cut from the sintered sample for data collection. Imaging plates were used to record the diffraction patterns over a period of 20 min. The depth of X-ray penetration into the composite depends on the grazing angle of incidence (a). The values of a used were 0.1, 0.3, 0.5, 0.7, 1.0, 3.0 and 5.0
346
M. Singh et al. / Materials Letters 55 (2002) 344–349
˚ . For calculating the and the wavelength used was 1 A depth of penetration of the beam into the sample, it is necessary to calculate first the critical angle for total external reflection (ac) to occur, as follows [13,15]; ac ð2dÞ1=2 ¼ 1:6 103 qk
ð1Þ
˚ , and where q is density in g/cm3, k is wavelength in A d = 1 n for refractive index (n). The penetration depth is then calculated from the following equations [15]. For angle of incidence a < ac, the penetration depth (l) is given by: l¼
k 2pða2c a2 Þ1=2
Information of near-surface depth profiles was studied qualitatively from the GISRD data. For k = 1 ˚ and q = 3.70 g/cm3, the critical angle was calculated A as ac = 0.339. As the index of refraction of X-rays is slightly less than unity, the peaks shift to higher angles than those calculated from the Bragg law. For small incidence angles the Bragg angle shift from refraction is given by [13]: D2H ¼
d sina sin2Href 2þ þ ð4Þ sin2Href sin2Href sina
ð2Þ dcorr ¼ dobs D2H
ð5Þ
and for a >ac, the penetration depth (l) is given by: 2a l¼ l where l is the linear attenuation coefficient [16].
ð3Þ
where 2Href, dcorr and dobs are the reference, corrected and observed values, respectively. After subtracting the refraction shift from observed value, the corrected d values (dcorr) are obtained. Residual strains arising from the thermal expansion mismatch between the
Fig. 1. GISRD patterns of functionally graded alumina/AT sample at different grazing incidence angles. Legend: x = alumina; o = AT; uk = unknown phase.
M. Singh et al. / Materials Letters 55 (2002) 344–349
347
phases in the composite can contribute to line-shifts and may then be calculated from: Strain ð%Þ ¼ 100ðdref dcorr Þ=dref
ð6Þ
where dref is the reference value of strain-free alumina or AT powder.
3. Results and discussion The near-surface diffraction profiles as obtained from GISRD at various grazing-incidence angles are shown in Fig. 1. The calculated penetration depths at the corresponding grazing-incidence angles are shown in Fig. 2. The results indicate that increasing a to just below ac increased the depth to some hundreds of angstroms. Above ac the penetration was inversely proportional to l and increased to several thousands of angstroms. Small changes in a cause large variations in depth interval, and depth profiling cannot be sensitively controlled at the nanometer scale. However, this provides a depth profile on a larger micrometer scale. Fig. 3 shows the ratio of peak intensity counts for the strongest peak of aluminium titanate (AT), to that of alumina (A). As the angle of incidence or depth increases, the overall AT features become less intense as compared to alumina. AT intensity ratio curve shows an initial rise which peaked at 0.3, followed by a sharp fall at between 0.3 and 0.5 and a leveling off at angles greater than 1.0. As shown in Fig. 2, above the critical angle f 0.3, there is a sharp increase in penetration depth. This may be indicative
Fig. 2. Variation of penetration depth as a function of grazingincidence angle.
Fig. 3. Ratio of intensity counts versus grazing angle of incidence.
of a gradation of composition in the microstructure of the material as would be expected (Fig. 4). The nearsurface layer (0 – 0.3 mm) is richer in AT, which peaked at about 30 – 50 mm followed by a rapid decrease in abundance with depth but levels off gradually at depths greater than 0.5 mm from the surface. This suggests that the maximum content of AT did not form on the as-sintered surface of the sample but at a depth of f 30– 50 mm below, as the microstructure in Fig. 4 clearly reveals. This revelation comes as a surprise from the processing perspective, which was not observed from our previous studies by X-ray diffraction on similar graded ceramic systems [1 –9]. The unique but powerful capability of GISRD as a tool for depth-profiling the near-surface information of graded materials is clearly demonstrated in this work.
Fig. 4. Back-scattered electron imaging showing the cross-sectional microstructure of graded alumina/AT composite. The light phase is AT and the matrix is alumina. The direction of infiltration is from left to right.
348
M. Singh et al. / Materials Letters 55 (2002) 344–349
Fig. 5. Variation of residual strain as function of grazing-incidence angle.
It is interesting to note the display of line-shifts associated with all peaks as the grazing angle increased from 0.1 to 5. As previously highlighted, this phenomenon can be partly attributed to the intrinsic refraction property of X-rays, causing the peaks to shift to higher angles than those calculated from the Bragg’s law. The presence of residual strains due to thermal expansion mismatch between the phases in the sample can also give rise to this lineshift. If the sample has a graded distribution of phases, the presence of graded residual strains will be manifested as line-shifts as the grazing angle or surface depth changes. Eqs. (4) and (5) were used to compute the corrected d-spacing taking into account the contribution of lineshift due to refraction. The reference dref and calculated dcorr values were used in conjunction with Eq. (6) to compute the amount of residual strains formed in the sample as a result of a large mismatch in the thermal expansion coefficient between that of alumina ( f 8 10 6/C) and AT ( f 1 10 6/C). Fig. 5 shows the plots of residual strain as a function of grazing-incidence angle. It is interesting to note that the variation of residual strains within the AT grains is nearly constant or depth-independent. In contrast, the residual strains within the alumina matrix are distinctly depth-dependent probably due to the graded distribution of AT in the sample.
4. Conclusion Near-surface depth profiling of phase composition and residual strains in a functionally graded alumina/
AT composite has been characterized using grazingincidence synchrotron radiation diffraction. Results show that the gradation of both composition and strains in the composite is continuous even at the nanometer scale. Depth profiling is possibly down to several thousand angstroms in steps of a few angstroms by changing the incidence angle. Above the critical angle, the depth of penetration is several thousand angstroms and depends on the wavelength as well as on the incident angle. The X-ray index of refraction shifts the peaks to higher Bragg angles and these shifts have been calculated using established model. Residual strains at different depths are calculated from line shift by comparing with d values of reference sample.
Acknowledgements The GISRD data for this work were acquired at the Australian National Beamline Facility at the Photon Factory, Tsukuba, Japan with financial support from the Australian Synchrotron Research Program (ANBF Proposal 99/2000-AB-26), which is funded by the Commonwealth of Australia under the major National Research Facilities Program. We thank Dr. Garry Foran and Dr. James Hester for technical assistance in data collection. Mr. S. Leung of ANSTO assisted with the work on scanning electron microscopy through funding from an AINSE grant.
References [1] I.M. Low, Mater. Res. Bull. 33 (10) (1998) 1475 – 1482. [2] I.M. Low, J. Aust. Ceram. Soc. 34 (2) (1998) 250. [3] I.M. Low, R.D. Skala, D. Asmi, P. Manurang, M. Singh, Proc. Int. Conf. on Powder Metallurgy (PM2000), 12 – 16 Nov. 2000, Kyoto, Japan, 2000, pp. 1469 – 1472. [4] D. Asmi, I.M. Low, S. Kennedy, R.A. Day, Mater. Lett. 40 (1999) 96 – 102. [5] R.D. Skala, PhD Thesis, Curtin University of Technology, Perth, Australia (2000). [6] D. Asmi, PhD Thesis, Curtin University of Technology, Perth, Australia (2001). [7] P. Manurung, PhD Thesis, Curtin University of Technology, Perth, Australia (2001). [8] S. Pratapa, I.M. Low, J. Mater. Sci. Lett. 15 (1996) 800 – 802. [9] S. Pratapa, MSc Thesis, Curtin University of Technology, Perth, Australia (1996). [10] B.R. Marple, D.J. Green, J. Mater. Sci. 28 (1993) 4637. [11] T. Wroblewski, Mater. Sci. Forum 79 (1991) 469 – 474.
M. Singh et al. / Materials Letters 55 (2002) 344–349 [12] R. Delhez, Th.H. Keijser, E.J. Mittemeijer, B.J. Thijsse, M.A. Hollanders, O.B. Loopstra, W.G. Sloof, Aust. J. Phys. 41 (1988) 261 – 282. [13] G. Lim, W. Parrish, C. Ortish, M. Belletto, M. Hart, J. Mater. Res. 2 (4) (1987) 471 – 477.
349
[14] W.C. Marra, P. Eisenberg, A.Y. Cho, J. Appl. Phys. 50 (1979) 697. [15] G.H. Vineyard, Phys. Rev. B 26 (8) (1982) 4146 – 4158. [16] ANBF local webpage XCOM, http://www.ansto.gov.au/natfac/ asrp1.html.
Materials Letters 55 (2002) 344 – 349 www.elsevier.com/locate/matlet
Depth profiling of near-surface information in a functionally graded alumina/aluminium titanate composite using grazing-incidence synchrotron radiation diffraction M. Singh, P. Manurung, I.M. Low * Materials Research Group, Department of Applied Physics, Curtin University of Technology, P.O. Box U1987 Perth 6845 W.A., Australia Received 24 August 2001; accepted 25 October 2001
Abstract Grazing-incidence synchrotron radiation diffraction (GISRD) has been successfully used for near-surface depth profiling of phase composition in a functionally graded alumina (A)/aluminium titanate (AT) composite prepared by an infiltration process. Depth profiling of near-surface layers both within the nanometer and micrometer range could be done by using angles of incidence below and above the critical angle (ac) for total external reflection. The penetration depth increased to several hundred angstroms as a approached ac. However, above ac there was a rapid increase in penetration depth to several thousand angstroms or more. As the penetration depth increased, the intensity of aluminium titanate peaks relative to those of alumina became less intense, indicating a graded change in the phase abundance. The presence of graded residual strains in the composite due to the thermal expansion mismatch between the phases has been computed and verified from the display of line shifts. The unique but powerful capability of GISRD as a tool for depth-profiling the near-surface information of graded materials has been demonstrated in this work. D 2002 Elsevier Science B.V. All rights reserved. Keywords: Depth profiling; Phase abundance; Grazing-incidence synchrotron radiation; Functionally graded; Alumina; Aluminium titanate
1. Introduction Functionally graded alumina-based composites with improved damage tolerance properties have been recently produced with success by an infiltration technique [1 –7]. These layered composites, comprising of a homogenous alumina layer and a heterogenous layer of alumina/aluminium titanate (AT) or alumina/calcium-hexaluminate (CA6), are formed by controlled
*
Corresponding author. Tel.: +61-8-92667544; fax: +61-892662377. E-mail address: [email protected] (I.M. Low).
infiltration of liquid precursors into a porous preform and subsequent firing to form the desired phases. The heterogeneous layer in the composite is designed to give functional gradation of hardness and fracture toughness so as to result in improved damage-tolerance properties. Depth profiling of phase composition in the composite is particularly of importance as it is expected to have a profound influence on both physical and mechanical properties of the composite as a whole. X-ray diffraction has been previously used for depth profiling of phase composition and abundance in various ceramic systems such as alumina/AT, alumina/mullite, alumina/mullite/zirconia, and alumina/ CA6 [1 –10]. This is usually done by gradual polishing
0167-577X/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 5 7 7 X ( 0 2 ) 0 0 3 9 0 - 7
M. Singh et al. / Materials Letters 55 (2002) 344–349
with emery paper or cutting the sample into thin slices of f 1 mm thick. However, the information obtained in these cases is from depths of few micrometers from the surface. In addition, the use of polishing or cutting may introduce undesirable attributes such as surface damage, grain pull-outs and residual strains. Hence, it is impossible to depth profile the compositions in these materials in a nanometer scale using conventional XRD. In order to allow in situ measurement of nearsurface information on a nanometer scale, grazing angles of incidence are used, which allow the penetration of the beam to be varied by variation of the angle of incidence [11– 13]. Marra et al. [14] has demonstrated that surfaces of solids can be studied to advantage by using a highly collimated beam of X-rays striking the surface at a grazing angle near or within the range at which total external reflection occurs. In this situation, only the near-surface layer is illuminated and a diffraction pattern is produced which reveals the structure of the surface region in preference to that of the interior. Synchrotron sources are particularly well suited for providing the highly collimated radiation with necessary intensity [15]. The wavelength can also be selected to avoid fluorescent background and to obtain the highest possible peak to background. The beam broadening due to asymmetric geometry will not lead to peak broadening if the diffraction angle is determined by a set of parallel foils or an analyser crystal in front of the detector [15]. Going from symmetric diffraction to lower angles of incidence reduces the penetration depth and thus contribution of the near-surface layers. At grazing incidence, not only the intensities but also the peak positions (due to refraction) and peak widths (due to total reflection) are changed. Furthermore, the drastic reduction of the penetration depth as the angle of total reflection leads to a steep decrease in intensity from the layers below this interface. In the present paper, we describe the use of grazingincidence synchrotron radiation diffraction (GISRD) as a powerful tool for depth profiling the near-surface composition and residual strains of a functionally graded alumina/AT composite both at the nanometer and micrometer range through the use of angles of incidence below and above the critical angle for total external reflection. The residual strains at different depths from the surface were computed from line shifts
345
due to thermal expansion mismatch between the phases in the graded composite.
2. Experimental procedure 2.1. Preparation of graded alumina/aluminium titanate composite Reactive alumina powder of 0.4 mm average particle size and 99.9% purity from Mandoval, UK was used to fabricate alumina preforms for infiltration processing. The as-received powder was uni-axially pressed into discs of 19-mm diameter and 2-mm thickness at a pressure of 75 MPa, and then presintered at a temperature of 1100 C for 2 h. The resultant preforms had a porosity of f 46%. The preforms were degassed in a vacuum chamber, and then immersed in TiCl4 solution for 30 min. Multiple infiltrations were performed for increased retention of the infiltrant. After each infiltration, the samples were fired to 600 C for 2 h for better adherence of the previous infiltrant. The green infiltrated discs were subsequently fired at a rate of 3 C/min to 1600 C for 2 h with 2 h of soaking at 1310 C in a high temperature furnace (Ceramic Engineering, Model HT 04/ 17), to facilitate the formation of aluminium titanate. The cooling rate was 1 C/min till room temperature. In order to observe the graded microstructure, the cross-section of sintered sample was cut, mounted in epoxy resin and polished with diamond paste to a surface finish of 1 mm. The polished was then thermally etched at 1400 C for 15 min, gold-coated and examined with back-scattered electrons using a JOEL scanning-electron microscope. 2.2. Depth profiling using grazing-incidence synchrotron radiation diffraction Near-surface depth profiling by GISRD was performed using the BIGDIFF instrument located at the Photon Factory in Tsukuba, Japan. Thin slices of f 1 mm thick were cut from the sintered sample for data collection. Imaging plates were used to record the diffraction patterns over a period of 20 min. The depth of X-ray penetration into the composite depends on the grazing angle of incidence (a). The values of a used were 0.1, 0.3, 0.5, 0.7, 1.0, 3.0 and 5.0
346
M. Singh et al. / Materials Letters 55 (2002) 344–349
˚ . For calculating the and the wavelength used was 1 A depth of penetration of the beam into the sample, it is necessary to calculate first the critical angle for total external reflection (ac) to occur, as follows [13,15]; ac ð2dÞ1=2 ¼ 1:6 103 qk
ð1Þ
˚ , and where q is density in g/cm3, k is wavelength in A d = 1 n for refractive index (n). The penetration depth is then calculated from the following equations [15]. For angle of incidence a < ac, the penetration depth (l) is given by: l¼
k 2pða2c a2 Þ1=2
Information of near-surface depth profiles was studied qualitatively from the GISRD data. For k = 1 ˚ and q = 3.70 g/cm3, the critical angle was calculated A as ac = 0.339. As the index of refraction of X-rays is slightly less than unity, the peaks shift to higher angles than those calculated from the Bragg law. For small incidence angles the Bragg angle shift from refraction is given by [13]: D2H ¼
d sina sin2Href 2þ þ ð4Þ sin2Href sin2Href sina
ð2Þ dcorr ¼ dobs D2H
ð5Þ
and for a >ac, the penetration depth (l) is given by: 2a l¼ l where l is the linear attenuation coefficient [16].
ð3Þ
where 2Href, dcorr and dobs are the reference, corrected and observed values, respectively. After subtracting the refraction shift from observed value, the corrected d values (dcorr) are obtained. Residual strains arising from the thermal expansion mismatch between the
Fig. 1. GISRD patterns of functionally graded alumina/AT sample at different grazing incidence angles. Legend: x = alumina; o = AT; uk = unknown phase.
M. Singh et al. / Materials Letters 55 (2002) 344–349
347
phases in the composite can contribute to line-shifts and may then be calculated from: Strain ð%Þ ¼ 100ðdref dcorr Þ=dref
ð6Þ
where dref is the reference value of strain-free alumina or AT powder.
3. Results and discussion The near-surface diffraction profiles as obtained from GISRD at various grazing-incidence angles are shown in Fig. 1. The calculated penetration depths at the corresponding grazing-incidence angles are shown in Fig. 2. The results indicate that increasing a to just below ac increased the depth to some hundreds of angstroms. Above ac the penetration was inversely proportional to l and increased to several thousands of angstroms. Small changes in a cause large variations in depth interval, and depth profiling cannot be sensitively controlled at the nanometer scale. However, this provides a depth profile on a larger micrometer scale. Fig. 3 shows the ratio of peak intensity counts for the strongest peak of aluminium titanate (AT), to that of alumina (A). As the angle of incidence or depth increases, the overall AT features become less intense as compared to alumina. AT intensity ratio curve shows an initial rise which peaked at 0.3, followed by a sharp fall at between 0.3 and 0.5 and a leveling off at angles greater than 1.0. As shown in Fig. 2, above the critical angle f 0.3, there is a sharp increase in penetration depth. This may be indicative
Fig. 2. Variation of penetration depth as a function of grazingincidence angle.
Fig. 3. Ratio of intensity counts versus grazing angle of incidence.
of a gradation of composition in the microstructure of the material as would be expected (Fig. 4). The nearsurface layer (0 – 0.3 mm) is richer in AT, which peaked at about 30 – 50 mm followed by a rapid decrease in abundance with depth but levels off gradually at depths greater than 0.5 mm from the surface. This suggests that the maximum content of AT did not form on the as-sintered surface of the sample but at a depth of f 30– 50 mm below, as the microstructure in Fig. 4 clearly reveals. This revelation comes as a surprise from the processing perspective, which was not observed from our previous studies by X-ray diffraction on similar graded ceramic systems [1 –9]. The unique but powerful capability of GISRD as a tool for depth-profiling the near-surface information of graded materials is clearly demonstrated in this work.
Fig. 4. Back-scattered electron imaging showing the cross-sectional microstructure of graded alumina/AT composite. The light phase is AT and the matrix is alumina. The direction of infiltration is from left to right.
348
M. Singh et al. / Materials Letters 55 (2002) 344–349
Fig. 5. Variation of residual strain as function of grazing-incidence angle.
It is interesting to note the display of line-shifts associated with all peaks as the grazing angle increased from 0.1 to 5. As previously highlighted, this phenomenon can be partly attributed to the intrinsic refraction property of X-rays, causing the peaks to shift to higher angles than those calculated from the Bragg’s law. The presence of residual strains due to thermal expansion mismatch between the phases in the sample can also give rise to this lineshift. If the sample has a graded distribution of phases, the presence of graded residual strains will be manifested as line-shifts as the grazing angle or surface depth changes. Eqs. (4) and (5) were used to compute the corrected d-spacing taking into account the contribution of lineshift due to refraction. The reference dref and calculated dcorr values were used in conjunction with Eq. (6) to compute the amount of residual strains formed in the sample as a result of a large mismatch in the thermal expansion coefficient between that of alumina ( f 8 10 6/C) and AT ( f 1 10 6/C). Fig. 5 shows the plots of residual strain as a function of grazing-incidence angle. It is interesting to note that the variation of residual strains within the AT grains is nearly constant or depth-independent. In contrast, the residual strains within the alumina matrix are distinctly depth-dependent probably due to the graded distribution of AT in the sample.
4. Conclusion Near-surface depth profiling of phase composition and residual strains in a functionally graded alumina/
AT composite has been characterized using grazingincidence synchrotron radiation diffraction. Results show that the gradation of both composition and strains in the composite is continuous even at the nanometer scale. Depth profiling is possibly down to several thousand angstroms in steps of a few angstroms by changing the incidence angle. Above the critical angle, the depth of penetration is several thousand angstroms and depends on the wavelength as well as on the incident angle. The X-ray index of refraction shifts the peaks to higher Bragg angles and these shifts have been calculated using established model. Residual strains at different depths are calculated from line shift by comparing with d values of reference sample.
Acknowledgements The GISRD data for this work were acquired at the Australian National Beamline Facility at the Photon Factory, Tsukuba, Japan with financial support from the Australian Synchrotron Research Program (ANBF Proposal 99/2000-AB-26), which is funded by the Commonwealth of Australia under the major National Research Facilities Program. We thank Dr. Garry Foran and Dr. James Hester for technical assistance in data collection. Mr. S. Leung of ANSTO assisted with the work on scanning electron microscopy through funding from an AINSE grant.
References [1] I.M. Low, Mater. Res. Bull. 33 (10) (1998) 1475 – 1482. [2] I.M. Low, J. Aust. Ceram. Soc. 34 (2) (1998) 250. [3] I.M. Low, R.D. Skala, D. Asmi, P. Manurang, M. Singh, Proc. Int. Conf. on Powder Metallurgy (PM2000), 12 – 16 Nov. 2000, Kyoto, Japan, 2000, pp. 1469 – 1472. [4] D. Asmi, I.M. Low, S. Kennedy, R.A. Day, Mater. Lett. 40 (1999) 96 – 102. [5] R.D. Skala, PhD Thesis, Curtin University of Technology, Perth, Australia (2000). [6] D. Asmi, PhD Thesis, Curtin University of Technology, Perth, Australia (2001). [7] P. Manurung, PhD Thesis, Curtin University of Technology, Perth, Australia (2001). [8] S. Pratapa, I.M. Low, J. Mater. Sci. Lett. 15 (1996) 800 – 802. [9] S. Pratapa, MSc Thesis, Curtin University of Technology, Perth, Australia (1996). [10] B.R. Marple, D.J. Green, J. Mater. Sci. 28 (1993) 4637. [11] T. Wroblewski, Mater. Sci. Forum 79 (1991) 469 – 474.
M. Singh et al. / Materials Letters 55 (2002) 344–349 [12] R. Delhez, Th.H. Keijser, E.J. Mittemeijer, B.J. Thijsse, M.A. Hollanders, O.B. Loopstra, W.G. Sloof, Aust. J. Phys. 41 (1988) 261 – 282. [13] G. Lim, W. Parrish, C. Ortish, M. Belletto, M. Hart, J. Mater. Res. 2 (4) (1987) 471 – 477.
349
[14] W.C. Marra, P. Eisenberg, A.Y. Cho, J. Appl. Phys. 50 (1979) 697. [15] G.H. Vineyard, Phys. Rev. B 26 (8) (1982) 4146 – 4158. [16] ANBF local webpage XCOM, http://www.ansto.gov.au/natfac/ asrp1.html.