Availableonlineatwww.sciencedirect.com Science Direct E噩≈RS ELSEVIER Journal of the European Ceramic Society 29(2009)539-550 www.elsevier.comlocate/jeurceramsoc Development of microstructure during creep of polycrystalline mullite and a nanocomposite mullite/5 vol %o sic S Gustafsson a, L K L. Falk * J.E. Pitchford, W.J. Clegg ,E. Liden, E. Carlstromc Department of Applied Physics, Chalmers University of Technology SE-412 96 Goteborg, Sweden Metallurgy, University of Cambridge, e Swedish Ceramic Institute. Swerea /VE Box 104. SE-431 22 MoIndal. Sweden Received 25 June 2008: accepted 27 June 2008 Available online 21 August 2008 The microstructures of as-sintered and creep tested polycrystalline mullite and mullite reinforced with 5 vol %o nano-sized Sic particles have been characterized by scanning and transmission electron microscopy. The dislocation densities after tensile creep testing at 1300 and 1400C were virtually unchanged as compared to the as-sintered materials which indicates diffusion-controlled deformation. Mullite matrix grain boundaries bending around intergranular Sic particles suggest that grain boundary pinning, in addition to a reduced mullite grain size, contributed to the increased creep resistance of the mullite/5 vol. SiC nanocomposite. Both materials showed pronounced cavitation at multi-grain junctions after creep testing at 1400C which suggests that unaccommodated grain boundary sliding, facilitated by softening of the intergranular glass, occurred at this temperature. This is consistent with the higher stress exponents at 1400C C 2008 Elsevier Ltd. all rights reserved. Keywords: Mullite; Nanocomposites; Grain boundaries; Electron microscopy; Creep 1. Introduction not fully understood.. 5. It has been suggested that the improved creep resistance of alumina/SiC nanocomposites is caused by The incorporation of nano-sized second-phase ceramic parti- the thermal mismatch between alumina and SiC 5 Internal com- cles into a ceramic matrix may lead to significant improvements pressive stresses are introduced at the alumina/SiC interface, and in the mechanical properties. -Ohji et al. 2 found that the this results in a stronger particle/matrix bonding and thereby an creep rate of alumina reinforced with 17 vol %o SiC nanopar- improved creep resistance. Itis,therefore, of interest toevaluate ticles was three orders of magnitude lower, and the creep life 10 different matrix materials with different thermal expansion coef- times longer, than that of single-phase alumina. Alumina rein- ficients, and to characterize particle/matrix interface structures forced with 5 vol. SiC nanoparticles, studied by Thompson and properties et al., showed an increase in creep resistance that was sim- Mullite, 3Al2O32SiO2, is one potential matrix material in ilar to Ohji's results, but the lower fraction of Sic particles nanocomposite ceramics Mullite has excellent high temperature resulted in a reduced number of intergranular creep cavities and properties, and its creep resistance is high compared to other a much longer creep life. As shown in several studies, a smaller oxide ceramics 9-5 In the work by Lessing et al. it was shown volume fraction of nanoparticles, typically around 5 vol %, is that the creep rate of polycrystalline mullite at 1450.Cwas one often sufficient in order to give a substantial improvement in the order of magnitude lower than that of polycrystalline alumina mechanical properties. .3. of the same grain size. The thermal mismatch between mullite The mechanism behind the pronounced improvements with and SiC is. however. smaller than that between alumina and smaller additions of a nano-sized second phase is, however, st sC.16.17 < s The present paper is focussed on the relationship between he fine-scale micro- and nanostructure and the creep deforma- ing author. Tel. +4631 772 3321 tion process in polycrystalline mullite and mullite reinforced ss: Iklfalk (@chalmers. se(L K.L. Falk) with 5 vol. nano-sized SiC particles. The microstructures 0955-2219/S-see front matter o 2008 Elsevier Ltd. All rights reserved. doi: 10. 1016/j-jeurceramsoc. 2008.06.036
Available online at www.sciencedirect.com Journal of the European Ceramic Society 29 (2009) 539–550 Development of microstructure during creep of polycrystalline mullite and a nanocomposite mullite/5 vol.% SiC S. Gustafsson a, L.K.L. Falk a,∗, J.E. Pitchford b, W.J. Clegg b, E. Lidén c, E. Carlströmc a Department of Applied Physics, Chalmers University of Technology, SE-412 96 Göteborg, Sweden b Department of Materials Science and Metallurgy, University of Cambridge, Pembroke Street, CB2 3QZ Cambridge, UK c Swedish Ceramic Institute, Swerea IVF, Box 104, SE-431 22 Mölndal, Sweden Received 25 June 2008; accepted 27 June 2008 Available online 21 August 2008 Abstract The microstructures of as-sintered and creep tested polycrystalline mullite and mullite reinforced with 5 vol.% nano-sized SiC particles have been characterized by scanning and transmission electron microscopy. The dislocation densities after tensile creep testing at 1300 and 1400 ◦C were virtually unchanged as compared to the as-sintered materials which indicates diffusion-controlled deformation. Mullite matrix grain boundaries bending around intergranular SiC particles suggest that grain boundary pinning, in addition to a reduced mullite grain size, contributed to the increased creep resistance of the mullite/5 vol.% SiC nanocomposite. Both materials showed pronounced cavitation at multi-grain junctions after creep testing at 1400 ◦C which suggests that unaccommodated grain boundary sliding, facilitated by softening of the intergranular glass, occurred at this temperature. This is consistent with the higher stress exponents at 1400 ◦C. © 2008 Elsevier Ltd. All rights reserved. Keywords: Mullite; Nanocomposites; Grain boundaries; Electron microscopy; Creep 1. Introduction The incorporation of nano-sized second-phase ceramic particles into a ceramic matrix may lead to significant improvements in the mechanical properties.1–8 Ohji et al.2 found that the creep rate of alumina reinforced with 17 vol.% SiC nanoparticles was three orders of magnitude lower, and the creep life 10 times longer, than that of single-phase alumina. Alumina reinforced with 5 vol.% SiC nanoparticles, studied by Thompson et al.,3 showed an increase in creep resistance that was similar to Ohji’s results, but the lower fraction of SiC particles resulted in a reduced number of intergranular creep cavities and a much longer creep life. As shown in several studies, a smaller volume fraction of nanoparticles, typically around 5 vol.%, is often sufficient in order to give a substantial improvement in the mechanical properties.1,3,7,8 The mechanism behind the pronounced improvements with smaller additions of a nano-sized second phase is, however, still ∗ Corresponding author. Tel.: +46 31 772 3321. E-mail address: lklfalk@chalmers.se (L.K.L. Falk). not fully understood.3,5,8 It has been suggested that the improved creep resistance of alumina/SiC nanocomposites is caused by the thermal mismatch between alumina and SiC.5 Internal compressive stresses are introduced at the alumina/SiC interface, and this results in a stronger particle/matrix bonding and thereby an improved creep resistance.5 It is, therefore, of interest to evaluate different matrix materials with different thermal expansion coef- ficients, and to characterize particle/matrix interface structures and properties. Mullite, 3Al2O3·2SiO2, is one potential matrix material in nanocomposite ceramics. Mullite has excellent high temperature properties, and its creep resistance is high compared to other oxide ceramics.9–15 In the work by Lessing et al.9 it was shown that the creep rate of polycrystalline mullite at 1450 ◦C was one order of magnitude lower than that of polycrystalline alumina of the same grain size. The thermal mismatch between mullite and SiC is, however, smaller than that between alumina and SiC.16,17 The present paper is focussed on the relationship between the fine-scale micro- and nanostructure and the creep deformation process in polycrystalline mullite and mullite reinforced with 5 vol.% nano-sized SiC particles. The microstructures 0955-2219/$ – see front matter © 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jeurceramsoc.2008.06.036
S. gi Joumal of the European Ceramic Society 29(2009 )539-550 of as-fabricated and creep tested specimens were character- zed by scanning and transmission electron microscopy (SEM. mull e TEM), and particular attention was paid to the grain bound 106 ary regions and the location and size distribution of the Sic particles. The creep tests, and theoretical modelling for the prediction of the creep behaviour of these ceramics, have uFau 1400° previously been carried out by Clegg and co-workers. 8, I9 Results from that work, relevant to the electron microscopy ivestigation presented in this paper, are reviewed shortly 1300 2. Review of creep test and modelling results 1010 STRESS(MPa) Polycrystalline mullite, and mullite reinforced with 5 vol %o SiC nanoparticles (in the following termed"the nanocompos- mu‖tesc ite"), have been subjected to tensile creep tests in air at 1300 and 1400C under stresses between 10 and 50 MPa. 18, 19 The oretical modelling of diffusion-controlled creep deformation of these two materials was also carried out. 18,19 u 2.1. Polycrystalline mullite 1300°C Creep tests of the polycrystalline mullite material at 1300C showed a stress exponent of n=1. 2 which implies that diffu sion processes(n=1)are controlling the creep deformation at this temperature, see Fig. la Creep tests performed at 1400C resulted in a higher stress exponent of n=2(Fig. la), which STRESS(MPa suggests that, in addition to diffusion processes, other creep Fig. 1. Experimentally determined steady-state creep rates of(a)the polycrys mechanisms become active at this temperature. talline mullite and(b) the mullite/5 vol %o SiC at 1300 and 1400C plotted The experimentally determined creep rates were compared as function of stress. The testing conditions for the specimens subjected to with creep rates expected for diffusion-controlled creep as the microstructural characterization are marked by circles. Diffusional creep experimental materials, are also shown plotted. Data taken from Pitchtordto 1402 from mullite creep data presented in the literature. 9, 12, 13, 20These where o is the applied stress, s2 the volume of the rate- values were then used in an estimate of the creep rate interval of ontrolling diffusing species, k the Boltzmann constant and d is the mullite ceramic in the present investigation(d=1.5 um). 18 the grain size. Deff is the effective diffusion coefficient. related The plots in Fig. la, based on data taken from Pitchford, show to the diffusion coefficients for lattice diffusion Di and grain that the experimentally determined creep rates at 1300 C, and at boundary diffusion Db according to Deff= DI (2) 2.2. Mullite reinforced with 5 vol %o SiC nanoparticles where 8 is the grain boundary width. Maximum and minimum The experimental creep rates of the nanocomposite at 1300 values of Defm S2 at the two test temperatures were calculated and 1400C are plotted in Fig. Ib. The creep tests at 1400C Table I The as-sintered and creep tested materials Material Test temperature(°C) Stress(MPa) Steady-state creep rate(s-) Grain size(um) Polycrystalline mullite 15 15x 1400 1.2×10 5 1300 9.5×10-9 Mullite/SiC nanocomposite As-sintered 1.9x 1400 2.9×10 3.2×10 0.8
540 S. Gustafsson et al. / Journal of the European Ceramic Society 29 (2009) 539–550 of as-fabricated and creep tested specimens were characterized by scanning and transmission electron microscopy (SEM, TEM), and particular attention was paid to the grain boundary regions and the location and size distribution of the SiC particles. The creep tests, and theoretical modelling for the prediction of the creep behaviour of these ceramics, have previously been carried out by Clegg and co-workers.18,19 Results from that work, relevant to the electron microscopy investigation presented in this paper, are reviewed shortly below. 2. Review of creep test and modelling results Polycrystalline mullite, and mullite reinforced with 5 vol.% SiC nanoparticles (in the following termed “the nanocomposite”), have been subjected to tensile creep tests in air at 1300 and 1400 ◦C under stresses between 10 and 50 MPa.18,19 Theoretical modelling of diffusion-controlled creep deformation of these two materials was also carried out.18,19 2.1. Polycrystalline mullite Creep tests of the polycrystalline mullite material at 1300 ◦C showed a stress exponent of n = 1.2 which implies that diffusion processes (n = 1) are controlling the creep deformation at this temperature, see Fig. 1a. Creep tests performed at 1400 ◦C resulted in a higher stress exponent of n =2 (Fig. 1a), which suggests that, in addition to diffusion processes, other creep mechanisms become active at this temperature. The experimentally determined creep rates were compared with creep rates expected for diffusion-controlled creep as given by ε˙ = 14σΩ kTd2 Deff (1) where σ is the applied stress, Ω the volume of the ratecontrolling diffusing species, k the Boltzmann constant and d is the grain size. Deff is the effective diffusion coefficient, related to the diffusion coefficients for lattice diffusion Dl and grain boundary diffusion Db according to Deff = Dl + πδ d Db (2) where δ is the grain boundary width. Maximum and minimum values of DeffΩ at the two test temperatures were calculated Fig. 1. Experimentally determined steady-state creep rates of (a) the polycrystalline mullite and (b) the mullite/5 vol.% SiC at 1300 and 1400 ◦C plotted as function of stress. The testing conditions for the specimens subjected to the microstructural characterization are marked by circles. Diffusional creep rate intervals of polycrystalline mullite, predicted for the grain sizes of the two experimental materials, are also shown plotted. Data taken from Pitchford18. from mullite creep data presented in the literature.9,12,13,20 These values were then used in an estimate of the creep rate interval of the mullite ceramic in the present investigation (d = 1.5m).18 The plots in Fig. 1a, based on data taken from Pitchford,18 show that the experimentally determined creep rates at 1300 ◦C, and at higher stresses at 1400 ◦C, were higher than the predicted values. 2.2. Mullite reinforced with 5 vol.% SiC nanoparticles The experimental creep rates of the nanocomposite at 1300 and 1400 ◦C are plotted in Fig. 1b. The creep tests at 1400 ◦C Table 1 The as-sintered and creep tested materials Material Test temperature (◦C) Stress (MPa) Steady-state creep rate (s−1) Grain size (m) Polycrystalline mullite As-sintered 1.5 1400 48.6 1.5 × 10−6 1.3 1400 13.0 1.2 × 10−7 1.5 1300 14.9 9.5 × 10−9 1.5 Mullite/SiC nanocomposite As-sintered 0.7 1400 50.0 1.9 × 10−6 0.7 1400 12.1 2.9 × 10−8 0.8 1300 14.4 3.2 × 10−9 0.8
S Gustafsson et al. /Joumal of the European Ceramic Sociery 29(2009)539-550 8 gm( 2um Fig. 2. Thermally etched surfaces of the polycrystalline mullite in the(a)as-sintered condition, and after creep testing under a stress of ( b)48.6 MPa at 1400C,(c) 130MPat1400°C,and(d)149 MPa at1300°C showed that the stress exponent increased with increasing stress; 3. Experimental procedures om around n=1.5 at stresses under 25 MPa to around n=4 at stresses above 25 MPa. This implies that the total strain was not 3.1. Materials caused by one single creep mechanism The creep rate intervals for diffusion-controlled creep of 3.1.1. Polycrystalline mullite polycrystalline mullite with a reduced average grain size The polycrystalline mullite material was produced by mxIn =0.7 um, corresponding to the average matrix grain size commercially available 3: 2 mullite powder(KM-10l, Kyoritsu, of the nanocomposite)were calculated as described in Sec- Japan) and an ammonium polyacrylate dispersant (Dispex tion 2. 1. 8,9 This was done in order to better assess the Allied Colloids, England) in water. The slurry was ball milled effect of the SiC particles, and these creep rate intervals for 24 h using zirconia ball milling beads. Green bodies were re also shown in Fig. 1b. As illustrated in Fig. Ib, the produced by slip casting and pressureless sintered in air at experimental creep rates of the nanocomposite tested at low 1650C for 3 h. The material was 97% dense as measured by stresses(<30 MPa)at 1400C were lower than the pre- Archimedean densitometry dicted diffusion creep rates of polycrystalline mullite with this grain size. At higher stresses, however, the creep rate of the 3. 1.2. Mullite/Sic nanocomposite predicted by the diffusion- The a-SiC starting powder (UF-45, H.C. Starck, Germany) controlled creep model. The two data points from creep tests had a specific surface area of around 45 m-/g. The larger particles at 1300C were within the predicted diffusion creep rate inter- and agglomerates that were difficult to break down by milling powder were removed by sedimentation. This resulted in nanocomposite was determined not only by the reduced mullite particle size(dso)of 0.22 will starting powder that had a mean drive self-diffusion in the low diffusivity SiC particles, so that aqueous suspension containing 95 vol. of the mullite pm grain size. It has been suggested that the extra work required to The nanocomposite material was then produced from they can move with the grain boundaries during creep, will lead der, 5 vol. of the milled and fractionated a-SiC powder, to a reduced creep rate as compared to polycrystalline mullite and 0.3 wt% of an ammonium polyacrylate dispersant(Dura of the same grain size max 3021, Rohm and Haas, Sweden). The suspension was
S. Gustafsson et al. / Journal of the European Ceramic Society 29 (2009) 539–550 541 Fig. 2. Thermally etched surfaces of the polycrystalline mullite in the (a) as-sintered condition, and after creep testing under a stress of (b) 48.6 MPa at 1400 ◦C, (c) 13.0 MPa at 1400 ◦C, and (d) 14.9 MPa at 1300 ◦C. showed that the stress exponent increased with increasing stress; from around n = 1.5 at stresses under 25 MPa to around n = 4 at stresses above 25 MPa. This implies that the total strain was not caused by one single creep mechanism. The creep rate intervals for diffusion-controlled creep of polycrystalline mullite with a reduced average grain size (d = 0.7m, corresponding to the average matrix grain size of the nanocomposite) were calculated as described in Section 2.1. 18,19 This was done in order to better assess the effect of the SiC particles, and these creep rate intervals are also shown in Fig. 1b. As illustrated in Fig. 1b, the experimental creep rates of the nanocomposite tested at low stresses (<30 MPa) at 1400 ◦C were lower than the predicted diffusion creep rates of polycrystalline mullite with this grain size. At higher stresses, however, the creep rate of the nanocomposite was in the range predicted by the diffusioncontrolled creep model. The two data points from creep tests at 1300 ◦C were within the predicted diffusion creep rate interval. The data presented in Fig. 1b indicate that the creep rate of the nanocomposite was determined not only by the reduced mullite grain size. It has been suggested that the extra work required to drive self-diffusion in the low diffusivity SiC particles, so that they can move with the grain boundaries during creep, will lead to a reduced creep rate as compared to polycrystalline mullite of the same grain size.19 3. Experimental procedures 3.1. Materials 3.1.1. Polycrystalline mullite The polycrystalline mullite material was produced by mixing commercially available 3:2 mullite powder (KM-101, Kyoritsu, Japan) and an ammonium polyacrylate dispersant (Dispex, Allied Colloids, England) in water. The slurry was ball milled for 24 h using zirconia ball milling beads. Green bodies were produced by slip casting and pressureless sintered in air at 1650 ◦C for 3 h. The material was 97% dense as measured by Archimedean densitometry. 3.1.2. Mullite/SiC nanocomposite The -SiC starting powder (UF-45, H.C. Starck, Germany) had a specific surface area of around 45 m2/g. The larger particles and agglomerates that were difficult to break down by milling the powder were removed by sedimentation. This resulted in a milled and fractionated SiC starting powder that had a mean particle size (d50) of 0.22m. The nanocomposite material was then produced from an aqueous suspension containing 95 vol.% of the mullite powder, 5 vol.% of the milled and fractionated -SiC powder, and 0.3 wt% of an ammonium polyacrylate dispersant (Duramax 3021, Rohm and Haas, Sweden). The suspension was
S Gustafsson et al. / Journal of the European Ceramic Society 29(2009)539-550 P (ah 500nm 200nm 500nm Fig. 3. The microstructure of the as-sintered polycrystalline mullite (TEM). a) Intergranular porosity(P), and a dislocation pile-up at a grain boundary (arrowed ).(b) Dislocation network(arrowed) associated with cavities on a larger 25 nm homogenised by milling for I h in a planetary mill using Si3N4 balls whereafter 3 wt% of a polyethylene glycol binder was added to the slip. In order to retain a homogeneous distribution of(c) the Sic nanoparticles, the slip was freeze granulated by sprayin into liquid nitrogen. The ice was removed by sublimation using a freeze dryer and the resulting granules were hot pressed into plates at 1600C for I h in an argon atmosphere at a maximum pressure of 40 MPa. Hot pressing has been widely used for pro- ducing dense nanocomposite materials since the nanoparticles may suppress full densification 24.6 The nanocomposite mate- rial in the present study reached nearly full density, 99.8%,as measured by a helium pycnometer. 3. 2. Microstructural characterization and instrumentation 20nm The as-sintered and creep tested materials included in the microstructural characterization are shown in Table 1. Polished Fig. 4. Amorphous pockets at triple grain junctions and glassy grain bound- and thermally etched (45 min at 1300C in argon)specimens ary films in the as-sintered polycrystalline mullite (TEM).(a) Glass containing were imaged in a SEM(Leo ULTRA 55)equipped with a field triple grain junctions(arrowed).(b)Diffuse dark field image of a thin glassy emission gun(FEG)in order to assess grain size and overall grain boundary film(arrowed) merging into an amorphous pocket. The glass homogeneity. The average grain size was determined by the tion Fresnel fringes(arrowed) extending along the grain boundaries reveal the mean linear intercept method, and the average intercept length presence of thin intergranular films merging into a pocket at the triple grain was multiplied by a factor of 1.5
542 S. Gustafsson et al. / Journal of the European Ceramic Society 29 (2009) 539–550 Fig. 3. The microstructure of the as-sintered polycrystalline mullite (TEM). (a) Intergranular porosity (P), and a dislocation pile-up at a grain boundary (arrowed). (b) Dislocation network (arrowed) associated with cavities on a larger elongated grain section. homogenised by milling for 1 h in a planetary mill using Si3N4 balls whereafter 3 wt% of a polyethylene glycol binder was added to the slip. In order to retain a homogeneous distribution of the SiC nanoparticles, the slip was freeze granulated by spraying into liquid nitrogen. The ice was removed by sublimation using a freeze dryer and the resulting granules were hot pressed into plates at 1600 ◦C for 1 h in an argon atmosphere at a maximum pressure of 40 MPa. Hot pressing has been widely used for producing dense nanocomposite materials since the nanoparticles may suppress full densification.2–4,6 The nanocomposite material in the present study reached nearly full density, 99.8%, as measured by a helium pycnometer. 3.2. Microstructural characterization and instrumentation The as-sintered and creep tested materials included in the microstructural characterization are shown in Table 1. Polished and thermally etched (45 min at 1300 ◦C in argon) specimens were imaged in a SEM (Leo ULTRA 55) equipped with a field emission gun (FEG) in order to assess grain size and overall homogeneity. The average grain size was determined by the mean linear intercept method, and the average intercept length was multiplied by a factor of 1.5. Fig. 4. Amorphous pockets at triple grain junctions and glassy grain boundary films in the as-sintered polycrystalline mullite (TEM). (a) Glass containing triple grain junctions (arrowed). (b) Diffuse dark field image of a thin glassy grain boundary film (arrowed) merging into an amorphous pocket. The glass appears with bright contrast. (c) Defocus Fresnel image of a triple grain junction. Fresnel fringes (arrowed) extending along the grain boundaries reveal the presence of thin intergranular films merging into a pocket at the triple grain junction
S Gustafsson et al. /Joumal of the European Ceramic Sociery 29(2009)539-550 Thin-foil specimens for TEM were prepared by standard composition of mullite grains and amorphous grain boundary pecimen preparation techniques; mechanical grinding and pol- regions was determined by fine probe eDX point analysis. Ele ishing, dimpling and final ion milling to electron transparency. mental profiles acquired by the EDX system attached to the The microstructures were characterized in a Philips CM200 FEGTEM were used for the evaluation of peak areas in the FEGTEM equipped with a Link ISis energy dispersive X-ray quantification of the EDX spectra Due to the uncertainty in (EDX) system and a Gatan Imaging Filter(GIF). The chemical oxygen quantification, only the relative amounts of aluminum -600nm 0 nm nm 3 nn +600nm 3.5 1.5 1000 3 nI Defocus(nm) Fig. 5. Assessment of grain film thickness in the as-sintered polycrystalline mullite using the defocus Fresnel imaging technique. (a) The underfocussed mage shows a set of dark frin to the grain boundary.(b)At Gaussian focus there are no fringes. (c)The fringe contrast is reversed in the overfocussed image so that two the boundary. (d) By plotting the fringe spacing as a function of defocus and extrapolating the data to Gaussian focus, it is possible to estimate the film thickness to, in this case, approximately 0.75 nm
S. Gustafsson et al. / Journal of the European Ceramic Society 29 (2009) 539–550 543 Thin-foil specimens for TEM were prepared by standard specimen preparation techniques; mechanical grinding and polishing, dimpling and final ion milling to electron transparency. The microstructures were characterized in a Philips CM200 FEGTEM equipped with a Link ISIS energy dispersive X-ray (EDX) system and a Gatan Imaging Filter (GIF). The chemical composition of mullite grains and amorphous grain boundary regions was determined by fine probe EDX point analysis. Elemental profiles acquired by the EDX system attached to the FEGTEM were used for the evaluation of peak areas in the quantification of the EDX spectra. Due to the uncertainty in oxygen quantification, only the relative amounts of aluminum Fig. 5. Assessment of grain boundary film thickness in the as-sintered polycrystalline mullite using the defocus Fresnel imaging technique. (a) The underfocussed image shows a set of dark fringes parallel to the grain boundary. (b) At Gaussian focus there are no fringes. (c) The fringe contrast is reversed in the overfocussed image so that two bright lines delineate the boundary. (d) By plotting the fringe spacing as a function of defocus and extrapolating the data to Gaussian focus, it is possible to estimate the film thickness to, in this case, approximately 0.75 nm
S Gustafsson et al. /Journal of the European Ceramic Sociery 29(2009)539-550 and silicon were determined. These were then converted into the but a limited number of larger and elongated sections was also equivalent mol fractions of alumina and silica. observed on the etched surfaces. The average grain size was Grain boundary films and amorphous regions were imaged in determined to 1. 5 um. A smaller amount of residual porosity diffuse dark field, and the presence of grain boundary films was was observed throughout the microstructure, consistent with the also established by defocus Fresnel imaging of edge-on grain density measurement (97% dense), see Fig. 3. Intergranular. boundaries.21-25The defocus Fresnel technique was used in esti- irregularly shaped, cavities with sizes in the range 0. 1-1 um mating the thickness of the amorphous grain boundary film in were present at some multi-grain junctions, while the grain around 20 randomly chosen grain boundaries in the as-sintered boundaries were free of cavities materials and in the mullite and nanocomposite specimens crept Largerelongated grain sections often contained faceted intra- at 1400C under a stress of 48.6 and 50.0 MPa, respectively. granular cavities, 50-500 nm in size, but only a limited number The grain boundaries were oriented in the edge-on position by of the equiaxed grain sections showed cavities. The intragranular oking at the reflection symmetry of the fringe intensity across cavities were often associated with single dislocations, or an odd the boundary. Through-focus series were then recorded and pro- dislocation network, as shown in Fig. 3b. The overall dislocation cessed with the Gatan DigitalMicrograph software. An intensity density in the microstructure was, however, low. Only occa profile across a boundary was obtained by integrating the image sional dislocation structures, such as pile-ups at grain boundaries over a distance of 10-15 nm along the boundary. The fringe sepa-(Fig 3)and low angle grain boundaries, were observed ration was then determined from the intensity profiles and plotted Thin glassy grain boundary films merging into amorphous as a function of defocus. The film thickness was estimated by pockets at triple grain junctions were present throughout the xtrapolating the data to Gaussian focus microstructure, see Fig 4. The amorphous triple grain junctions d a diameter(corresponding to the diameter of a circle of equivalent area)in the range 30-70 nm. All analysed amorphous grain boundary films were found to be in the range 0.6-0.9nm 4.1. The as-sintered polycrystalline mullite thick. The result of a TEM through-focus series is shown in The general microstructure of the as-sintered mullite material The mullite grains did not have a perfect 3: 2 mullite compo- is shown in Figs 2a and 3. Most grain sections were equiaxed, sition; the mol fraction Al2O3 was 57.6+2.0%, slightly lower 多 (c) Fig. 6. Thermally etched surfaces of the nanocomposite in the(a)as-sintered condition, and after creep testing under a stress of(b)50.0 MPa at 1400C, (c)12.1 MPa at1400°C,and(d)144 MPa at1300°C
544 S. Gustafsson et al. / Journal of the European Ceramic Society 29 (2009) 539–550 and silicon were determined. These were then converted into the equivalent mol fractions of alumina and silica. Grain boundary films and amorphous regions were imaged in diffuse dark field, and the presence of grain boundary films was also established by defocus Fresnel imaging of edge-on grain boundaries.21–25 The defocus Fresnel technique was used in estimating the thickness of the amorphous grain boundary film in around 20 randomly chosen grain boundaries in the as-sintered materials and in the mullite and nanocomposite specimens crept at 1400 ◦C under a stress of 48.6 and 50.0 MPa, respectively. The grain boundaries were oriented in the edge-on position by looking at the reflection symmetry of the fringe intensity across the boundary. Through-focus series were then recorded and processed with the Gatan DigitalMicrograph software. An intensity profile across a boundary was obtained by integrating the image over a distance of 10–15 nm along the boundary. The fringe separation was then determined from the intensity profiles and plotted as a function of defocus. The film thickness was estimated by extrapolating the data to Gaussian focus. 4. Results 4.1. The as-sintered polycrystalline mullite The general microstructure of the as-sintered mullite material is shown in Figs. 2a and 3. Most grain sections were equiaxed, but a limited number of larger and elongated sections was also observed on the etched surfaces. The average grain size was determined to 1.5m. A smaller amount of residual porosity was observed throughout the microstructure, consistent with the density measurement (97% dense), see Fig. 3. Intergranular, irregularly shaped, cavities with sizes in the range 0.1–1 m were present at some multi-grain junctions, while the grain boundaries were free of cavities. Larger elongated grain sections often contained faceted intragranular cavities, 50–500 nm in size, but only a limited number of the equiaxed grain sections showed cavities. The intragranular cavities were often associated with single dislocations, or an odd dislocation network, as shown in Fig. 3b. The overall dislocation density in the microstructure was, however, low. Only occasional dislocation structures, such as pile-ups at grain boundaries (Fig. 3) and low angle grain boundaries, were observed. Thin glassy grain boundary films merging into amorphous pockets at triple grain junctions were present throughout the microstructure, see Fig. 4. The amorphous triple grain junctions had a diameter (corresponding to the diameter of a circle of equivalent area) in the range 30–70 nm. All analysed amorphous grain boundary films were found to be in the range 0.6–0.9 nm thick. The result of a TEM through-focus series is shown in Fig. 5. The mullite grains did not have a perfect 3:2 mullite composition; the mol fraction Al2O3 was 57.6 ± 2.0%, slightly lower Fig. 6. Thermally etched surfaces of the nanocomposite in the (a) as-sintered condition, and after creep testing under a stress of (b) 50.0 MPa at 1400 ◦C, (c) 12.1 MPa at 1400 ◦C, and (d) 14.4 MPa at 1300 ◦C.
S Gustafsson et al. /Joumal of the European Ceramic Sociery 29(2009)539-550 than the expected 60%0. The glassy pockets at multi-grain junc- rich in silicon. Impurities were not detected in any of these tions were rich in silicon. The mol fraction SiO2 in the glass was analyses 93.0+3.2%. Some of the analysed glassy pockets showed an alumina content close to the eutectic Al2O3-SiO2 composition 4.2. The as-sintered mullite/SiC nanocomposite of approximately 5 mol%Al2O3, reported in several studies. 26,2 EDX also showed that the amorphous grain boundary films were The microstructure of the as-sintered nanocomposite material shown in Figs 6a and 7. Most grain sections were equiaxed (a) and an average matrix grain size of 0.7 um was determined from SEMimages(Figs 6 and 7a). The1 granular cavities also in this microstructure, and these cavities were usually faceted and less than 100 nm in diameter(Fig. 7b) The majority(around 80%)of the Sic particles were located at grain boundaries and multi-grain junctions, see Fig. 7. The siz of these particles was in the range 30-90 nm, and a number of them formed agglomerates as shown in Fig. 7. Particle agglom- erates present in multi-grain junctions were generally associated with cavities, see Fig 7c. The intragranular SiC particles did not form clusters, and were smaller, typically 10-50 nm, see Fig. 7a. Strain contours were only occasionally observed in the mullite matrix immediately surrounding a SiC particle 200nm Some dislocations were pinned by, or emerging from, the rity of the observed dislocations were. as in the mullite microstructure associated with intragranular cavities, see Fig. 7b. Although the overall dislocation density was low, dislocation pile-ups at triple grain junctions or at grain boundaries were observed in a few areas Thin glassy films were observed in the matrix grain bound- aries along with smaller volumes of glass separating both intragranular and intergranular SiC particles from the surround ing mullite grains, see Figs. 8 and 9. The thickness of the amorphous grain boundary films varied between 0.6 and 0.9nm Very small (10-15 nm) amorphous pockets were observed at some multi-grain junctions, see Fig 8. The Al?O3 mol f 100nm to 55.9+1.5%,1.e also in this case lower than for 3: 2 mullite The intergranular glass was enriched in silicon, although the 200nm 50 nm Fig. 7. The microstructure of the as-sintered nanocomposite (TEM).(a)SiC par- ticles present in both inter- and intragranular positions(black and white arrows, Fig. 8. Thin glassy grain boundary films(white arrows) merging into a glas respectively).(b) Dislocations pinned by intragranular cavities(C)and SiC par- containing triple grain junction(black arrow) in the as-sintered nanocomposite ticles(P).(c)Clusters of SiC particles(arrowed), associated with intergranular material. The glass appears with bright contrast in the diffuse TEM dark field
S. Gustafsson et al. / Journal of the European Ceramic Society 29 (2009) 539–550 545 than the expected 60%. The glassy pockets at multi-grain junctions were rich in silicon. The mol fraction SiO2 in the glass was 93.0 ± 3.2%. Some of the analysed glassy pockets showed an alumina content close to the eutectic Al2O3–SiO2 composition of approximately 5 mol% Al2O3, reported in several studies.26,27 EDX also showed that the amorphous grain boundary films were Fig. 7. The microstructure of the as-sintered nanocomposite (TEM). (a) SiC particles present in both inter- and intragranular positions (black and white arrows, respectively). (b) Dislocations pinned by intragranular cavities (C) and SiC particles (P). (c) Clusters of SiC particles (arrowed), associated with intergranular porosity. rich in silicon. Impurities were not detected in any of these analyses. 4.2. The as-sintered mullite/SiC nanocomposite The microstructure of the as-sintered nanocomposite material is shown in Figs. 6a and 7. Most grain sections were equiaxed, and an average matrix grain size of 0.7 m was determined from SEM images (Figs. 6 and 7a). The mullite grains contained intragranular cavities also in this microstructure, and these cavities were usually faceted and less than 100 nm in diameter (Fig. 7b). The majority (around 80%) of the SiC particles were located at grain boundaries and multi-grain junctions, see Fig. 7. The size of these particles was in the range 30–90 nm, and a number of them formed agglomerates as shown in Fig. 7. Particle agglomerates present in multi-grain junctions were generally associated with cavities, see Fig. 7c. The intragranular SiC particles did not form clusters, and were smaller, typically 10–50 nm, see Fig. 7a. Strain contours were only occasionally observed in the mullite matrix immediately surrounding a SiC particle. Some dislocations were pinned by, or emerging from, the intra- and intergranular SiC particles, but the majority of the observed dislocations were, as in the mullite microstructure, associated with intragranular cavities, see Fig. 7b. Although the overall dislocation density was low, dislocation pile-ups at triple grain junctions or at grain boundaries were observed in a few areas. Thin glassy films were observed in the matrix grain boundaries along with smaller volumes of glass separating both intragranular and intergranular SiC particles from the surrounding mullite grains, see Figs. 8 and 9. The thickness of the amorphous grain boundary films varied between 0.6 and 0.9 nm. Very small (∼10–15 nm) amorphous pockets were observed at some multi-grain junctions, see Fig. 8. The Al2O3 mol fraction of the mullite matrix was determined to 55.9 ± 1.5%, i.e. also in this case lower than for 3:2 mullite. The intergranular glass was enriched in silicon, although the Fig. 8. Thin glassy grain boundary films (white arrows) merging into a glass containing triple grain junction (black arrow) in the as-sintered nanocomposite material. The glass appears with bright contrast in the diffuse TEM dark field image.
S Gustafsson et al. / Journal of the European Ceramic Society 29(2009)539-550 SiC 500nm Fig. 10. A grain showing strain contours(arrowed) in the TEM thin-foil speci- men of the polycrystalline mullite crept under a stress of 13.0 MPa at 1400C 25 nm Strain contours were generally observed in the TEM bright field images of the microstructures of the crept mullite speci mens as shown in Figs. 10 and 11. These contours reflect the stresses that built up in the material during creep deformation An increased cavitation at multi-grain junctions was observed in some areas of the crept microstructures, see Fig. 11. This type of cavitation was most pronounced in the specimen that had been subjected to the highest stress (48.6 MPa) at 1400C, but so evident in th SiC at 1300C. The microstructure of the specimen crept under a stress of 13.0 MPa at 1400C resembled. on the other hand. that of the as-sintered material, with only a limited increase in the density of cavi A number of the intragranular cavities changed shape or size during creep testing, and in some cases thin channels connecting Sic nearby cavities had developed, see Fig. 12. The size of these cavities ranged from several hundred nanometers up to 1 um. The overall dislocation density was low also creep tested mullite specimens. Dislocations were present 25 nm Fig 9. Thin amorphous films separating intragranular SiC particles from the surrounding mullite grain in the as-sintered nanocomposite. (a) Defocus Fres- nel fringes(arrowed ).(b) Diffuse dark field image. The glassy films(arrowed) appear with bright contrast. glass volumes in the thin-foil TEM specimens were too small for quantitative analysis 4.3. The polycrystalline mullite after creep testing 200nm The average grain sizes of the creep tested mullite specimens were virtually the same as that of the as-sintered material, Fig. 11. Cavity formation in a multi-grain junction in the polycrystalline mullite Table 1. This indicates that grain growth was not significa creep tested under a stress of 48.6 MPa at 1400C. One of the grains shows strain during creep testing, see Fig. 2
546 S. Gustafsson et al. / Journal of the European Ceramic Society 29 (2009) 539–550 Fig. 9. Thin amorphous films separating intragranular SiC particles from the surrounding mullite grain in the as-sintered nanocomposite. (a) Defocus Fresnel fringes (arrowed). (b) Diffuse dark field image. The glassy films (arrowed) appear with bright contrast. glass volumes in the thin-foil TEM specimens were too small for quantitative analysis. 4.3. The polycrystalline mullite after creep testing The average grain sizes of the creep tested mullite specimens were virtually the same as that of the as-sintered material, see Table 1. This indicates that grain growth was not significant during creep testing, see Fig. 2. Fig. 10. A grain showing strain contours (arrowed) in the TEM thin-foil specimen of the polycrystalline mullite crept under a stress of 13.0 MPa at 1400 ◦C. Strain contours were generally observed in the TEM bright field images of the microstructures of the crept mullite specimens as shown in Figs. 10 and 11. These contours reflect the stresses that built up in the material during creep deformation. An increased cavitation at multi-grain junctions was observed in some areas of the crept microstructures, see Fig. 11. This type of cavitation was most pronounced in the specimen that had been subjected to the highest stress (48.6 MPa) at 1400 ◦C, but also evident in the specimen crept under a stress of 14.9 MPa at 1300 ◦C. The microstructure of the specimen crept under a stress of 13.0 MPa at 1400 ◦C resembled, on the other hand, that of the as-sintered material, with only a limited increase in the density of cavities. A number of the intragranular cavities changed shape or size during creep testing, and in some cases thin channels connecting nearby cavities had developed, see Fig. 12. The size of these cavities ranged from several hundred nanometers up to 1 m. The overall dislocation density was low also in the creep tested mullite specimens. Dislocations were present Fig. 11. Cavity formation in a multi-grain junction in the polycrystalline mullite creep tested under a stress of 48.6 MPa at 1400 ◦C. One of the grains shows strain contours (arrowed).
S Gustafsson et aL. /Joumal of the European Ceramic Sociery 29(2009)539-550 400nm 200nm (a) Fig. 12. Intragranular cavities in the polycrystalline mullite crept under a stress of 14.9 MPa at 1300C. Thin channels(arrowed) are connecting, or extending out from. the enlarged cavities predominantly in the larger grain sections, and in general asso- ciated with intragranular cavities, as in the as-sintered material Glass pockets at multi-grain junctions and thin amorphous grain boundary films were present also in the crept microstruc tures. Measurements of the film thickness in the sample crept at 1400C under a stress of 48.6 MPa did not reveal any pronounced changes in the thickness as compared to the as- sintered mullite material; the films had a thickness of, typically, 0.6-0.9 nm also after creep testing 4.4. The nanocomposite after creep testing 200nm Grain size measurements on the creep tested mullite Fig growth during creep deformation, see Table 1 and Fig. 6. The 50.0 MPa at 14002 the nanocomposite after creep testing under a stress of nanocomposite specimens did not show any evidence of grain grain junctions (b) of location and distribution of the SiC particles was also appar ently unchanged, see Fig. 13. Around 80% of the Sic particles were located at the grain boundaries and multi-grain junctions in the specimen that had been crept at 1400C under a stress of 50.0 MPa. The mullite/mullite grain boundaries were often observed to bend near the intergranular SiC particles, see Fig. I 200mm An increased number of the multi-grain junctions, pre- dominantly junctions containing intergranular Sic particles, contained cavities after creep testing. This was particularly pro- nounced in the specimen crept under the highest stress, 50 MPa, at 1400 C, see Fig. 13b. Cavitation associated with the Sic par- ticles at the grain boundaries was rarely observed. The cavities (a) inside the mullite matrix grains seemed to have retained their ize and shape during creep testing. (200nm The dislocation densities in the creep tested nanocompos- e specimens were low and seemingly unchanged as compared to the as-sintered material. A limited number of grains with a locally increased dislocation density were observed in the crept microstructures. but these areas were not different from similar areas in the as-sintered nanocomposite microstructure. Thin intergranular glassy films were present also after creep testing Measurements of the grain boundary film thickness in Fig 14 Mullite matrix grain boundaries bending(arrowed)around intergranular the sample crept at 1400C under a stress of 50 MPa, showed SiC particles after creep testing under a stress of 50.0 MPa at 1400oC
S. Gustafsson et al. / Journal of the European Ceramic Society 29 (2009) 539–550 547 Fig. 12. Intragranular cavities in the polycrystalline mullite crept under a stress of 14.9 MPa at 1300 ◦C. Thin channels (arrowed) are connecting, or extending out from, the enlarged cavities. predominantly in the larger grain sections, and in general associated with intragranular cavities, as in the as-sintered material. Glass pockets at multi-grain junctions and thin amorphous grain boundary films were present also in the crept microstructures. Measurements of the film thickness in the sample crept at 1400 ◦C under a stress of 48.6 MPa did not reveal any pronounced changes in the thickness as compared to the assintered mullite material; the films had a thickness of, typically, 0.6–0.9 nm also after creep testing. 4.4. The nanocomposite after creep testing Grain size measurements on the creep tested mullite nanocomposite specimens did not show any evidence of grain growth during creep deformation, see Table 1 and Fig. 6. The location and distribution of the SiC particles was also apparently unchanged, see Fig. 13. Around 80% of the SiC particles were located at the grain boundaries and multi-grain junctions in the specimen that had been crept at 1400 ◦C under a stress of 50.0 MPa. The mullite/mullite grain boundaries were often observed to bend near the intergranular SiC particles, see Fig. 14. An increased number of the multi-grain junctions, predominantly junctions containing intergranular SiC particles, contained cavities after creep testing. This was particularly pronounced in the specimen crept under the highest stress, 50 MPa, at 1400 ◦C, see Fig. 13b. Cavitation associated with the SiC particles at the grain boundaries was rarely observed. The cavities inside the mullite matrix grains seemed to have retained their size and shape during creep testing. The dislocation densities in the creep tested nanocomposite specimens were low and seemingly unchanged as compared to the as-sintered material. A limited number of grains with a locally increased dislocation density were observed in the crept microstructures, but these areas were not different from similar areas in the as-sintered nanocomposite microstructure. Thin intergranular glassy films were present also after creep testing. Measurements of the grain boundary film thickness in the sample crept at 1400 ◦C under a stress of 50 MPa, showed Fig. 13. The general microstructure (a) and cavity formation (arrowed) at multigrain junctions (b) of the nanocomposite after creep testing under a stress of 50.0 MPa at 1400 ◦C. Fig. 14. Mullite matrix grain boundaries bending (arrowed) around intergranular SiC particles after creep testing under a stress of 50.0 MPa at 1400 ◦C
S. gi Joumal of the European Ceramic Society 29(2009 )539-550 film widths in the range 0.6-0.9nm. The creep deforma and accumulated in grain boundaries under tension. 3 The mea- process, did, hence, not have any pronounced effect on surements of intergranular film widths presented in this paper distribution of the intergranular glass do not, however, show any clear evidence of glass redistribution during these creep tests. The thickness of the investigated grain 5. Discussion boundary films was in the range 0.6-0.9 nm in both as-sintered and crept specimens, and a possible redistribution might, there- 5.1. The polycrystalline mullite fore, be difficult to detect. The virtually unchanged low dislocation densities in the 5.2. The mullite/SiC nanocomposite crept mullite specimens strongly suggest that dislocation glide was not active to any significant extent under the applied The addition of Sic nanoparticles resulted in a reduced testing conditions. This is in agreement with previous exper- mullite grain size (Table 1). This indicates that the nanopar- imental studies which indicated a very limited (if any at all) ticles suppressed grain growth during sintering through grain dislocation mobility during plastic deformation of both single boundary pinning. The SiC particles located at the grain bound crystal and polycrystalline mullite specimens. 10, 12, 13,28A previ- aries were larger(30-90 nm)than the intragranular particles ous TEM investigation of dislocations in mullite by Gustafsson (10-50 nm). A critical Sic particle size for effective mullite and Falk29,30 revealed comparatively large Burgers vectors of grain boundary pinning would, hence, be in the range 30-50nm the type b=, , and . This may, during hot pressing under a pressure of 40 MPa at 1600C together with the complex mullite crystal structure, result in a Strain contours were generally not observed in the mullite limited dislocation activity. matrix around the Sic particles in the as-sintered specimen. This The changes observed in the microstructures of the crept mul- indicates that the thermal expansion mismatch between mul- lite specimens indicate that lattice diffusion is an active creep lite(a=5.3 x 10-6oC-)6 and SiC (a=4. 7 x 10-6oC-)7is deformation mechanism, both at 1300 and 1400C. The elon- too low for the introduction of compressive residual stresses gation and enlargement of some intragranular cavities, and the of any significant magnitude at the SiC/mullite interface. In development of thin intragranular cavity channels, indicate dif- addition, the analysed Sic/mullite interfaces contained amor fusion activity within the grains, and thus a contribution from phous films or pockets(Fig 9). The SiC/mullite interfaces are, Nabarro-Herring creep. In addition, the presence of a continuous hence, not likely to be significantly more rigid than the glass intergranular glassy phase at the grain boundaries(Figs. 4 and 5) containing mullite/mullite grain boundaries. This is in contrast would provide rapid diffusion paths and thereby promote creep to SiC reinforced alumina nanocomposites where the thermal deformation by grain boundary diffusion. The intergranular expansion mismatch between alumina(a=88x 10-6oC-)7 glass may, hence, contribute to an increased creep rate as com- and Sic puts the particle/matrix interface under compression pared to the model discussed in Section 2, see Fig. la during cooling from the sintering temperature. It has been pro- Creep deformation by diffusion, either through the grains or posed that this leads to a more rigid interface bonding and through the grain boundary glass, would give a stress exponent thereby to a reduced creep rate due to the suppression of of n=l. This is in good agreement with the stress exponent of the nucleation and annihilation of point defects during creep n=1. 2 that was determined from the creep tests performed at deformation. The proposed rigid interface bonding was sup- 1300.C8 Solution-reprecipitation creep, limited by the trans- ported by theoretical calculations and TEM observations of glass port of matter through the glassy grain boundary films, is also free SiC/alumina interfaces. This interface mechanism would, expected to give a stress exponent of n=l. There is, however, hence, not be active to any significant extent in the present no microstructural evidence for such a process mulliteSiC nanocomposite material Diffusion creep only would, however, not result in the The fraction of Sic particles located at mullite grain bound stress exponent n=2 determined from the creep tests carried aries and multi-grain junctions was unchanged(80%)in the out at 1400C.A large number of cavities had formed at specimen that had been creep tested under a stress of 50.0 MPa multi-grain junctions in these specimens(Fig. 11), and the sur- at 1400C. This suggests that moving grain boundaries dragged rounding grains showed strain contrast(Figs. 10 and 11). This the SiC particles during creep deformation. This process requires was most pronounced in the specimen that had been tested at a directional flow of atoms from one side of the particle to the 486MPa. These observations suggest that rigid grain bound- other. 3 The diffusional migration may occur along three dif- ary sliding, facilitated by softening of the intergranular glass, ferent paths: through the SiC particle, in the thin glassy film contributed to the strain during creep testing. Such a process separating the SiC particle from the mullite matrix or around the values of n>1. A stress exponent close to n=2, and evidence interfacial reactions, may be rate controlling 33nprocesses, or would increase the stress exponent, since cavitation alone gives particle through the matrix. One of these diffusion processes,or of cavitation and grain boundary sliding, have been observed A model proposed by Clegg and co-workers 8, suggests that previously in other creep studies of glass containing mullite the creep rate would be limited by self-diffusion through the low diffusivity SiC particles. It was assumed that the particles would It may be expected that the amorphous grain boundary phase move with a velocity equal to that of the grain boundary at which would redistribute during rigid grain sliding because the glass they are situated. It was also assumed that the global defor- may be squeezed out from grain boundaries under compression mation of the body would be caused by self-diffusion within
548 S. Gustafsson et al. / Journal of the European Ceramic Society 29 (2009) 539–550 film widths in the range 0.6–0.9 nm. The creep deformation process, did, hence, not have any pronounced effect on the distribution of the intergranular glass. 5. Discussion 5.1. The polycrystalline mullite The virtually unchanged low dislocation densities in the crept mullite specimens strongly suggest that dislocation glide was not active to any significant extent under the applied testing conditions. This is in agreement with previous experimental studies which indicated a very limited (if any at all) dislocation mobility during plastic deformation of both single crystal and polycrystalline mullite specimens.10,12,13,28 A previous TEM investigation of dislocations in mullite by Gustafsson and Falk29,30 revealed comparatively large Burgers vectors of the type b = , , and . This may, together with the complex mullite crystal structure, result in a limited dislocation activity.12 The changes observed in the microstructures of the crept mullite specimens indicate that lattice diffusion is an active creep deformation mechanism, both at 1300 and 1400 ◦C. The elongation and enlargement of some intragranular cavities, and the development of thin intragranular cavity channels, indicate diffusion activity within the grains, and thus a contribution from Nabarro-Herring creep. In addition, the presence of a continuous intergranular glassy phase at the grain boundaries (Figs. 4 and 5) would provide rapid diffusion paths and thereby promote creep deformation by grain boundary diffusion. The intergranular glass may, hence, contribute to an increased creep rate as compared to the model discussed in Section 2, see Fig. 1a. Creep deformation by diffusion, either through the grains or through the grain boundary glass, would give a stress exponent of n = 1. This is in good agreement with the stress exponent of n = 1.2 that was determined from the creep tests performed at 1300 ◦C.18 Solution-reprecipitation creep, limited by the transport of matter through the glassy grain boundary films, is also expected to give a stress exponent of n = 1. There is, however, no microstructural evidence for such a process. Diffusion creep only would, however, not result in the stress exponent n = 2 determined from the creep tests carried out at 1400 ◦C.18 A large number of cavities had formed at multi-grain junctions in these specimens (Fig. 11), and the surrounding grains showed strain contrast (Figs. 10 and 11). This was most pronounced in the specimen that had been tested at 48.6 MPa. These observations suggest that rigid grain boundary sliding, facilitated by softening of the intergranular glass, contributed to the strain during creep testing. Such a process would increase the stress exponent, since cavitation alone gives values of n > 1.31 A stress exponent close to n = 2, and evidence of cavitation and grain boundary sliding, have been observed previously in other creep studies of glass containing mullite ceramics.10,12,13 It may be expected that the amorphous grain boundary phase would redistribute during rigid grain sliding because the glass may be squeezed out from grain boundaries under compression and accumulated in grain boundaries under tension.32 The measurements of intergranular film widths presented in this paper do not, however, show any clear evidence of glass redistribution during these creep tests. The thickness of the investigated grain boundary films was in the range 0.6–0.9 nm in both as-sintered and crept specimens, and a possible redistribution might, therefore, be difficult to detect. 5.2. The mullite/SiC nanocomposite The addition of SiC nanoparticles resulted in a reduced mullite grain size (Table 1). This indicates that the nanoparticles suppressed grain growth during sintering through grain boundary pinning. The SiC particles located at the grain boundaries were larger (30–90 nm) than the intragranular particles (10–50 nm). A critical SiC particle size for effective mullite grain boundary pinning would, hence, be in the range 30–50 nm during hot pressing under a pressure of 40 MPa at 1600 ◦C. Strain contours were generally not observed in the mullite matrix around the SiC particles in the as-sintered specimen. This indicates that the thermal expansion mismatch between mullite (α = 5.3 × 10−6 ◦C−1) 16 and SiC (α = 4.7 × 10−6 ◦C−1) 17 is too low for the introduction of compressive residual stresses of any significant magnitude at the SiC/mullite interface. In addition, the analysed SiC/mullite interfaces contained amorphous films or pockets (Fig. 9). The SiC/mullite interfaces are, hence, not likely to be significantly more rigid than the glass containing mullite/mullite grain boundaries. This is in contrast to SiC reinforced alumina nanocomposites where the thermal expansion mismatch between alumina (α = 8.8 × 10−6 ◦C−1) 17 and SiC puts the particle/matrix interface under compression during cooling from the sintering temperature. It has been proposed that this leads to a more rigid interface bonding and thereby to a reduced creep rate due to the suppression of the nucleation and annihilation of point defects during creep deformation.5 The proposed rigid interface bonding was supported by theoretical calculations and TEM observations of glass free SiC/alumina interfaces.5 This interface mechanism would, hence, not be active to any significant extent in the present mullite/SiC nanocomposite material. The fraction of SiC particles located at mullite grain boundaries and multi-grain junctions was unchanged (80%) in the specimen that had been creep tested under a stress of 50.0 MPa at 1400 ◦C. This suggests that moving grain boundaries dragged the SiC particles during creep deformation. This process requires a directional flow of atoms from one side of the particle to the other.33 The diffusional migration may occur along three different paths: through the SiC particle, in the thin glassy film separating the SiC particle from the mullite matrix or around the particle through the matrix. One of these diffusion processes, or interfacial reactions, may be rate controlling.33 A model proposed by Clegg and co-workers18,19 suggests that the creep rate would be limited by self-diffusion through the low diffusivity SiC particles. It was assumed that the particles would move with a velocity equal to that of the grain boundary at which they are situated. It was also assumed that the global deformation of the body would be caused by self-diffusion within