REFRACTORY METALS HARD MATERIALS ELSEVIER International Journal of Refractory Metals Hard Materials 19(2001)425-435 www.elsevier.com/locate/ijrmhm Processing and microstructure of Sic laminar composites C. Reynaud F. Thevenot a,, T. Chartier b Dept Ceramiques Speciales, Ecole Nationale Superieure des mines de St-Etienne, 42023, Saint-Etienne Cedex 2, france S.P. C.T.S., U.M.R CNR. 6638,ENS. de Ceramiques Industrielles, 87065, Limoges, france Received 5 March 2001; accepted 14 August 2001 Abstract Porous laminar materials and alternate laminates of dense and porous layers in silicon carbide have been elaborated by tape casting and liquid phase sintering (YAG-alumina eutectic)processing. Porosity was introduced by the incorporation of pore forming agents(PFA)(5-50 vol% )in the slurry. Two types of PFA with a narrow size distribution have been used(corn starch and polyamide powders). The effects of size, content, type of PFA on tape casting processing, sintering characteristics, porosity control and microstructure are investigated. For each PFA, the porosity attains a maximum value dependent on the PFA nature(41 vol% with corn starch). Only for starch, the volumetric shrinkage was unaffected by the PFA content up to 45 vol%. Homogeneou distribution of porosity has been obtained for both monolithic and composite laminates. An equiaxed and homogeneous silicon carbide microstructure has been obtained and was unaffected by PFA. Layered structures without defects have been obtained with parallel layers and uniform thickness(dense layer: 70 um; porous layer: 80 um).@ 2001 Published by Elsevier Science Ltd. Keywords: Alternate laminate; Pore forming agent; SiC; Liquid phase sintering: Tape casting 1. ntroduction stresses introduced during cooling, and may be better resistant to oxidation in air compared to graphite(de- Fracture toughness of ceramic materials can be im- pending on the intrinsic ceramic properties). A study proved by designing laminar structures with weak in- performed on solid phase sintered silicon carbide/porous terlayers that promote crack-deflection mechanisms and silicon carbide laminar composites by Blanks et al. [3] hence increase fracture energy. Clegg et al. [l] have has shown the porosity level required to ensure reliably elaborated laminar materials in alpha silicon carbide crack deflection is 37 vol%. Porous interlayers appear with graphite interfaces with an apparent toughness and the most straightforward approach to elaborate weak a fracture energy, respectively, 5 and 200 times higher interfaces for reinforced laminar materials used in severe than the typical value measured for monolithic sintered conditions a-SiC. Many systems have been studied Sic/C [1, 21, SiC/ This paper describes the elaboration of laminar ma- SiC [3]. AlO3/AlO3 [4], Si3 N4/BN [5, 6], Si3 N4/Si3 N4 terials made of silicon carbide layers, obtained by tape whiskers [7], ZrO2 toughened AlO3 [8, 9], AlO3/Sic casting, and cosintered in liquid phase processing. A [10] and have exhibited higher fracture energy than their liquid phase sintering processing has been chosen due to respective monolithic form. Several processings can be lower sintering temperature and the possible design of sed to elaborate multilayered structures: tape casting duplex microstructure. The porosity is controlled by [3, 4, 8, 9], extrusion [10] or electrophoretic deposition [2]. adding pyrolysable particles (pore forming agents Weak interfaces can be constituted by graphite [1, 2], (PFA)) in tape casting slurries as reviewed in numerous boron nitride [5, or porous ceramic material [3, 4, 7, 10]. studies [3, 4, 11, 12]. If the PFa diameters are sufficiently Porous interlayers offer some advantages such as large, they form after burnout pores thermodynamically emical compatibility between layers, no thermal stable that shrink by the same amount as the ceramic material surrounding them 3]. However, two problems Corresponding author. Tel. +33-4-7742-0020: fax: +33-4-7742. can appear with the liquid phase sintering process.On the one hand, the pore mobility is higher than in solid E-inail address: thevenot@emse. fr(F. Thevenot) phase sintering and can lead to higher pore coalescence see front matter c 2001 Published by Elsevier Sc PI:S0263-4368(01)00055-5
! " " ##$ % $&#' ( ) !!"# "$ % !!" & '(%)* + , '%+ ' - ! .+ , % , , / ' , + 0 / % % 1 % '$" . , + 2 3 , 0 % $ . 4 % & 1 , 0 % 5 , , 3 ' 6 7! # 6 8! + !!" 9 5 *+, - % # # # 5& # * , , 3 * 3* : ;"% 2> ;$?$@? ; A?$>?$ ,3 ;7 ;8C ;"!$8C$7"!7 . , , 3 * & % & , 1 ' '%++ , ;>$""" ,,, = =I 6 J>>*$*77$ *!! !# 16 J>>*$*77$ * !!!! - K ' + ! A>*$>A8=!"=L * !!" 9 5 EE6 ! A > * $ > A 8 ' ! " + ! ! ! *
426 C. Reynaud et al. International Journal of Refractory Metals Hard Materials 19(2001)425-435 On the other hand, pores left after burnout may be filled Table 2, and are compared with compositions reviewed by the liquid phase. The effects of size, content, type of in the literature; the organics contents are significantly 'A on tape casting processing, sintering characteris- lower in our case. The volume fraction of PFA incor tics, porosity control and microstructure are presented porated in the slurry is referred to the theoretical volume of the dry gree e ceramic volume tape casting organic volume PFA volume) 2. Processing and experimental procedures Two different types of PFAs: polyamide powders ( PAl2, Orgasol Elf-Atochem-France) and corn starch A fine alpha silicon carbide powder has been chosen (Roquette-France), with a spheroidal shape and a nar- (Sika Tech FCP13, Norton-Norway). Alpha polytype row particle size distribution have been used. PFA was preferred to beta polytype to avoid excessive grain characteristics are listed in Table I growth during beta-alpha transformation. An amount The incorporation of polymer particles in the suspen- of 5 wt%(related to all ceramic powders) of the YAG- sion without precautions has led after casting to the alumina eutectic(60 wt% Al2O3-40 wt%Y203) com- elaboration of brittle and less flexible tapes. Three strat position was added. Characteristics of raw materials are egies have been envisaged to take into account the Pfa listed in Table 1 () Volume compensation: a part of ceramic volume Tape casting slurries have been elaborated in four was replaced by the PFA. Consequently the volume tages. In the first stage, ceramic powders were mixed, fraction (ceramic PFA)versus organic additives by using ball milling alumina media, during 4 h in a volume was constant [4]. This method led to a high MEK-ethanol azeotrope solvent(60 vol% butanone-2/ nts: the tapes we 40 vol% ethanol) containing 0.6 wt% of a phosphate ened. Therefore, the pyrolysis of organic components ester as dispersant (Beycostat C213, CECA-France) was too problematic(cracking and swelling) accord Then(second stage), an acrylic binder and a phthalate ing to Corbin and Apte [12] plasticiser were added to the suspension and mixed (i) Weight compensation: the weight ratio between during 14-16 h. The third stage corresponds to the in- (ceramic+ PFA) and organic additives was kept con- A s oration of the PFA into the slurry mixing time of stant [12]. A cracking behaviour was obtained during h leads to an uniform distribution of the Pfa. fi the drying stage nally slurry was de-aired at a slow rotation speed during (ini) No compensation: the organic content was inde- 24h. Typical compositions of suspensions are given in pendent of the PFA. This method was retained. The Table I Characteristic parameters of the ceramic powders and of the different PFAs incorporated into the slurry Raw ceramic materi P-SiC FCP13 Alumina Cr15 Yttrium oxide dso(um)a 0.3 BET S(m/g) PFA PA-A PA-B PA-C Corn starch Chemical nature olyamide-I Polyamide-12 Polyamide-I Starch .5 122 Aspect rati 1.37 1.29 a Laser grain sizer analysis. Data obtained by image analysis. Table 2 Concentration of components in the dry tape for our tapes compared to compositions reviewed in the literature Refs Our study Corbin and Apte [12] Corbin and Apte [12 Davis et al. [4] Volume fraction of PFA referred 0.55 0 to ceramic volume Ceramic volume fraction"(vol % 68.4 37 38.7 31.5 fraction(vol % PFA volume fraction"(vol % 0 The PFA volume fraction incorporated is referred to the total volume composed of ceramic material+ pore forming agent(PFA). The volume fractions are referred to the theoretical volume of the dry green tape (i.e. ceramic volume tape casting organic volume PFA volume)
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
C. Reynaud et al. International Journal of Refractory Metals d Hard Materials 19(2001)425-435 ormulation must be optimised for 40 vol% of PFA polished microstructure. Polished surfaces were etched incorporated in the slurry during 40 min in a plasma apparatus(Plasmod March) Rheological characterisations of the slurry have shown a under an atmosphere composed of CF4 +8 vol%O shear-thinning behaviour. Furthermore, slurries con- Same parameters listed above of 1200 ceramic grains taining polyamide powders as PFA exhibit higher vis- have been measured osity than the slurries using corn starch. Stronger chemical interactions between polyamide particles and tape casting organics may explain this result Then the slurries were tape cast onto a Mylar film by 3. Results and discussion the way of a moving double blade device at speed be- tween 0.17 and 0.26 m/min on a laboratory tape casting 3.1. PFA thermal degradation bench(Cerlim Equipement, Limoges-France). Strong and soft tapes with smooth surfaces, uniform thickness A chemical analysis of the gas resulting from the between 100 and 150 um were obtained. thermal degradation of the PFA has been investigated Disks(diameter 20 mm) and sheets (50 x 50 mm by using a Fourier Transformation-Infra Red (FTIR were punched in green tapes. Stack of two types com- spectrometer(Equinox 55, BRUKER) coupled to a posed of 40-45 layers (total thickness 4-5 mm)were thermogravimetric analyzer (TG 209, NETZSCH). Two gas atmospheres(Oz, Ar) have been used to study the monolithic blocks of same stacked layers, effect of the oxidant character of the atmosphere on the symmetric multilayered materials constituted by an thermal degradation. Indeed, during the burn-out under alternative stacking of dense and porous layers air, the gas atmosphere near the sample surface may be After stacking, specimens were pressed under 60 MPa mainly composed of air whereas the atmosphere cor has bee perature near 65 C. The burnout of organics position inside the sample is unknown due to the inward performed in air by heating with an heating diffusion of oxygen and nitrogen atoms and to the rate of o i oCImin between 20 and 550 C with inter- outward diffusion of the thermal decomposition prod mediate dwells according to thermogravimetric analysis ucts in the opposite way. An inert atmosphere may be of laminar materials assumed in the core of the sample Sintering of the specimens was conducted in a Polyamide powders appears more sensitive to the oxidant character of the atmosphere than corn starch nosphere during Ih at a temperature of 1950 %C. The (Fig. I). The thermal degradation of polyamide under specimens were placed in capped graphite crucible con- argon occurs close to 460C for a straight range of taining a powder bed made of a coarse alpha Sic temperature. The complete decomposition is achieved in owder (110 um)(Sika Nor Iv F120, Norton As only one stage. Whereas the polyamide thermal degra Norway) and 5 wt% of alumina(Ls3, Alcan Chemical dation under oxygen is divided into three stages dis to promote a protective atmosphere [13] tributed on a wide temperature range (300-550C) Dimensions and weights of green and sintered lami nar monolithic materials have been measured to deter mine green and sintered densities, linear and volume Te-CS under O2 shrinkage. Porosity(size, repartition, volume)has beer characterised by mercury porosimetry(AutoPore IV 三 PA-C under Ar 9500, Micromeritics-USA). The surface tension of 心 PA-C under C mercury and the contact angle between mercury and porous LPS-SiC have been, respectively, taken as 2 60 0.485 J/m and 130. Bulk density (ceramic mate- rial+ open porosity closed porosity) and apparent (only ceramic material closed porosity) of sintered samples have been also measured orosity has been characterised by image analysis performed on polished samples via an optical micro- cope. Nearly 1500 pores were taken into account. The ollowing parameters have been measured: Feret'di ameter, equivalent diameter, aspect ratio and surface 300400500 area Temperature (C) Grain size has also racterised by image Fig. 1. Variation of the PFA mass during thermal degradation(20C/ analysis performed on min) under O, or Ar gases(25 ml/min)
$! . % , * * , % 1 * , 1 , M , * , !"7 ! A = ' 9& 5 * + , 3 , "!! " ! , O3 ' ! + ' ! ! + , 3 , * $!-$ ' 3 $- + , 6 3 3 3 % 3 , A! A , !" = , ! ! , * , , 'P% + * " "C ! , * , , '""! + '3 ? EP " !? % - ?, + , . '5> % + ;">! @3 ' * J J + ' J + * ? " !! , 3 , 6 R * & ) / 9 , $! ' + $ 8 . 2 " !! 6 ; 0 ) % % *E 'E+ '9&1 @QN9+ / ') !C ?9B4+ , '2 %+ 0 1 E * , * 3, , 0 1 , 0 * , % , 1 ' "+ $A! D * 1 * , '>!!- ! + " P % ' ! = + 2 % ' = + "+ 1 % 2 3 "3+ 4 5 67 869 ':' $ 7
C. Reynaud et al. International Journal of Refractory Metals Hard Materials 19(2001)425-435 The PF thermal decomposition has led to the for to the difference between the values of the ir absorptio mation of carbon dioxide, carbon monoxide, water and coefficient of each product, which furthermore varies some organic molecules containing C-H bond(alcane with the temperature, a rigorous quantitative compari for PA, monomer, dimer, etc. )in various quantities. The son between the different emitted gases cannot be per intensity of the characteristic absorption IR band of formed. Nevertheless, a semi-quantitative comparison each compound has been monitored during the thermal for each gas between the two PFA and the two atmo- related to the a mo,nt of the sas molecule however. due of the polyamide pea leads to higher amounl of 3. 3. 2. Green sample characteristics of monolithic blocks Carbon dioxide Fig. 4 indicates the influence of PFA nature and content on the bulk green density of the tape and the variation is compared to the idealised packing density of binary mixture derived by Furnas [14]. The PFA content is expressed on the basis of the ceramic volume tape ne PFa volume After thermocompaction, some residual porosity remains in the tape(16-24 vol%). However the bulk 全 green density increases with the PFA content and is higher than the bulk green density of the free PFA Eq.(1)derived by Furnas [14], where FIxture is theo- PFA bulk green density and f A is the PFa content Fig.2. Intensity of the gas IR-bands, related to I mg of PA-C, during PF PFA the thermal degradation under argon Carbon monoxide CS under O2 CS under Ar PA-C under O2 PA-C under Ar PFA and atmosphere of thermal degradation Fig. 3. Amount of gases produced during the thermal degradation(20C/min) of corn starch(CS)and PA-C under O2 or Ar gases(25 mI/min)
% * 1 1 , -4 ' % + & E ' + 4, 0 , E F , , & * , 0 * ? *& , , % , * % > $ : % 3 3 1 ;"$ % ' ! = + '+ %* 2 % ' = + $ 8 "+ 1 % 2 3 "3+ 4 5 67 869 ':'
C. Reynaud et al. International Journal of Refractory Metals d Hard Materials 19(2001)425-435 46}|·日PB X-PA-C 102030 PFA content Fig.4. Effects of the PFA size and nature on the bulk density of green Fig. 5. Efects of the PF. and nature on the skeletal density of tape(thermocompressed). green tape. Skeletal density is calculated from Eq.(2): Pskelet a quite similar evolution between bulk density for corn starch as PFA and predicted values is shown. Corn starch leads to very low additional voids. On the op- posite, Orgasol PFA which has a similar effect intro- duces higher additional voids in the tape. The higher additional void in the tape may be attributed to the more difficult rearrangement of the ceramic and PFa particles in the organic phase during the drying of the tape, likely due to the chemical interactions between 不“A polyamide particles and tape casting additives The skeletal density of the green thermocompressed 8品号5 ape has been calculated according to eq (2) 日PAB A·· Com starch It is assumed that no casting organics are adsorbed on the PFA surface, the casting organics are entirely 30405060 distributed in the matrix. Fig. 5 shows the variation of PFA content(vol%) the green skeletal density versus PFA content. The skeletal density is quite constant for corn starch PFA Fig. 6. Effect of the PFA nature on the volumetric shrinkage during Consequently a constant volumetric shrinkage can be sintering expected for porous materials with corn starch as PFA to the annihilation of the small porosity which is ther 3.3. Sintering characteristics of monolithic block modynamically unstable, or the filling by the liquid phase of the finest porosity effects of PFa on volumetric shrinkage are Fig. 7 illustrates the evolution of the volumetric shown in Fig. 6. As expected, porous material using shrinkage during sintering of the green skeletal density corn starch as PFA exhibits a quite constant volumetric of the tape compared to the theoretical prediction in the shrinkage up to 45 vol% similar to free PFA volumetric case of a full densification of the ceramic skeleton, or shrinkage. Surprisingly, PA-C has a sligh partial densification of the skeleton(98 and 96%). The volumetric shrinkage and it is very closed to corn stard bulk density of free PFA laminar specimen reaches 97 one. On the opposite, PA-A porous materials present a 99% of the theoretical bulk density. Most of the exper- significant increase of volumetric shrinkage. This is due imental values are included between the lower and upper
% & , 3 % , , 2 * 2 % , 0 * F % 3 , 3 9& ' + 3 % E % 1 , 3 % 3 & % & 3 1 , % ) 0 3 0 => 0 % 3 , A % 1 % 1 & 3 $ . % 3 %* 0 3 2 %*% M 3 , * M & M 7 3 3 M 3 M 3 'C8 CA.+ 3 % C7- CC. 3 1* , , $ 90 % / 3 ' + 90 % / 3 3 9& ' +6 3 % A 90 % 3 "+ 1 % 2 3 "3+ 4 5 67 869 ':' $ C
C. Reynaud et al. International Journal of Refractory Metals Hard Materials 19(2001)425-435 tween the geometric level of porosity calculated from the weight and dimensions of sintered specimens and bulk porosity measured by mercury porosimetry is attained On the other hand, similar trends between bulk porosity free PFA and PFA content are obtained. Linear relationships fuly densification between porosity and PFa with a slope less than I have partial densification-98% been obtained for low PFA concentrations(see Table 3) For higher PFA contents, the porosity values are almost constant and maximal. a higher maximum of porosity has been obtained by using starch. Only laminates using corn starch present a level of porosity higher than the X金 theoretical level of porosity required to ensure crack deflection The porosity distribution between open and closed porosity is presented in Fig. 9. Results of only two PFA which represent the two behaviours observed are plotted for clarity reasons. Firstly, the volume fractions of closed porosity for low PFA content (5-10 vol%)are lower than usually expected. Corbin et al. [ 12] have seen of volumetric shrinkage during sintering with the this phenomenon and have attributed it to an insufi- green ske nsity for the ditterent PFA nature and content tested. cient homogenisation of the PFa into the slurry.Sec ondly, PA-C and PA-B have an irregular behaviour; whereas, the trend of the variation of open and closed porosity for corn starch and PA-A is similar to the po- rosity distribution usually observed on starch(geometric density 3.4. Microstructure of monolithic block Homogeneous microstructure has been obtained Figs. 10(a)and(b) shows optical micrographs of the different pore structures elaborated. Pore size measured by mercury porosimetry versus PFA content is plotted in Fig. 11. A similar trend in- dependent of the PFA type was observed(sigmoidal e ). For low PFA content (rA40 vol%), significant greater volumetric clusions connects one side to the other. For an inclu shrinkages are obtained than theoretical predictions as sion volume fraction equals 0. 10, many pairs of seen previously, and suggested some pores left after inclusions are created. And on the other hand, every burn-out of pa-a have been annihilated inclusion is a part of the same continuous network if Fig. 8 shows the porosity of sintered specimens versus the inclusion volume fraction is higher than 0. 22 [15] PFA content On the one hand a good agreement be Fig 12 which shows the cumulative volume fraction of
M %* % ' % $! .+ M 3 * %*% 8 , % 2 * , , 3 2 , 3 % 5 , % , " , % ' >+ % 1 % 1 2 & 3 : , C , % , , , % ' -"! .+ , 1 ;" 4 "!' + ' + , 0 / % "" % * % , ' + , % ' % " .+ & * % C "!# , D % , " . * % . & % / E 3 * 1 , !"A ,3 * * & !"! % ,3 ! ;" ! "+ 1 % 2 3 "3+ 4 5 67 869 ':'
C. Reynaud et al. International Journal of Refractory Metals d Hard Materials 19(2001)425-435 Table 3 Characteristics of sintered monolithic materials Pore forming agent or ceramic material PA-A P PA-C Corn starch ure Sic (free PFA) PFA contents(PFA volume/volume of fully dens ape i.e. SiC+ casting organics+ PFA) Domain of linear shrinkage during sintering (% l8-25 l8-25 l8-25 l8-20 18-19 Linear domain of pfa content versus final 040 0-45 porosity(vol%) Slope of the linear domain 0.68 0.9 Maximal porosity of sintered materials(vol%) 0±2 31.5±1.5 4士1 41.5±1.5 porosity versus pore size for porous material usin corn starch. illustrates this behaviour. The channel size between two pores in contact can be 100 estimated according to the eq. 3)derived by Zok et al. [15] 中四m=(1-LS)×GS×12m-1 where linear shrinkage(LS)is taken as 0.2, and grain size(GS) is taken as I um. The calculated values, listed in Table 4, are somewhat higher than the experimental 3.5. Porosity control The porosity control can be mainly affected by the thermodynamic stability of pores and the pore filling by the liquid phase The thermodynamical stability of pores is directly PFA contentⅳo%) linked to the dihedral angle in solid phase sintering Fig. 9. Open and closed porosities versus different PFA (corn starch [16]. This value of dihedral angle o determines the and PA-C)contents. value of the critical pore coordination number Nc. If Fig. 10. Porous microstructure obtained after liquid phase sintering of a laminar composite composed of same layers elaborated by tape casting 40 vol% of PFA PA-C was incorporated in the slurry, and a volume fraction of porosity of 33 vol% was reached;(b)25 vol% of corn staro corporated in the slurry led to 21. 5 vol% of porosity
/ / , , 9& '>+ B3 ;" , 3 '5+ 3 ! / ')+ 3 " $ , 1 ' + 0 M & 3 ;"A> . , # ' + . * " . > %*% %*@ %* ' %+ % '% = J J %+ '.+ -$ -$ -$! - ! O 3 '.+ "8- "8- "8- "8- ! "8-"C 5 % M '.+ !- !-$! !-> !-$ !A8 !7 !C !8 1 '.+ ! >" " >$ " $" " C 2 0 % ' %*+ "+ 1 % 2 3 "3+ 4 5 67 869 ':' $>"
C. Reynaud et al. International Journal of Refractory Metals Hard Materials 19(2001)425-435 the pore coordination number is higher than the crit ical value Nc, the pore is thermodynamically stable, the pore surface curvature is concave and it will grow △ Corn starch For lower coordination number, the pore is unstable and will shrink(annihilation) due to its convex shape For a dihedral angle value of 120(typical value ob- served for solid phase sintering), the critical number n extende analysis in space, carried out by Kingery and francois [17 indicates a value of Nc= 12. If we adapt those 。o results to our liquid phase sintering, pores left after burnout of the pfa for PA-B. PA-c and corn starch △△△ are thermodynamically stable owing to the high ratio between the resulting pore diameter(PA-C: 12 um, Corn starch: 9 um, cf. Table 5)and Sic grain size (1.2 um). For PA-A(diameter=3.8 um), some pores may PFA content(vol% The pore filling is connected to the contact angle tween ceramic grain and liquid phase. For a total ig. 11. Mode of pore size distribution with different PFA contents. For a given PFA content, the mode of pore size distribution corre- wetting, the pores filling occurs when the liquid menis sponds to the pore diameter for which the largest fraction of pores is cus radius is equal to the pore radius according to [18]: pore filling is easier for low contact angles. However low liquid phase volume fraction and large pores limit this phenomenon due to the preferential localisation of the o 5vol% liquid phase along the smaller spaces between higher capillary pressure). The pore filling must be limited in our materials. Moreover, no refilled pore has been observed 25wc% 3.6. Microstructure of multilayered composites () Layered structure: Layered structures have been 9E>9 obtained with parallel layers and uniform thickness um; porous layer: 80 um)(Fig 13(a)). No migrations of pores from porous to dense 1.88金aege layers have been observed. Excepted for alternate laminates using PA-A as PFA, no defects and no del aminations have been shown These defects for PA-a ight be attributed to internal stresses due to the mis- match between the shrinkage of a dense and a porous re diameter(um) layer or caused by an excessive and sudden gas emis- Fig. 12. Cumulative volume of porosity versus pore diameter for dif- sion during burnout(Fig. 13(b)) Trent corn starch contents (i) Grain Size: Image analysis was performed on three different multilayered specimens with a stack ing of dense and porous layers. The PFA volume Table 4 Comparison between the plateau value reached by the mode of the contents of porous layer were: 45 vol% of corn pore size distribution with theoretical prediction calculated according arch. 45 and 25 vol% of PA-A. the values mea the Eq 3 derived by Zok et al. (15 sured are plotted in Fig. 14 for the two multilayered PFA type of the Mode of the composites with porous layers containing 45 vol% of materia distribution(um) PFA. No influence of PFA type and content on grain size is shown. Dense layers and skeletons in porous layers are constituted by the same homo PA-C geneous, fine and equiaxed microstructure (mean Corn starch Ceramic grain size has been taken as I um pea 1 25 um, mean aspect ratio= 1.7)(see Figs. 13(c)and (d))
* , , , , 3 ' + 1 " ! ' * + & A , % 1 N ;"78 + M , & , M , & * & ;"8' ++ ? 91 %*% % * , %*% * , 3 1 * ' ">' ++ '+ '+ ' ++ " * "" / , 0 % % / * , $ , / , 9& > B3 ;" !8 %*@ > 8 "A %* !8 $"A / 3 " $> "+ 1 % 2 3 "3+ 4 5 67 869 ':'
C. Reynaud et al. International Journal of Refractory Metals d Hard Materials 19(2001)425-435 Table 5 Pores characteristics, measured by image analysis, of the monolithic laminates and of the laminar composites. Nearly 1500 pores have been taken Into account PFA Content Laminar composites (vo%) Equivalent diameter(um) Aspect ratIo Equivalent diameter(um) Corn starch 9±1.5 7.5±1.5 PA-C 12.8±1.5 l17±1 l1.3±1.5 8.9±1.5 PA-A 3.8±1.5 5±1.5 秘论没学叫 验就染静感 Fig. 13. Microstructure of symmetric multilayered sayers containing PA-A;(c)dense layer; (d)porous layer with equiaxed grains(bar =10 um)for licon carbide materials:(a) laminar composite with 45 vol% of corn starch incorporated in the a laminar composite containing 45 vol/ of corn starch. (iii)Pores: Porous layers exhibit the same pore fea- 4. Conclusion ures(identical mean values); i.e. pore characteris- tics are unaltered by the position (core or shell) Laminar silicon carbide composites have been elab- of the layer in the laminar composite. Moreover, orated by the tape casting- liquid phase sintering the pore characteristics are the same as those (YAG-alumina eutectic) processing. Porosity was in obtained for corresponding monolithic blocks (see troduced by the incorporation of PFAs(5-50 vol%,re- Table 5) ferred to the theoretical volume of the dry tapes)in the
'+ 6 1 * ' +# * ' + 3 ' + 5 * - & '(%)* + , * % ' - ! . * + "> 6 ' + , $ . # ' + %*%# '+ # ' + , & 1 ' "! + $ . ? " !! 3 % '.+ 5 9& ' + % 9& ' + % C " " 7 " " 7 %* $! " 8 " " A ""7 " "A ""> " "A 8C " "A %*% $ >8 " "A " "A$ > " "A >C " "A "+ 1 % 2 3 "3+ 4 5 67 869 ':' $>>
C. Reynaud et al. International Journal of Refractory Metals Hard Materials 19(2001)425-435 average feret diameter Layered structures without defects have been obtained with parallel layers and uniform thickness(dense layer eigthed mean 70 um; porous layer: 80 um) The last step of our research will be the study rupture behaviour This process, using PFAs to elaborate laminar com- posites, can be transposed to other materials and ce- ramIcs 2 Acknowledgements The authors would like to thank r. Gibert (SPIN center, ENS Mines St-Etienne) for carrying out the !∴: pectrometry of gases produced during the ther mal degradation of PFA 8g88 analysis performed on dense and porous layers of [Clegg wJ, Kendall K, Alford NMcN, Button TW A simple way to make tough ceramics. Nature 1990 wo svm layered composites containing, respectively. 45 and 45 vol% of PA-A. CS corresponds to corn 2Vandeperre L, Van Der Biest O. Composite PL and IPL, are, respectively, the abbreviations of terlayers for crack deflection in ceramic laminates. Silic Ind external dense layer, internal dense layer, external porous layer, in 8:63:39-43 ernal porous laye 3 Blanks Ks, Kristoffersson A, Carlstrom E, Clegg, WJ. Crack deflection in ceramic laminates using porous interlayers. J Eur eram Soc 1998;18:1945-51. 4 Davis JB, Kristoffersson A, Carlstrom E, Clegg, WJ. Fabrication durry. Different PFA with a narrow size distribution and crack deflection in ceramic laminates with porous interlayers. have been used(corn starch and powders of polyamide [5 Liu H, Hsu SM. fracture behavior of multilayer silicon nitride/ 12). Two types of samples have been prepared: mono- boron nitride ceramics. J Am Ceram Soc 1996: 79: 2452-7 lithic blocks and alternate laminates (laminar compo 6 Kovar D, Thouless MD, Halloran Jw. Crack deflection and ites) of dense and porous layers propagation in layered silicon nitride/boron nitride cer J Am Ceram Soc 1998: 81: 1004-12 For each PFA, the porosity attains a maximum value Ohji T, Shigegaki Y, Miyajima T, Kanzaki S. Fracture resistance dependent on the PFA size and nature(41 vol% with behavior of multilayered silicon nitride. J Am Ceram Soc corn starch) 199780:991-4 Only laminates using corn starch present a level of 8 Chartier T, Merle D, Besson JL. Laminar ceramic composites. orosity higher than the theoretical level of porosity J Eur Ceram Soc 1995: 15: 101-7. required to ensure crack deflection. Furthermore the 9 Chartier T, Rouxel T. Tape cast alumina-zirconia laminates: processing and mechanical properties. J Eur Ceram Soc 1997; 17 volumetric shrinkage during sintering for porous mate- 299-308 rial using corn starch is quite constant and is similar to [10] She J, Inoue T, Kazuo U. MultilayerAlOj/SiC cs with that of free PFA materials for PFA content up to 45 proved mechanical behavior. J Eur Ceram Soc 2000: 20: 1771-5 vol% due to the lowest chemical interaction between I Lyckfeldt O, Liden E, Carlsson R. Processing of thermal corn starch and the tape casting additives. No residual insulation materials with controlled porosity. In: Linton DP, Limaye SY, editors. Ceramic transactions Low-exp thermal stresses induced during cooling can be expected ials. The American Ceramic Society: 1995. p. 217- The thermal degradation of corn starch produces a [12] Corbin SF, Apte Ps. Engineered porosity via lower amount of gas than polyamide PFA and is less lamination and the percolation of pyrolysable particulates. J Am Ceram Soc1999:82:1693-701 ensitive to the nature of the gas used during the burn- [13] Baud S, Thevenot F Effect of powder bed composition on liquid ing-out. Consequently, corn starch appears the most phase sintered silicon carbide. In: Vincenzini P, editor. Ninth favourite candidate to elaborate laminar composites Cimtec- World Ceramic Congress Ceramics Getting into the 2000s with engineered porosity Part B. Faenza: Techna: 1999. p. 805-12 Homogeneous distribution of porosity has beer [14 Furnas CC. The relations between specific tained for both monolithic and composites laminates An equiaxed and homogeneous silicon carbide micro- Rep Invest 1928: 2894: 1-10. [5 Zok F, Lange FF, Porter JR. Packing density of composite structure has been obtained and was unaffected by PFA owder mixtures. J Am Ceram Soc 1991: 74: 1880-5
O0 % , , / ' , " + , 6 * 3 ' * + % 1 % / '$" . , + 2 & 3 : 3 * & % % $ . , , ? 1 , % * * & , 4 * % & 1 * , 0 % 5 , , 3 ' 6 7! # 6 8! + , % * * , 3 3 ) 'E? 9? *9 + E * % ;"$76$ -7 ; 6>C*$> ;>6 >AC-7$ ; !8 ;"!= , G 9 !!!# !6"77"- ;""-7!" ;">$ "+ 1 % 2 3 "3+ 4 5 67 869 ':'