《复合材料 Composites》课程教学资源（学习资料）第五章 陶瓷基复合材料_fibrous monoliths-23

ULTRA-HIGH TEMPERATURE CERAMICS JOURNAL OF MATERIALS SCIENCE 39(2004)5951-5957 Processing and characterization of ZrB2-based ultra-high temperature monolithic and fibrous monolithic ceramics W.G. FAHRENhOlTZG.E. HILMAS A.L. CHAMBERLAIN, J.W. ZIMMERMANN Department of Ceramic Engineering, University of Missouri-Rolla, Rolla, MO 65409, USA E-mail: billf@umr. edu Web:www.umr.edu/billf Zirconium diboride(zrB2)based ultra-high temperature ceramics either unmodified or with SiC particulate additions of 10, 20, or 30 volume percent were prepared by conventional hot pressing The ZrB2-SiC compositions had improved four-point bend strength compared to the Zr B2 prepared in our laboratory as well as other reported ZrB2 or ZrB2-SiC materials Strength and toughness increased as the amount of SiC increased Measured strengths ranged from 550 MPa for ZrB2 to over 1000 MPa for ZrB2-30% SiC. Likewise, toughness increased from 3. 5 MPa to more than 5 MPa over the same composition range. The ddition of sic also improved oxidation resistance compared to pure ZrB2 Co-extrusion processing was used to produce ZrB2-based ultra-high temperature ceramics with a fibrous monolithic structure Samples had dense ZrB2-30 vol% SiC cells approximately 100 um in diameter surrounded by porous Zr B2 cell boundaries approximately 20 um thick. ZrB2-based fibrous monoliths had four point bend strength of 450 MPa, about half of a conventional ZrB2-Sic ceramic with the cell composition Preliminary analysis of fracture behavior found that ZrB2-based fibrous monoliths did not exhibit graceful failure because the difference in strength between the cell and cell boundary of the current materials was not sufficient. c 2004 Kluwer Academic Publishers 1. Introduction that have a deleterious effect on high temperature per- Transition metal borides and carbides such as ZrB ,, formance. As an example, the strength of ZrB2 contain ZrC, HfB2, HfC, and TaC have melting temperatures ing 3 volume percent Ni has been shown to drop by in excess of 3000C making them candidates for use about 60% between 800oC and 1000oC [5]. Although a as structural materials at temperatures above 1800C ZrB2-Ni phase diagram has not been reported, the eu [1]. Within the family of ultra-high temperature ceram- tectic temperature is likely similar to that for the ZrC-Ni ics(UHTCs), ZrB2 has the lowest theoretical density system, which was reported to be 1290oc[ll]. Presum- (6.09 g/cm), which makes it attractive for aerospace ably, more refractory compounds would have less of applications[2]. Combined with its high melting tem- an effect on high temperature performance. Reinforce- perature, ZrB2 is reported to have excellent resis- ment with Sic particulates has been shown to improve tance to thermal shock and oxidation compared both the room temperature strength and oxidation re- other non-oxide structural ceramics [2]. In addition sistance of Zr B2 without sacrificing high temperature to aerospace applications, ZrB2 has also been con- strength [12]. The first part of this study focused on the sidered for use as electrode or crucible materials for preparation, characterization, and testing of ZrB2-SiC molten metal contact [1-3]. Physical and mechanical UHTCs prepared without oxide or low-melting temper- property values reported in the recent technical lit- ature impurities. erature for ZrB2-based materials are summarized in Beyond the control of impurities and the addition Table I [1, 2, 4, 5] of particulates, the thermo-mechanical performance of 32-based ceramics can be prepared from commer- high temperature materials such as Zrb cial powders or synthesized using a variety of reaction- lored by manipulating the meso-scale(10-500 um) based methods[4-9]. Unintentional impurities [10] or structure. Co-extrusion of ceramic powder loaded ther- processing additives [5, 91, which affect both the sin- moplastic polymers has been used to fabricate so tering behavior and high temperature performance, called fibrous monolithic structures with meso-scale often incorporated during processing. Additives or im architectures designed to improve thermo-mechanical purities can form low melting eutectics with UHTCs performance [13, 14]. Using relatively inexpensive, Corresponding author: Dr. Bill Fahrenholtz, Assistant Professor of Ceramic Engineering. 222 McNutt Hall, University of Missouri-Rolla, Rolla, MO 65409,USA. 0022-2461 2004 Kluwer Academic Publishers 5951

ULTRA-HIGH TEMPERATURE CERAMICS JOURNAL OF MATERIALS SCIENCE 3 9 (2 004) 5951 – 5957 Processing and characterization of ZrB2-based ultra-high temperature monolithic and fibrous monolithic ceramics W. G. FAHRENHOLTZ, G. E. HILMAS, A. L. CHAMBERLAIN, J. W. ZIMMERMANN Department of Ceramic Engineering, University of Missouri-Rolla, Rolla, MO 65409, USA E-mail: billf@umr.edu Web: www.umr.edu/∼billf Zirconium diboride (ZrB2) based ultra-high temperature ceramics either unmodified or with SiC particulate additions of 10, 20, or 30 volume percent were prepared by conventional hot pressing. The ZrB2-SiC compositions had improved four-point bend strength compared to the ZrB2 prepared in our laboratory as well as other reported ZrB2 or ZrB2-SiC materials. Strength and toughness increased as the amount of SiC increased. Measured strengths ranged from ∼550 MPa for ZrB2 to over 1000 MPa for ZrB2-30% SiC. Likewise, toughness increased from 3.5 MPa to more than 5 MPa over the same composition range. The addition of SiC also improved oxidation resistance compared to pure ZrB2. Co-extrusion processing was used to produce ZrB2-based ultra-high temperature ceramics with a fibrous monolithic structure. Samples had dense ZrB2-30 vol% SiC cells approximately 100 µm in diameter surrounded by porous ZrB2 cell boundaries approximately 20 µm thick. ZrB2-based fibrous monoliths had four point bend strength of ∼450 MPa, about half of a conventional ZrB2-SiC ceramic with the cell composition. Preliminary analysis of fracture behavior found that ZrB2-based fibrous monoliths did not exhibit graceful failure because the difference in strength between the cell and cell boundary of the current materials was not sufficient. C 2004 Kluwer Academic Publishers 1. Introduction Transition metal borides and carbides such as ZrB2, ZrC, HfB2, HfC, and TaC have melting temperatures in excess of 3000◦C making them candidates for use as structural materials at temperatures above 1800◦C [1]. Within the family of ultra-high temperature ceram￾ics (UHTCs), ZrB2 has the lowest theoretical density (6.09 g/cm3), which makes it attractive for aerospace applications [2]. Combined with its high melting tem￾perature, ZrB2 is reported to have excellent resis￾tance to thermal shock and oxidation compared to other non-oxide structural ceramics [2]. In addition to aerospace applications, ZrB2 has also been con￾sidered for use as electrode or crucible materials for molten metal contact [1–3]. Physical and mechanical property values reported in the recent technical lit￾erature for ZrB2-based materials are summarized in Table I [1, 2, 4, 5]. ZrB2-based ceramics can be prepared from commer￾cial powders or synthesized using a variety of reaction￾based methods [4–9]. Unintentional impurities [10] or processing additives [5, 9], which affect both the sin￾tering behavior and high temperature performance, are often incorporated during processing. Additives or im￾purities can form low melting eutectics with UHTCs Corresponding author: Dr. Bill Fahrenholtz, Assistant Professor of Ceramic Engineering, 222 McNutt Hall, University of Missouri-Rolla, Rolla, MO 65409, USA. that have a deleterious effect on high temperature per￾formance. As an example, the strength of ZrB2 contain￾ing ∼3 volume percent Ni has been shown to drop by about 60% between 800◦C and 1000◦C [5]. Although a ZrB2-Ni phase diagram has not been reported, the eu￾tectic temperature is likely similar to that for the ZrC-Ni system, which was reported to be 1290◦C [11]. Presum￾ably, more refractory compounds would have less of an effect on high temperature performance. Reinforce￾ment with SiC particulates has been shown to improve both the room temperature strength and oxidation re￾sistance of ZrB2 without sacrificing high temperature strength [12]. The first part of this study focused on the preparation, characterization, and testing of ZrB2-SiC UHTCs prepared without oxide or low-melting temper￾ature impurities. Beyond the control of impurities and the addition of particulates, the thermo-mechanical performance of high temperature materials such as ZrB2 can be tai￾lored by manipulating the meso-scale (10–500 µm) structure. Co-extrusion of ceramic powder loaded ther￾moplastic polymers has been used to fabricate so￾called fibrous monolithic structures with meso-scale architectures designed to improve thermo-mechanical performance [13, 14]. Using relatively inexpensive, 0022–2461 C 2004 Kluwer Academic Publishers 5951

ULTRA-HIGH TEMPERATURE CERAMICS TABLE I Reported room temperature bend strength, Young's Modu- 2.2. Powder preparation and binder lus hardness, and fracture toughness values for ZrBi-based ceramics burnout for co-extrusion processing Strength Modulus Hardness Toughness ZrB2-based fibrous monolithic ceramics were prepared Composition (GPa)(MPa.m1/)Reference by co-extrusion processing. Dense ZrB2-SiC contain ing 30 vol% SiC was selected as the cell phase with 480 [1 275-305343-500179-22 porous ZrB2 as the desired cell boundary phase For the rade B)and SiC (st arck UF ZrB2-20% SiC 50 2 10) powders were attrition milled to reduce the average ZrB2-3% Ni 370 14.4 ZrB2 particle size from x2 to <l um and uniformly disperse the SiC. Unmilled ZrB2(Grade A, H C Starck, commercially available powders as precursors, multi- phase After the precursor powders and milling condi- pass ram co-extrusion of core-shell type preforms has tions were selected, separate binder formulations were been used to fabricate ceramics that fail"gracefully" developed for the cell and cell boundary materials us- (i.e, provide some load retention after the initial failure ing an iterative process. Precise data regarding the shear event). The performance of fibrous monoliths is sim- thinning behavior of the thermoplastic binder-powder ilar to the performance observed in continuous fiber mixtures were collected using a torque rheometer(CW. ceramic composites produced from more expensive Brabender, South Hackensack, NJ). Proper binder for- g more complicated processing routes. mulations are essential for development of the flow Typical fibrous monolithic structures consist of a ma- characteristics necessary to maintain the desired fibrous jor phase(80-90 vol% ) formed into strong, brittle monolith architecture during the two-pass co-extrusion cells"that are surrounded by a thin(10-50 um), process. The behavior of the thermoplastics is generally continuous cell boundary phase(10-20 vol%)of a shear thinning and can be described by an empirical weaker material such as BN. By controlling structure power law relations similar to Equation 1[15] on this scale, co-extrusion processing can be used to manipulate the fracture behavior of ceramics without [=Ky changing the intrinsic properties of the material or ignificantly altering the composition of the compo- where t is shear stress, K is the consistency index, y nent. For the second part of this study, ZrB2-based fi- is the shear rate, and n is the shear thinning exponent brous monolithic structures were fabricated and char- The shear thinning exponent is a measure of the relative acterized. The UHTC fibrous monoliths were made change in viscosity with shear rate [15]. If the viscosi- up of dense ZrB2-SiC cells and porous ZrB2 cell ties of the cell and cell boundary batches are similar during batching and if the shear thinning exponent of The purpose of this paper is to report the pro- the cell boundary is somewhat higher than the cell mate cessing and characterization of ZrB2-based UHTCs. rial, then mixing at the cell-cell boundary interface will Microstructure. me echanical properties, and oxida- be minimal as the preform is co-extruded. Basically, the tion behavior are reported for ZrB2-SiC prepared lower viscosity material(cell boundary) will maintain by conventional hot pressing. In addition, the pro- its position at the periphery of the co-extrudate. In both cessing conditions and resulting microstructures of the cell and the cell boundary phases, ethylene ethyl a novel ZrB, -based fibrous monolithic ceramic are acrylate(EEA; DPDA-6182, Union Carbide, Danbury, reported CT) was used as the binder. For the cells, the binder was plasticized by adding methoxypolyethyleneglycol (MPEG-500: Union Carbide, Danbury, CT). For the 2. Experimental cell boundaries, the binder was plasticized by addition 2.1. Powder preparation for conventional of heavy mineral oil (HMO, Aldrich, Milwaukee, WI) ZrB2 UHTCs The final binder amounts for the cell and cell boundary Conventional ZrB2 and ZrB2-SiC ultra-high tempe phases, along with the shear thinning exponents, are summarized in Table ll ature ceramics were prepared from commercial ZrB2 Multi-core filaments were formed using a two Grade B, H.C. Starck, Newton, MA, 2 um) and SiC (UF-10, H.C. Starck, Newton, MA, 0.7um).As pass co-extrusion process. The initial feedrod was received powders were batched to produce ZrB2 or ZrB2 containing 10, 20, or 30 vol% SiC particulates. TABLE II Binder formulations and rheological constants for the cells After batching, powders were attrition milled(Model and cell boundaries of ZrB2-based fibrous monoliths HD-O1, Union Process, Akron, OH) to reduce particle size and promote intimate mixing A 750 ml polymer Cell boundary lined bucket was charged with approximately 3000 g of Compositic wC media, 130 to 150 g of powder depending on theBinder SiC content, and 250 ml of hexane. The mixture was Binder content (vol%) milled for 2 h at 600 rpm using a polymer-coated spin mineral oil dle Solvent was removed by rotary evaporation(Roto- Plasticizer content(vol%) 7 MPEG-500 vapor R-124, Buchi, Flawil, Switzerland) to minimize Shear thinning exp. 0.30 segregation during drying. 5952

ULTRA-HIGH TEMPERATURE CERAMICS TABLE I Reported room temperature bend strength, Young’s Modu￾lus, hardness, and fracture toughness values for ZrB2-based ceramics Strength Modulus Hardness Toughness Composition (MPa) (GPa) (GPa) (MPa·m1/2) Reference ZrB2 480 490 21 3.7 [1] ZrB2 275–305 343–500 17.9–22 – [2] ZrB2-20% SiC 506 – 21 4.0 [4] ZrB2-3% Ni 370 496 14.4 3.4 [5] commercially available powders as precursors, multi￾pass ram co-extrusion of core-shell type preforms has been used to fabricate ceramics that fail “gracefully” (i.e., provide some load retention after the initial failure event). The performance of fibrous monoliths is sim￾ilar to the performance observed in continuous fiber ceramic composites produced from more expensive precursors using more complicated processing routes. Typical fibrous monolithic structures consist of a ma￾jor phase (80–90 vol%) formed into strong, brittle “cells” that are surrounded by a thin (10–50 µm), continuous cell boundary phase (10–20 vol%) of a weaker material such as BN. By controlling structure on this scale, co-extrusion processing can be used to manipulate the fracture behavior of ceramics without changing the intrinsic properties of the material or significantly altering the composition of the compo￾nent. For the second part of this study, ZrB2-based fi- brous monolithic structures were fabricated and char￾acterized. The UHTC fibrous monoliths were made up of dense ZrB2-SiC cells and porous ZrB2 cell boundaries. The purpose of this paper is to report the pro￾cessing and characterization of ZrB2-based UHTCs. Microstructure, mechanical properties, and oxida￾tion behavior are reported for ZrB2-SiC prepared by conventional hot pressing. In addition, the pro￾cessing conditions and resulting microstructures of a novel ZrB2-based fibrous monolithic ceramic are reported. 2. Experimental 2.1. Powder preparation for conventional ZrB2 UHTCs Conventional ZrB2 and ZrB2-SiC ultra-high temper￾ature ceramics were prepared from commercial ZrB2 (Grade B, H.C. Starck, Newton, MA, ∼2 µm) and SiC (UF-10, H.C. Starck, Newton, MA, ∼0.7µm). As￾received powders were batched to produce ZrB2 or ZrB2 containing 10, 20, or 30 vol% SiC particulates. After batching, powders were attrition milled (Model HD-01, Union Process, Akron, OH) to reduce particle size and promote intimate mixing. A 750 ml polymer￾lined bucket was charged with approximately 3000 g of WC media, 130 to 150 g of powder depending on the SiC content, and 250 ml of hexane. The mixture was milled for 2 h at 600 rpm using a polymer-coated spin￾dle. Solvent was removed by rotary evaporation (Roto￾vapor R-124, Buchi, Flawil, Switzerland) to minimize segregation during drying. 2.2. Powder preparation and binder burnout for co-extrusion processing ZrB2-based fibrous monolithic ceramics were prepared by co-extrusion processing. Dense ZrB2-SiC contain￾ing 30 vol% SiC was selected as the cell phase with porous ZrB2 as the desired cell boundary phase. For the cell phase, ZrB2 (Starck Grade B) and SiC (Starck UF- 10) powders were attrition milled to reduce the average ZrB2 particle size from ∼2 to <1 µm and uniformly disperse the SiC. Unmilled ZrB2 (Grade A, H.C. Starck, Newton, MA, ∼6 µm) was used as the cell boundary phase. After the precursor powders and milling condi￾tions were selected, separate binder formulations were developed for the cell and cell boundary materials us￾ing an iterative process. Precise data regarding the shear thinning behavior of the thermoplastic binder-powder mixtures were collected using a torque rheometer (C.W. Brabender, South Hackensack, NJ). Proper binder for￾mulations are essential for development of the flow characteristics necessary to maintain the desired fibrous monolith architecture during the two-pass co-extrusion process. The behavior of the thermoplastics is generally shear thinning and can be described by an empirical power law relations similar to Equation 1 [15]. τ = Kγ n (1) where τ is shear stress, K is the consistency index, γ is the shear rate, and n is the shear thinning exponent. The shear thinning exponent is a measure of the relative change in viscosity with shear rate [15]. If the viscosi￾ties of the cell and cell boundary batches are similar during batching and if the shear thinning exponent of the cell boundary is somewhat higher than the cell mate￾rial, then mixing at the cell-cell boundary interface will be minimal as the preform is co-extruded. Basically, the lower viscosity material (cell boundary) will maintain its position at the periphery of the co-extrudate. In both the cell and the cell boundary phases, ethylene ethyl acrylate (EEA; DPDA-6182, Union Carbide, Danbury, CT) was used as the binder. For the cells, the binder was plasticized by adding methoxypolyethyleneglycol (MPEG-500; Union Carbide, Danbury, CT). For the cell boundaries, the binder was plasticized by addition of heavy mineral oil (HMO, Aldrich, Milwaukee, WI). The final binder amounts for the cell and cell boundary phases, along with the shear thinning exponents, are summarized in Table II. Multi-core filaments were formed using a two pass co-extrusion process. The initial feedrod was TABLE II Binder formulations and rheological constants for the cells and cell boundaries of ZrB2-based fibrous monoliths Component Cell Cell boundary Composition ZrB2-30% SiC ZrB2 Binder EEA EEA Binder content (vol%) 36 45 Plasticizer Heavy mineral oil MPEG-500 Plasticizer content (vol%) 7 4 Shear thinning exp. 0.30 0.54 5952

ULTRA-HIGH TEMPERATURE CERAMICS approximately 2.5 cm in diameter and 10 cm long. values are the average of at least ten samples. Oxida Feedrods consisted of a fine particle size ZrB2-30% tion studies were conducted using thermal gravimetric Sic core and a coarse particle size ZrB2 shell. Both analysis (TGA) following a proposed ASTM standard the core and shell structures were prepared by mix- for oxidation testing [16]. The sample size was approxi ng the powder and binder in a high shear mixer(C w. mately l mm by 1.5 mm by 1.5 mm. Surface areas were Brabender, South Hackensack, NJ) above the softening calculated from external sample dimensions measured temperature of the polymer. The UHTC fibrous mono- to 0.Ol mm. The weight changes for composites liths were formed by extrusion at 0.6 mm/s, 110C, and measured under flowing air at a ramp rate of 10oC/min 39 MPa. The drawdown ratio was approximately 10: 1 up to 1500oC without an isothermal hold. The resultin for all extrusion passes. After the first pass, the thin fil- oxide layer was examined using SEM and EDS aments were bound, pressed, and re-extruded to form the multi-core filament. after the second extrusion the multi-core filaments were placed in a die and pressed 3. Results and discussion 1000C, 10 MPa, 30 min) to form a rectangular plate 3.1. Conventional ZrB2-SiC UHTCs 58 mm by 28 mm by 5 mm). The rectangular plate un- Archimedes, density measurements for conventional derwent binder burnout to prepare it for densification The binder burnout cycle was 154 h long with a max ZrB2-based UHTCs indicated that little or no poros imum temperature of 1054 C. The atmosphere during tion of samples containing 10% SiC, the density of all as fo to 28 of the samples was greater than 100% relative den by a change to argon sity based on the as-batched ZrB2 and Sic contents CTable IID) and the assumed theoretical densities of ZrB2(6.09 g/cm )and SiC (3.21 g/cm ). The higher 2. 3. Hot pressing conditions than predicted density of the hot pressed composites Hot pressing was used to densify both conventional was attributed to the introduction of wC impurities and fibrous monolithic UHTCs Either milled ZrB2-Sic during attrition milling. Comparison of bulk densities powder mixtures or burned-out ZrB2-SiC/ZrB2 fibrous determined by the Archimedes'method to estimated monolith preforms were placed in BN-coated, graphite true densities determined by helium pycnometry in- foil-lined graphite dies. Samples were densified by hot dicated that 1. 4 to 2.2 vol% wC(theoretical density pressing at 1900C for 45 min at 32 MPa. The hot press- 15.0 g/cm for wC bonded with 6% Co) was present atmosphere was argon in the hot pressed composites. The presence of wC was confirmed by XRD and sEM/EDS(discussed be- low). The low density of the 10% SiC sample(93% 24. Characterization and mechanical relative) manifests itself as large, isolated pores. Such testing pores might be formed by agglomeration of the starting After hot pressing, the bulk density of hot pressed billets powder or by an unknown chemical effect during hot was measured using the Archimedes technique with water as the immersing medium. Helium pycnome Examination of ZrB2-SiC UHTCs by SEM showed try(1305 Multivolume, Micromeretics, Norcross, GA) that the Sic particles were well dispersed in the ZrB was employed to estimate the true density of the hot matrix(Fig. I). The SiC particle less than a micron to approximately 5 um. In addi 325 mesh using a high purity alumina mortar and pes- tion to ZrB2 and Sic, two other phases were observed tle to expose as much of the closed porosity as possible. X-ray diffraction analysis(Fig. 2)and EDS confirmed the presence of wc (along with the possible presence ing scanningelectron microscopy (SEM;: S5 70, Hitachi, of wCl-r and/or W2 C) in the form of discrete particles Tokyo). Samples were prepared by cutting cross sec- (Fig. I). An unidentified phase was observed by SEM tions either parallel or perpendicular to the hot pressing By EDS, it contained w, Zr andB.However,a reliable direction. Chemical analysis was performed simultane- composition could not be determined by EDS nor could ously with SEM using energy dispersive spectroscopy the low intensity, unindexed peaks in the XRD pattern (EDS: AAT, X-ray Optics, Gainsville, FL). Vickers hardness was determined from a minimum of t sing a load of 500 g and a TABLE III Density of hot pressed ZrB2-based UHTCs determined by Archimedes method, calculated from ZrB2 and Sic contents, and dwell time of 30 s. The elastic constants were mea- determined by He pycnometry along with the WC content of hot pressed sured following the ASTM standard C1259-01 for im- composites pulse excitation of vibration. The fracture toughness Archimedes Theoretical Helium wC using a 30 kg load to form radial-median cracks fol- density density ensity Content value reported for each composition. Flexure strength zB a lowed by fracture in four-point bending. A minimum g/cm) (g/cm)(vol%) of seven bars were tested and averaged to obtain the ZrB2-10% was measured in four-point using a fully articulated ZrB7-20% test fixture according to astm standard cl161-o2a for ZrB,-30% SiC 5.43 5.46 type a bars(25mm×2mm×1.5mm). The reported 5953

ULTRA-HIGH TEMPERATURE CERAMICS approximately 2.5 cm in diameter and 10 cm long. Feedrods consisted of a fine particle size ZrB2-30% SiC core and a coarse particle size ZrB2 shell. Both the core and shell structures were prepared by mix￾ing the powder and binder in a high shear mixer (C.W. Brabender, South Hackensack, NJ) above the softening temperature of the polymer. The UHTC fibrous mono￾liths were formed by extrusion at 0.6 mm/s, 110◦C, and 39 MPa. The drawdown ratio was approximately 10:1 for all extrusion passes. After the first pass, the thin fil￾aments were bound, pressed, and re-extruded to form the multi-core filament. After the second extrusion, the multi-core filaments were placed in a die and pressed (100◦C, 10 MPa, 30 min) to form a rectangular plate (58 mm by 28 mm by 15 mm). The rectangular plate un￾derwent binder burnout to prepare it for densification. The binder burnout cycle was 154 h long with a max￾imum temperature of 1054◦C. The atmosphere during binder burnout was flowing air up to 280◦C, followed by a change to argon. 2.3. Hot pressing conditions Hot pressing was used to densify both conventional and fibrous monolithic UHTCs. Either milled ZrB2-SiC powder mixtures or burned-out ZrB2-SiC/ZrB2 fibrous monolith preforms were placed in BN-coated, graphite foil-lined graphite dies. Samples were densified by hot pressing at 1900◦C for 45 min at 32 MPa. The hot press￾ing atmosphere was argon. 2.4. Characterization and mechanical testing After hot pressing, the bulk density of hot pressed billets was measured using the Archimedes’ technique with water as the immersing medium. Helium pycnome￾try (1305 Multivolume, Micromeretics, Norcross, GA) was employed to estimate the true density of the hot pressed billets. For pycnometry, billets were ground to −325 mesh using a high purity alumina mortar and pes￾tle to expose as much of the closed porosity as possible. The microstructure of the composites was examined us￾ing scanning electron microscopy (SEM; S570, Hitachi, Tokyo). Samples were prepared by cutting cross sec￾tions either parallel or perpendicular to the hot pressing direction. Chemical analysis was performed simultane￾ously with SEM using energy dispersive spectroscopy (EDS; AAT, X-ray Optics, Gainsville, FL). Vickers’ hardness was determined from a minimum of ten in￾dents that were formed using a load of 500 g and a dwell time of 30 s. The elastic constants were mea￾sured following the ASTM standard C1259-01 for im￾pulse excitation of vibration. The fracture toughness was measured according to ASTM standard C1421, using a 30 kg load to form radial-median cracks fol￾lowed by fracture in four-point bending. A minimum of seven bars were tested and averaged to obtain the value reported for each composition. Flexure strength was measured in four-point using a fully articulated test fixture according to ASTM standard C1161-02a for type A bars (25 mm × 2 mm × 1.5 mm). The reported values are the average of at least ten samples. Oxida￾tion studies were conducted using thermal gravimetric analysis (TGA) following a proposed ASTM standard for oxidation testing [16]. The sample size was approxi￾mately 1 mm by 1.5 mm by 1.5 mm. Surface areas were calculated from external sample dimensions measured to ±0.01 mm. The weight changes for composites were measured under flowing air at a ramp rate of 10◦C/min up to 1500◦C without an isothermal hold. The resulting oxide layer was examined using SEM and EDS. 3. Results and discussion 3.1. Conventional ZrB2-SiC UHTCs Archimedes’ density measurements for conventional ZrB2-based UHTCs indicated that little or no poros￾ity was present after hot pressing. With the excep￾tion of samples containing 10% SiC, the density of all of the samples was greater than 100% relative den￾sity based on the as-batched ZrB2 and SiC contents (Table III) and the assumed theoretical densities of ZrB2 (6.09 g/cm3) and SiC (3.21 g/cm3). The higher than predicted density of the hot pressed composites was attributed to the introduction of WC impurities during attrition milling. Comparison of bulk densities determined by the Archimedes’ method to estimated true densities determined by helium pycnometry in￾dicated that 1.4 to 2.2 vol% WC (theoretical density ∼15.0 g/cm3 for WC bonded with 6% Co) was present in the hot pressed composites. The presence of WC was confirmed by XRD and SEM/EDS (discussed be￾low). The low density of the 10% SiC sample (∼93% relative) manifests itself as large, isolated pores. Such pores might be formed by agglomeration of the starting powder or by an unknown chemical effect during hot pressing. Examination of ZrB2-SiC UHTCs by SEM showed that the SiC particles were well dispersed in the ZrB2 matrix (Fig. 1). The SiC particle size ranged from less than a micron to approximately 5 µm. In addi￾tion to ZrB2 and SiC, two other phases were observed. X-ray diffraction analysis (Fig. 2) and EDS confirmed the presence of WC (along with the possible presence of WC1−x and/or W2C) in the form of discrete particles (Fig. 1). An unidentified phase was observed by SEM. By EDS, it contained W, Zr and B. However, a reliable composition could not be determined by EDS nor could the low intensity, unindexed peaks in the XRD pattern TABLE III Density of hot pressed ZrB2-based UHTCs determined by Archimedes’ method, calculated from ZrB2 and SiC contents, and determined by He pycnometry along with the WC content of hot pressed composites Archimedes’ Theoretical Helium WC density density density Content Composition (g/cm3) (g/cm3) (g/cm3) (vol%) ZrB2 6.26 6.09 6.27 1.9 ZrB2-10% SiC 5.54 5.80 5.94 1.4 ZrB2-20% SiC 5.69 5.51 5.74 2.3 ZrB2-30% SiC 5.43 5.23 5.46 2.2 5953

ULTRA-HIGH TEMPERATURE CERAMICS Sic C Unknown phase 5 Figure I Polished cross section of a ZrB2-20% SiC composite Strength and Toughness of zrB-SIC 1000 8 。§强 600 8 nN 一 Strength(MPa gness(MPa"m") 20 Two Theta( Degrees Figure 3 Four point bend strength and fracture toughness of ZrB2-SiC Figure 2 XRD pattern of ZrB2-SiC composites after hot pressing, whi a function of sic content indicated that ZrB2(4), SiC (S), and wC or related phases (W) noted for the 10% SiC specimen, as would be expected be attributed to known Zr, b, C, Si, or w-containing for a porous material. Otherwise, no trend in modulus was observed as a function of composition. Likewise phases.A previous investigation of phase equilibria no compositional trend was noted in hardness as all of in the Zr-B-W system postulated that a ternary phase might form, but the investigators could not confirm the the average values were 24 GPa(Table IV).Hardness was not adversely affected by the porosity of the 10% transmission electron microscopy study is underway SiC sample, perhaps because none of the indents used in our laboratories to determine the composition and structure of this phase relatively large, isolated pores. In contrast to modulus and hardness, the addition of Sic to ZrB2 improved The hardness, Youngs modulus, strength, and tough ess were determined for ZrB,-SiC as a function of Sic the bend strength and fracture toughness(Fig. 3). Pure content. The modulus values were between 450 and ZrB, had an average strength of 565 MPa. Strength increased as sic content increased to a maximum of 490GPa(Table IV). a slight decrease in modulus was 1089 MPa for the 30% SiC sample. The increase in strength has been attributed to a combination of grain TABLE IV Hardness and modulus measured for ZrB , -SiC size reduction and the wC impurities incorporated dur- Composition Hardness (GPa Modulus(GPa) ng milling [10]. Fracture toughness also increased as Sic content increased. For ZrB,. a toughness of 3.5 MPa mi/2 was determined. Toughness increased to ZrB2.10% Si 4士0.9 5.3 MPa for the 30% SiC sample. To understand the ZrB2-20% SiC toughening mechanism, the path of indentation cracks ZrB2-30% SiC 4士0.7 was examined in the sem. As shown in Fig. 4. the Sic particles appear to bridge the cracks, which should 5954

ULTRA-HIGH TEMPERATURE CERAMICS Figure 1 Polished cross section of a ZrB2-20% SiC composite. Figure 2 XRD pattern of ZrB2-SiC composites after hot pressing, which indicated that ZrB2 (Z), SiC (S), and WC or related phases (W) were present. be attributed to known Zr, B, C, Si, or W-containing phases. A previous investigation of phase equilibria in the Zr-B-W system postulated that a ternary phase might form, but the investigators could not confirm the composition or structure of the compound [17, 18]. A transmission electron microscopy study is underway in our laboratories to determine the composition and structure of this phase. The hardness, Young’s modulus, strength, and tough￾ness were determined for ZrB2-SiC as a function of SiC content. The modulus values were between 450 and 490 GPa (Table IV). A slight decrease in modulus was TABLE IV Hardness and modulus measured for ZrB2-SiC Composition Hardness (GPa) Modulus (GPa) ZrB2 23 ± 0.9 489 ZrB2-10% SiC 24 ± 0.9 450 ZrB2-20% SiC 24 ± 2.8 466 ZrB2-30% SiC 24 ± 0.7 484 Figure 3 Four point bend strength and fracture toughness of ZrB2-SiC as a function of SiC content. noted for the 10% SiC specimen, as would be expected for a porous material. Otherwise, no trend in modulus was observed as a function of composition. Likewise, no compositional trend was noted in hardness as all of the average values were ∼24 GPa (Table IV). Hardness was not adversely affected by the porosity of the 10% SiC sample, perhaps because none of the indents used to calculate the average value were in proximity to the relatively large, isolated pores. In contrast to modulus and hardness, the addition of SiC to ZrB2 improved the bend strength and fracture toughness (Fig. 3). Pure ZrB2 had an average strength of 565 MPa. Strength increased as SiC content increased to a maximum of 1089 MPa for the 30% SiC sample. The increase in strength has been attributed to a combination of grain size reduction and the WC impurities incorporated dur￾ing milling [10]. Fracture toughness also increased as SiC content increased. For ZrB2, a toughness of 3.5 MPa·m1/2 was determined. Toughness increased to 5.3 MPa for the 30% SiC sample. To understand the toughening mechanism, the path of indentation cracks was examined in the SEM. As shown in Fig. 4, the SiC particles appear to bridge the cracks, which should 5954

ULTRA-HIGH TEMPERATURE CERAMICS ZrB2 WC Crack 10 Figure 4 Path of an indentation crack in ZrB2-20% SiC showing possible crack bridging increase to ughness. The toughness increase observed 3.2. Co-extruded UHTCs for ZrB2-SiC was nearly identical to the 2 MPa in- The bulk density of fibrous monoliths with ZrB2-30% crease predicted by Ohji et al. for an analogous system SiC cells and porous ZrB2 cell boundaries was 5.11 (same relative elastic moduli and thermal expansion g/cm after hot pressing. Based on the helium pycnom coefficients) that exhibited similar fracture behavior etry results on this composition, the relative density of [19J the overall fibrous monolith was 91%. Assuming fully The addition of Sic to ZrB2 also improved its dense ZrB2-30% SiC cells and a theoretical density of oxidation resistance. Analysis by TGA found that 6.1 g/cm for the ZrB2 cell boundaries, the relative den weight gain began at "700C when ZrB2 or ZrB2-Sic sity of the cell boundaries was estimated to be 72% samples were heated in air. The rate of weight gain using a rule of mixtures calculation. Examination of was similar for all compositions up to-1200C Above the microstructure by SEM revealed that the desired 1200C, the oxidation rate of ZrB2 increased dramat- cellular architecture(high density ZrB2-SiC cells with ically resulting in a weight gain of 13.0 mg/cm by porous ZrB 2 grain boundaries)was maintained through 1500C. The addition of Sic decreased the normalized the co-extrusion and densification steps(Fig. 6). Fur- weight gain(Table V)for ZrB2 UHTCs; the sample con- ther, SEM analysis revealed that the cell boundaries taining 30% SiC gained the least weight at 1.8 mg/em- were made up of grains that had an average size of by 1500C. Examination by SEM/EDS showed that the nearly 10 um with interconnected open porosity.The higher oxidation resistance of the Sic-containing sam- SiC inclusions were well distributed within the ZrB ples was imparted by a SiOr-rich layer 4 um thick matrix of the ZrB2-SiC cells. Some closed porosity was that formed during heating(Fig. 5). No such layer was observed in the cells, but the porosity was estimated to observed on Zrb2 that did not contain sic be less than 5 vol%o and was not considered in the cell boundary density calculation. Some of porosity appar- ent on the polished surface has been attributed to grain TABLE V Weight gain for ZrB2-SiC as determined by TGA pullout during polishing Composition Analyses of hardness, modulus, and strength have Weight gain(mg/cm") been completed for ZrB2-based fibrous monoliths 13.0 (Table Vi). The values were all less than the values mea ZrB,-10% SiC sured for conventional ZrB-30% Sic that was used ZrB2-20% SiC as the cell material. The bend strength of the fibrous ZrB,-30% SiC monolith was 450 MPa, compared to-1000 MPa fo the conventional UHTC with the same composition 5955

ULTRA-HIGH TEMPERATURE CERAMICS Figure 4 Path of an indentation crack in ZrB2-20% SiC showing possible crack bridging. increase toughness. The toughness increase observed for ZrB2-SiC was nearly identical to the 2 MPa in￾crease predicted by Ohji et al. for an analogous system (same relative elastic moduli and thermal expansion coefficients) that exhibited similar fracture behavior [19]. The addition of SiC to ZrB2 also improved its oxidation resistance. Analysis by TGA found that weight gain began at ∼700◦C when ZrB2 or ZrB2-SiC samples were heated in air. The rate of weight gain wassimilar for all compositions up to ∼1200◦C. Above 1200◦C, the oxidation rate of ZrB2 increased dramat￾ically resulting in a weight gain of 13.0 mg/cm2 by 1500◦C. The addition of SiC decreased the normalized weight gain (Table V) for ZrB2 UHTCs; the sample con￾taining 30% SiC gained the least weight at 1.8 mg/cm2 by 1500◦C. Examination by SEM/EDS showed that the higher oxidation resistance of the SiC-containing sam￾ples was imparted by a SiO2-rich layer ∼4 µm thick that formed during heating (Fig. 5). No such layer was observed on ZrB2 that did not contain SiC. TABLE V Weight gain for ZrB2-SiC as determined by TGA Composition Weight gain (mg/cm2) ZrB2 13.0 ZrB2-10% SiC 4.3 ZrB2-20% SiC 3.1 ZrB2-30% SiC 1.8 3.2. Co-extruded UHTCs The bulk density of fibrous monoliths with ZrB2-30% SiC cells and porous ZrB2 cell boundaries was 5.11 g/cm3 after hot pressing. Based on the helium pycnom￾etry results on this composition, the relative density of the overall fibrous monolith was 91%. Assuming fully dense ZrB2-30% SiC cells and a theoretical density of 6.1 g/cm3 for the ZrB2 cell boundaries, the relative den￾sity of the cell boundaries was estimated to be 72% using a rule of mixtures calculation. Examination of the microstructure by SEM revealed that the desired cellular architecture (high density ZrB2-SiC cells with porous ZrB2 grain boundaries) was maintained through the co-extrusion and densification steps (Fig. 6). Fur￾ther, SEM analysis revealed that the cell boundaries were made up of grains that had an average size of nearly 10 µm with interconnected open porosity. The SiC inclusions were well distributed within the ZrB2 matrix of the ZrB2-SiC cells. Some closed porosity was observed in the cells, but the porosity was estimated to be less than 5 vol% and was not considered in the cell boundary density calculation. Some of porosity appar￾ent on the polished surface has been attributed to grain pullout during polishing. Analyses of hardness, modulus, and strength have been completed for ZrB2-based fibrous monoliths (Table VI). The values were all less than the values mea￾sured for conventional ZrB2-30% SiC that was used as the cell material. The bend strength of the fibrous monolith was ∼450 MPa, compared to ∼1000 MPa for the conventional UHTC with the same composition as 5955

TRA-HIGH TEMPERATURE CERAMICS OrB SiC Oxide layer Oxygen not Oxygen present present 20m 20 Figure 5(a)An SEM image and (b)an oxygen map for ZrB2-30% SiC after heating to 1500C in ai ZrB-SiC Cell ZrB, Cell oundary 200p 10 um (b) Figure 6 SEM micrographs of (a)a fibrous monolith with, (b)dense ZrB2-SiC cells, and (c)porous ZrB2 cell boundaries. the cell phase. The measured elastic modulus of the mixing model assumed that the volume fraction of cell fibrous monolith was 416 GPa. The measured modu- boundary phase was 0.3; it incorporated the measured lus was nearly identical to a modulus of 415 GPa modulus of the cell phase(484 GPa for ZrB2-30% SiC that was predicted using a series mixing model. The from Table IV), and used a calculated modulus of the porous ZrB2 cell boundary phase of 253 GPa. The mod- TABLE VI Average density, four-point bend strength, hardness, and ulus of the porous cell boundary phase was estimated and cell-boundary phases and a conventional ZrB2-30% Sic cerami Equation 2 [20] trength Hardness modulus Ep= Eo(1-P) 5.11 ZrB,-SiC cell ZrB2 cell boundary 4.00 where Ep is the modulus of a porous ceramic, Eo is the ZrB2-30% SiC 1031 modulus of a dense ceramic of the same composition, and p is the volume fractio 5956

ULTRA-HIGH TEMPERATURE CERAMICS Figure 5 (a) An SEM image and (b) an oxygen map for ZrB2-30% SiC after heating to 1500◦C in air. Figure 6 SEM micrographs of (a) a fibrous monolith with, (b) dense ZrB2-SiC cells, and (c) porous ZrB2 cell boundaries. the cell phase. The measured elastic modulus of the fibrous monolith was 416 GPa. The measured modu￾lus was nearly identical to a modulus of ∼415 GPa that was predicted using a series mixing model. The T A B L E V I Average density, four-point bend strength, hardness, and modulus of ZrB2-based fibrous monoliths compared to values for the cell and cell-boundary phases and a conventional ZrB2-30% SiC ceramic Density Strength Hardness Modulus Composition (g/cm3) (MPa) (GPa) (GPa) Fibrous monolith 5.11 459 – 416 ZrB2-SiC cell 5.23 – 25 – ZrB2 cell boundary 4.00 – 7 – ZrB2-30% SiC 5.43 1031 24 484 mixing model assumed that the volume fraction of cell boundary phase was 0.3; it incorporated the measured modulus of the cell phase (484 GPa for ZrB2-30% SiC from Table IV), and used a calculated modulus of the porous ZrB2 cell boundary phase of 253 GPa. The mod￾ulus of the porous cell boundary phase was estimated using a relation for porous materials reported by Clegg, Equation 2 [20]. EP = Eo(1 − P) 2 (2) where EP is the modulus of a porous ceramic, Eo is the modulus of a dense ceramic of the same composition, and P is the volume fraction porosity. 5956

ULTRA-HIGH TEMPERATURE CERAMICS Based on the interaction of a fracture crack with the Acknowledgements cellular architecture, the apparent fracture toughness This work was supported by the Ceramics and Non- of fibrous monolithic ceramics can increase by a fac- Metallic Materials Program in the Air Force Office tor of 10 or more compared to the measured fracture of Scientific Research on Grant Number F49620-03-1 toughness of the cell material [14]. The fracture tough- 0072. One of the authors, ALC, is supported by a grad ness of the cell material used in this fibrous monolith uate Assistance in Areas of National Need (GAANN) was 5.25 MPa- mI/(Fig 3). However, four point bend Fellowship sponsored by the U.S. Department of Ed testing of ZrB2-based fibrous monoliths found no indi- ucation and from a grant from the University of Mis- cation of load retention after the initial fracture event. souri research board. the authors wish to thank ms as has been observed in other fibrous monolith sys- Michelle Schaeffler for technical assistance. Finally,we tems. Thus, only a minimal improvement in fracture would like to thank H.C. Starck for donating the Grade ysis of laminate structures by Clegg et al. suggests that 2rE toughness would be expected in these materials. Anal- A ZrB2 powder used in this work a porosity level of 40% is necessary to sufficiently weaken a porous layer so that it can promote crack de- References flection [20], compared to less than 30% porosity in 1. R. TELLE, L. S. SIGL and K. TAKAGI, in"Handbook of Ce. the porous cell boundaries in the fibrous monolith pre ramic Hard Materials"edited by R. Riedel (Wiley-VCH, Weinheim, pared in this study. Currently, specimen preparation and 2000p.803 2. R.A. CUTLER, in"Ceramics and Glasses, Engineered Materi- densification procedures are being modified to increase als Handbook", edited by S.J. Schneider, Jr (ASM International the porosity in the cell boundary to weaken it without Materials Park, OH, 1991)p 787 adversely affecting the density of the cells. The first ap- 3.N. KAJI, H. SHIKANO and I. TANAKA, Taikabutsu Over- proach will be to evaluate the effect of using larger ZrB2 4 G,- ZHANG, 2-Y, DENG N, KONDOL-F. YaNG and T. OHJI, J. Amer Ceram Soc. 83(2000)2330 sinter less effectively, resulting in a lower relative den- 5. F. MONTEVERDE, A. BELLoSI ands. BUICCIARDI,J. sity of the cell boundary. The second approach will be to Euro. Ceram Soc. 22(2002)2 add an inert polymeric filler to the cell boundary mate C. MRoZ."Zirconium Diboride". Amer: Ceram. Soc. Bull. rial(zrB2-thermoplastic polymer) prior to co-extrusion (1994)141. to decrease the solids loading of the cell boundary. Th 7.A. GOLDSTEIN. Y. gEFFEn and A. GOLD J Amer Ceran. Soc. 84 (2001)642 cell boundary porosity would be increased when the 8 H. ZHAO, Y. HE and Z. JIN, ibid. 78(1995)2534 filler is removed during binder burnout. 9.S MISHRA, S. K. DAS, A. K. RAY and P RAMACHANDRARAO, ibid. 85(2002)2846. 10.A G. FAHRENHOLTZ. G. E HILMAS and D. T. ELLERBY, High Strength ZrB2-Based 4. Summary and conclusions Ceramics, "J. Amer. Ceram Soc. 87(6)(2004)1170. Conventional ZrB2-based UHTCs were prepared by hot Figure 9020 in "Phase Equilibria Diagrams--Phase Diagrams for Ceramists Vol. X: Borides, Carbides, and Nitrides", edited by A.E. pressing. The addition of SiC improved mechanical McHale(The American Ceramic Society, 1994) performance and oxidation resistance of ZrB2-based 12 12. E. CLOUGHERTY, R. HILL, W. RHODES and E UHTCS Four-point bend strength increased from 560 PETERS, " Research and Development of Refractory Oxidation- MPa for pure ZrB2 to more than 1000 MPa for ZrB2 Resistant Diborides, Part Il, Vol. ll: Processing and Characteriza- 30% SiC. Likewise, fracture toughness increased fro tion", Technical Report No. AFML-TR-68-190(1970) 13. D. POPOVICH, J. W. HALLORAN, G. E. HILMAS, G. 3.5 MPa-mi/2 for pure ZrB2 to more than 5 MPa.mi/2 A. BRADY, G. ZYWICKI, S. SOMERS and A. BARDA. for ZrB2-30% SiC. The increase in toughness was at "Process for Preparing Textured Ceramic Composites", U.S. Patent No. 5,645, 781, issue date July 8, 1997 tributed to crack bridging by dispersed Sic particles 14. G.E. HILMAS,.A BRADY andJ.W.HALLORAN,in in the zrB matrix Oxidation behavior. examined us- ing TGA, found that the weight gain decreased from Ceramic Transactions, Vol. 51, Fifth Int'I Conf on Ceramic Process- ing Science and Technology, edited by H. Hausner, G L Messing 0 mg/cm for rB, to 1. 8 mg/cm- for ZrB 30%SiC Improved oxidation resistance relative to pure p.609 rB2 was imparted by the formation of a refractory 15.J. S. REED, in"Principles of Ceramics Processing"(ohn wiley SiO2-based glassy layer on composites containing SiC ZrB2-based fibrous monoliths were formed by co- 16. High Temperature Oxidation Exposure Testing of Non-Oxide Ad- vanced Ceramics at Atmospheric Pressure and Low Gas Velocities ZrB2-30% SC ng. The fibrous monoliths had dense under consideration by ASTM C-28 extrusion proces y porous ZrB2 17. Figure 8851 in"Phase Equilibria Diagrams--Phase Diagrams for cell boundaries. From density measurements, the cell Ceramists Vol. X: Borides, Carbides, and Nitrides A E. MchALE (The American Ceramic Society, 1994). boundarieswere-72%dense(28%porosity).The18.e.rudy,"partv.compendIumofphAsediagramData",temary average strength of the ZrB2-based fibrous monoliths Phase Equilibria in Transition Metal-Boron-Carbo was 450 MPa, compared to m1000 MPa for a con- tems, Technical Report AFML-TR-65-2, Air Force als lab. ventional ZrB2-30% SiC ceramic. Examination of the oratory, Wright-Patterson Air Force Base, OH(1969) fracture behavior showed no load retention after the 19. T. OHJI.Y. K. JEONG. Y.H. CHOA and K. NIIHARA initial failure event in four point bend testing, indicat- J. Amer Cera. Soc. 81(1998)1453 20. K.S. BLANKS. A. KRISTOFFERSSON. E. CARLSTOM ing that further weakening of the grain boundary phase CLEGG, J. Euro Ceram Soc. 18(1998)1945 is needed to develop high apparent fracture toughness and promote graceful failure in ZrB2-SiC/porous ZrB2 Received 9 October 2003 and accepted 15 March 2004 5957

ULTRA-HIGH TEMPERATURE CERAMICS Based on the interaction of a fracture crack with the cellular architecture, the apparent fracture toughness of fibrous monolithic ceramics can increase by a fac￾tor of 10 or more compared to the measured fracture toughness of the cell material [14]. The fracture tough￾ness of the cell material used in this fibrous monolith was 5.25 MPa·m1/2 (Fig. 3). However, four point bend testing of ZrB2-based fibrous monoliths found no indi￾cation of load retention after the initial fracture event, as has been observed in other fibrous monolith sys￾tems. Thus, only a minimal improvement in fracture toughness would be expected in these materials. Anal￾ysis of laminate structures by Clegg et al. suggests that a porosity level of ∼40% is necessary to sufficiently weaken a porous layer so that it can promote crack de- flection [20], compared to less than 30% porosity in the porous cell boundaries in the fibrous monolith pre￾pared in this study. Currently, specimen preparation and densification procedures are being modified to increase the porosity in the cell boundary to weaken it without adversely affecting the density of the cells. The first ap￾proach will be to evaluate the effect of using larger ZrB2 particles for the cell boundaries. Larger particles should sinter less effectively, resulting in a lower relative den￾sity of the cell boundary. The second approach will be to add an inert polymeric filler to the cell boundary mate￾rial (ZrB2-thermoplastic polymer) prior to co-extrusion to decrease the solids loading of the cell boundary. The cell boundary porosity would be increased when the filler is removed during binder burnout. 4. Summary and conclusions Conventional ZrB2-based UHTCs were prepared by hot pressing. The addition of SiC improved mechanical performance and oxidation resistance of ZrB2-based UHTCs. Four-point bend strength increased from ∼560 MPa for pure ZrB2 to more than 1000 MPa for ZrB2- 30% SiC. Likewise, fracture toughness increased from ∼3.5 MPa·m1/2 for pure ZrB2 to more than 5 MPa·m1/2 for ZrB2-30% SiC. The increase in toughness was at￾tributed to crack bridging by dispersed SiC particles in the ZrB2 matrix. Oxidation behavior, examined us￾ing TGA, found that the weight gain decreased from 13.0 mg/cm2 for pure ZrB2 to 1.8 mg/cm2 for ZrB2- 30% SiC. Improved oxidation resistance relative to pure ZrB2 was imparted by the formation of a refractory SiO2-based glassy layer on composites containing SiC. ZrB2-based fibrous monoliths were formed by co￾extrusion processing. The fibrous monoliths had dense ZrB2-30% SiC cells surrounded by porous ZrB2 cell boundaries. From density measurements, the cell boundaries were ∼72% dense (∼28% porosity). The average strength of the ZrB2-based fibrous monoliths was ∼450 MPa, compared to ∼1000 MPa for a con￾ventional ZrB2-30% SiC ceramic. Examination of the fracture behavior showed no load retention after the initial failure event in four point bend testing, indicat￾ing that further weakening of the grain boundary phase is needed to develop high apparent fracture toughness and promote graceful failure in ZrB2-SiC/porous ZrB2 fibrous monoliths. Acknowledgements This work was supported by the Ceramics and Non￾Metallic Materials Program in the Air Force Office of Scientific Research on Grant Number F49620-03-1- 0072. One of the authors, ALC, is supported by a Grad￾uate Assistance in Areas of National Need (GAANN) Fellowship sponsored by the U.S. Department of Ed￾ucation and from a grant from the University of Mis￾souri Research Board. The authors wish to thank Ms. Michelle Schaeffler for technical assistance. Finally, we would like to thank H.C. Starck for donating the Grade A ZrB2 powder used in this work. References 1. R. TELLE, L. S . SIGL and K. TAKAGI, in “Handbook of Ce￾ramic Hard Materials” edited by R. Riedel (Wiley-VCH, Weinheim, 2000) p. 803. 2. R.A. CUTLER, in “Ceramics and Glasses, Engineered Materi￾als Handbook”, edited by S. J. Schneider, Jr. (ASM International, Materials Park, OH, 1991) p. 787. 3. N. KAJI, H. SHIKANO and I. TANAKA, Taikabutsu Over￾seas 44 (1992) 387. 4. G.-J. ZHANG, Z.-Y. DENG, N. KONDO, J.- F . YANG and T. OHJI, J. Amer. Ceram. Soc. 83 (2000) 2330. 5. F . MONTEVERDE, A. BELLOSI and S . BUICCIARDI, J. Euro. Ceram. Soc. 22 (2002) 279. 6. C. MROZ, “Zirconium Diboride”, Amer. Ceram. Soc. Bull. 73 (1994) 141. 7. A. GOLDSTEIN, Y. GEFFEN and A. GOLDENBERG, J. Amer. Ceram. Soc. 84 (2001) 642. 8. H. ZHAO, Y. H E and Z. JIN, ibid. 78 (1995) 2534. 9. S . K. MISHRA, S . K. DAS , A. K. RAY and P . RAMACHANDRARAO, ibid. 85 (2002) 2846. 10. A. L. CHAMBERLAIN, W. G. FAHRENHOLTZ, G. E. HILMAS and D. T. ELLERBY, “High Strength ZrB2-Based Ceramics,” J. Amer. Ceram. Soc. 87(6) (2004) 1170. 11. Figure 9020 in “Phase Equilibria Diagrams—Phase Diagrams for Ceramists Vol. X: Borides, Carbides, and Nitrides”, edited by A. E. McHale (The American Ceramic Society, 1994). 12. E. CLOUGHERTY, R. HILL, W. RHODES and E. PETERS , “Research and Development of Refractory Oxidation￾Resistant Diborides, Part II, Vol. II: Processing and Characteriza￾tion”, Technical Report No. AFML-TR-68-190 (1970). 13. D. POPOVICH, J. W. HALLORAN, G. E. HILMAS , G. A. BRADY, G. ZYWICKI, S . SOMERS and A. BARDA, “Process for Preparing Textured Ceramic Composites”, U.S. Patent No. 5,645,781, issue date July 8, 1997. 14. G. E. HILMAS , G. A. BRADY and J. W. HALLORAN, in Ceramic Transactions, Vol. 51, Fifth Int’l Conf. on Ceramic Process￾ing Science and Technology, edited by H. Hausner, G. L Messing and S.-I. Hirano (American Ceramic Society, Westerville, OH, 1995) p. 609. 15. J. S . REED, in “Principles of Ceramics Processing” (John Wiley and Sons, New York, 1995) p. 280. 16. “High Temperature Oxidation Exposure Testing of Non-Oxide Ad￾vanced Ceramics at Atmospheric Pressure and Low Gas Velocities”, under consideration by ASTM C-28. 17. Figure 8851 in “Phase Equilibria Diagrams—Phase Diagrams for Ceramists Vol. X: Borides, Carbides, and Nitrides”, edited by A. E. McHALE (The American Ceramic Society, 1994). 18. E. RUDY, “Part V. Compendium of Phase Diagram Data”, Ternary Phase Equilibria in Transition Metal-Boron-Carbon-Silicon Sys￾tems, Technical Report AFML-TR-65-2, Air Force Materials Lab￾oratory, Wright-Patterson Air Force Base, OH (1969). 19. T. OHJI, Y. K. JEONG, Y. H. CHOA and K. NIIHARA, J. Amer. Ceram. Soc. 81 (1998) 1453. 20. K. S . BLANKS , A. KRISTOFFERSSON, E. CARLSTO M¨ and W. J. CLEGG, J. Euro. Ceram. Soc. 18 (1998) 1945. Received 9 October 2003 and accepted 15 March 2004 5957

Copyright of journal of Materials Science is the property of Kluwer Academic Publishing/ Academic and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder,'s express written permission However, users may print, download, or email articles for individual use