16l6 Journal of the American Ceramic Society- dotter and halloran Vol. 89. No. 5 Table L. Composition and Densities of the fabricated Si3N4/ Table ll. Flexural Strength and WOF of Si3 N,/BN FMs with BN FMS Ytterbium oxide(FMYb)and Si3,/BN FMs with Lanthanum oxide(FmLa) Ytterbium oxide(wt%) oxide(wt%) Work-of-fracture (J/m") FMYB 8 3.09+0.1 FMYB FMLA 295+0.1 Average 2126 Standard FMLA used in place of the reactive La2 O3. Heating La(oH) in dry air Average Standard 61 our fabrication, the La(oh)3 then dehydrated during hot press- ing and transformed into LazO. both billets consisted of fila- ments stacked up, forming a 3D structure with x250 um Si3N4 ells(MIl. H.C. Starck, MA)uniaxially aligned and separated the FMla and FMyB The Si3 N4/BN FMs with by x15 um BN cell boundaries(6003, Advanced Ceramics Yb2O3 showed average hig ength and graceful fail orp, OH). The densities of the specimens were measured ure and comparable strengt ommercially available uniax using the Archimedes method, and the theoretical density of N FMs with 6 wt%Y,O3 and 2 wt% Al,O3, with the specimen was estimated by the rule of mixture where the the average flexural strength reported as 510+87 MPa. On the vol% of the bn phase was estimated to be 20 and the Si3 N4 other hand, the Si3 N4/BN FMs with La,O3 showed lower flex phase to be 80. The densities and compositions of the billets are ural strength and brittle failure in the majority of the samples. The FMYB had 97.8% theoretical density and FMLa had 94.6% theoretical density; this is not believed to be a large (2) Flexural Testing enough difference to explain completely the poor mechanical Flexural bars were prepared for four-point flexural testing. The Behavior of the FMLA samples. Figure I shows an example ofa billets were first ground with a 220 grit--diamond whe flexural response, a graceful failure, of the FMYB sample in a then cut into 2.2 mm x4.2 mmx 49 mm bars. The sides of the stress was 448 MPa and there was a large load drop where the sides of the bars were polished down to I um using a medium J/m2. Figure 2 shows an SEM micrograph of the side view of the room temperature in laboratory air using a computer-control corresponding FMYB sample, where the tensile initiation of the led, screw-driven, testing machine(Model 4483, Instron Corp, fracture can be seen along with crack defection and subsequent Canton, MA). The specimens were tested using a four-poin delamination cracking and sliding along the side surface. The flexural testing fixture with an inner and an outer span of 20 ane two fracture modes tensile and shear initiation were observed 40 mm and at a cross-head speed of O 5 mm/min. Load versus example of stress versus cross-head displacement curve of an eported here. Flexural strength is defined as the apparent flex FMLA sample that fractured by shear initiation and failed ural stress at load drop. Energy absorption capability gracefully. The apparent peak stress was 233 MPa for this sam- acterized by the woF, calculated by determining the are ple and the load drop was less compared with the tensile-initi ated fracture shown in Fig. 1. The retained apparent stress was 160 MPa and the WOF was 1304 J/m. Figure 4 shows the ross-sectional area of the sample. Scanning electron 1 d failure initiated in the flexural bar between the outer and inner amination cracking and sliding on the side surfaces of the tested flexural bars higher retained apparent stress than the tensile- initiated fracture samples. It was observed that all the samples that failed grace- (3) Oxidation Testing fully(non-catastrophically) required extensive crack interaction 1400c in dry ai for eo hn The fur nace was heated at a heati ng The average woF for FMLa was lower than for the FMri rate of 100C/min and then maintained at 1400.C for 10 h. Be- oxidation of the samples,2.2mm×4.2mm×20 manner and therefore there was were polished down to I um using a medium polishing diamond fewer delamination and A brick deflection, resulting in less energy disk. The specimens were then ultrasonically cleaned in acetone and dried before oxidation. The materials were characterized by X-ray diffractometry(XRD)before and after the oxidation test. position and morphology of an oxide layer produced after oxidation were characterized by sEM and X-ray energy 400 dispersive spectroscopy (XEDS) II. Results and discussion (1) Mechanical Properties FMs are laminates and can fail in two different modes durin flexural testing by shear initiation, where the shear stress be- tween the inner and outer loading pins in the middle of the beam exceeds the shear strength of the material, and by tensile initi- ation, where the tensile stress in the outer layer of the tensile 0.000080.170250.330420.500.580.67 urface exceeds the tensile strength of the material. Here we re- Crosshead Displacement [mm] port fracture of the FMs by two modes: tensile initiation and Fig 1. Flexural response of Si3N4/BN FMs with ytterbium oxide shear initiation. Table il gives the average strength and woF of(FMYBused in place of the reactive La2O3. Heating La(OH)3 in dry air led to a progressive dehydration of the La(OH)3 to La2O3. 18 In our fabrication, the La(OH)3 then dehydrated during hot press￾ing and transformed into La2O3. Both billets consisted of fila￾ments stacked up, forming a 3D structure with B250 mm Si3N4 cells (M11, H.C. Starck, MA) uniaxially aligned and separated by B15 mm BN cell boundaries (6003, Advanced Ceramics Corp., OH). The densities of the specimens were measured using the Archimedes method, and the theoretical density of the specimen was estimated by the rule of mixture where the vol% of the BN phase was estimated to be 20 and the Si3N4 phase to be 80. The densities and compositions of the billets are shown in Table I. (2) Flexural Testing Flexural bars were prepared for four-point flexural testing. The billets were first ground with a 220 grit—diamond wheel and then cut into 2.2 mm
4.2 mm
49 mm bars. The sides of the bars were chamfered to minimize machining flaws. The tensile sides of the bars were polished down to 1 mm using a medium polishing diamond disk. The flexural strength was measured at room temperature in laboratory air using a computer-control￾led, screw-driven, testing machine (Model 4483, Instron Corp., Canton, MA). The specimens were tested using a four-point flexural testing fixture with an inner and an outer span of 20 and 40 mm and at a cross-head speed of 0.5 mm/min. Load versus cross-head deflection response and work of fracture (WOF) are reported here. Flexural strength is defined as the apparent flex￾ural stress at load drop. Energy absorption capability is char￾acterized by the WOF, calculated by determining the area under the load–cross-head deflection curve and dividing it by twice the cross-sectional area of the sample. Scanning electron micros￾copy (SEM) was used for examining crack deflection, and de￾lamination cracking and sliding on the side surfaces of the tested flexural bars. (3) Oxidation Testing Oxidation studies were conducted in a vertical tube furnace at 14001C in dry air for 10 h. The furnace was heated at a heating rate of 1001C/min and then maintained at 14001C for 10 h. Be￾fore oxidation of the samples, 2.2 mm
4.2 mm
20 mm bars were polished down to 1 mm using a medium polishing diamond disk. The specimens were then ultrasonically cleaned in acetone and dried before oxidation. The materials were characterized by X-ray diffractometry (XRD) before and after the oxidation test. The composition and morphology of an oxide layer produced after oxidation were characterized by SEM and X-ray energy￾dispersive spectroscopy (XEDS). III. Results and Discussion (1) Mechanical Properties FMs are laminates and can fail in two different modes during flexural testing by shear initiation, where the shear stress be￾tween the inner and outer loading pins in the middle of the beam exceeds the shear strength of the material, and by tensile initi￾ation, where the tensile stress in the outer layer of the tensile surface exceeds the tensile strength of the material. Here, we re￾port fracture of the FMs by two modes: tensile initiation and shear initiation. Table II gives the average strength and WOF of the FMLA and FMYB samples. The Si3N4/BN FMs with Yb2O3 showed average high flexural strength and graceful fail￾ure and comparable strength of commercially available uniax￾ially Si3N4/BN FMs with 6 wt% Y2O3 and 2 wt% Al2O3, with the average flexural strength reported as 510787 MPa.4 On the other hand, the Si3N4/BN FMs with La2O3 showed lower flex￾ural strength and brittle failure in the majority of the samples. The FMYB had 97.8% theoretical density and FMLA had 94.6% theoretical density; this is not believed to be a large enough difference to explain completely the poor mechanical behavior of the FMLA samples. Figure 1 shows an example of a flexural response, a graceful failure, of the FMYB sample in a stress versus cross-head displacement curve. The apparent peak stress was 448 MPa and there was a large load drop where the retained apparent stress was B50 MPa and the WOF was 3037 J/m2 . Figure 2 shows an SEM micrograph of the side view of the corresponding FMYB sample, where the tensile initiation of the fracture can be seen along with crack deflection and subsequent delamination cracking and sliding along the side surface. The two fracture modes, tensile and shear initiation, were observed for both the FMYB and FMLA samples. Figure 3 shows an example of stress versus cross-head displacement curve of an FMLA sample that fractured by shear initiation and failed gracefully. The apparent peak stress was 233 MPa for this sam￾ple and the load drop was less compared with the tensile-initi￾ated fracture shown in Fig. 1. The retained apparent stress was B160 MPa and the WOF was 1304 J/m2 . Figure 4 shows the side view of the corresponding FMLA sample where the shear failure initiated in the flexural bar between the outer and inner loading pins. The entire shear-initiated fractured samples had higher retained apparent stress than the tensile-initiated fracture samples. It was observed that all the samples that failed grace￾fully (non-catastrophically) required extensive crack interaction such as crack deflection, delamination cracking, and sliding. The average WOF for FMLA was lower than for the FMYB (Table II) because of the fact that the majority of the FMLA samples fractured in a brittle manner and therefore there was fewer delamination and crack deflection, resulting in less energy 0 100 200 300 400 500 0.00 0.08 0.17 0.25 0.33 0.42 0.50 0.58 0.67 Stress [MPa] Crosshead Displacement [mm] Fig. 1. Flexural response of Si3N4/BN FMs with ytterbium oxide (FMYB). Table I. Composition and Densities of the fabricated Si3N4/ BN FMs Ytterbium oxide (wt%) Lanthanum oxide (wt%) Measured r (g/cm3 ) Theoretical r (g/cm3 ) FMYB 8 — 3.0970.1 3.16 FMLA — 8 2.9570.1 3.12 Table II. Flexural Strength and WOF of Si3N4/BN FMs with Ytterbium oxide (FMYB) and Si3N4/BN FMs with Lanthanum oxide (FMLA) Stress (MPa) Work-of-fracture (J/m2 ) FMYB Average 340 2126 Standard 91 940 FMLA Average 298 1287 Standard 61 384 1616 Journal of the American Ceramic Society—Karlsdottir and Halloran Vol. 89, No. 5