《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_4Fracture toughness of concentric Si3N4-based laminated structures

Availableonlineatwww.sciencedirect.com Science Direct E噩≈RS ELSEVIER Joumal of the European Ceramic Society 29(2009)1825-1829 www.elsevier.comlocate/jeurceramsoc Fracture toughness of concentric Si3 N4-based laminated structures Zoran Krstic. Vladimir d. krstic Centre for Manufacturing of Advanced Ceramics and Nanomaterials, Queen's Universiry, Nicol Hall, Kingston, Ontario, Canada K7L 3N6 Received 1 May 2008; received in revised form 10 October 2008; accepted 16 October 2008 Available online 25 November 2008 A new concentric rectangular laminated structure was designed and fabricated by slip-casting method and densified by pressureless process. On of laminates consists of layers of Si3 N4 with 7 wt. Y2O3 and 3 wt %o Al2O3 as sintering aids, and of interlayers consisting of 50wt% BN wt. Al2O3 designated as SN-(BN+AlO3). The other class of laminates has the same Si, N4 layer composition but different interlayer composition of 90wt. BN and 10 wt %o Si3 N4 designated as SN-(BN+SN). The objective of this paper is to investigate the effects of the number of layers and their thickness on apparent fracture toughness of these laminates. The interfacial layer composition was discussed in terms of its role in toughening of the laminates. For the SN-(BN+Al2O,)laminates the highest apparent fracture toughness of 22 MPam was found in the samples with 7 Si3 N4 layers and for the SN-(BN+SN)laminates the highest apparent fracture toughness of 19.5 MPamwas found 2008 Elsevier Ltd. All rights reserved Keywords: Silicon nitride; Slip-casting: Fracture toughness; Laminated structures 1. Introduction The basic concept in this new design is that, as the crack prop- agates through the base material, it deflects at the weak interfaces duce structures with crack resistance capabilities approaching path of crack propagation in the circular and rectangular cylinder those of fiber composites As a result of this effort, various planar structures is shown schematically in Fig. 2. laminated structures were designed and fabricated possessing an The toughening in these ceramic/ceramic laminates is the apparent fracture toughness and work of fracture significantly crack deflection at the weak interface such that no catastrophic higher that those of monolithic counterpart. Although the failure occurs. This condition is achieved when the strength of improvements in fracture resistance in these planar laminates the interface is sufficiently weak to allow the deflected crack to were sufficient to ensure their safe use in many structural appli- propagate a long distance before changing its direction. cations, delamination and easy crack propagation along the weak It has been shown by Zhang and Krstic that both the numbers interface between the two layers has been th malor int edi- of layers and their thickness play an important role in toughening ment for wider use of these structures(Fig. 1). To overcome and strengthening of the planar laminates. In an attempt to model this unwanted delamination/peeling problem associated with the fracture behaviour of the planar laminates, Clegg assumed the plate-form laminates(see Fig. 1)a concentric rectangular that the fracture toughness of the planar laminates is related to the design has been developed and fabricated in which the potential strength of the laminate and the beam thickness(d)as expressed delamination direction is completely eliminated. 5,6 In addition by the equation of eliminating the direction of easy crack propagation, this new structure exhibits fracture resistance characteristics far beyond those of monolithic ceramics or planar laminates where Kic is the fracture toughness of the laminate, of is the fracture strength, c is the notch length and Yis a constant. Clearly Corresponding author. the toughening in Eq. (1)is based on the level of strength of the E-mail address: krsticz@ queensu ca(Z. Krstic). laminate which is a reasonable assumption considering that the 0955-2219 front matter@ 2008 Elsevier Ltd. All rights reserved. doi: 10.1016/j-jeurceramsoc 2008 10.007

Available online at www.sciencedirect.com Journal of the European Ceramic Society 29 (2009) 1825–1829 Fracture toughness of concentric Si3N4-based laminated structures Zoran Krstic ∗, Vladimir D. Krstic Centre for Manufacturing of Advanced Ceramics and Nanomaterials, Queen’s University, Nicol Hall, Kingston, Ontario, Canada K7L 3N6 Received 1 May 2008; received in revised form 10 October 2008; accepted 16 October 2008 Available online 25 November 2008 Abstract A new concentric rectangular laminated structure was designed and fabricated by slip-casting method and densified by pressureless sintering process. One class of laminates consists of layers of Si3N4 with 7 wt.% Y2O3 and 3 wt.% Al2O3 as sintering aids, and of interlayers consisting of 50 wt.% BN and 50 wt.% Al2O3 designated as SN-(BN + Al2O3). The other class of laminates has the same Si3N4 layer composition but different interlayer composition of 90 wt.% BN and 10 wt.% Si3N4 designated as SN-(BN + SN). The objective of this paper is to investigate the effects of the number of layers and their thickness on apparent fracture toughness of these laminates. The interfacial layer composition was discussed in terms of its role in toughening of the laminates. For the SN-(BN + Al2O3) laminates the highest apparent fracture toughness of 22 MPa m1/2 was found in the samples with 7 Si3N4 layers and for the SN-(BN + SN) laminates the highest apparent fracture toughness of 19.5 MPa m1/2 was found in the samples with 4 Si3N4 layers. © 2008 Elsevier Ltd. All rights reserved. Keywords: Silicon nitride; Slip-casting; Fracture toughness; Laminated structures 1. Introduction Over the last decade there has been concentrated effort to pro￾duce structures with crack resistance capabilities approaching those of fiber composites. As a result of this effort, various planar laminated structures were designed and fabricated possessing an apparent fracture toughness and work of fracture significantly higher that those of monolithic counterpart.1–4 Although the improvements in fracture resistance in these planar laminates were sufficient to ensure their safe use in many structural appli￾cations, delamination and easy crack propagation along the weak interface between the two layers has been the major impedi￾ment for wider use of these structures (Fig. 1). To overcome this unwanted delamination/peeling problem associated with the plate-form laminates (see Fig. 1) a concentric rectangular design has been developed and fabricated in which the potential delamination direction is completely eliminated.5,6 In addition of eliminating the direction of easy crack propagation, this new structure exhibits fracture resistance characteristics far beyond those of monolithic ceramics or planar laminates.7 ∗ Corresponding author. E-mail address: krsticz@queensu.ca (Z. Krstic). The basic concept in this new design is that, as the crack prop￾agates through the base material, it deflects at the weak interfaces oriented transversely to the direction of crack propagation. The path of crack propagation in the circular and rectangular cylinder structures is shown schematically in Fig. 2. The toughening in these ceramic/ceramic laminates is the crack deflection at the weak interface such that no catastrophic failure occurs. This condition is achieved when the strength of the interface is sufficiently weak to allow the deflected crack to propagate a long distance before changing its direction. It has been shown by Zhang and Krstic8 that both the numbers of layers and their thickness play an important role in toughening and strengthening of the planar laminates. In an attempt to model the fracture behaviour of the planar laminates, Clegg9 assumed that the fracture toughness of the planar laminates is related to the strength of the laminate and the beam thickness (d) as expressed by the equation9: KIC = σfY √c 1 − c d 2 (1) where KIC is the fracture toughness of the laminate, σf is the fracture strength, c is the notch length and Y is a constant. Clearly, the toughening in Eq. (1) is based on the level of strength of the laminate which is a reasonable assumption considering that the 0955-2219/$ – see front matter © 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.jeurceramsoc.2008.10.007

Z Krstic, V.D. Krstic /Jounal of the European Ceramic Sociery 29(2009)1825-1829 Fig. 1.(a) Peeling and (b) delamination in the plate-form laminates. Fig. 3. Sample configuration for three-point bending test with corresponding dimensions powder was used as a raw material. Sub-micron size Al2O3 (A-16, Alcoa) and Y203(Alpha Aesar) powders were used as sintering aids BN powder(Carborundum Co., grade HPP-325) with the addition of Al,O3 and Si3 N4 was used for casting weak interlayers. After drying, pressureless sintering was done in a Fig. 2. Schematics of crack deflection in(a)circular and (b)rectangular cross- graphite resistance furnace (Vacuum Industries, USA)at tem- sectioned concentric laminate structures peratures ranging from 1740C to 1800C for I h under static primary crack is blunted and reinitiated at the next layer. This N2 gas atmosphere. Fracture toughness was measured using the three-point bend- nakes the laminate's apparent fracture toug ss Insensitive to ing test at room temperature with a straight-through notch the crack radius introduced in the mid-section of the samples( Fig 3). The notch Shanches-Herencia et al. have showed that the crack exten- was introduced by a 500 um thick diamond wheel through the sion within the layer occurs only when the layer thickness first or first two layers and its depth(o750-1220 um) was exceeds a critical value which is directly related to the critical measured under an optical microscope with 50x or 100x mag- strain energy release rate or fracture toughness cifications. The initial crack radius does not play significant role GeE 0.34 (1-2)2 2) SInce the crack has to be reinitiated on the next layer creating 2) an inherently sharp crack. The test was carried out on an Instron machine(Model 8502 FIB, Instron Co., Canton, USA)using a jig where te is the critical layer thickness, Eis the Young's modulus, with the span of 26 mm and the crosshead speed of 0.06 mm/min v is the Poison's ratio, Ge is the critical strain energy release rate Five samples were tested per data point. The fracture toughness and or is the residual stress at the surface of the layer. Also, Philips et al. I0 showed that the crack deflection in the was calculated using the equation1 interfacial critical strain energy release rate Gic and the bulk KIc= P 301/2 plate-form laminates does not occur when the ratio between the Y critical strain energy release rate GB exceeds unity. Accordin to Philips et al. 0, this condition is achieved when the interfacial where Kic is the fracture toughness, P is the maximum load at toughness is high enough and the performance of the laminate fracture, S is the span, B is the sample width, W is the sample would revert to that of the monolithic materials. The relationship height, a is the coefficient(a=a/w; a the notch depth)and r is between the interfacial fracture toughness(Gic), the numbers of the stress intensity factor coefficient, which is expressed by the layers(T)and the layer thickness(8)is given by the equation: equation Y=1.9887-1.326a-(3.49-0.68a+1.35a2( G (1+a) where oc is the critical stress for failure of the next layer and e is the Youngs modulus 3. Results and discussion 2. Experimental procedure There are several important mechanical parameters which Concentric Si3 N4/BNlaminates are fabricated by slip-casting determine the engineering application of any material. In the alternate layers of Si3 N4 and Bn employing the previously present work, the emphasis was placed on the fracture tough- developed modified slip-casting method. High purity a-Si3N4 ness. Fig 4 shows the change of the apparent fracture toughness

1826 Z. Krstic, V.D. Krstic / Journal of the European Ceramic Society 29 (2009) 1825–1829 Fig. 1. (a) Peeling and (b) delamination in the plate-form laminates. Fig. 2. Schematics of crack deflection in (a) circular and (b) rectangular cross￾sectioned concentric laminate structures. primary crack is blunted and reinitiated at the next layer. This makes the laminate’s apparent fracture toughness insensitive to the crack radius. Shanches-Herencia et al.3 have showed that the crack exten￾sion within the layer occurs only when the layer thickness exceeds a critical value which is directly related to the critical strain energy release rate or fracture toughness: tc = GcE 0.34 (1 − ν2)σ2 r (2) where tc is the critical layer thickness, E is the Young’s modulus, ν is the Poison’s ratio, Gc is the critical strain energy release rate and σr is the residual stress at the surface of the layer. Also, Philips et al.10 showed that the crack deflection in the plate-form laminates does not occur when the ratio between the interfacial critical strain energy release rate GIC and the bulk critical strain energy release rate GBC exceeds unity. According to Philips et al.10, this condition is achieved when the interfacial toughness is high enough and the performance of the laminate would revert to that of the monolithic materials. The relationship between the interfacial fracture toughness (GIC), the numbers of layers (T) and the layer thickness (δ) is given by the equation: GIC = σc δ 18E  T −  (T − 1)3 T 2  where σc is the critical stress for failure of the next layer and E is the Young’s modulus. 2. Experimental procedure Concentric Si3N4/BN laminates are fabricated by slip-casting alternate layers of Si3N4 and BN employing the previously developed modified slip-casting method.8 High purity -Si3N4 Fig. 3. Sample configuration for three-point bending test with corresponding dimensions. powder was used as a raw material. Sub-micron size Al2O3 (A-16, Alcoa) and Y2O3 (Alpha Aesar) powders were used as sintering aids. BN powder (Carborundum Co., grade HPP-325) with the addition of Al2O3 and Si3N4 was used for casting weak interlayers. After drying, pressureless sintering was done in a graphite resistance furnace (Vacuum Industries, USA) at tem￾peratures ranging from 1740 ◦C to 1800 ◦C for 1 h under static N2 gas atmosphere. Fracture toughness was measured using the three-point bend￾ing test at room temperature with a straight-through notch introduced in the mid-section of the samples (Fig. 3). The notch was introduced by a 500 m thick diamond wheel through the first or first two layers and its depth (∼750–1220m) was measured under an optical microscope with 50× or 100× mag￾nifications. The initial crack radius does not play significant role since the crack has to be reinitiated on the next layer creating an inherently sharp crack. The test was carried out on an Instron machine (Model 8502 FIB, Instron Co., Canton, USA) using a jig with the span of 26 mm and the crosshead speed of 0.06 mm/min. Five samples were tested per data point. The fracture toughness was calculated using the equation11: KIC = P BW1/2 · S W · 3α1/2 2(1 − α) 3/2 · Y (4) where KIC is the fracture toughness, P is the maximum load at fracture, S is the span, B is the sample width, W is the sample height, α is the coefficient (α = a/W; a the notch depth) and Y is the stress intensity factor coefficient, which is expressed by the equation: Y = 1.9887 − 1.326α − (3.49 − 0.68a + 1.35α2)α(1 − α)(1 + α) −2 (5) 3. Results and discussion There are several important mechanical parameters which determine the engineering application of any material. In the present work, the emphasis was placed on the fracture tough￾ness. Fig. 4 shows the change of the apparent fracture toughness

Z Krstic, V.D. Krstic/ Joumal of the European Ceramic Society 29(2009)1825-1829 1827 E22g月5g 0 tion of apparent fracture toughness of SN-(BN+Al2O3) with the Number of SiN4 layer umber of Si] N4 layers. Fig. 5. The effect of the number of Si3 N4 layers on apparent fracture toughness ith the number of Si3N4 layers for SN-(BN+Al2O3)lami- nates. Here, the term apparent fracture toughness is used instead gating crack. It is also found that, as the number of the layers in of fracture toughness to indicate that the fracture toughness mea- the sample exceeds some critical number the ability of the crack surements were done on the materials where the crack initiation to deflect becomes smaller and smaller. This trend was found to was responsible for fracture rather then the propagation of sharp exist in both laminates SN-(BN+ SN) and SN-(Bn+Al2O3),as crack. Even though the sharp crack was not introduced initially, shown in Fig. 5 it has been observed that this primary crack is deflected by the Again, as with other laminates, there is an increase in the weak interface and the new crack is initiated at the next Si3N4 apparent fracture toughness with the number of Si3N4 layers layer, making the entire system independent on the sharpest reaching a maximum at a certain number of layers, followed by radius of the primary crack. As shown in Fig. 4, the apparent a decrease of the fracture toughness. In the SN-(BN+SN) lami fracture toughness increases with an increase of number of Si3N4 nates, the highest apparent fracture toughness of 19.5 MPam layers up to 7 layers and then the fracture toughness decreases. was measured in the sample having 4 Si3N4 layers. When The highest apparent fracture toughness of 22 MPam was this value for the fracture toughness is compared with those measured in samples having 7 Si3 N4 layers. The lowest appar- of SN-(BN+Al2O3) it is found that the fracture toughness of ent fracture toughness of MPa m"was found in samples with the SN-(BN+ Al2O3) laminates is higher than that of the SN 19 Si3N4 layers. The initial increase in the apparent fracture (BN+ SN) laminates. This difference in the fracture toughness toughness in samples with 5-7 layers is due to the fact that a between the two composites is associated with the ability of certain minimum number of layers is required to avoid the crack the interface to deflect the propagating crack. When it comes to reaching the core of the sample which was made of monolithic Si3N4 layers, the toughness of Si3N4 layer must be sufficiently i3N4(non-laminated) which has fracture toughness of only high to prevent an easy crack initiation at the surface of the next 9MPam/ It appears that, for the samples with the number of layer. An example of the microstructure showing a weak and layers between 5 and 7, the ability of the interface to deflect porous interface and a dense and tough Si3 N4 layers is shown in the crack is the highest, thus leading to the maximum toughen- Fig. 6. The microstructure shown in Fig. 6 is the one that consists ing. Once the crack is deflected, it can propagate only a certain of a weak and/or porous interface and a dense and strong Si3N4 distance along the interface before its stops. Its initiation at the layers and exhibits the highest fracture toughness(Fig. 6(b)) surface of the next Si3N4 layer requires higher stress compared It is also interesting to note from a)that there are a to that one for the propagation of the existing crack. Thus, in this number of B-Si3N4 grains sticking out from the of Si3N4 lay region, the fracture toughness increases with the number of lay- ers indicating that the interface provides the vehicle for growth ers As shown in Fig 4, after a certain number of layers(7 in the of the elongated grains from one layer of Si3 N4 to another present work) the overall toughness decreases to -8MPam crossing the weak interface. It is believed that this interfacial which is very close the level of the fracture toughness measured bridging contributes to high apparent fracture toughness in these in monolithic Si3N4. A possible reason for the decrease in laminates the apparent fracture toughness for the samples having more The effect of thickness of the Si3N4 layers on the fracture than 7 Si3N4 layers is the decrease in thickness of the Si3N4 toughness is shown in Figs. 7 and 8. The highest apparent layers and BN-based interfacial layer as the number of layers is fracture toughness was found for the laminates with the layer increased. The measurement of the layer thickness revealed that, thickness between 200 um and 250 um. The apparent fracture as the number of the Si3 N4 layers increases, the layer thickness toughness as high as 22 MPa m"was obtained in the laminates decreases. When the thickness of the Si3N4 layer becomes too having an average layer thickness of -230 um These maximum small, the laminates tend to exhibit its fracture behaviour closer in the fracture toughness at a particular Si3 Na layer thickness, to that of a monolithic ceramic(KIC 9 MPam 2), measured by occurs in all samples where the ratio of Si3N4 layer thickness to the three-point bending test, with no ability to deflect the propa- BN interlayer thickness(12-15 um) is around 20

Z. Krstic, V.D. Krstic / Journal of the European Ceramic Society 29 (2009) 1825–1829 1827 Fig. 4. Variation of apparent fracture toughness of SN-(BN + Al2O3) with the number of Si3N4 layers. with the number of Si3N4 layers for SN-(BN + Al2O3) lami￾nates. Here, the term apparent fracture toughness is used instead of fracture toughness to indicate that the fracture toughness mea￾surements were done on the materials where the crack initiation was responsible for fracture rather then the propagation of sharp crack. Even though the sharp crack was not introduced initially, it has been observed that this primary crack is deflected by the weak interface and the new crack is initiated at the next Si3N4 layer, making the entire system independent on the sharpest radius of the primary crack. As shown in Fig. 4, the apparent fracture toughness increases with an increase of number of Si3N4 layers up to 7 layers and then the fracture toughness decreases. The highest apparent fracture toughness of 22 MPa m1/2 was measured in samples having 7 Si3N4 layers. The lowest appar￾ent fracture toughness of 8 MPa m1/2 was found in samples with 19 Si3N4 layers. The initial increase in the apparent fracture toughness in samples with 5–7 layers is due to the fact that a certain minimum number of layers is required to avoid the crack reaching the core of the sample which was made of monolithic Si3N4 (non-laminated) which has fracture toughness of only 9 MPa m1/2. It appears that, for the samples with the number of layers between 5 and 7, the ability of the interface to deflect the crack is the highest, thus leading to the maximum toughen￾ing. Once the crack is deflected, it can propagate only a certain distance along the interface before its stops. Its initiation at the surface of the next Si3N4 layer requires higher stress compared to that one for the propagation of the existing crack. Thus, in this region, the fracture toughness increases with the number of lay￾ers. As shown in Fig. 4, after a certain number of layers (7 in the present work) the overall toughness decreases to ∼8 MPa m1/2 which is very close the level of the fracture toughness measured in monolithic Si3N4. 12 A possible reason for the decrease in the apparent fracture toughness for the samples having more than 7 Si3N4 layers is the decrease in thickness of the Si3N4 layers and BN-based interfacial layer as the number of layers is increased. The measurement of the layer thickness revealed that, as the number of the Si3N4 layers increases, the layer thickness decreases. When the thickness of the Si3N4 layer becomes too small, the laminates tend to exhibit its fracture behaviour closer to that of a monolithic ceramic (KIC ∼9 MPa m1/2), measured by the three-point bending test, with no ability to deflect the propa￾Fig. 5. The effect of the number of Si3N4 layers on apparent fracture toughness of SN-(BN + SN). gating crack. It is also found that, as the number of the layers in the sample exceeds some critical number the ability of the crack to deflect becomes smaller and smaller. This trend was found to exist in both laminates SN-(BN + SN) and SN-(BN + Al2O3), as shown in Fig. 5. Again, as with other laminates, there is an increase in the apparent fracture toughness with the number of Si3N4 layers reaching a maximum at a certain number of layers, followed by a decrease of the fracture toughness. In the SN-(BN + SN) lami￾nates, the highest apparent fracture toughness of 19.5 MPa m1/2 was measured in the sample having 4 Si3N4 layers. When this value for the fracture toughness is compared with those of SN-(BN + Al2O3) it is found that the fracture toughness of the SN-(BN + Al2O3) laminates is higher than that of the SN- (BN + SN) laminates. This difference in the fracture toughness between the two composites is associated with the ability of the interface to deflect the propagating crack. When it comes to Si3N4 layers, the toughness of Si3N4 layer must be sufficiently high to prevent an easy crack initiation at the surface of the next layer. An example of the microstructure showing a weak and porous interface and a dense and tough Si3N4 layers is shown in Fig. 6. The microstructure shown in Fig. 6 is the one that consists of a weak and/or porous interface and a dense and strong Si3N4 layers and exhibits the highest fracture toughness (Fig. 6(b)). It is also interesting to note from Fig. 6(a) that there are a number of -Si3N4 grains sticking out from the of Si3N4 lay￾ers indicating that the interface provides the vehicle for growth of the elongated grains from one layer of Si3N4 to another crossing the weak interface. It is believed that this interfacial bridging contributes to high apparent fracture toughness in these laminates. The effect of thickness of the Si3N4 layers on the fracture toughness is shown in Figs. 7 and 8. The highest apparent fracture toughness was found for the laminates with the layer thickness between 200 m and 250m. The apparent fracture toughness as high as 22 MPa m1/2 was obtained in the laminates having an average layer thickness of ∼230m. These maximum in the fracture toughness at a particular Si3N4 layer thickness, occurs in all samples where the ratio of Si3N4 layer thickness to BN interlayer thickness (∼12–15m) is around 20.

Z Krstic, V.D. Krstic /Jounal of the European Ceramic Sociery 29(2009)1825-1829 etched surface showing the direction of the crack propagation(pointed by the black arrow)and the bridging grain pull out at of the crack at the weak interface(the white arrow ). (b) A weak/porous interf M450 um. The evaluation of the relationship between Si3N4 and BN layer thickness reveals that there is an optimum thick- ness ratio for Si3 N4/BN of 30 for the SN-(Bn SN) laminate and 20 for SN-(BN+Al2O3)laminates. The lowest apparent fracture toughness of 8 MPamis observed with the samples having the ratio of Si3N4 layers to BN layers thickness of 6 which is almost identical to that of SN-(bn Al2O3) laminates. It is worth noting that difficulties were experienced in keeping the thickness of the Si3 N4 layers constant as the number of the layers increased. This difficulty stems from the fact that number of the layers increases, so does the wall thickness leading SiN Layer Thickness [uml to reduced rate of the particle deposition. In order to eliminate Fig. 7. Effect of Sig Na layers thickness on fracture toughness of SN. the effect of number of the layers on the fracture toughness, a series of tests were conducted where the number of the si3na layers was kept constant. The results are shown in Fig. 9 for The decrease in the fracture toughness at higher Si3 N4 layer the SN-(BN+Al2O3)laminates. A strong effect of the layer thickness is considered to be caused by a decrease in the number thickness on the apparent fracture toughness was observed for of interfaces available for crack deflection and responsible for all thickness up to -230 um, followed by a slight decrease in toughening and strengthening he apparent fracture toughness(Fig. 9). This decrease in the Fig. 8 depicts the effect of the Si3N4 layers thickness on apparent fracture toughness above -230 um is not clear at this the apparent fracture toughness of the SN-(BN + SN) laminates. point. One possible explanation could be the decrease in strength Unlike the SN-(BN +Al2O3) laminates, which show maximum of the interface as its thickness becomes smaller compared to the in the apparent fracture toughness at Si3 N4 layer thickness of thickness of the Si3 N4 layers 230 Hm, the SN-(BN+ SN) laminates exhibit maximum in the apparent fracture toughness of 19.5 MPam at thickness of SigNa Layer Thickness [um SisN, Layer Thickness [H Fig8. Effect of Si3 N4 layers thickness on fracture toughness of SN-(BN +SN) Fig 9. Apparent fracture toughness vs SiaN4 layer thickness in(BN+Al2O3)

1828 Z. Krstic, V.D. Krstic / Journal of the European Ceramic Society 29 (2009) 1825–1829 Fig. 6. (a) Micrograph of an etched surface showing the direction of the crack propagation (pointed by the black arrow) and the bridging grain pull out at the surface of the crack at the weak interface (the white arrow). (b) A weak/porous interfaces between dense and strong Si3N4 layers. Fig. 7. Effect of Si3N4 layers thickness on fracture toughness of SN- (BN + Al2O3) laminated structure. The decrease in the fracture toughness at higher Si3N4 layer thickness is considered to be caused by a decrease in the number of interfaces available for crack deflection and responsible for toughening and strengthening. Fig. 8 depicts the effect of the Si3N4 layers thickness on the apparent fracture toughness of the SN-(BN + SN) laminates. Unlike the SN-(BN + Al2O3) laminates, which show maximum in the apparent fracture toughness at Si3N4 layer thickness of ∼230m, the SN-(BN + SN) laminates exhibit maximum in the apparent fracture toughness of 19.5 MPa m1/2 at thickness of Fig. 8. Effect of Si3N4 layers thickness on fracture toughness of SN-(BN + SN) laminated structure. ∼450m. The evaluation of the relationship between Si3N4 and BN layer thickness reveals that there is an optimum thick￾ness ratio for Si3N4/BN of ∼30 for the SN-(BN + SN) laminates and ∼20 for SN-(BN + Al2O3) laminates. The lowest apparent fracture toughness of ∼8 MPa m1/2 is observed with the samples having the ratio of Si3N4 layers to BN layers thickness of ∼6 which is almost identical to that of SN-(BN + Al2O3) laminates. It is worth noting that difficulties were experienced in keeping the thickness of the Si3N4 layers constant as the number of the layers increased. This difficulty stems from the fact that, as the number of the layers increases, so does the wall thickness leading to reduced rate of the particle deposition. In order to eliminate the effect of number of the layers on the fracture toughness, a series of tests were conducted where the number of the Si3N4 layers was kept constant. The results are shown in Fig. 9 for the SN-(BN + Al2O3) laminates. A strong effect of the layer thickness on the apparent fracture toughness was observed for all thickness up to ∼230m, followed by a slight decrease in the apparent fracture toughness (Fig. 9). This decrease in the apparent fracture toughness above ∼230m is not clear at this point. One possible explanation could be the decrease in strength of the interface as its thickness becomes smaller compared to the thickness of the Si3N4 layers. Fig. 9. Apparent fracture toughness vs. Si3N4 layer thickness in (BN + Al2O3) laminates with 7 layers.

Z Krstic, V.D. Krstic/ Joumal of the European Ceramic Society 29(2009)1825-1829 4. Conclusion 2. She, J, Inoe, T. and Ueno, K, Multilayer Al2O3/SiC havior. J. Eur Ceram. Soc. Pull out and crack deflection along the weak interfac re found to be the dominant mechanisms of toughening of 3. Sanchez-Herencia, A.J., Pascual, C, He, J and Lange, F. F, Zro2/ZrO Layered composites for crack bifurcation. J. Am. Ceram. Soc., 1999, 82, the concentric Si3 N4/BN laminated structures. The highest 1512-1 apparent fracture toughness of 22MPam2 was found in SN- 4. Mawdsley, I, Kover, D and Halloran, J w. Fracture behavior of alu- (BN+Al2O3)laminates having 7 Si3N4 layers with an average mina/monazite multilayer laminates. J Am. Ceram. Soc., 2000, 83 802 thickness of 230 um In the SN-(BN + SN) laminates, the high est apparent fracture toughness of 19.5 MPamwas found with 5. Yu, Z, Krstic, Z and Krstic, V. D, Laminated Si3 N4/SiC composites with self-sealed structure. Key Eng. Mater, 2005, 280-283, 1873-1876. the samples having 4 Si3N4 layers with an average thickness of 6. Krstic, Z and Krstic, V D, Young's Modulus, density and phase composi- 430 um. Due to the presence of the weak interfaces and repeated n of pressureless sintered self-sealed Si3 N4/BN laminated structures.J. crack initiation across each Si3 N4 layers, the laminated struc Eur Ceram Soc.,2008,28,1723-1730. tures exhibit no notch width sensitivity which, in the past, was 7. Liu, H and Hsu, S M., Fracture behavior of multilayer silicon nitride/boron nitride cera J. Am. ceram 79,2452-2457 found only in fibre 8. Zhang, L and Krstic, V.D., H silicon carbide/graphite lam- Crack deflection and its reinitiation at the next Si3 N4 lay inar composite by slip casting Fract.Mec,1995,24,13- ers are found to be the two dominant factors which control the apparent fracture toughness. High density of the Si3 N4 layers 9. Clegg, W.J., The fracture and failure of laminar ceramic composites. Acta which resists crack initiation on its surface is favourable condi- 10. Phillips. A. ]. Clegg. W.J. and Clyne. T. W. Fracture behavior of ceramic tion for a crack deflection at the interface and it is responsible laminates in bending I: modeling of crack propagation. Acta Metall. Mater for the high fracture toughness in SN-(BN+Al2O3) laminates. 1993,41(3),805-817 11. Kubler, J, Fracture toughness of ceramics using the sevnb method: prelim- References inary results. Ceram. Eng. Sci. Proc., 1997, 18(4), 155-162. 12. Krstic, Z, Yu, Z. and Krstic, V.D.. Effect of grain width and aspect 1. Clegg. w.J. Kandall, K, Alford, N. M. Birchall, D and Button. T. w.A ratio on mechanical properties of Si3 N4 ceramics. J. Mater. Sci., 2007, 42, mple way to make tough ceramics. Nature, 1990, 347, 455-457. 5431-5436

Z. Krstic, V.D. Krstic / Journal of the European Ceramic Society 29 (2009) 1825–1829 1829 4. Conclusion Pull out and crack deflection along the weak interface are found to be the dominant mechanisms of toughening of the concentric Si3N4/BN laminated structures. The highest apparent fracture toughness of 22 MPa m1/2 was found in SN- (BN + Al2O3) laminates having 7 Si3N4 layers with an average thickness of 230m. In the SN-(BN + SN) laminates, the high￾est apparent fracture toughness of 19.5 MPa m1/2 was found with the samples having 4 Si3N4 layers with an average thickness of 430m. Due to the presence of the weak interfaces and repeated crack initiation across each Si3N4 layers, the laminated struc￾tures exhibit no notch width sensitivity which, in the past, was found only in fibre composites. Crack deflection and its reinitiation at the next Si3N4 lay￾ers are found to be the two dominant factors which control the apparent fracture toughness. High density of the Si3N4 layers which resists crack initiation on its surface is favourable condi￾tion for a crack deflection at the interface and it is responsible for the high fracture toughness in SN-(BN + Al2O3) laminates. References 1. Clegg, W. J., Kandall, K., Alford, N. M., Birchall, D. and Button, T. W., A simple way to make tough ceramics. Nature, 1990, 347, 455–457. 2. She, J., Inoe, T. and Ueno, K., Multilayer Al2O3/SiC Ceramics with improved mechanical behavior. J. Eur. Ceram. Soc., 2000, 20, 1771– 1775. 3. Sanchez-Herencia, A. J., Pascual, C., He, J. and Lange, F. F., ZrO2/ZrO2 Layered composites for crack bifurcation. J. Am. Ceram. Soc., 1999, 82, 1512–1518. 4. Mawdsley, J., Kover, D. and Halloran, J. W., Fracture behavior of alu￾mina/monazite multilayer laminates. J. Am. Ceram. Soc., 2000, 83, 802– 808. 5. Yu, Z., Krstic, Z. and Krstic, V. D., Laminated Si3N4/SiC composites with self-sealed structure. Key Eng. Mater., 2005, 280–283, 1873–1876. 6. Krstic, Z. and Krstic, V. D., Young’s Modulus, density and phase composi￾tion of pressureless sintered self-sealed Si3N4/BN laminated structures. J. Eur. Ceram. Soc., 2008, 28, 1723–1730. 7. Liu, H. and Hsu, S. M., Fracture behavior of multilayer silicon nitride/boron nitride ceramics. J. Am. Ceram. Soc., 1996, 79, 2452–2457. 8. Zhang, L. and Krstic, V. D., High toughness silicon carbide/graphite lam￾inar composite by slip casting. Theor. Appl. Fract. Mech., 1995, 24, 13– 19. 9. Clegg, W. J., The fracture and failure of laminar ceramic composites. Acta Metall. Mater., 1992, 40, 3085–3093. 10. Phillips, A. J., Clegg, W. J. and Clyne, T. W., Fracture behavior of ceramic laminates in bending I: modeling of crack propagation. Acta Metall. Mater., 1993, 41(3), 805–817. 11. Kübler, J., Fracture toughness of ceramics using the sevnb method: prelim￾inary results. Ceram. Eng. Sci. Proc., 1997, 18(4), 155–162. 12. Krstic, Z., Yu, Z. and Krstic, V. D., Effect of grain width and aspect ratio on mechanical properties of Si3N4 ceramics. J. Mater. Sci., 2007, 42, 5431–5436.