e 1996 Elsevier Science Printed in Great Britain. All rights 0955-2219(95)00163-8 955-22199631500 Fatigue Crack Growth Rate and Fracture Toughness of 25 wt% Silicon Carbide Whisker Reinforced Alumina Composite with Residual porosity AK.Ray, E.R. Fuller&S. Banerjeec "National Metallurgical Laboratory, Jamshedpur-831007, Bihar, India National Institute of Standards and Technology, Gaithersburg, MD 20899, USA Research and Development Centre for Iron and Steel, SAIL, Ranchi-834002, Bihar, India (Received 8 December 1994; revised version received 20 July 1995; accepted 28 August 1995) Abstract and cyclic loading which produces crack extension Accordingly, in the present investigation, the fatigue The main purpose of this study was to determine the crack growth bchaviour(FCGR) and the fracture fracture toughness and the fatigue crack growth toughness(K,c) of a high density, 25 wt% silicon rate behaviour of 25 wt%o silicon carbide whisker carbide whisker reinforced alumina composite reinforced alumina ceramic composite. The fracture have been studied toughness values determined using the indentation In general, the determination of FCGR and kic technique depended significantly on the crack length of such ceramic materials is difficult since the produced at the corners of the indentation which, in specimens are small, Young's modulus of such turn, depended on the load used for the indentation materials is rather high and the material is brittle and anisotropy in orientation of whiskers in the Consequently, the load, displacement and crack matrix. However, the fracture toughness values length which are all required to be measured determined using the precracked four-point bend for the determination of Kic and FCGR--are specimens were in general higher than that obtained very small and their precise measurement poses by the indentation technique and the value was some problems. In addition, the permissible 596+0. 15 MPa m. The fatigue crack growth dimcnsional tolerance of the specimens and those behaviour in this material was similar to that in of the grips and fixtures used to test the speci the case of metals. However, the exponent for the mens, have to be very close. Particularly difficult is fatigue crack growth rate was 15.5, significantly the precracking of ceramic specimens since these higher than that usually observed in metals. The materials have very low toughness. Moreover, the likely micromechanism of crack growth under crack initiation in such materials often requires a monotonic and cyclic loading in this composite has load which is higher than that required for crack been identified from fractography of fatigue failed extension. Therefore specimens fail before crack samples. growth is achieved in a controlled manner, because he precision and dimensional tolerance of the fixture used to precrack the specimens are not 1 Introduction adequate to avoid spurious loading, that is load ing in modes other than in mode I Silicon carbide whiskers have been incorporated in such ceramic materials as alumina to improve the general mechanical propertiesand the resis- 2 Material and Specimen Orientation ce to catastrophic failure in particular. These ceramic composite materials have potential appli- The ceramic composite material was prepared by cation in the production of structural components mixing a-alumina powder of particle size l um used at elevated temperatures, 0-12i e in high effi- with 25 wt% B-silicon carbide whiskers the aver- cy heat engines and heat recovery systems, age whisker diameter was 0.45-065 um and the and for making cutting tools to machine special length ranged from 10 to 80 um. This mixture materials. When used in such applications, these was hot-pressed at 1700 to 1850 C under a pressure ceramic components often encounter monotonic of 25 MPa for 30 min to produce a preformed billet
Journul (I/ the Europcw~ Ckrwi~ SocictJ: 16 ( 1996) 503-5 13 0 1996 Elsevier Science Limited 0955-2219(95)00163-8 Printed in Great Britain. All rights reserved 0955-2219/96/$15.00 Fatigue Crack Growth Rate and Fracture Toughness of 25 wt% S:ilicon Carbide Whisker Reinforced Alumina Composite with Residual Porosity A. K. Ray,” E. R. Fullerb & S. Banerjee” “National Metallurgical Laboratory, Jamshedpur-831007, Bihar, India ‘National Institute of Standards and Technology, Gaithersburg, MD 20899, USA “Research and Development Centre for Iron and Steel, SAIL, Ranchi-834002, Bihar, India (Received 8 December 1994; revised version received 20 July 1995; accepted 28 August 1995) Abstract The main purpose of this study was to determine the fracture toughness and the fatigue crack growth rate behaviour of 25 wt% silicon carbide whisker reinforced alumina ceramic composite. The fracture toughness values determined using the indentation technique depended sign$cantly on the crack length produced at the corners of the indentation which, in turn, depended on the load used for the indentation and anisotropy in orientation of whiskers in the matrix. However, the fracture toughness values determined using the precracked four-point bend specimens were in general higher than that obtained by the indentation technique and the value was 5.96 f 0.15 MPa m ‘I2 . The fatigue crack growth behaviour in this material was similar to that in the case of metals. How(ever, the exponent for the fatigue crack growth rate was 15.5, sigru~cantly higher than that usually observed in metals. The likely micromechanism of crack growth under monotonic and cyclic loading in this composite has been identtjied from fractography of fatigue failed samples. 1 Introduction Silicon carbide whiskers, have been incorporated in such ceramic materials as alumina to improve the general mechanical properties’-’ and the resistance to catastrophic failure in particular. These ceramic composite materials have potential application in the production of structural components used at elevated tempera.tures,‘@i2 i.e. in high efficiency heat engines and heat recovery systems, and for making cutting tools to machine special materials. When used in such applications, these ceramic components 0fl:en encounter monotonic and cyclic loading which produces crack extension. Accordingly, in the present investigation, the fatigue crack growth behaviour (FCGR) and the fracture toughness (K,,) of a high density, 25 wt% silicon carbide whisker reinforced alumina composite have been studied. In general, the determination of FCGR and K,, of such ceramic materials is difficult since the specimens are small, Young’s modulus of such materials is rather high and the material is brittle. Consequently, the load, displacement and crack length - which are all required to be measured for the determination of K,, and FCGR - are very small and their precise measurement poses some problems. In addition, the permissible dimensional tolerance of the specimens and those of the grips and fixtures used to test the specimens, have to be very close. Particularly difficult is the precracking of ceramic specimens since these materials have very low toughness. Moreover, the crack initiation in such materials often requires a load which is higher than that required for crack extension. Therefore specimens fail before crack growth is achieved in a controlled manner, because the precision and dimensional tolerance of the fixture used to precrack the specimens are not adequate to avoid spurious loading, that is loading in modes other than in mode I. 2 Material and Specimen Orientation The ceramic composite material was prepared by mixing a-alumina powder of particle size < 1 pm with 25 wt% p-silicon carbide whiskers. The average whisker diameter was 0~450.65 pm and the length ranged from 10 to 80 pm.8 This mixture was hot-pressed at 1700 to 1850°C under a pressure of 25 MPa for 30 min to produce a preformed billet. 503
A.K. Ray, E.R. Fuller, S. Banerjee The grain size of the matrix varied between I and 4 um. Details of fabrication, processing and micro- structural characterization are reported elsewhere. s The composite material had a porosity of 89%, Youngs modulus of 340 GPa, fracture strength of 559 MPa and a hardness of 20 GPa,as determined by National Institute of Standards and Technology(NIST), USA, and has been Ref. 8 mm×4mm×50mm, were sliced from the pre- formed billet and were supplied to us(National Metallurgical Laboratory, NML, India) by nist. a sketch of the billet showing the longitudinal (L), long transverse(LT)and the short transverse Hot (ST) planes is presented in Fig. 1(a). Opposite sides of the four-point flexure specimens were dia mond-ground fat and parallel with a 30 um dia- mond wheel, and the prospective tensile surface was polished with 9 um diamond paste. The specimens were soaked and rinsed in ethyl alcohol ALL DIMENSIONS ARE IN MM to remove the wax needed to mount them for pol ishing, and then dried in a hot air flow The location and orientation of the flexure mens in the billet is shown in Fig. 1(a). The 3 mm X 50 mm faces of the specimen were parallel to the st plane. The crack plane introduced later in the specimen was parallel to the lt plane and the crack propagation direction was parallel to the 4 mm dimension of the specimen. Thus the direction of crack propagation was perpendicular to the hot-pressing direction. Accordingly, the crack front was parallel to the hot-pressing direction [se Fig. 1(a)]. On the other hand, in the R-curve studies undertaken earlier by one of the authors while the plane of crack propagation was parallel to the lt plane like in the present investigation, the direction of crack propagation was parallel to HoT PTESSING the 3 mm dimension A montage of the microstructures of the 25 wt% silicon carbide whisker reinforced alumina ceramic composite, as shown in Fig. 1(b), revealed a three Fig. 1.(a) Billet p dimensional 3D distribution pattern of the whisker reinforce whiskers in the L, the Lt and the st planes of the microstructures she billet. As can be seen from Fig. 1(b), the distribu the three planes tion of the whiskers in the l plane was non-uni- form and heterogeneous. The cross-section of of the whiskers seemed to be embedded with ran these whiskers measured -0.45 um in diameter. dom orientation; and the others were observed to During hot-pressing, the whiskers which were not be embedded normal to this plane. Probably, the normal to the longitudinal plane became further friction of the walls, which in the ST planes were clined and therefore some of the whiskers very close to each other, prevented easy material appeared to be randomly oriented in the l plane flow and the alignment of all the whiskers normal [Fig. 1(b). However, a majority of the whiskers to the ST plane. tended to be oriented normal to the hot-pressing According to Becher and Wei, 3 whisker orien direction. Since the material flow along this plane tation during processing of hot-pressed Sic was high, whiskers were aligned normal to this whisker reinforced alumina leads to anisotropy in plane. On the other hand, in the St plane, a few both fracture toughness and fracture strength of
504 A. K. Ray, E. R. The grain size of the matrix varied between 1 and 4 pm. Details of fabrication, processing and microstructural characterization are reported elsewhere.* Fuller, S. Banerjee The composite material had a porosity of 4.89%, Young’s modulus of 340 GPa, fracture strength of 559 MPa and a hardness of 20 GPa, as determined by National Institute of Standards and Technology (NIST), USA, and has been reported in Ref. 8. Four-point flexure specimens, of dimensions 3 mm X 4 mm X 50 mm, were sliced from the preformed billet and were supplied to us (National Metallurgical Laboratory, NML, India) by NIST. A sketch of the billet showing the longitudinal (L), long transverse (LT) and the short transverse (ST) planes is presented in Fig. l(a). Opposite sides of the four-point flexure specimens were diamond-ground flat and parallel with a 30 pm diamond wheel, and the prospective tensile surface was polished with 9 pm diamond paste.* The specimens were soaked and rinsed in ethyl alcohol to remove the wax needed to mount them for polishing, and then dried in a hot air flow. The location and orientation of the flexure specimens in the billet is shown in Fig. l(a). The 3 mm X 50 mm faces of the specimen were parallel to the ST plane. The crack plane introduced later in the specimen was parallel to the LT plane and the crack propagation direction was parallel to the 4 mm dimension of the specimen. Thus the direction of crack propagation was perpendicular to the hot-pressing direction. Accordingly, the crack front was parallel to the hot-pressing direction [see Fig. l(a)]. On the other hand, in the R-curve studies’ undertaken earlier by one of the authors, while the plane of crack propagation was parallel to the LT plane like in the present investigation, the direction of crack propagation was parallel to the 3 mm dimension. A montage of the microstructures of the 25 wt% silicon carbide whisker reinforced alumina ceramic composite, as shown in Fig. l(b), revealed a threedimensional 3D distribution pattern of the whiskers in the L, the LT and the ST planes of the billet. As can be seen from Fig. l(b), the distribution of the whiskers in the L plane was non-uniform and heterogeneous. The cross-section of these whiskers measured -0.45 pm in diameter. During hot-pressing, the whiskers which were not normal to the longitudinal plane became further inclined and therefore some of the whiskers appeared to be randomly oriented in the L plane [Fig. l(b)]. However, a majority of the whiskers tended to be oriented normal to the hot-pressing direction. Since the material flow along this plane was high, whiskers were aligned normal to this plane. On the other hand, in the ST plane, a few Fig. 1. (a) Billet prepared from the 25 wt% silicon carbide whisker reinforced alumina composite; (b) montage of the microstructures showing the distribution of the SIC whiskers along the three planes. of the whiskers seemed to be embedded with random orientation; and the others were observed to be embedded normal to this plane. Probably, the friction of the walls, which in the ST planes were very close to each other, prevented easy material flow and the alignment of all the whiskers normal to the ST plane. According to Becher and Wei,13 whisker orientation during processing of hot-pressed SiCwhisker reinforced alumina leads to anisotropy in both fracture toughness and fracture strength of ALL DIMENSlONS ARE IN MM (a) (b)
FCGR and fracture toughness of 25 wt% SiC /Al2O, composite these composites. In other words, their fracture 3.2 Crack length measurement strengths are limited by the non-unifor The crack starter indentation at the mid-point of tribution of the whiskers, i.e. by the ability to dis- the upper face of the specimen had cracks at all erse the Sic whiskers. They also found that four of its corners(Fig. 2). During precracking, dispersion of the whiskers was improved by using the crack length of the two corner cracks growing finer alumina powder, resulting in an increase in the thickness of the specimen, which is of he fracture strength of the composite Neverthe- prime importance for the present investigation less, they have clearly observed 3, 4 that, similar to was measured using the micron marker in the our composite under investigation [Fig. 1(b), the SEM at a magnification of 40x. The tip of the whiskers were preferentially aligned perpendicular crack was located through observation at higher to the hot-pressing axis. this type of distribution magnification, which often showed evidence of of whiskers suggests that a great deal of rear- crack branching. The branch extending the fur- rangement of whiskers and powders occurred in thest was considered to be the crack tip and the the initial stage of densification of the composites crack length was measured accordingly and/or the matrix material underwent consider. At first, the crack length of these two crack able deformation or creep during hot-pressing growing on the tensile (3 mm X 50 mm) surface of the specimen from the two opposite corners of the indentation as crack starter, was monitored during 3 Experimental Procedure precracking. Later, after these two cracks had spanned right across the specimen thickness, the 3.1 Precracking of the specimens crack lengths were measured on both the side The standard bridge technique is normally used surfaces(4 mm X 50 mm planes)from the respec for precracking four-point bend specimens. How- tive upper edges. The average of the crack length ever, it did not produce satisfactory precracking measured on the side surfaces gave the crack and all the specimens loaded for precracking in length a this fixture(24 out of the 30 specimens supplied) The crack length a was measured in this manner were lost due to premature crack extension. This during the precracking and also during the FCGR is probably related to fixture stiffness more than determination which preceded the Kic testing. anything else. An examination of the changing After the crack had advanced with regular incre- crack path trajectory on the fractured surfaces of ments of number of cycles, the specimen was these lost specimens showed that, instead of pure unloaded from the MTS-880 servohydraulic test mode I loading, the specimens experienced a com- machinc, and the current crack length was mca bined mode loading with mode III loading playing sured using the SEM. Since the determination of a significant role in the crack extension. The FCGr required precise measurement of au, bridge technique was, therefore, modified to avoid the crack length during FCGR studies was mea mode III loading during precracking. Accordingly, sured with special care taken in locating the crack new articulated precracking bridge fixtures were tip. designed and fabricated. These fixtures gave excel lent results--achieving 100% success in precrack- ing. The articulated bridge fixtures have been Extemal span described elsewhere A Vicker's indentation was produced at 0.8 kN pad at the mid-point of the upper face (3 X 50 mm surface) of the four-point bend specimen which acted as the crack starter, Bcforc indention all the specimens were coated with aluminium in a acuum evaporator by physical vapour deposition technique to -0-03 um thickness, to facilitate loca- tion of the crack tip in scanning electron microscope. Precracking was accomplished using the articulated bridge fixture at a force of 4 to 5 kN, load ratio(R)of 0.1 and frequency () of 20 Hz. The crack growth was measured at first on the top surface, and later on the two side surfaces ALL DIMENSIONS ARE N MK of the four-point bend specimen using the micron Fig. 2. Indented and precracked specimen for four-point marker in a Jeol JSM 840A scanning electron micro- bend loading. Cracks are located at the four corners of the scope(SEM)
FCGR and fracture toughness of 25 wt% SiC,,/Alz03 composite 505 these composites. In olther words, their fracture strengths are limited by the non-uniformity of distribution of the whiskers, i.e. by the ability to disperse the SIC whiskers. They also found that dispersion of the whiskers was improved by using finer alumina powder, resulting in an increase in the fracture strength of the composite. Nevertheless, they have clearly observed’3s’4 that, similar to our composite under investigation [Fig. l(b)], the whiskers were preferentially aligned perpendicular to the hot-pressing axis. This type of distribution of whiskers suggests that a great deal of rearrangement of whiskers and powders occurred in the initial stage of densification of the composites and/or the matrix material underwent considerable deformation or creep during hot-pressing. 3 Experimental Procedure 3.1 Precracking of the specimens The standard bridge technique is normally used for precracking four-point bend specimens. However, it did not produce satisfactory precracking and all the specimens loaded for precracking in this fixture (24 out of the 30 specimens supplied) were lost due to premature crack extension. This is probably related to fixture stiffness more than anything else. An examination of the changing crack path trajectory on the fractured surfaces of these lost specimens showed that, instead of pure mode I loading, the specimens experienced a combined mode loading with mode III loading playing a significant role in the crack extension. The bridge technique was, therefore, modified to avoid mode III loading during precracking. Accordingly, new articulated precracking bridge fixtures were designed and fabricated. These fixtures gave excellent results-achieving 100% success in precracking. The articulated bridge fixtures have been described elsewhere.15 A Vicker’s indentation was produced at 0.8 kN load at the mid-point of the upper face (3 X 50 mm surface) of the flour-point bend specimen, which acted as the crack starter. Before indention, all the specimens were coated with aluminium in a vacuum evaporator by Iphysical vapour deposition technique to -0.03 pm thickness, to facilitate location of the crack tip in the scanning electron microscope. Precracking was accomplished using the articulated bridge hxture at a force of 4 to 5 kN, load ratio (R) of 01.1 and frequency u> of 20 Hz. The crack growth was measured at first on the top surface, and later on the two side surfaces, of the four-point bend specimen using the micron marker in a Jeol JSM 841DA scanning electron microscope @EM). 3.2 Crack length measurement The crack starter indentation at the mid-point of the upper face of the specimen had cracks at all four of its corners (Fig. 2). During precracking, the crack length of the two corner cracks growing across the thickness of the specimen, which is of prime importance for the present investigation, was measured using the micron marker in the SEM at a magnification of 40X. The tip of the crack was located through observation at higher magnification, which often showed evidence of crack branching. The branch extending the furthest was considered to be the crack tip and the crack length was measured accordingly. At first, the crack length of these two cracks growing on the tensile (3 mm X 50 mm) surface of the specimen from the two opposite corners of the indentation as crack starter, was monitored during precracking. Later, after these two cracks had spanned right across the specimen thickness, the crack lengths were measured on both the side surfaces (4 mm X 50 mm planes) from the respective upper edges. The average of the crack length measured on the side surfaces gave the crack length a. The crack length a was measured in this manner during the precracking and also during the FCGR determination which preceded the K,, testing. After the crack had advanced with regular increments of number of cycles, the specimen was unloaded from the MTS-880 servohydraulic test machine, and the current crack length was measured using the SEM. Since the determination of FCGR required precise measurement of Au, the crack length during FCGR studies was measured with special care taken in locating the crack tip. ALL DIMENSIONS ARE Y MM. Fig. 2. Indented and precracked specimen for four-point bend loading. Cracks are located at the four corners of the indentation
A.K. Ray, E.R. Fuller, S. Banerjee 3.3 Determination of fatigue crack growth rate monotonic loading; the ramp rate was 0.25 N s FCGR The load value corresponding to the onset of fast FCGR was determined after the crack had grown fracture was used in eqn (1)to obtain the K ignificantly on the side surfaces during precrack value. Since four- point bend specimens were used 0.05 to 0. 1. The four-point bend specimen is not to the mcthod givcn clscwhcrc 6/S in all respects ing and achieved length a equivalent to a/w Kic testing followed here conforme recommended in Astm Standard E647 and In addition to this procedure, the indentation therefore, the k values of the specimen were cal- fracture toughness Kc of this material was deter culated using the standard formula reported else- mined from the indentation technique at various (referred to as the ASTM STP 410 loads(0 63, 0-8, 1-0 and 1.2 kN using the follow method). Except for this aspect and the use of an ing equation proposed by anstis et al. 8 indentation as a crack starter, the procedure used here to determine fcgr conformed to the recom- K=0061E.P h a mendations given in ASTM Standard E647 The tests were conducted in an MTS-880 servo- where E is the Young's modulus, H is the hardness, hydraulic test machine using a I kN load cell P is the load and a the crack length. Many emper- under four-point bend loading, in laboratory cal expressions to evaluate indentation fracture atmosphere and at ambient temperature. The toughness have been reported in the literature. 19 loading rate was 0.25 s. The frequency wa However, in the case of toughened Hz and r= 01. Typically, the specimens were adial cracks are emanating at the four corners of cycled within the load range between 11 and 11 the Vickers indentation, 9 the above model [eqn(2)1 N, when a/w=0-1. The crack lengths were mes has been used by Anstis et al. to determine Ko sured at regular increments of number of cycles giving 0-05 to 0.1 mm of crack growth. While 3.5 SEM studies measuring the crack length, the specimen was first The fracture surfaces of the test specimens were unloaded from the servohydraulic machine and coated with a thin film of gold (thickness 0-02 um) the current crack length was measured with the and then examined g the scanning electron help of the micron marker in the SEM. Thus the microscope to identify the characteristic fracto fatigue cracking was interrupted after a predeter- graphic features of the fatigue and fast fracture mined number of load cycles. This was continued regions in this material. The identification of these until the crack length increased to a value giving features gives a clue as to the likely mechanisms of a/W-045to0.5 fracture for this material The test data of crack length were plotted in At first, the SEM examination of the fatigue tcrms of crack length a vS. numbcr of cycles N. fracture and fast fracture zones was carried out at The values of da/dN were generated from the a vs low magnification of 30X. Thereafter, each of N plot at any given a and plotted against AK these zones was scanned at 4500x and 7500x (stress intensity range). With load and a known, The fatigue crack growth at the low AK region K values were calculated from and the fast fracture region(due to monotonic K - Y3P(L L,)v/2bw () tify and distinguish between the characteristic where fractographic features Y=199-247(a/W)+1297(a/W) 2317(a/W)3+2480(a/W 4 Results P=load; L, external span; L2 internal span b= thickness of specimen;W= depth or width 4.1 Fracture toughness data of specimen; a= crack length. The values of Kmax Fracture toughness values determined in this pro and Kmin were calculated using eqn(1). The AK ject, using both the indentation technique as well value is given by kmax -kmi as the ASTM STP 410 method with precracked specimens, are reported in Table 1. Indentation 3.4 Fracture toughness(Kic) testing fracture toughness test data determincd with a After the FCGr determination was complete and Vickers indentation at various loads of 0-63,0-80, he crack had grown to a level of a/w =0-45 1- 0 and 1. 2 kN yielded Kc values of 537, 5 45, 5.5 0.5, Kc was determined by subjecting the pre- and 5 6 MPa m", respectively. Figure 3 shows the variation of indentation fracture to oughness values loading. The test record of load vs. mid-point dis- with the square root of the corresponding crack placement of the specimen was obtained during lengths. It was observed (see Fig. 3)that the
506 A. K. Ray, E. R. Fuller, S. Banerjee 3.3 Determination of fatigue crack growth rate (FCGR) FCGR was determined after the crack had grown significantly on the side surfaces during precracking and achieved length a equivalent to a/W = 0.05 to 0.1. The four-point bend specimen is not recommended in ASTM Standard E647 and, therefore, the K, values of the specimen were calculated using the standard formula reported elsewhere’6,‘7 (referred to as the ASTM STP 410 method). Except for this aspect and the use of an indentation as a crack starter, the procedure used here to determine FCGR conformed to the recommendations given in ASTM Standard E647. The tests were conducted in an MTS-880 servohydraulic test machine using a 1 kN load cell under four-point bend loading, in laboratory atmosphere and at ambient temperature. The loading rate was 0.25 N s?. The frequency was 1 Hz and R = 0.1. Typically, the specimens were cycled within the load range between 11 and 111 N, when a/W = 0.1. The crack lengths were measured at regular increments of number of cycles giving - 0.05 to 0.1 mm of crack growth. While measuring the crack length, the specimen was first unloaded from the servohydraulic machine and the current crack length was measured with the help of the micron marker in the SEM. Thus the fatigue cracking was interrupted after a predetermined number of load cycles. This was continued until the crack length increased to a value giving a/W = 0.45 to 0.5. The test data of crack length were plotted in terms of crack length a vs. number of cycles N. The values of da/dN were generated from the a vs. N plot at any given a and plotted against AK (stress intensity range). With load and a known, KI values were calculated from16,‘7 KI= Y3P(L,-L,)da/2bW2 (1) where Y = 1.99 - 2.47 (a/W) + 12.97 (a/v2 - 23.17 (alW’j3 + 24.80 (a/w4 P = load; L1 = external span; L, = internal span; b = thickness of specimen; W = depth or width of specimen; a = crack length. The values of Km,, and Kmin were calculated using eqn (I). The AK value is given by Km,, - Kmi,. 3.4 Fracture toughness (K,,) testing After the FCGR determination was complete and the crack had grown to a level of a/W = 0.45 to 0.5, K,, was determined by subjecting the precracked four-point bend specimens to monotonic loading. The test record of load vs. mid-point displacement of the specimen was obtained during monotonic loading; the ramp rate was 0.25 N ss’. The load value corresponding to the onset of fast fracture was used in eqn (1) to obtain the K,, value. Since four-point bend specimens were used, K,, testing followed here conformed in all respects to the method given elsewhere.‘6,‘7 In addition to this procedure, the indentation fracture toughness Kc of this material was determined from the indentation technique at various loads (0.63, 0.8, 1 .O and 1.2 kN) using the following equation proposed by Anstis et al.‘* Kc = 0.016 1/ $. -$ where E is the Young’s modulus, H is the hardness, P is the load and a the crack length. Many emperical expressions to evaluate indentation fracture toughness have been reported in the literature.]’ However, in the case of toughened ceramics where radial cracks are emanating at the four corners of the Vickers indentation,” the above model [eqn (2)] has been used by Anstis et al.‘* to determine Kc. 3.5 SEM studies The fracture surfaces of the test specimens were coated with a thin film of gold (thickness 0.02 pm) and then examined using the scanning electron microscope to identify the characteristic fractographic features of the fatigue and fast fracture regions in this material. The identification of these features gives a clue as to the likely mechanisms of fracture for this material. At first, the SEM examination of the fatigue fracture and fast fracture zones was carried out at a low magnification of 30X. Thereafter, each of these zones was scanned at 4500X and 7500X. The fatigue crack growth at the low AK region and the fast fracture region (due to monotonic loading) were carefully examined in order to identify and distinguish between the characteristic fractographic features. 4 Results 4.1 Fracture toughness data Fracture toughness values determined in this project, using both the indentation technique as well as the ASTM STP 410 method with precracked specimens, are reported in Table 1. Indentation fracture toughness test data determined with a Vickers indentation at various loads of 0.63, 0.80, 1.0 and 1.2 kN yielded Kc values of 5.37, 5.45, 5.5 and 5.6 MPa m1’2, respectively. Figure 3 shows the variation of indentation fracture toughness values with the square root of the corresponding crack lengths. It was observed (see Fig. 3) that the
FCGR and fracture toughness of 25 wt% SiC., composite Table 1. Fracture toughness determined by indentation and by using precracked specimen for 25 wt% SiC reinforced 01.f=1H ughness(MPu m) TM STP 410 60 Average:596±015 548±0-08 12.64S/C-AlO, [21 Stress Intensity Range, AK(MP a/M) 5 Fig. 4. Fatigue crack propagation rate data 0521215222252323.524245 Square Root of (rock Length, va Ivpm) Fig. 3. Dependence of indentation fracture toughness Ke on re root of crack length Fig. 5. At low AK(0-8 to 1.8 MPa m")region, a majority of indentation fracture toughness of this composite the whiskers failed with a square fracture without evidence of when determined using the indentation technique large-scale pull-out increased as a function of the square root of the crack length 4.2 FCGR studies Fatigue crack propagation behaviour of this ceramic material is reported in Fig 4. The FCGR data of this material were fitted to the usual paris equation da/dN= A(4K,20-22 and the fatigue crack growth rate da/dN (in m/cycle) increased linearly with AK(in MPa m )in a log-log plot The plot yields a value of n=15 5 and A =3 4 X 10-(m/cycle(MPa m-m 4.3 SEM studies SEM fractographs of the fracture surface at Fig. 6. In the fast fracture region(KIC 5.9 MPa m), due to low AK cyclic and at monotonic loading(fast monotonic loading, whiskers failed predominantly by pull-out ture region) are presented in Figs 5 to ll. It is evi- dent that, at low AK(0-8 to 1-8 MPa m),a majority of the whiskers fail by producing a fat out wherein the whiskers stick out of the genera fracture surface which has a vertical level, same as fracture surface of the composite(see Fig. 6) the general fracture surface of the composite(see The matrix material failed predominantly by Fig. 5). On the other hand, in the fast fracture transgranular fracture (see Fig. 7)at low AK region, the whiskers fail predominantly by pull- fatigue(0 8 to 1- 8 MPa m). During monotonic
FCGR and fracture toughness of 25 wt’% SiC,,/AI,O, composite 507 Table 1. Fracture toughness determined by indentation and by using precracked specimen for 25 wt% Sic reinforced A&O3 composite Fracture toughness (MPa rnlf2) Precracked specimen (ASTM STP 410 method) Indentation technique Indentationstrength method 8 6.1 537 535 f 0.17 5.8 5.45 6.0 5.5 5.6 Average: 5.96 + 0.15 5.48 f 0.08 p 5.6sr-------j G 0 5.60. Y f t 5x8- 5 a ; 5.50- 2 i / . u’ 5.l.5, 5.35 I I I I I I I I 19.5 20 20.5 21 21.5 22 22.5 23 23.5 24 2L.5 Square Root of Crack Length. dTi I qm’ Fig. 3. Dependence of indentation fracture toughness K, on the square root of crack length a”*. indentation fracture toughness of this composite when determined using the indentation technique increased as a function of the square root of the crack length. 4.2 FCGR studies Fatigue crack propagation behaviour of this ceramic material is reported in Fig. 4. The FCGR data of this material were fitted to the usual Paris equation da/dN = A (AlK)n,2C22 and the fatigue crack growth rate da/ldN (in m/cycle) increased linearly with AK (in MPa m1’2) in a log-log plot. The plot yields a value of 12 = 155 and A = 3.4 X lo-l5 [m/cycle (MPa ml”)-“]. 4.3 SEM studies SEM fractographs of the fracture surface at the low AK cyclic and a$ monotonic loading (fast fracture region) are presented in Figs 5 to 11. It is evident that, at low AK (04 to 1.8 MPa m1’2), a majority of the whiskers fail by producing a flat fracture surface which has a vertical level, same as the general fracture surface of the composite (see Fig. 5). On the other hand, in the fast fracture region, the whiskers faril predominantly by pull- ~ R= 0.1 to 0.4, / f-50 Hz UI \ c IO7 - t a -8 6 10 - % MAlERIAL f i 169 - 0 25 Sic - Al201 &t x s :: 0 MgO PSZ nq & 10 -10 - I I I lo6 - R = 0.1, f z 1Hz 3 4 5 Stress Intensity Range, AK(MP am) Fig. 4. Fatigue crack propagation rate data. Fig. 5. At low AK (0.8 to 1.8 MPa ml’*) region, a majority of the whiskers failed with a square fracture without evidence of large-scale pull-out. Fig. 6. In the fast fracture region (K,, = 5.9 MPa ml”), due to monotonic loading, whiskers failed predominantly by pull-out mechanism. out wherein the whiskers stick out of the general fracture surface of the composite (see Fig. 6). The matrix material failed predominantly by transgranular fracture (see Fig. 7) at low AK fatigue (0.8 to 1.8 MPa m1’2). During monotonic