504 FABER Matr coating/fiber contact area Fiber/matrix contact area Thin ce Fibe Thick Coating Fib Figure 2 Schematic illustration of the effect of fiber coatings on fiber sliding behavior. The coatings control the degree of asperity interactions by providing varying amounts of separation of the fiber and matrix phase. As the coating thickness is increased, from top to bottom, the asperity interaction(and the resulting radial misfit strain) is reduced (courtesy of D Mumn
P1: ARK/MBL/rkc P2: MBL/vks QC: MBL/agr T1: MBL May 16, 1997 13:47 Annual Reviews AR034-16 504 FABER Figure 2 Schematic illustration of the effect of fiber coatings on fiber sliding behavior. The coatings control the degree of asperity interactions by providing varying amounts of separation of the fiber and matrix phase. As the coating thickness is increased, from top to bottom, the asperity interaction (and the resulting radial misfit strain) is reduced (courtesy of D Mumm)
CERAMIC COMPOSITE INTERFACES 2 Uncoated Increasing Coatin Thickness 6070 COD (um) Figure 3 Force-displacement curve for pullout testing of four SiC-glass composites with carbon erlayers of increasing thickness(from 16) sliding and attributed it to the reduction in residual stress with fiber loading Alternatively, compliant interphase layers of varying thicknesses also can serve to accomodate large residual thermal mismatch stresses (22, 23) MEASURING INTERFACE PROPERTIES Over the past fifteen years, a variety of interfacial tests for ceramic compos ites either have evolved from evaluation methods used in the polymer-matrix composite field or have developed anew for brittle-brittle composites. Shown in Figure 4 is a sampling of tests for the mechanical evaluation of interfaces (24, 25). Of the tests shown, the bimaterial bend test( Figure 4a), the bimaterial cantilever beam test( Figure 4e ), the single-edge notch beam test(Figure 4), the Brazilian disk test( Figure 4g), the double-cleavage drilled compression tests (Figure 4h), the vickers indentation test(Figure 4)), and Hertzian indentation tests(Figure 4) are occasionally used in screening tests to ascertain whether
P1: ARK/MBL/rkc P2: MBL/vks QC: MBL/agr T1: MBL May 16, 1997 13:47 Annual Reviews AR034-16 CERAMIC COMPOSITE INTERFACES 505 Figure 3 Force-displacement curve for pullout testing of four SiC-glass composites with carbon interlayers of increasing thickness (from 16). sliding and attributed it to the reduction in residual stress with fiber loading. Alternatively, compliant interphase layers of varying thicknesses also can serve to accomodate large residual thermal mismatch stresses (22, 23). MEASURING INTERFACE PROPERTIES Over the past fifteen years, a variety of interfacial tests for ceramic composites either have evolved from evaluation methods used in the polymer-matrix composite field or have developed anew for brittle-brittle composites. Shown in Figure 4 is a sampling of tests for the mechanical evaluation of interfaces (24, 25). Of the tests shown, the bimaterial bend test (Figure 4a), the bimaterial cantilever beam test (Figure 4e), the single-edge notch beam test (Figure 4f ), the Brazilian disk test (Figure 4g), the double-cleavage drilled compression tests (Figure 4h), the Vickers indentation test (Figure 4i), and Hertzian indentation tests (Figure 4j) are occasionally used in screening tests to ascertain whether
FABER 干千7 Figure Schematic of test geometries to measure interfacial mechanical properties crocomposite test, (c) single-fiber double cantilever beam, (n)single- beam, (g) Brazilian age drilled compression test() Vickers inde O Hertzian indentation(after 24, 25) a given materials pair is compatible. In each of these tests, the reinforcement must be of a monolithic form. Consequently, they rarely contain the exact sur- face chemistry, microstructure, or residual stress profile of the true fiber-matrix pair. Therefore, the discussion here is limited to those geometries in which the fiber and matrix can be made identically to those in an actual composite rather than to those geometries that allow materials only similar to the mate- rials used in the composites. The former include push-in, push-through tests
P1: ARK/MBL/rkc P2: MBL/vks QC: MBL/agr T1: MBL May 16, 1997 13:47 Annual Reviews AR034-16 506 FABER Figure 4 Schematic of test geometries to measure interfacial mechanical properties: (a) bimaterial bend test, (b) concentric cylinder tensile test or microcomposite test, (c) single-fiber pullout test, (d ) fiber pullout or push-down test, (e) bimaterial double cantilever beam, (f ) single-edge notched beam, (g) Brazilian disk, (h) double-cleavage drilled compression test (i) Vickers indentation, and ( j) Hertzian indentation (after 24, 25). a given materials pair is compatible. In each of these tests, the reinforcement must be of a monolithic form. Consequently, they rarely contain the exact surface chemistry, microstructure, or residual stress profile of the true fiber-matrix pair. Therefore, the discussion here is limited to those geometries in which the fiber and matrix can be made identically to those in an actual composite, rather than to those geometries that allow materials only similar to the materials used in the composites. The former include push-in, push-through tests
CERAMIC COMPOSITE INTERFACES 507 (Figure 4d) relying on an indenter for loading, pullout tests(Figure 4c), and microcomposite tests(Figure 4b) Indentation Push-in and Push-through Techniques Indentation tests receiving the greatest attention due to their simplicity involve a sharp(26, 27)or blunt(27)indenter that is used to push in a fiber in a composite or push through a fiber in a composite of thin cross-section. First developed by Marshall (26), a sharp indenter was used to displace a fiber into a matrix, and the residual displacement could be ascertained from impressions left in the matrix in he near vicinity of the fiber. Marshall Oliver(28)used a nanoindenter for the same purpose and instrumented the test to provide a continuous measure of the force and displacement during loading, unloading, and load cycling. Analysis of the force-displacement results allowed upper bound estimates of the debond fracture energy and frictional sliding stress, in contrast to the original push-in or push-through test, which was limited to frictional stress evaluation. A further variation of the push-through technique uses a cylindrical indenter that allows no contact with the matrix(29) Elegant analysis of the experiment has been presented by Zhou Mai( 32), who include the radial constraint imposed by neighboring fibers on their analysis of stress transfer and frictional push-out in such a test and have recently included roughness effects(33). Not surprisingly, the frictional push-out stress increases with reinforcement volume fraction. and radial constraints of sur- increase as the embedded length increases More recently, the push-out test has been used for arrays of fibers by Mackin zok(34). protruding fibers, 10 to 15 um in height, trolled etching of the matrix and form the push surface. A displacement piston n the underside of the sample measures displacement. The and the number of fibers displaced Pullout Techniques Conventional pullout tests are prepared with an end of the fiber protruding from the matrix material, which is gripped directly to the loading apparatus (35), a variation of Figure 4c. The free length of the fiber provides a processing challenge, an alignment challenge during mechanical testing, and an enhanced compliance in the system that may prove undesirable for unstable debond crack initiation. The matrix crack is replaced by a matrix surface, which may pos an artificial barrier to debond crack initiation and, hence, unstable debonding Despite this, evaluations by Bright et al (36) showed that sliding resistance measurements on a SiC fiber-reinforced borosilicate glass made via pullout and push-out tests were equivalent. However, constant friction shear stress over the
P1: ARK/MBL/rkc P2: MBL/vks QC: MBL/agr T1: MBL May 16, 1997 13:47 Annual Reviews AR034-16 CERAMIC COMPOSITE INTERFACES 507 (Figure 4d) relying on an indenter for loading, pullout tests (Figure 4c), and microcomposite tests (Figure 4b). Indentation Push-in and Push-through Techniques Indentation tests receiving the greatest attention due to their simplicity involve a sharp (26, 27) or blunt (27) indenter that is used to push in a fiber in a composite or push through a fiber in a composite of thin cross-section. First developed by Marshall (26), a sharp indenter was used to displace a fiber into a matrix, and the residual displacement could be ascertained from impressions left in the matrix in the near vicinity of the fiber. Marshall & Oliver (28) used a nanoindenter for the same purpose and instrumented the test to provide a continuous measure of the force and displacement during loading, unloading, and load cycling. Analysis of the force-displacement results allowed upper bound estimates of the debond fracture energy and frictional sliding stress, in contrast to the original push-in or push-through test, which was limited to frictional stress evaluation. A further variation of the push-through technique uses a cylindrical indenter that allows no contact with the matrix (29). Elegant analysis of the experiment has been presented by Zhou & Mai (30– 32), who include the radial constraint imposed by neighboring fibers on their analysis of stress transfer and frictional push-out in such a test and have recently included roughness effects (33). Not surprisingly, the frictional push-out stress increases with reinforcement volume fraction, and radial constraints of surrounding fibers increase as the embedded length increases. More recently, the push-out test has been used for arrays of fibers by Mackin & Zok (34). Protruding fibers, 10 to 15 µm in height, are created through a controlled etching of the matrix and form the push surface. A displacement piston on the underside of the sample measures displacement. The average interface sliding stress is determined from the applied load, the measured displacement, and the number of fibers displaced. Pullout Techniques Conventional pullout tests are prepared with an end of the fiber protruding from the matrix material, which is gripped directly to the loading apparatus (35), a variation of Figure 4c. The free length of the fiber provides a processing challenge, an alignment challenge during mechanical testing, and an enhanced compliance in the system that may prove undesirable for unstable debond crack initiation. The matrix crack is replaced by a matrix surface, which may pose an artificial barrier to debond crack initiation and, hence, unstable debonding. Despite this, evaluations by Bright et al (36) showed that sliding resistance measurements on a SiC fiber-reinforced borosilicate glass made via pullout and push-out tests were equivalent. However, constant friction shear stress over the