Grain Bridging Locations of Monolithic Silicon Carbide by Means of Focused Ion Beam Milling

Grain bridging locations of monolithic silicon carbide by means of focused ion beam milling technique

N. Al Nasiri[1], E. Saiz, F. Giuliani and L.J. Vandeperre

Centre for Advanced Structural Ceramics, Department of Materials, Imperial College London, South Kensington Campus, London SW7 2BP, United Kingdom

Abstract

A slice and view approach using a focused ion beam (FIB) milling technique was employed toinvestigate grain bridging near the tip of cracks in four silicon carbide (SiC) based materials with different grain boundary chemistries and grain morphologies. Using traditional observationsintergranular fracture behaviour and hence clear evidence of grain bridging was found for SiC based materials sintered with oxide additives.More surprisingly, in large grain materials, the FIB technique reveals evidence of grain bridging irrespective of the grain boundary chemistry, i.e. also in materials which macroscopically fail by transgranular failure.This helps to explain why the toughness of large grained materials is highereven if failure is transgranular.

Keywords: SiC, FIB, grain bridging and grain boundary chemistry.

Introduction

The strong covalent bonding between silicon and carbon atoms gives SiC materials an excellent oxidation resistance, strength retention to high temperatures, high wear resistance, high thermal conductivity and good thermal shock resistance. Nowadays, SiC is recognized as one of the leading structural ceramics for gas turbines [1], for mechanical seals and brakes [2] and for bioengineering application (i.e. implants) [3, 4]. On the other hand, there is a limitation to the usefulness of silicon carbide materials due to their low fracture toughness and high strength dispersion, i.e. an inhomogeneous and large distribution of flaws and defects.

During the last 30 years a lot of attention has been devoted to improving the toughness of ceramic materials to ensure that the damage during processing and post-processing can be acceptable without compromising the structure’s reliability. Ceramic materials have been toughened using different approaches. A very successful approach consists of crack tip shielding through grain bridging in the wake of the crack tip [5]. The toughness increase arises because crack deflection leaves grains connected to both crack faces and these grains transmit a closing force across crack walls trying to prevent the crack from further extension by reducing the applied stress intensity at the crack tip [6]. The resistance to crack propagation due to grain bridging in general increases with increasing crack length, which leads to the so called Resistance-curve (R-curve) [7-9].During the last 30 years, the R-curve behaviour of ceramic materials has been extensively studied for long cracks. What may happen for short cracks (< 100 µm) and its effect on toughness has until recently largely been overlooked. A material with flat R-curve may still have a steep rise at the beginning of the test as suggested by Fett et al. [10]. In materials with a steep rising R-curve, one may expect that toughness will behave differently for short cracks than for long cracks. The understanding of this concept is highly relevant as ceramic materials are generally produced with small natural defects in the order of few microns[11]. While there is general agreement that the action of grains a long distance behind the crack tip is a consequence of friction between the grains and the matrix from which they are being pulled, the steep rise in R-curve at the onset of failure is increasingly attributed to elastic bridging, i.e. where the grain bridging the crack faces are still to some extent fully bonded to both fracture surfaces. Different approaches were considered to study the effect of elastic bridging i.e. intact grains still bonded to both crack faces bridging the crack near its tip. For example Foulk et al. [12] proposed a 2D model of bridging by a single elongated grain based on finite element analysis during the early stage of the rising R-curve. This increase in crack resistance in very short cracks before forming grain bridges is in line with Fett et al. [10].Fünfschilling and co-workers[13] observed that silicon nitride materials sintered with different additives show different R-curve behaviour ranging from flat, moderate and rising R-curve. However, the main common factor between the studied microstructures is that they all have a steep rising R-curve in the first 10 µm of crack extension. Still, direct evidence from observation of such elastic grain bridging sites close to the tip of the crack has so far not been reported.

Experimental methods

In this work, four silicon carbide based materials were produced: two using solid state sintering with boron and carbon (SS)and two using liquid phase sinteringusing alumina and yttria as sintering additives (LP), see Figure 1. The processing details are described in[14]. The LP-Fine consists of homogenous fine equiaxed grains with average grain size of 1.0 ± 0.1 µm while LP-Coarse has elongated grains with a length of 16 ± 1 µm and a width of 5 ± 0.3 µm. For the SS-Fine, the grains have an average size of 4.0 ± 0.4 µm and the grains in the SS-Coarse have a length of 18 ± 1 µm and a width of 5 ± 0.3µm.The samples were ground and polished up to 1µm using diamond suspension. A crack is introduced by placing an indent of 30 kg on the polished surface.Once the crack tip was found an area with dimensions of 4µm by 5 µm by 5 µm, Figure 2, was selected for investigation with a slice and view approach, i.e. progressive milling in small steps (20 nm) along the X-axis and taking micrographs in between each milling stage to create a 3D information set of the microstructure and crack.

Figure 1: Scanning electron micrographs of chemically etched silicon carbide materials of this work.

Figure 2: Sketch showing the selected coordinates to identify planes and region of interest for the FIB slice and view process.

The R-curve measurements were carried out insitu in the SEM (vacuum level of 1 x 10-3 Pa) using a constant moment double cantilevered test rig, built in line with the set-up proposed by Sorensen et al.[15]. The silicon carbide specimens were 65 mm by 10 mm by 5 mm. One side of the samples was polished to facilitate the observation of the crack tip. The speed of the stage motor was 0.1mmmin1. Every time crack propagation was observed, the applied load was recorded and a SEM micrograph was taken.

Results and discussion

As expected in the LP-sintered materials, cracks are deflected at grain boundaries, whereas in the SS-sintered materials, the cracks grow much straighter and do not appear to be influenced strongly by grain boundaries, see Figure 3. However, a closer look at the coarse grained materials revealed that close to the tip of the crack, fracture is slightly complicated. While fracture continues to be along the grain boundaries in LP-Coarse and more cleavage-like in SS-Coarse, ligaments of material which remain solidly bonded to both sides of the fracture surface can be found in both materials.

Figure 3: FIB micrographs showing crack deflection for LP-Fine and SS-Fine and elastic bridging for LP-Coarse and SS-Coarse. Squares represent the area of interest, where elastic bridging takes place.

In LP-Coarse sequence it is clear that the main crack is found to change from propagating on the right of a grain initially but changes to propagating on the left of a grain as the distance from the crack tip increases as noted at the bottom box in Figure 3. For the SS-Coarse, it is clear that the crack front has split at a grain and is moving forward either side of it, while leaving a sliver of material connecting the two fracture surfaces.

The evolution of another elastic bridge in LP-Coarse is shown in Figure 4. It was not possible to pin point the exact location of the crack tip during slice and view due to crack propagation driven by a relaxation of the residual stresses when the material surrounding the crack was milled away. However, by taking the actual crack tip location as a reference (X = 0 nm), it can be stated that the crack propagated an unknown distanceX0 beyond the first slice obtained. The location of subsequent slices has been determined from the thickness of each slice (20 nm) and as a function of the X0 distance, indicating the propagation of the crack up to the crack tip. The crack starts splitting at a distance of X0 + 2200 nm and it reaches its maximum separation between its branches at a distance of X0, which is the closest slice obtained from the crack tip.

Figure 4: FIB micrographs and corresponding sketches of the crack path at different distances approaching the crack tip of LP-Coarse.

At a distance of X0+2280nm, the crack bumps into a grain, which is shown as a small curve comparing to the rest of the crack path. Part of the crack path is not visible at a distance of X0+2200nm, because the grain is holding it from either side to prevent further propagation. Until X0+1800nm, the separation between the crack faces is small and occurs in a different direction (perpendicular to the crack front). A noticeable big separation is observed at a distance of X0+1500nm. Further splitting of crack faces takes place, and the separation between the crack segments is largest at a distance equal to X0 + 300 nm, hence near the crack tip.

The lack of grain bridges near the crack tip in SS-Fine SiC relative to SS-Coarse can explain why the long crack toughness of the latter is i.e. ~ 1 MPa m1/2 higher while both materials have flat R-curves in the long crack regime (Figure 5). Similarly, the R-curve of LP-Coarse SiC quickly rises to 5 MPa m1/2 in the first 100’s of micrometre of crack propagation, whereas LP-Fine shows a slow rising R-curve starting from ~ 2MPa m1/2. Hence, the elastic bridging near the tip of the crack seems to require the presence of large grains. It is likely that this is the result of residual stresses due to differential thermal expansion, leading to protective compressive stresses which prevent the crack from penetrating the large grains, or the presence of microcracks which help the crack front to split.

Figure 5: R-curves of the four SiC materials measured by DCB measurements under vacuum level of 1x10-3 Pa. Note the rising R-curves in LP-SiC compared to flat R-curves in SS-SiC.

Conclusion

The work provides evidence that elastic grain bridging occurs in coarse grained silicon carbide regardless of the grain boundary chemistry, whereas no evidence of such elastin ligaments was found in fine grained microstructures. The presence and absence of elastic bridges as observed here correlates with toughness values as measures for short extension of cracks elsewhere.

Acknowledgments

The authors would like to thank the Engineering and Physical Sciences Research Council (EPSRC) of the UK for the funding of this project through grant EP/F033605/1.

References

1.L. Jiang, A review of silicon carbide development in MEMS applications. International journal computational materials science and surface engineering, 2009. 2(3/4): p. 225-240.

2.X. Zhao, Y. Liu, Q. Wen, and Y. Wang, Frictional performance of silicon carbide under different lubrication conditions Friction Springer Link, 2013. 2(1): p. 58-63.

3.J. Will, A. Hoppe, F. Müller, C. Raya, J. Fernández, and P. Greil, Bioactivation of biomorphous silicon carbide bone implants. Acta Biomaterialia, 2010. 6(12): p. 4488-4494.

4.J. Chevalier, P. Taddei, L. Gremillard, S. Deville, G. Fantozzi, J. Bartolomé, C. Pecharroman, J. Moya, L. Diaz, and R. Torrecillas, Reliability assessment in advanced nanocomposite materials for orthopaedic applications. Journal of the Mechanical Behavior of Biomedical Materials, 2011. 4(3): p. 303-314.

5.R. Steinbrech, Toughening mechanisms for ceramic materials. Journal of the European Ceramic Society, 1992. 10(13): p. 131-142.

6.R. Ritchie, Mechanisms of fatigue crack propagation in metals, ceramics and composites: Role of crack tip shielding. . Materials Science and Engineering, 1988. A103(1): p. 15-28.

7.P. Swanson, C. Fairbanks, B. Lawn, Y. Mai, and B. Hockey, Crack-interface grain bridging as a fracture resistance mechanism in ceramics: I, experimental study on alumina. Journal of the American Ceramic Society, 1987. 70(4): p. 279-289.

8.P. Swanson, Crack-interface traction: A fracture resistance mechanism in brittle polycrystals. Journal of the American Ceramic Society, 1988. 22(1): p. 135-155.

9.R. Steinbrech and O. Schmenkel, Crack-resistance curves of surface cracks in alumina Journal of the American Ceramic Society, 1988. 71(5): p. C-271-C-273.

10.T. Fett, S. Fünfschilling, M. Hoffmann, R. Oberacker, H. Jelitto, and G. Schneider, R-Curve determination for the initial stage of crack extension in Si3N4. Journal of the American Ceramic Society, 2008. 91(11): p. 3638-3642.

11.S. Fünfschilling, T. Fett, R. Oberacker, M. Hoffmann, G. Schneider, P. Becher, and J. Kruzic, Crack-tip toughness from vickers crack-tip opening displacements for materials with strongly rising R-cuves. Journal of the American Ceramic Society, 2011. 94(6): p. 1884-1892.

12.J. Foulk, R. Cannon, G. Johnson, P. Klein, and R. Ritchie, A micromechanical basis for partitioning the evolution of grain bridging in britle materials. Journal of the Mechanics and Physics of Solids, 2007. 55(2): p. 719-743.

13.S. Fünfschilling, T. Fett, M. Hoffmann, R. Oberacker, T. Schwind, J. Wippler, T. Böhlke, H. Özcoban, G. Schneider, P. Becher, and J. Kruzic, Mechanisms of toughening in silicon nitrides: The roles of crack bridging and microstructure. Acta Materialia, 2011. 59(10): p. 3978-3989.

14.N. Al Nasiri, N. Ni, E. Saiz, J. Chevalier, F. Giuliani, and L. Vandeperre, Effect of microstructure and grain boundary chemistry on slow crack growth in silicon carbide at ambient conditions. The European Ceramic Society, 2014. 35(8): p. 2253-2260.

15.B. Sorensen, A. Horsewell, O. Jorgensen, and A. Kumar, Fracture Resistance measurement method for in situ observation of crack mechanisms. Journal of the American Ceramic Society, 1998. 81(3): p. 661-669.

1 Corresponding author. Tel: +44 207 5895111

Email address: