Improvement of the Contact Strength of Al 2 O 3 / SiC by a Combination of Shot Peening and Crack-Healing

Al 2 O 3 /SiC composite ceramics with high crack-healing ability were subjected to shot peening (SP) using zirconiumoxide shots with several peening pressures and shot diameters. Specimens subjected to SP were heat-treated in air to heal the surface cracks induced by SP. The residual stress, the apparent fracture toughness, and the Weibull distribution of the contact strength were investigated, revealing that the combination of SP and crack-healing is effective for increasing the contact strength and decreasing the scatter of the contact strength of Al 2 O 3 /SiC.


Introduction
Shot peening (SP) is a procedure commonly utilized to increase the fatigue strength of metals.The compressive residual stress generated by SP prevents fatigue crack propagation.Recent studies have shown that the near-surface strength of ceramics can be improved by SP [1,2].Pfeiffer and Frey [1] reported that high compressive residual stress of up to more than 1 GPa could be introduced near the surface of alumina (Al 2 O 3 ) and silicon nitride (Si 3 N 4 ) substrates treated by SP using 650 m tungsten carbide shots.Furthermore, Tanaka et al. [2] reported that compressive residual stress of up to 1.5 GPa was introduced near the surface of Si 3 N 4 that was subjected to SP using 50 m high-strength steel shots and 1100 m tungsten carbide shots and that the compressive residual stress increased the apparent fracture toughness.Additionally, Moon et al. [3] reported that the residual stress which is introduced to Al 2 O 3 by SP is attributed to microcracks and dislocation.Based on the results obtained in previous studies, SP treatment is a promising technique for increasing the strength of ceramics in the surface region.Strengthening the surface layer increases the contact strength of ceramic components such as bearings and cutting tools, for which higher contact strengths are desirable.However, unintentional introduction of cracks during the SP process may compromise the reliability of the ceramics components.
Certain structural ceramics exhibit crack-healing ability [4][5][6][7].Thus, if crack-healing can be combined with SP, the surface strength and reliability of ceramics can be increased.Takahashi et al. [8,9] investigated the effects of SP and crackhealing on the contact strength of an Si 3 N 4 /SiC composite.The Si 3 N 4 /SiC was subjected to SP using 300 m zirconium oxide (ZrO 2 ) shots, and the contact strength of the Si 3 N 4 /SiC was reportedly increased by SP treatment [8].It was also reported that the scatter of the contact strength in the SP specimens of the Si 3 N 4 /SiC could be improved by crackhealing in air at 1100 ∘ C for 5 h given that the surface cracks induced by SP were reduced by crack-healing.However, the combined effects of SP and crack-healing on the Al 2 O 3 /SiC composite have not yet been investigated.
In this study, the effects of shot peening and crack-healing on the residual stress, the apparent fracture toughness, and the Weibull distribution of contact strength are investigated for the Al 2 O 3 /SiC composite.

Test Procedure
2.1.Materials and Specimens.Silicon carbide reinforced alumina (Al 2 O 3 /SiC) was selected as a test material because it has excellent crack-healing ability [5,6].The Al 2 O 3 powder (AKP-20, Sumitomo Chemical Co. Ltd., Japan) used in this study has an average particle size of 0.5 m.The SiC powder (ultrafine grade, Ibiden Co. Ltd., Japan) used has a mean particle size of 0.27 m.The samples were prepared using a mixture of Al 2 O 3 with 15 vol.%SiC powder.Alcohol was then added to the solution, and the mixture was thoroughly homogenized by blending for 48 h.The mixture was placed in an evaporator to extract the solvent and then vacuum-treated to produce a dry powder.The mixture was subsequently hotpressed at 1700 ∘ C and 35 MPa for 2 h under N 2 atmosphere.The hot-pressed plate was then cut into test specimens measuring 3 × 4 × 40 mm.The specimens were polished to a mirror finish on one face.These specimens are hereinafter referred to as "Smooth" specimens.

Crack-Healing and Shot Peening Conditions.
The "Smooth" specimens were crack-healed in air at 1300 ∘ C for 1 h.These specimens are denoted "CH" (crack-healed) specimens.This crack-healed condition in which surface cracks of about 100 m long can be healed fully was selected based on the results of a previous study [5,6].
Shot peening (SP) was carried out on the "Smooth" specimens using a direct pressure peening system.Commercial zirconium oxide (ZrO 2 ) shots with a diameter of 180 m were used.The Vickers hardness of the ZrO 2 shots was 1250 HV.The peening pressure selected was 0.1 MPa.The specimens were shot peened for 40 s.The peening coverage was 300%.The specimens subjected to SP are denoted as "SP" specimens.Following SP, the specimens were heat treated in air at 950 ∘ C for 100 h.The oxide film of 1-2 m depth was then removed by polishing the specimens with diamond abrasives with a diameter of 1 m.These specimens are denoted "SP + CH" specimens.This crack-healed condition was selected based on the results of a preliminary experiment [10].Surface cracks introduced by SP can be healed while maintaining the compressive residual stress introduced by SP.

Residual Stress, Surface Roughness, and Apparent Fracture
Toughness.The residual stresses on the surface were measured for Smooth, CH, SP, and SP + CH specimens via the Xray diffraction method.The conditions for X-ray diffraction are shown in Table 1.The surface roughness was measured at five points on each specimen using a stylus-type surface roughness tester.
Smooth, CH, SP, and SP + CH specimens were prepared for measurement of the apparent fracture toughness (  ) at the subsurface.The   was evaluated by the indentation fracture (IF) method, in which   can be estimated using the following equation in accordance with the Japan Industry Standards (JIS) [11]: where  is Young's modulus,  is the indent load,  is half the diagonal length of the indentation, and  is half the surface crack length.Vickers indentations were introduced on the polished surface by applying an indent load of 49 N for 20 s.The Young's modulus of the material is 324 GPa, as determined by the strain measurements in the bending test.

Measurement of Contact Strength.
Smooth specimens, CH specimens, SP specimens, and SP + CH specimens were subjected to contact strength measurement.The contact strength was measured using sphere indentation tests.
Figure 1 shows the testing system.Indentations were made on the surfaces of the specimens using tungsten carbide (WC) spheres with a diameter of 4 mm.The sphere indentation tests were carried out at a crosshead speed of 0.2 kN/min using a universal testing machine.The ring crack initiation load ( max ) was determined using acoustic emission (AE).
Loading was interrupted at the load at which crack initiation was detected by the AE system.The loads were then removed, and the indented surfaces were observed using optical microscopy to identify the point of crack initiation.Figure 2 shows an example of a ring crack.The Weibull distributions of the crack initiation load were determined for each specimen.The distance between indentations was 4 mm, and thus the indentations were fully separated.In this study, the ring crack initiation loads ( max ) were statistically analyzed by a two-parameter Weibull function.In these analyses, the failure probability, , was calculated using the median rank estimate.These results were plotted using the two-parameter Weibull function, expressed by the following equation: where  max is the ring crack initiation load,  is the scale parameter, and  is the shape parameter.The scale parameter  describes the crack initiation load when ( max ) = 63.2%.The shape parameter  describes the width of the distribution of the crack initiation load.

Residual Stress.
Compressive residual stress on the surface layer may reflect the volume expansion induced by microcrack initiation within crystallites [3]. Figure 3 shows the residual stress distributions for the SP (◻) and SP + CH specimens (⧫).The in-depth residual stress distribution was investigated after removing the surface layers by polishing the specimens with diamond abrasives of 9.0, 3.0, and 0.5 m.
The residual stress induced by polishing was also investigated by measuring the residual stress after polishing.A maximum compressive residual stress of 300 MPa was observed on the surface of the SP specimen.The compressive residual stress decreased in the depth direction.Compressive residual stress of up to approximately 10 m was generated.The compressive residual stress decreased in the case of the SP + CH specimen, whereas this sample still exhibited a compressive residual stress of 200 MPa.Thus, the induction of compressive residual stress and crack-healing occurred simultaneously.
The lattice strain was relieved due to thermal expansion as a result of heat treatment.Crack-healing due to oxidation of SiC occurs above 800 ∘ C. The oxidation of SiC is accompanied by approximately 80% volume expansion of the condensing phase.The oxidation products fill the microcracks, which leads to volume expansion at the surface.Thus, the relaxation     the extent of this increase was quite small.Thus, it can be deduced that SP treatment can be effectively applied to ceramic components.The surface roughness increased after heat treatment due to oxidation of the surface layer by heat treatment in air.The surface roughness values of the SP + CH specimens were comparable to those obtained for the Smooth specimens given that the oxide film was removed by polishing.However, it is estimated that contact strength is not affected by surface roughness because the variation of surface roughness is small.

Apparent Fracture Toughness.
Figure 4 shows the apparent fracture toughness (  ) at the sub-surface measured by the IF method.The   increased by 96% in the case of the SP specimens in contrast with the   for the Smooth specimens.Subsequent to heat treatment, the   of the SP specimens decreased.Nevertheless, the   of the SP specimen after heat treatment was 50% higher than that of the Smooth specimen.Thus, SP in combination with crack-healing can improve   while healing the surface crack.

Contact Strength.
Figure 5 shows the effects of shot peening and crack-healing on the Weibull distribution of the  max .The shape parameter and scale parameter of the  max obtained from the Weibull distribution are summarized in Table 3.

Effects of Crack-Healing.
Good linearity of the Weibull distribution of  max was achieved for the Smooth (I) and CH specimens (•).The scale parameter of CH specimens increased by 20% in contrast with that of the Smooth specimens.Thus, the value of  max of the CH specimens increased in contrast with that of the Smooth specimens.The shape parameter of CH specimens increased by 24.4% in contrast with that of the Smooth specimens.Thus, the scatter of the  max decreased in the case of the CH specimens in contrast with that of the Smooth specimens.This is due to the healing of the minute surface cracks introduced by the machining of the Smooth specimens.

Effects of Shot
Peening.The scale parameter of SP specimens increased by 84%.Thus, the mean value  max of the SP specimens increased in contrast with that of the Smooth specimens.However, the shape parameter of SP specimens decreased by 44.8% in contrast with that of the Smooth specimens.Moreover, the Weibull distribution of  max for the SP specimens (◻) was not characterized by good linearity in contrast with that of the Smooth specimens (I).Possible reasons for the nonlinearity of the Weibull distribution of the SP specimens are considered as follows.
In the lower fracture probability region, the  max values of the SP specimens were smaller than those of the Smooth specimens given that the surface cracks induced by SP were present near the sphere indentation site.The  max depends on the size and distribution of the cracks induced by SP.When a large crack is present near the sphere indentation site, the  max tended to decline.
In the higher fracture probability regions, the  max values of the SP specimens were larger than those of the Smooth specimens.It was presumed that the size of the surface cracks near the sphere indentation site was small.Thus, the crack initiation load increased because of the effect of the compressive residual stress.

Effects of Combination of Shot Peening and Crack-
Healing.The scale parameter of SP + CH specimens increased by 84% and 53% in contrast with that of the Smooth and CH specimens, respectively.Thus, it is clear that the SP + CH specimens exhibited higher  max than the Smooth and CH specimens.The shape parameter of SP + CH specimens increased by 87.8% in contrast with that of the SP specimens.Thus, it is also clear that the  max values of the SP + CH specimens were characterized by a smaller scatter than those of the SP specimens.Thus, SP in combination with crack-healing is a useful technique for improving the contact strength of ceramics.
In the lower fracture probability region, the  max values of the SP + CH specimens (⧫) were higher than those of the SP specimens (◻), which is attributed to the healing of the cracks induced by SP for the former samples.
In the higher fracture probability region, the  max values of the SP + CH specimens were larger than those of the Smooth (I) and CH specimens (•) given that the minute surface cracks introduced during machining of the Smooth specimens were healed, and a compressive residual stress of 200 MPa remained on the specimen surface.However, the  max of the SP + CH specimens was slightly lower than that of the SP specimens given that the compressive residual stress decreased due to the heat treatment, as discussed in Section 3.1.

Conclusions
The effects of shot peening (SP) and crack-healing on the residual stress, the apparent fracture toughness, and the Weibull distribution of contact strength were investigated using an Al 2 O 3 /SiC composite.The results are summarized as follows.
(1) Compressive residual stresses of about 300 MPa were observed at the surface of the SP specimens.(2) The apparent fracture toughness (  ) at the subsurface of the SP specimens increased by 96% for the SP specimen in comparison with the   for the Smooth specimens.(3) The compressive residual stress and the   of the SP + CH specimens decreased in contrast with the SP specimens.However, the   value of the SP + CH specimens was 50% higher than that of the Smooth specimen.Thus, it could be deduced that the induction of compressive residual stress and crackhealing occurred simultaneously.(4) The SP specimens exhibited the highest crack initiation load ( max ) of the evaluated specimens.However, the scatter of the  max values was larger in the case of the SP specimens than that of the values of the other specimens because cracks induced during the SP process were present near the specimen surface.(5) The scatter of the  max values of the SP specimens could be improved by crack-healing in air at 950 ∘ C for 100 h given that the surface cracks induced by SP were effectively healed.(6) On the basis of the results discussed in points (1) to (5), it can be concluded that SP in combination with crack-healing is a useful technique for improving the contact strength of ceramics.

Figure 1 :
Figure 1: Schematic illustration of the sphere indentation test, ring cracks on the surface, and AE signal.

Figure 2 :
Figure 2: A ring crack after the sphere indentation test.

Figure 3 :
Figure 3: Relationship between residual stress and depth from the surface.

Figure 5 :
Figure 5: Effects of shot peening and crack-healing on the Weibull distribution of crack initiation load.

Table 1 :
Conditions for X-ray diffraction.

Table 2 :
Surface roughness of each test specimen.
3.2.Surface Roughness.Table2shows the average value of the surface roughness for each specimen.It can be seen that the surface roughness increased after SP treatment.However,

Table 3 :
Shape parameter  and scale parameter  of contact strength.