Abstract
The effects of shot peening (SP) on the fatigue strength of partially stabilised zirconia were studied. Smooth specimens and specimens containing a surface pre-crack with depths in the range 35–110 μm were subjected to SP. The cyclic fatigue tests were performed using a three-point bending setup. SP introduced compressive residual stresses on the specimens and improved their fatigue strengths. The shot peened specimens with pre-crack depths ≤ 50 μm fractured outside the pre-crack area and exhibited considerably high fatigue limits, equivalent to those of the shot peened smooth specimens. Therefore, the pre-cracks with depths ≤ 50 μm could be rendered harmless by SP, which was confirmed by the theoretical estimations based on fracture mechanics.
Introduction
Zirconia can exhibit superior fracture toughness than other ceramics due to the stress-induced transformation from the tetragonal to the monoclinic phase, 1 which makes it suitable for applications such as mechanical components and dental materials. The mechanical components are subjected to repeated stress during their use. Accordingly, it is important to assess the fatigue strength characteristics of zirconia, and many studies have been carried out in this regard.2–5 In the process of manufacturing of ceramic components, machining procedures such as grinding which may initiate the formation of surface cracks, are carried out. This leads to reduction in the fatigue strength if the size of the crack is larger than the acceptable size.3,4
Shot peening (SP) is a surface treatment process that involves impacting a surface with shots. The main purpose of SP is to increase the fatigue strength and prevent the stress corrosion cracking of metals. It has been reported that SP on the steels containing crack-like surface defects improves the fatigue limit by increasing the acceptable defect size, in spring steels,6–9 low-alloy steels 10 and carbon steels.11,12 In other words, the surface defects are rendered harmless by SP, from the viewpoint of the fatigue limit.
Recent studies have shown that SP introduces compressive residual stresses near the surfaces of ceramics such as silicon nitride,13–16 alumina,13,17 and zirconia.18–20 Moon et al. carried out SP on alumina and silicon nitride and reported that the fracture resistance of the sub-surface regions increased. 13 Pfeiffer and Frey reported that large compressive residual stresses up to 2000 MPa could be introduced into the near surface of alumina and silicon nitride ceramics by SP. 14 They showed that the static and cyclic load capacities increased by a factor of four. 14 Takahashi et al. reported that the combination of SP and crack healing increased the static contact strength of silicon nitride15,16 and alumina composites. 17 Ito et al. reported that the flexural strength and the Weibull coefficient of partially stabilised zirconia (PSZ) could be improved by the beneficial effects of soft SP using aluminium shots. 18 Shukla and Lawrence applied micro-SP using glass beads and found that the method was effective in increasing the fracture toughness of zirconia-advanced ceramics. 19 Takahashi et al. reported that a compressive residual stress of 1800 MPa introduced in PSZ can improve both the wear resistance 21 and the fracture toughness. 20 It was also shown that the flexural strengths of the pre-cracked specimens were significantly improved and the pre-cracks with a depth less than 60 µm were rendered harmless by SP. 20 However, the effects of SP on the cyclic fatigue strength of zirconia have not yet been studied.
This study was carried out to analyse the effects of SP on the cyclic fatigue strength of PSZ containing a surface pre-crack. Smooth specimens and specimens containing a surface pre-crack were subjected to SP and cyclic fatigue tests were performed. The depth of the pre-crack that can be rendered harmless by SP was studied in terms of the cyclic fatigue limit.
Materials and methods
Specimen preparation
The PSZ ceramic used in this study was ZR1 (Japan Fine Ceramics Center), which consists of 3 mol-% Y2O3 as a stabiliser. Figure 1a shows the shape and the dimensions of the specimens used for the cyclic fatigue tests. The tensile surface of the specimen (4 mm wide side) was precision polished to achieve a mirror-finish. These specimens are referred to as the ‘smooth’ specimens. Pre-cracks were introduced at the centre of the tensile surface of the smooth specimen by Vickers indentation and these specimens are referred to as the ‘pre-cracked’ specimens. Figure 1b shows the optical micrograph of the surface of the pre-cracked specimen. The Vickers indentation loads of 30, 50 and 100 N with a loading time of 20 s resulted in surface crack lengths (2c) of approximately 80, 110 and 250 μm, respectively. The aspect ratio (a/c) of the pre-crack was 0.9. Thus, the pre-crack depths (a) were approximately 35, 50 and 110 μm, respectively. Five specimens were used for the measurements of the Vickers hardness (HV). The average values of the Vickers hardness for the SP and the non-SP specimens were 1310 ± 6.3 and 1290 ± 13.3, respectively. We believe that the hardness increased due to the plastic deformation near the surface layer.
a The shape and the dimension of the specimen used in the present study (unit: mm) and the schematic of the pre-crack position. b Optical micrograph of the surface of the pre-cracked specimen (indentation load: 50 N, pre-crack length 2c ≅ 110 μm, pre-crack depth a ≅ 50 μm)
SP procedure and measurement of the residual stress
Conditions used for the SP of PSZ
The residual stresses on the surfaces of the non-SP and SP specimens were measured using the X-ray diffraction method with a Cu–Kα beam X-ray spectrum. The details of the measurement of the residual stress are explained in the literature. 20 A large compressive residual stress of approximately 1400 MPa was introduced on the surface of the PSZ specimen, reaching a maximum of 1800MPa at a depth of 20 µm. 20 As the depth increased to 50 µm, the compressive residual stress decreased to 100 MPa. 20 The depth of the compressive residual stress is important because the pre-crack size rendered harmless by SP is related to the depth of the compressive stress layer. The large compressive residual stress was induced by the local plastic deformation and the transformation of zirconia from the tetragonal to the monoclinic phase. 20 The surface textures before and after SP were observed using optical microscopy and scanning probe microscopy (SPM) techniques.
Measurement of cyclic fatigue strength
The cyclic bending fatigue tests were carried out using a hydraulic testing machine with a stress ratio R = 0.1 and a nominal frequency of 10 Hz, at room temperature and room humidity by the three-point bending method. The span length for the three-point bending test was 16 mm, which was identical to that of the previous study. 20 The origins of the fractures of all the tested specimens were observed using an optical microscope and a scanning electron microscope (SEM).
Results and discussion
Surface texture after SP
Figure 2 shows the optical micrographs of the surfaces of the smooth and the smooth + SP specimens. The surface roughness was determined by the maximum height of the profile (Ry) according to a profilometer scan of the specimen surface. The average values of Ry for five scans were 0.36 ± 0.04 and 0.47 ± 0.07 µm for the smooth and the smooth + SP specimens, respectively. The surface roughness after SP was slightly higher due to the plastic deformation caused by the impact of the shot media. Figure 3 shows the SPM images of the surfaces of the smooth and the smooth + SP specimens, respectively. Nano-sized dimples were observed on the surface of the smooth + SP specimen, instead of the scratch marks on the surface of the smooth specimen. The heights of the dimples were in the range of 30–50 nm, and they had a negligible effect on the fatigue strength, as discussed in the next section.
Optical micrographs of the surfaces of the a smooth and the b smooth + SP specimens SPM images and profiles of the surfaces of the a smooth and the b smooth + SP specimens
Effect of SP on the cyclic bending fatigue strength
Figure 4 shows the S–N curves, showing the relationship between the maximum bending stress during the cyclic fatigue tests (σmax) and the number of cycles to failure (Nf). The bending strengths for each specimen
20
are also plotted in Fig. 4. The asterisks (*) indicate that the specimens fractured outside the pre-crack area. The arrows indicate that the specimens endured 2 × 106 fatigue cycles. The fatigue limit (σf0) was defined as the maximum value of σmax at which the specimen endured 2 × 106 fatigue cycles. Figure 5 shows the comparison of the values of σf0. The fatigue limit of the smooth specimens increased by 19% after SP. The fatigue limit of the pre-cracked specimens decreased with increase in the depth of the pre-crack. However, the values of σf0 for the pre-crack + SP specimens increased by 200–540% compared to those of the pre-cracked specimens not subjected to SP. Thus, SP is effective in increasing the fatigue limit of PSZ. It is surprising that the fatigue limit of the pre-crack + SP specimens with a ≅ 35 or 50 µm were close to that of the smooth specimen. Similar results have been reported for metals6–9,12; however, this is the first report with regard to ceramics.
S–N curves for the following specimens: a Smooth, b Pre-cracked with a ≅ 35 μm, c Pre-cracked with a ≅ 50 μm, and d Pre-cracked with a ≅ 110 μm. The open and solid symbols represent the non-SP and SP specimens, respectively Comparison of the fatigue limits for the smooth specimens, and for the pre-cracked specimens with varying crack depth a
Experimental evaluation of the pre-crack size rendered harmless by SP
Figure 6 shows the SEM micrographs of the fractured surfaces after the fatigue test. All the pre-cracked specimens without SP fractured at the pre-crack region regardless of the pre-crack depth. The Vickers indentations were observed on the fracture surfaces. Among the pre-crack + SP specimens, however, those with a ≅ 110 µm fractured at the pre-crack location, while those with a ≅ 35 or 50 µm fractured outside the pre-crack region. There was no crack propagation from the pre-crack in the pre-crack + SP specimens with a ≅ 35 or 50 µm, owing to the introduction of the compressive residual stress in the near-surface region by SP.
If the fatigue test result of a pre-crack + SP specimen meets either one of the following two conditions, the pre-crack is considered to be rendered harmless by SP.8,9
Condition (a): The fatigue limit increased to more than 95% of that of the smooth + SP specimen.
Condition (b): More than half of the specimens fractured outside the pre-crack zone.
SEM micrographs of the fractured surfaces obtained after the fatigue tests, showing the origin of fracture in the different specimens
In the pre-crack + SP specimens with a ≅ 50 µm, two out of the three specimens fractured outside the pre-crack zone, as shown in the data with the asterisks in Fig. 4c. Thus, based on condition (b), it can be interpreted that the pre-cracks with depths below a ≅ 50 µm can be rendered harmless by SP.
Theoretical estimation of the pre-crack size rendered harmless by SP
The method to evaluate the crack depth, which is rendered harmless by SP, is based on the stress intensity factors. Here, we demonstrate that the estimated value of the crack depth is consistent with the experimental results. It was assumed that a positive value of the stress intensity factor contributes to the propagation of the fatigue crack. The apparent range of the stress intensity factor (ΔKT), which is a driving force for the crack propagation is expressed as follows
6
:
Distribution of the stress intensity factors as a function of the crack depth
The method for determining the size of the pre-crack which is rendered harmless is described as follows. If ΔKT is less than the threshold stress intensity factor range, ΔKth, the surface crack is considered to be harmless. The values of ΔKth were calculated from the fatigue limits (σf0) of the pre-cracked specimens and are shown in Fig. 8. The figure also shows the relationship between ΔKT and a. The fatigue cracks are initiated at the deepest point in the pre-crack (point A) and the values of ΔKT,A increase with the crack depth. Thus, the intersection between ΔKT,A and ΔKth gives the maximum defect size amax, that can be rendered harmless by SP. From Fig. 8, the value of amax was estimated to be 53 µm. The experimental results indicated that the surface cracks with depth a ≤ 50 µm were rendered harmless by SP. Thus, the theoretical result obtained using fracture mechanics is consistent with our experimental result. SP is therefore a useful technique for improving the fatigue strength of PSZ and rendering the detrimental surface cracks harmless.
Estimation of the maximum crack size that can be rendered harmless by SP
Conclusion
The effects of SP on the fatigue strength of PSZ and the pre-crack depth that could be rendered harmless by SP were investigated. The following conclusions were obtained:
The fatigue limits of the pre-cracked specimens after SP were significantly increased, by 200–540%, compared to those of the pre-cracked specimens without SP. The compressive residual stress induced by the local plastic deformation and the transformation of zirconia from the tetragonal to the monoclinic phase contributed to the improvement in the fatigue limits. In the pre-crack + SP specimens with a ≅ 35 and 50 µm, majority of the specimens fractured outside the pre-crack zone. Thus, according to the experimental results, the surface cracks with a depth ≤ 50 µm could be rendered harmless by SP. The maximum crack depth (amax) that could be rendered harmless by SP was also estimated based on fracture mechanics. The value of amax was found to be 53 µm, which was consistent with our experimental results. SP is a useful surface treatment method for improving the fatigue properties of PSZ and rendering detrimental surface cracks harmless.
