Evaluation of fatigue properties under four-point bending and fatigue crack propagation in austenitic stainless steel with a bimodal harmonic structure

Austenitic stainless steel (JIS-SUS304L) with a bimodal harmonic structure, which is defined as a coarse-grained structure surrounded by a network of fine grains, was fabricated using powder metallurgy to improve both the strength and ductility. Four-point bending fatigue tests and Kdecreasing tests were conducted in air at room temperature under a stress ratio R of 0.1 to investigate fatigue crack propagation in SUS304L. The fatigue limit of this harmonic-structured material is higher than that of the material with a homogeneous coarse-grained structure. This is attributable to the formation of fine grains by mechanical milling and to the suppression of pore formation. In contrast, the threshold stress intensity range, Kth, for the harmonicstructured material is lower than that for the homogeneous coarse-grained material, while the crack growth rates, da/dN, are higher at comparable K. These results can be attributed to a reduction in the effective threshold stress intensity range, Keff,th, due to the presence of fine grains in the harmonic structure.


INTRODUCTION
ustenitic stainless steel has been widely used in various engineering fields because of its high heat resistance, high corrosion resistance, and excellent formability [1].Recent years have seen a rise in the demand for improvement in the mechanical properties of structural steels, including the present austenitic stainless steel.Thus, increasing the structural reliability of steels has become important.The microstructures and mechanical properties of stainless steels can be controlled by surface modification [2][3][4][5][6][7], the addition of different alloying elements [7,8], and grain refinement [2,5,9,10].Grain refinement through severe plastic deformation is particularly effective in strengthening metallic materials; however, it leads to a decrease in ductility [9,11,12].In order to suppress the decrease in ductility due to the formation of a homogeneous fine-grained structure, new microstructural designs were proposed [11][12][13][14][15][16][17].Our research group has designed a harmonic structure using powder metallurgy to sinter mechanically milled stainless steel powders [18][19][20][21][22][23], which improve both their strength and ductility by suppressing necking during tensile deformation [20].In particular, we have focused on the fatigue properties under fourpoint bending [24][25][26][27] and near-threshold fatigue propagation of long cracks [28][29][30] in titanium-based materials with a bimodal harmonic structure.To achieve sufficient performance with the newly developed harmonic structured stainless steels, their fatigue properties and fatigue crack propagation behavior need to be examined.The purpose of the present study is to elucidate the mechanism of fatigue fracture in bimodal harmonic-structured austenitic stainless steel under four-point bending, and to examine the effects of the bimodal harmonic structure on the fatigue crack propagation in austenitic stainless steel.

Material and specimens
his study employed austenitic stainless steel (JIS-SUS304L) containing 19.35% Cr, 9.18% Ni, 1.83% Mn, 0.25% Si, and 0.02% C (all by mass, with the balance being Fe).This material was made into a powder (particle diameter: 120 µm) through a plasma rotating electrode process that can be used to fabricate spherical particles with negligible contamination by impurities such as oxygen or nitrogen gas [31].A bimodal microstructural design using mechanical milling (MM) and spark plasma sintering (SPS) was introduced for the formation of the harmonic-structured SUS304L.MM was performed for 180 ks in Ar at room temperature for the SUS304L powders using a planetary ball mill (Fritch P-5) with a tungsten carbide vessel and steel ball bearings to form fine grains on the particle surfaces.The rotation speed was 200 rpm, and the ball-to-powder mass ratio was 2:1.The powders were subsequently consolidated by SPS at 1223 K for 3.6 ks under vacuum (less than 15 Pa) and applied pressure (50 MPa) using a 25-mm internal diameter graphite die to produce the specimens, hereafter referred to as the "MM series."A second set of specimens was prepared by sintering the as-received powders (hereafter, the "untreated series") for comparison.The tensile strength of the MM series was higher than that of the untreated series, but the elongations of the two series were almost the same [19,20].The MM series has a Vickers hardness of 169.6±2.5 HV, as measured for a polished surface with an indentation force of 1.961 N and a load holding time of 10 s (n = 30).Hardness value of the MM series is higher than that of the untreated series (125.2±3.6 HV).The sintered materials were sliced into disks approximately 1.5 mm thick and machined into a blunt-notched specimen for four-point bending fatigue tests [26,27].After machining, the specimen surface was polished with emery paper (#240 to #4000) to a thickness of 1 mm and polished in a SiO 2 suspension to obtain a mirror finish.The notch roots of the specimen were also polished with emery paper (#240) to remove the electro-discharge machined layer.

Testing
Four-point bending fatigue tests were performed in an electrodynamic fatigue testing system under a stress ratio R of 0.1.The frequency of stress cycling was 10 Hz, and the tests were conducted in ambient conditions.Fatigue tests were interrupted after a certain number of cycles, and acetyl cellulose films were placed on the specimen surface using the replica method to examine fatigue crack initiation and propagation.Once the crack length was measured using optical microscopy, the stress intensity range, K, was calculated [27,28,32]; the aspect ratio, c/a, for small cracks was estimated as follows: c/a = 1 -1.607(a/t) + 1.080(a/t) 2 -0.2149(a/t) 3  for a/t < 1 and ( 1) A T where a is the crack length on the surface, c is the crack length along the thickness direction, and t is the specimen thickness.K-decreasing tests were also conducted to approach the fatigue threshold.The present study employed disk-shaped compact (DC(T)) specimens (2 mm thick) [28][29][30] in accordance with the ASTM standard.The tests were conducted in the ambient laboratory atmosphere under a stress ratio R of 0.1.The fatigue threshold, K th , is defined as the maximum value of K under a crack growth rate of 10 -10 m/cycle.Crack lengths were monitored by the unloading elastic compliance method [33].
The magnitude of crack closure was also monitored; the closure stress intensity, Kcl, was obtained from the closure load, Pcl.
After the fatigue tests, fracture surfaces and crack profiles were observed using scanning electron microscopy (SEM), and the microstructure around the crack paths was analyzed using electron backscattered diffraction (EBSD) at an acceleration voltage of 20 kV.

Microstructural characterization
he microstructure of sintered compacts was characterized using EBSD.The image quality (IQ) map and grain boundary map obtained by EBSD analysis for the MM series is shown in Fig. 1.In the present study, the grain boundary is defined as the high-angle grain boundaries greater than 15º.The MM series contained regions of fine equiaxed grains and regions with a coarse microstructure.The regions of fine equiaxed grains formed a continuous connected three-dimensional network that surrounded the coarse-grained structure.This network is referred to as a harmonic structure in the present study.

Fatigue properties determined by four-point bending
Fig. 2 shows the results of four-point bending fatigue tests for the sintered compacts (untreated and MM series); a stress amplitude, a, was applied to the specimen surface as a function of the number of cycles to failure, Nf.In this figure, those plots with an arrow represent the run-out specimens without failure at N = 10 7 cycles.The fatigue limit, which was defined as the average value of the maximum stress amplitude without fatigue failure and the minimum stress amplitude with fatigue failure, for the MM series (260 MPa) with high hardness was higher than that for the untreated series (105 MPa).Sufficient data could not be obtained for the MM series because some MM series specimens did not exhibit fatigue failure at stress amplitudes above 280 MPa owing to plastic deformation.Fig. 3 shows SEM fractographs of the MM series (N f = 3.4×10 5 cycles), which failed at  a = 280 MPa, and of the untreated series (Nf = 5.6×10 6 cycles), which failed at a = 120 MPa.In all micrographs, the surface subjected to tensile stress is the upper surface.Macroscopic observation revealed only one fatigue crack near the specimen surface, which gradually propagated across the cross-section of the specimen.The fracture surface of the MM series specimen is divided into two regions by a clear boundary, as is shown in Fig. 3(a).In contrast, a characteristic, powder-like microstructure is observed on the surface of the untreated series specimen (see Fig. 3(b)).These results indicate that the MM suppresses the formation of pores in SUS304L during the subsequent SPS process owing to plastic deformation [34] of the SUS304L powder surface.
The formation of pores tends to be suppressed with increasing SPS temperature [35]; thus, the SPS temperature (1223 K) is not high enough for the initial SUS304L powder in the present study.
To elucidate the mechanism of fatigue crack initiation in the harmonic-structured SUS304L, EBSD analysis was conducted on the specimens after fatigue testing.Figs. 4 shows the inverse pole figure (IPF) map obtained by EBSD for the fracture surface near the crack initiation site of the MM series and a schematic in which the red square indicates the analyzed region in the MM series.A fatigue crack was initiated in the coarse-grained (> 10 µm) structure of the harmonic-structured MM series.In addition, the fatigue crack profile is not influenced by the harmonic structure, and propagates perpendicular to the loading direction.These same effects have also been observed with a Ti-6Al-4V alloy [24,25] and CP titanium [27].Furthermore, Zhang et al. [23] reported that the fatigue limit for the harmonic-structured JIS-SUS316L austenitic stainless steel tends to increase as the MM time increases under uniaxial stress loading, because the areal fraction of the fine-grained structure tends to increase as the MM time increases [25,36].The harmonic structure increases the fatigue limit of SUS304 owing to grain refinement and the suppression of pore formation during the SPS process.

Fatigue crack propagation behavior
To examine the effect of pores on the fatigue properties of SUS304L, small fatigue crack propagation was examined by the replica method for the untreated series with pores.In the present study, small fatigue crack propagation in the MM series is not examined because the MM series have no pores.Fig. 5 shows optical micrographs of the surface of the untreated series tested at  a = 160 MPa and observed after a given number of loading cycles.Fig. 5 reveals that a 17-µm-long fatigue crack was initiated from a pore after 8.5×10 4 cycles (Fig. 5(b)).Subsequently, multiple cracks were initiated from pores and gradually propagated (Figs.5(c)-(g)).Fatigue cracks, which were initiated from pores, coalesced at several instants and the final fracture occurred at 2.08×10 5 cycles.Fig. 6(a) plots the crack growth rate, da/dN, of long cracks for the untreated and MM series against K.In both series, da/dN decreases with decreasing K and the da/dN values for the MM series are consistently higher than those for the untreated specimens at comparable K levels.Furthermore, the K th value for the untreated series is higher than that for the MM series.The long fatigue crack growth resistance of SUS304L is reduced owing to the presence of the harmonic structure.Furthermore, the da/dN against K eff for the untreated and MM series are also plotted in Fig. 6(a), which reveals that the Keff,th value for the untreated series is higher than that for the MM series.This indicates that the effect of the harmonic structure at near-threshold levels is attributable not to crack closure but to grain size, i.e., the effect of slip-band growth resistance by grain boundaries and misorientations [37,38].This same effect, in which a bimodal harmonic structure affects Keff,th, has also been observed for a Ti-6Al-4V alloy [29].
To compare the small and long crack propagation behaviors of SUS304L, the crack growth rate, da/dN, under four-point bending is calculated using the data shown in Fig. 5. Fig. 6(b) plots the dependence of the crack growth rate, da/dN, on K for small and long cracks in the untreated series.For the small crack, the da/dN values are slightly higher than those for the long crack under comparable K.In addition, a small crack in the untreated series propagates at K values lower than the K eff,th for a long crack.The results show that the K th value for small cracks in the untreated series is lower than that for long cracks with the elimination of crack closure.The same effect has been observed for metallic materials [39,40].Consequently, the harmonic-structured SUS304L exhibits a lower resistance to fatigue crack growth owing to grain refinement, and the formation of pores is suppressed by MM.In the future, the three-dimensional fatigue crack shape [41][42][43][44][45][46] and the misorientations near the crack initiation site [47][48][49] will be measured for SUS304L with fatigue damage using synchrotron radiation to describe the mechanism of fatigue crack initiation and propagation in greater detail.

CONCLUSIONS
he effect of a bimodal harmonic structure on the fatigue properties under four-point bending and fatigue crack propagation in austenitic stainless steel (JIS-SUS304L) was investigated.The main conclusions obtained can be summarized as follows: (1) Mechanical milling suppresses the formation of pores in SUS304L during the subsequent sintering process, owing to the plastic deformation of SUS304L powder surface.(2) The harmonic-structured SUS304L exhibits a higher fatigue limit than the sintered compact prepared from as-received powders, which had pores and a coarse microstructure.(3) The da/dN values for harmonic-structured SUS304L are higher than those for steel fabricated from as-received powders with coarse grains for comparable values of K.(4) The K th for harmonic-structured SUS304L is lower than that for steel fabricated from coarse-grained as-received powders.This difference is attributable to a reduction in Keff,th, resulting from the presence of fine grains in the harmonic structure.

Figure 1 :
Figure 1: Image quality (IQ) map and grain boundary map obtained by EBSD analysis for MM series.

Figure 2 :
Figure 2: Results of four-point bending fatigue tests, showing stress amplitude as a function of cycles to failure.

Figure 4 :
Figure 4: Inverse pole figure (IPF) map obtained by EBSD analysis for MM series that failed at a stress amplitude of 280 MPa (N f = 3.4×10 5 cycles).

Figure 6 :
Figure 6: Relationship between (a) crack growth rate and stress intensity range for the untreated series and MM series, and (b) crack growth rate and stress intensity range for the untreated series with a long crack and a small crack.