Fatigue Fracture Process of Extruded and Heat-Treated Al-Zn-Mg Alloys under Uniaxial Cyclic Loading with Positive Stress Ratio

Abstract

Fatigue fracture process of extruded and heat-treated Al-Zn-Mg alloys, i.e., 7075- and 7150-alloys, under uniaxial cyclic loading with positive stress ratio, i.e., R = 0.25, has been investigated. Several tens’ micrometers wide preexisting “pits” on surfaces have been confirmed as the preferential fatigue crack initiation sites of the original 7000 series alloys. And, after secondary surface treatment, i.e., surface layer removal, metallic compounds or scratches induced during specimen machining have become possible fatigue crack initiation sites. Cycles to failure of the surface-layer-removed specimens were much longer than the specimens before the surface removal. We have classified the entire fatigue process of the original extruded and heat-treated 7000 series alloys as follows. Crack Initiation Stage: Cleavage crack starts from the interface between several tens’ micrometers wide preexisting “pit” and matrix. Striation Incubation Stage: Cleavage crack without remarkable striation propagates into the specimen. Striation Stage: Striations propagate and create the several millimeters wide semicircle “Fatigue Crack Zone”. Final Fracture Stage: Ductile dimple fracture stage with the final one cycle follows the “Fatigue Crack Zone” and leads the specimen to failure. We have quantitatively estimated the period of each stage and revealed that the lengthiest stage is the combined stage of the crack initiation and the striation incubation. When the cycle to failure reaches six digits number, the period of the combined stage becomes around 90% of the cycle to failure.

Fig. 4 High-speed camera visions of the fatigue fracture sequence of the dialogue-ballooned specimen of Fig. 3. LD means the loading direction. Each frame shows crack appearances at (a) 0, (b) 40,005, (c) 58,095, (d) 65,003, (e) 69,696 and (f) 69,697 loading cycles, respectively. (online color)

1. Introduction

High strength aluminum alloys are widely used as structural materials of aircraft due to their high specific strength originated from the precipitation hardening created through adequate heat treatment. Regarding commercial jets, Al-Zn-Mg alloys represented by 7000 series alloys and Al-Cu-Mg alloys represented by 2000 series ones are mainly used. As for the constituent materials of the aircraft, fatigue strength is one of the most required mechanical properties. However, despite their static high strength, aluminum alloys are generally inferior to steels because of their poor fatigue related properties. Fatigue properties of 7000 series alloys which have the highest static strength among the aluminum alloys have been investigated by many researchers [1–15]. It is well known that 7000 series alloys are not appropriate materials such as the lower wing skins which suffer fatigue damage under tensile stress. Therefore, 7000 series alloys are used for the upper wing skins or fuselage girder materials which suffer only compressive stress and 2000 series alloys are currently used for the lower wing skins [11]. Although Payne et al. [9] expressed that the fatigue cracks of 7075 aluminum alloys emanate from cracked Fe-rich inclusions, the detailed origin of the inferiority of 7000 series alloys regarding the fatigue process under tensile stressing has not been clarified yet.

A huge amount of fundamental research on the fatigue phenomena of the face-centered-cubic metals to which aluminum alloys belong have been carried out [16–52]. Forsyth [18, 19], who had revealed the appearance of the “extrusion” during fatigue process for the first time [16], classified the fatigue process into the two stages based on a study of various aluminum alloy fracture surfaces. According to the classification of Forsyth [18, 19], “Stage I” is the result of dislocation movement and fracture occurs along the slip plane and “Stage II” is consisting of the cleavage crack and the following striation stage. Xue et al. [8] and Shiraiwa et al. [15] classified the entire stage of the fatigue process as “fatigue crack initiation”, “small crack growth” and “long crack growth” under the negative stress ratio, i.e., R = minimum stress/maximum stress < 0.

Investigation of the fatigue crack initiation is the most important and difficult task among the fatigue related research. Although many authors, e.g., Kinnunen et al. [54], analyzed the fatigue crack growth process numerically by the Paris-Erdogan law [53], the crack initiation must be discussed based on the precise microstructure observations. Shiraiwa et al. [15] showed a prediction of the crack initiation life of 7075-alloys based on the criterion of Tanaka and Mura [55] and indicated a good agreement between their prediction and a data of Tokaji et al. [1] However, actual microstructure of the crack initiation site was not indicated in the literature of Shiraiwa et al. [15] Tokaji et al. [1] conducted fatigue tests for 7075 aluminum alloys by changing the stress ratio, i.e., R = minimum stress/maximum stress, as R = −2, −1 and 0 and showed the different initial crack paths between R = −2, −1 and 0. Although the crack paths under R = −2 and −1 were the 45° facet against the loading direction, the crack paths of R = 0 were initiated from inclusions and straitly extended into the specimen without the 45° facet. The result of Tokaji et al. [1] suggests that the fatigue crack initiation under the positive stress ratio is not driven by dislocation movement. Forsyth [19] suggested that the slip band extrusions are encouraged by a compressive mean stress but absent under conditions of repeated tensile stress. Tokuno et al. [56] observed the well grown extrusions for the pure aluminum polycrystal of which plastic deformation proceeds with the dislocation channeling [57–61] under the negative stress ratio, i.e., R = −1.

Forsyth [18] also emphasized that specimen surface is supposed to be the primary initiation site of the fatigue crack. Chen et al. [14] investigated fatigue process of extruded 7N01 aluminum alloys by focusing on their surface conditions. They controlled surface treatment as, original surface, original surface + electropolishing and original surface + sandblasting, and revealed the fatigue life improvement for the sandblasted specimens due to an induced compressive stress.

Besides, a few research on the influences of the pitting corrosion on the fatigue strength of 7000 series aluminum alloys have been conducted [3, 4, 6, 7]. Pitting corrosion is characterized by the formation of small pits on the surface of the aluminum alloys. At the surface of aluminum alloys, pitting corrosion starts from the defects on their surface, such as grain boundaries or constituent particles. However, although the pitting corrosion has been recognized as the primary initiation site of the fatigue cracking, detailed research regarding the entire crack propagation process leading to the final failure has not been clarified.

By overviewing the several eventful research on the fatigue process of 7000 series aluminum alloys, we have extracted the following issues as should-be-solved ones.

1. 1)    Identification of the preferential microstructure of the fatigue crack initiation at the surface.
2. 2)    Classification of the entire fatigue process by focusing on the crack initiation and propagation under the positive stress ratio, i.e., R > 0.

For the present study, we have investigated the total fatigue process of Al-Zn-Mg alloys, i.e., 7075 and 7150 series alloys, in terms of the identification of the crack initiation site and the classification of the total fatigue stages under the positive stress ratio. In order to confirm the generality of the experimental results, two types of 7000 series alloys were selected and their commonality was explored.

2. Experimental Procedure

Table 1 shows chemical composition of the alloying elements of the Al-Zn-Mg alloys, i.e., 7075 and 7150 series alloys, for the present investigation. 145 mm diameter ingots of those materials were hot extruded to 2.4 mm thick plates. The extruded plates were solution-heat-treated, i.e., 470°C × 30 min followed by water-quenching, and aged by three steps, i.e., 120°C × 24 hr + 200°C × 0.5 hr + 120°C × 24 hr. After the heat-treatment, those plates were machined to the shapes as shown in Fig. 1. LD means the loading direction.

Table 1 Chemical composition of the alloying elements of the specimens used for the present investigation. (online color)

Fig. 1

Plate-shaped specimen machined from heat-treated materials. LD means the loading direction.

We slightly polished the specimen surfaces by water-resistant abrasive papers followed by buff polishing with 0.05 µm particles. For checking the surface condition, one of the specimens was chemically etched with solutions of HF: 0.5 cm³, HCl: 1.5 cm³, HNO₃: 2.5 cm³, H₂O: 95.5 cm³. Figure 2 shows an optical micrograph of the etched specimen’s cross section. About 50 µm thick recrystallized layer remained even after the surface polishing. This recrystallized layer is considered to be created due to rearrangement of dislocations generated through the hot extrusion followed by the heat-treatment. Sizes of grains inside the recrystallized layer were widely scattering between 10∼100 µm, and below the recrystallized layer, elongated pancake-shaped extremely fine fiber-structures were formed. Creations of the similar recrystallized surface layers of 7N01 alloys were reported by Chen et al. [14]. Figure 2 also shows the 50 g-weight micro-Vickers hardness of the recrystallized layer, 176, was almost same with that of the fiber-structures, 174, and roughly equivalent to tensile strength of 565 MPa.

Fig. 2

Optical micrograph showing the microstructure of the etched surface of a specimen. About 50 µm thick recrystallized layer remained even after the surface polishing. Below the recrystallized layer, elongated pancake-shaped extremely fine fiber-structures were formed. 50 g-weight micro-Vickers hardnesses of the recrystallized layer and the fiber-structures were 176 and 174, respectively.

The push-and-pull cyclic loadings were performed for the machined specimens toward the final failures in a servo-hydraulic machine in air and at room temperature. The stress values were determined by dividing the loads with the minimum cross section of the specimen, i.e., 8 × 2.4 mm². The loading cycle was a sinewave with 5 or 10 Hz and the stress-ratio (R) was controlled to be 0.25. In the present research, we performed the investigation of the total fatigue process of the plate-shaped Al-Zn-Mg alloy specimens by combining macroscopic high-speed camera, optical-microscope and scanning-electron-microscope visions.

3. Fatigue Crack Initiation

Figure 3 shows the S-N plots with the 5 Hz loading cycles, i.e., the relationship between maximum stress (σ_max) and cycles to failure (N_f), of the 7075-alloy specimens of the present investigation. The dialogue-balloon indicates the specimen fractured at 69,697 loading cycles under the maximum stress of 320 MPa. Figure 4 shows high-speed camera visions of the fatigue fracture sequence of the dialogue-ballooned specimen of Fig. 3. LD means the loading direction. Each frame shows crack appearances at (a) 0, (b) 40,005, (c) 58,095, (d) 65,003, (e) 69,696 and (f) 69,697 loading cycles, respectively. The arrows, A and B, shown in Fig. 4, indicate extremely vague preexisting defects on the specimen surface. Because of the poor resolution of the high-speed camera visions, further details of those defects were not identified here. At 40,005 cycles, i.e., Fig. 4(b), obvious crack growth had not been observed. At 58,095 and 65,003 cycles, i.e., Figs. 4(c) and 4(d), defects A and B, grew to cracks which were perpendicular direction of LD, i.e., Crack 1 and Crack 2, respectively. At 69,696 cycles, which was just one cycle before the final fracture, i.e., Fig. 4(e), the Crack 1 had grown to be about 3.3 mm long crack. And at 69,697 cycle (final cycle), i.e., Fig. 4(f), the Crack 1 had reached the specimen’s both side surfaces and the specimen had been split into two parts. Figure 5(a) shows a scanning-electron-microscope (SEM) image of the fractured surface of the specimen shown in Fig. 4(f). A semicircle shaped “Fatigue Crack Zone” spread inside the enclosed area. Width of the “Fatigue Crack Zone”, about 3.3 mm, almost matches with the length of the crack at just one cycle before the final fracture shown in Fig. 4(e). Therefore, “Final Fracture Zone (Dimple Zone)” following the “Fatigue Crack Zone” was a ductile dimple fracture zone formed by the final one cycle. Figure 5(b) shows an enlarged view of the “Fatigue Crack Zone”. The origin of the crack inside the zone indicated by the arrow can be easily traced back along the tear ridges as shown in Fig. 5(b). Figure 6 shows a further enlarged view showing the arrowed crack initiation site of the specimen. About 45 µm wide “pit” spreading under the surface has been identified as the crack initiation site. This “pit” was supposed to be the defect A shown in Fig. 4 and very similar to the corrosion pits of 7075-alloy reported by several authors, e.g., Pao et al. [3] and Sankaran et al. [4]. Cleavage crack started from the interface between the “pit” and matrix. On the other hand, traces of dislocation related slips have not been observed.

Fig. 3

S-N plots with the 5 Hz loading cycles of the 7075-alloy specimens. The dialogue-balloon indicates the specimen fractured at 69,697 loading cycles under the maximum stress of 320 MPa. (online color)

Fig. 4

High-speed camera visions of the fatigue fracture sequence of the dialogue-ballooned specimen of Fig. 3. LD means the loading direction. Each frame shows crack appearances at (a) 0, (b) 40,005, (c) 58,095, (d) 65,003, (e) 69,696 and (f) 69,697 loading cycles, respectively. (online color)

Fig. 5

(a) SEM image of the fractured surface of the specimen shown in Fig. 4(f). (b) Enlarged view of the “Fatigue Crack Zone” of Fig. 5(a). The origin of the crack inside the zone is indicated by the arrow. (online color)

Fig. 6

An enlarged view showing the arrowed crack initiation site of Fig. 5(b). About 45 µm wide “pit” spreading under the surface has been identified as the crack initiation site. (online color)

Figure 7 shows the S-N plots with the 10 Hz loading cycles of the 7150-alloy specimens of the present investigation. The dialogue-balloon indicates the specimen fractured at 14,386 loading cycles under the maximum stress of 430 MPa. Figure 8(a) shows a high-speed camera vision of a plate surface of the dialogue-ballooned specimen shown in Fig. 7 before applying the load. The arrows, A, B, C and D, indicate preexisting defects on the specimen surface. By an enlarged view, defect A was identified to be three small defects, A-1, A-2 and A-3, as shown in Fig. 8(b). However, because of the poor resolution of the high-speed camera visions, further details of those defects were not identified at this moment. Figure 9 shows high-speed camera visions of the fatigue fracture sequence of the specimen of Fig. 8. LD means the loading direction. Each frame shows crack appearances at (a) 85, (b) 9,035, (c) 13,686, (d) 14,290, (e) 14,385 and (f) 14,386 loading cycles, respectively. At 85 cycles, i.e., Fig. 9(a), obvious crack growth had not been observed. At 9,035, 13,686 and 14,290 cycles, i.e., Figs. 9(b), 9(c) and 9(d), defects A, B, C and D, especially A, grew to the perpendicular direction of LD. At 14,385 cycles, which was just one cycle before the final fracture, i.e., Fig. 9(e), the defect A had grown to be about 2.5 mm wide long crack. And at 14,386 cycle (final cycle), i.e., Fig. 9(f), the crack had reached the specimen’s both side surfaces and the specimen had been split into two parts. Figure 10(a) shows a SEM image of the fractured surface of the specimen shown in Fig. 9(f). A semicircle shaped “Fatigue Crack Zone” spread from the enclosed area. Width of the “Fatigue Crack Zone”, about 2.5 mm, almost matches with the length of the crack at just one cycle before the final fracture shown in Fig. 9(e). Therefore, “Final Fracture Zone (Dimple Zone)” following the “Fatigue Crack Zone” was a ductile dimple fracture zone formed at the final one cycle. Figure 10(b) shows an enlarged view of the enclosed area of Fig. 10(a). The origin of the crack inside the enclosed area can be easily traced back along the tear ridges as shown in Fig. 5 of the 7075-alloy. Figure 11(a) was a SEM image taken by tilting the specimen of Fig. 10 by 30° and shows that the defects A-1, A-2 and A-3 were preexisting “pits” spreading under the surface, i.e., inside the recrystallized layer, and the origin of the fatigue crack. Figure 11(b) shows an enlarged view of the largest defect A-1 which was about 80 µm wide. Figure 11 indicates that those defects were very similar to the “pit” shown in Fig. 6 of the 7075-alloy.

Fig. 7

S-N plots with the 10 Hz loading cycles of the 7150-alloy specimens. The dialogue-balloon indicates the specimen fractured at 14,386 loading cycles under the maximum stress of 430 MPa. (online color)

Fig. 8

(a) High-speed camera vision of a plate surface of the dialogue-ballooned specimen shown in Fig. 7 before applying the load. The arrows, A, B, C and D, indicate preexisting defects on the specimen surface. (b) Enlarged view of the defect A of Fig. 8(a). (online color)

Fig. 9

High-speed camera visions of the fatigue fracture sequence of the specimen of Fig. 8. LD means the loading direction. Each frame shows crack appearances at (a) 85, (b) 9,035, (c) 13,686, (d) 14,290, (e) 14,385 and (f) 14,386 loading cycles, respectively. (online color)

Fig. 10

(a) SEM image of the fractured surface of the specimen shown in Fig. 9(f). (b) Enlarged view of the enclosed area of Fig. 10(a). The origin of the crack inside the enclosed area can be easily traced back along the tear ridges as shown in Fig. 5 of the 7075-alloy. (online color)

Fig. 11

(a) SEM image taken by tilting the specimen of Fig. 10 by 30° and shows that the defects A-1, A-2 and A-3 were the origin of the fatigue crack. (b) Enlarged view of the largest defect A-1 which was about 80 µm wide. (online color)

Figure 12 shows the S-N plots with the 5 Hz loading cycles of the 7075-alloy specimens, i.e., same plots of Fig. 3, and the dialogue-balloon indicates the specimen fractured at 34,272 loading cycles under the maximum stress of 460 MPa. Figure 13(a) shows an optical micrograph of the fractured plate surface of the dialogue-ballooned specimen of Fig. 12. LD means the loading direction. Besides the main fractured part, several tiny cracks were observed in Fig. 13(a). Enlarged SEM images of the arrowed cracks are shown in Figs. 13(b) and 13(c). Cracks developed toward the dotted-arrowed directions had been initiated from the small “pits”. Those “pits” were supposed to be the same kind “pits” observed in Fig. 6 and Fig. 11 but not the main crack initiation sites leading the specimen toward the final failure because of their limited sizes.

Fig. 12

S-N plots with the 5 Hz loading cycles of the 7075-alloy specimens. The dialogue-balloon indicates the specimen fractured at 34,272 loading cycles under the maximum stress of 460 MPa. (online color)

Fig. 13

(a) Optical micrograph of the fractured plate surface of the dialogue-ballooned specimen shown in Fig. 12. LD means the loading direction. Besides the main fractured part, several tiny cracks were observed. (b) Enlarged SEM image of the arrowed crack shown in Fig. 13(a). (c) Another enlarged SEM image of the arrowed crack shown in Fig. 13(a). (online color)

Through the present observations, we have confirmed that one of the preferential fatigue crack initiation sites of the 7000 series alloys are the several tens’ micrometers wide preexisting corrosion “pits”, and very limited number of those “pits” may lead the alloys toward total failures. Although we have not examined the formation process of those preexisting “pits” which become the origin of the crack, those are supposed to be preferentially formed at large grain regions inside the recrystallized layers at the ambient atmosphere after the heat-treatment. Accumulation of the strain due to the cyclic loading could enhance the stress concentration at the interface of the preexisting “pits” and matrices. Under the above circumstances, transgranular fracture traverses inside the specimens. Because the lattice orientations of the crack path were not examined in the present study, further investigations are needed in future studies.

On the other hand, Chen et al. [14] indicated that the fatigue life of sandblasted 7N01-alloy had been improved. By adding the conclusion of Chen et al. [14] into our consideration, we have investigated the influence of surface removal, i.e., removal of the preexisting “pits”, on the fatigue crack initiation.

Figure 14 shows the S-N plots with the 5 Hz loading cycles of the 7075-alloy specimens. Data of the round-points are the same plots of Fig. 3 and Fig. 12. And data of the triangular-points are the plots of the 7075-alloy specimens of which surfaces, i.e., about 50 µm thick recrystallized layers, had been removed by the water-resistant abrasive papers and buff polishing before the loading which has been conducted immediately after the surface polishing. The cycles to failure of the surface-layer-removed specimens were much longer than the specimens before the surface removal. For example, under the maximum stress of around 400 MPa, cycles to failure of the surface-layer-removed specimens have been increased by a factor of 10. Figure 15(a) shows a SEM image of the fractured surface of the dialogue-ballooned specimen shown in Fig. 14. Semi-elliptic shaped “Fatigue Crack Zone” spread from the arrowed area and “Final Fracture Zone (Dimple Zone)” was following the “Fatigue Crack Zone”. Figure 15(b) shows an enlarged view of the arrowed area of Fig. 15(a). Tear ridges were extending from the enclosed area. Figure 16(a) shows an enlarged view of the enclosed area of Fig. 15(b) and has revealed that the cleavage crack initiation site of this specimen was the constituent inclusion highlighted by the dotted line. By an energy dispersive spectroscopy (EDS) analysis shown in Figs. 16(b) and 16(c), this inclusion was identified as a metallic compound consisting of Mg and Si. As Payne et al. indicated in their literature [9], this compound was supposed to be Mg₂Si. Figure 17 shows EDS analysis of the crack initiation site of the other surface-layer-removed specimen fractured at 41,417 loading cycles under the maximum stress of 440 MPa. Figure 17(a)∼(c) shows the EDS analysis of the fractured surface and Fig. 17(d)∼(f) shows the other side of the fractured surface. The cleavage crack initiation site of the specimen was the arrowed metallic compound consisting of Mg and Si. And, because this compound was observed on both separate surfaces, this compound had been cracked into two parts and became the crack initiation site. Payne et al. [9] expressed that the soft Mg₂Si was not effective crack initiation site but only hard cracked-inclusion such as Fe-rich compound can become the crack initiation site. However, we have confirmed that the Mg₂Si also becomes the initiation site as the Fe-rich compound does.

Fig. 14

S-N plots with the 5 Hz loading cycles of the 7075-alloy specimens. Data of the round-points are the same plots of Fig. 3 and Fig. 12. Data of the triangular-points are the plots of the 7075-alloy specimens of which surfaces, i.e., about 50 µm thick recrystallized layers, had been removed. The dialogue-balloon indicates the specimen fractured at 139,164 loading cycles under the maximum stress of 440 MPa. (online color)

Fig. 15

(a) SEM image of the fractured surface of the dialogue-ballooned surface-layer-removed 7075-alloy specimen shown in Fig. 14. (b) Enlarged view of the arrowed area of Fig. 15(a). Tear ridges were extending from the enclosed area. (online color)

Fig. 16

(a) Enlarged view of the enclosed area of Fig. 15(b). The cleavage crack started from the constituent inclusion. EDS analysis results of this inclusion are shown in (b) and (c). (b) EDS analysis showing Mg. (c) EDS analysis showing Si. (online color)

Fig. 17

(a)∼(c) EDS analysis of the fractured surface of the surface-layer-removed 7075-alloy specimen fractured at 41,417 loading cycles under the maximum stress of 440 MPa. (d)∼(e) EDS analysis of the other side of the fractured surface of (a)∼(c). (online color)

Figure 18 shows the same S-N plots of Fig. 14, and the dialogue-balloon indicates the specimen fractured at 94,772 loading cycles under the maximum stress of 430 MPa. Fractography of the dialogue-ballooned specimen shown in Fig. 18 is shown in Fig. 19. Figure 19(a) shows a macroscopic view of the fractured specimen of which facet surface was slanted along A to B. LD means the loading direction. The slant line AB is inclined from the loading direction LD by about 20°. Figure 19(b) shows a SEM image of the fractured surface of the specimen. The crack initiation site was not located at the wide surface but at the arrowed point of the side surface A. The “Fatigue Crack Zone” was spreading inside the specimen and the “Final Fracture Zone (Dimple Zone)” followed the crack. However, as shown in Fig. 19(c), a “pit” or a metallic inclusion was not observed at the initiation site. Instead, a mechanical scratch which was supposed to be induced during the machining process of the specimen. Because we did not take special care of the side surfaces during the machining, scratches could remain. And the initiation site of the other specimen fractured at 23,056 loading cycles under the maximum stress of 475 MPa was also found to be its side surface. When a “pit” or a metallic inclusion was not located at the wide surface, those scratches of the side surfaces could become one of the preferential crack initiation sites.

Fig. 18

The same S-N plots of Fig. 14, and the dialogue-balloon indicates the 7075-alloy specimen fractured at 94,772 loading cycles under the maximum stress of 430 MPa. (online color)

Fig. 19

(a) Macroscopic view of the fractured dialogue-ballooned specimen shown in Fig. 18 of which facet surface was slanted along A to B. LD means the loading direction. (b) SEM image of the fractured surface A of the specimen of Fig. 19(a). (c) Enlarged view of the arrowed initiation site of Fig. 19(b). (online color)

Sankaran et al. [4] conducted fatigue tests of the 7075-T6 sheet specimens with 15 Hz loading cycles under the positive stress ratio, i.e., R = 0.02. In order to investigate the influence of the pitting corrosion, they also tested the specimens after the dilute electrolyte cycle fog-dry spray treatment and revealed that the created pitting corrosion decreased fatigue lives by a factor of 6 to 8. Although the heat-treatment of the 7075-alloy are different between the investigation of Sankaran et al. [4] and ours, influence of the surface conditions on the fatigue life has been well-matched.

4. Entire Fatigue Process

As mentioned in the Introduction section, Forsyth [18, 19] classified the fatigue process as “Stage I” which is the result of dislocation movement and fracture occurs along the slip plane and “Stage II” which is consisting of the cleavage crack and the following striation stage. Xue et al. [8] and Shiraiwa et al. [15] classified the entire stage of the fatigue process as “fatigue crack initiation”, “small crack growth” and “long crack growth”. However, the classification of Forsyth [18, 19] was based on the dislocation movement accompanying the “extrusion” which would be absent under the repeated tensile stress. And the classification of Xue et al. [8] and Shiraiwa et al. [15] were suggested under the negative stress ratio. On the other hand, we have investigated the entire fatigue process of the 7000-alloys under the positive stress ratio, i.e., R > 0.

Figure 20 shows a brief schematic showing the fatigue crack propagation process inside the semicircle shaped “Fatigue Crack Zone”. We assume that cleavage crack starts from the interface between the several tens’ micrometers wide preexisting “pit” and matrix, and striation appears after its incubation period. We call the zone where the striations are not observed the “Striation Incubation Zone”. Striations appear at x₀, i.e., edge of the “Striation Incubation Zone”, propagate into the specimen, and eventually form the several millimeters wide “Fatigue Crack Zone”. The “Fatigue Crack Zone” terminates at x_t and the “Final Fracture Zone (Dimple Zone)” follows the crack propagation. Through the SEM observation, we tracked the striation propagation and measured the widths of the striations, y, at several discrete points, x. As well known, widths of the striations increase with their propagation.

Fig. 20

A brief schematic showing the fatigue crack propagation process inside the semicircle shaped “Fatigue Crack Zone”. We assume that cleavage crack starts from the interface between the preexisting “pit” and matrix, and striation appears after its incubation period. Striations propagate into the specimen, and eventually the “Fatigue Crack Zone” is formed.

Figure 21 shows the S-N plots with the 5 Hz loading cycles of the 7075-alloy specimens, i.e., same plots of Fig. 3 and Fig. 12, and the dialogue-balloon indicates the specimen fractured at 28,146 loading cycles under the maximum stress of 440 MPa. Figure 22(a) shows the SEM image of the fractured surface of the specimen. A semicircle shaped “Fatigue Crack Zone” spread inside the enclosed area. The “Final Fracture Zone (Dimple Zone)” follows the “Fatigue Crack Zone”. Figure 22(b) shows an enlarged view of the arrowed crack initiation site. The origin of the crack inside the zone, i.e., preexisting “pit”, indicated by the arrow can be easily traced back along the tear ridges. Figure 23(a)∼(c) shows the SEM images of the striations observed at roughly (a) x = 200 µm, (b) x = 600 µm and (c) x = 1 mm inside the “Fatigue Crack Zone” of the specimen of Fig. 21. Bald arrows show propagating directions of the striations. Widths of those striations were (a) y = 63 nm, (b) y = 256 nm and (c) y = 434 nm, respectively.

Fig. 21

S-N plots with the 5 Hz loading cycles of the 7075-alloy specimens. The dialogue-balloon indicates the specimen fractured at 28,146 loading cycles under the maximum stress of 440 MPa. (online color)

Fig. 22

(a) SEM image of the fractured surface of the dialogue-ballooned specimen of Fig. 21. (b) Enlarged view of the crack initiation site. (online color)

Fig. 23

SEM images of the striations observed at roughly (a) x = 200 µm, (b) x = 600 µm and (c) x = 1 mm inside the “Fatigue Crack Zone” of the specimen of Fig. 21. Bald arrows show propagating directions of the striations.

Figure 24 shows the relationship between the distance from the “pit”, x, and the measured striation width, y, of the dialogue-ballooned specimen shown in Fig. 21. Relationship between x and y can be described as an exponential approximation function as below.

  

\begin{equation} \boldsymbol{y} \cong \boldsymbol{a}\boldsymbol{e}^{\boldsymbol{b}\boldsymbol{x}} \end{equation} (1)

a and b in eq. (1) are fitting constants. In Fig. 24, solid line is the best-fitted curve of the plotted data. On the other hand, because the striation width is known to be equivalent with the crack propagation rate during the striation stage, y can be also described as below.

  

\begin{equation} \boldsymbol{y} = \boldsymbol{d}\boldsymbol{x}/\boldsymbol{d}\boldsymbol{N} \end{equation} (2)

Here, N is the number of the loading cycle. Therefore, total loading cycle for the striation formation inside the “Fatigue Crack Zone”, N_s, can be calculated by the following equation.

  

\begin{equation} \boldsymbol{N}_{\boldsymbol{s}} = \int_{\boldsymbol{N}_{\mathbf{0}}}^{\boldsymbol{N}_{\boldsymbol{t}}} \boldsymbol{d}\boldsymbol{N} = \int_{\boldsymbol{x}_{\mathbf{0}}}^{\boldsymbol{x}_{\boldsymbol{t}}} \frac{\mathbf{1}}{\boldsymbol{y}}\boldsymbol{d}\boldsymbol{x} = \int_{\boldsymbol{x}_{\mathbf{0}}}^{\boldsymbol{x}_{\boldsymbol{t}}} \frac{\mathbf{1}}{\boldsymbol{a}\boldsymbol{e}^{\boldsymbol{b}\boldsymbol{x}}}\boldsymbol{d}\boldsymbol{x} = \frac{- \boldsymbol{e}^{-\boldsymbol{b}\boldsymbol{x}}}{\boldsymbol{a}\boldsymbol{b}} \bigg|_{\boldsymbol{x}_{\mathbf{0}}}^{\boldsymbol{x}_{\boldsymbol{t}}} \end{equation} (3)

Here, N₀, N_t, x₀ and x_t are the cycle before the striation formation starts, the cycle at which striation formation terminates, the distance between the “pit” and the starting point of the striation, and the distance between the “pit” and the terminal point of the striation, respectively. We have obtained x₀ = 100 µm and x_t = 1,319 µm for the specimen of Fig. 24 through the SEM observation over the “Fatigue Crack Zone”. By the integration of eq. (3) with the above values, the striation formation cycle, N_s, is obtained. Therefore, the cycle before the striation formation starts, N₀, is estimated as follows.

  

\begin{equation} \boldsymbol{N}_{\mathbf{0}} = \boldsymbol{N}_{\boldsymbol{f}} - \boldsymbol{N}_{\boldsymbol{s}} - \boldsymbol{N}_{\boldsymbol{d}} \end{equation} (4)

Here, N_f is the cycle to failure and N_d is the cycle for the final fracture stage. From the high-speed camera visions of the fatigue fracture sequence of Fig. 4 and Fig. 9, we can assume N_d is equal to one. By applying eq. (3) and eq. (4) on other four specimens of which fatigue cracks also started from the interface between the preexisting “pits” and matrices, we obtained each stage cycle of the 7075-alloy specimens of the present investigation as described in Table 2. The cycles before the striation formation starts, N₀, were much longer than N_s + N_d. Especially, when the cycle to failure, N_f, was six digits number, N₀ became around 90% of the cycle to failure.

Fig. 24

Relationship between the distance from the “pit”, x, and the measured striation width, y, of the dialogue-ballooned specimen shown in Fig. 21. Solid line is the best-fitted curve of the plotted data. (online color)

Table 2 Estimation results of the fatigue crack growth obtained by applying eq. (3) and eq. (4) to the data of the five specimens shown in Fig. 24. The cycles before the striation formation, N₀, were much longer than N_s + N_d. When the cycle to failure, N_f, was six digits number, N₀ became around 90% of the cycle to failure. (online color)

Figure 25 shows the S-N plots with the 10 Hz loading cycles of the 7150-alloy specimens, i.e., same plots of Fig. 7, and the dialogue-balloon indicates the specimen fractured at 210,066 loading cycles under the maximum stress of 275 MPa. Figure 26(a) shows the SEM image of the fractured surface of the specimen. A semicircle shaped “Fatigue Crack Zone” spread under the surface and the “Final Fracture Zone (Dimple Zone)” was following the “Fatigue Crack Zone”. Figure 26(b) is an enlarged view of the arrowed fatigue crack initiation site, i.e., preexisting “pit” observed by tilting the specimen by 30°. The cleavage crack started from the interface between this “pit” and matrix. Width of the “pit” was about 43 µm. Figure 26(c) shows the SEM image of the striations observed at x = 1.3 mm inside the “Fatigue Crack Zone”. Bald arrow shows propagating direction of the striation. Widths of the striation was 510 nm. Figure 27 shows the relationship between the distance from the “pit”, x, and the measured striation width, y, of the specimen shown in Fig. 25. We obtained x₀ = 245 µm and x_t = 2,649 µm for the specimen through the SEM observation. By conducting the same procedure of Fig. 24, we estimated the cycle at which striation formation started, N₀, as 191,844 which was 93.2% of the cycle to failure. This result matches that of the 7075-alloys.

Fig. 25

S-N plots with the 10 Hz loading cycles of the 7150-alloy specimens. The dialogue-balloon indicates the specimen fractured at 210,066 loading cycles under the maximum stress of 275 MPa. (online color)

Fig. 26

(a) SEM image of the fractured surface of the dialogue-ballooned specimen shown in Fig. 25. (b) Enlarged view of the arrowed crack initiation site by tilting the specimen by 30°. (c) SEM image of the striations observed at x = 1.3 mm. (online color)

Fig. 27

Relationship between the distance from the “pit”, x, and the measured striation width, y, of the specimen shown in Fig. 25. Solid line is the best-fitted curve of the plotted data. (online color)

Through the present investigations, we have confirmed that majority of fatigue life is occupied by the cycle before the striation formation starts, N₀. However, we have not distinguished the period of the cleavage crack initiation and the period after the cleavage crack initiation inside the “Striation Incubation Zone”. And the entire fatigue process of the surface-layer-removed specimens has not been investigated either. Further precise and careful investigation should be performed in future.

5. Conclusion

Fatigue fracture process of extruded and heat-treated Al-Zn-Mg alloys, i.e., 7075- and 7150-alloys, under uniaxial cyclic loading with positive stress ratio has been investigated. The push-and-pull cyclic loadings with the stress ratio of 0.25 were performed in air and at room temperature for the plate-shaped specimens and the following results have been obtained.

1. (1)    The several tens’ micrometers wide preexisting “pits” on the surfaces have been confirmed as the preferential fatigue crack initiation sites of the original 7000 series alloys. And, after the secondary surface treatment, i.e., surface layer removal, metallic compounds or scratches induced during machining have become possible fatigue crack initiation sites. Cycles to failure of the surface-layer-removed specimens were much longer than the specimens before the surface removal.
2. (2)    We have classified the entire fatigue process of the original extruded and heat-treated 7000 series alloys as follows.
   • Crack Initiation Stage: Cleavage crack starts from the interface between several tens’ micrometers wide preexisting “pit” and matrix.
   • Striation Incubation Stage: Cleavage crack without remarkable striation propagates into the specimen.
   • Striation Stage: Striations propagate and create the several millimeters wide semicircle “Fatigue Crack Zone”.
   • Final Fracture Stage: Ductile dimple fracture stage with the final one cycle follows the “Fatigue Crack Zone” and leads the specimen to failure.
3. (3)    We have quantitatively estimated the period of each stage and revealed that the lengthiest stage is the combined stage of the crack initiation and the striation incubation. When the cycle to failure reaches six digits number, the period of the combined stage becomes around 90% of the cycle to failure.

Acknowledgements

The authors sincerely express their appreciation to UACJ Corporation for providing the high-strength aluminum alloys for the present study.

REFERENCES

• 1)   K.  Tokaji,  T.  Ogawa and  Y.  Kameyama: The effect of stress ratio on growth behaviour of small fatigue cracks in an aluminum alloy 7075-T6 (With special interest in stage I crack growth), J. Soc. Mater. Sci. Jpn. 38 (1989) 1019–1025. doi:10.2472/jsms.38.1019
• 2)   M.A.  Przystupa,  R.J.  Bucci,  P.E.  Magnusen and  A.J.  Hinkle: Microstructure based fatigue life predictions for thick plate 7050-T7451 airframe alloys, Int. J. Fatigue 19 (1997) 285–288. doi:10.1016/S0142-1123(97)00026-1
• 3)   P.S.  Pao,  S.J.  Gill and  C.R.  Feng: On fatigue crack initiation from corrosion pits in 7075-T7351 aluminum alloy, Scr. Mater. 43 (2000) 391–396. doi:10.1016/S1359-6462(00)00434-6
• 4)   K.K.  Sankaran,  R.  Perez and  K.V.  Jata: Effects of pitting corrosion on the fatigue behavior of aluminum alloy 7075-T6: modeling and experimental studies, Mater. Sci. Eng. A 297 (2001) 223–229. doi:10.1016/S0921-5093(00)01216-8
• 5)   B.B.  Verma,  J.D.  Atkinson and  M.  Kumar: Study of fatigue behaviour of 7475 aluminium alloy, Bull. Mater. Sci. 24 (2001) 231–236. doi:10.1007/BF02710107
• 6)   Q.Y.  Wang,  N.  Kawagoshi and  Q.  Chen: Effect of pitting corrosion on very high cycle fatigue behavior, Scr. Mater. 49 (2003) 711–716. doi:10.1016/S1359-6462(03)00365-8
• 7)   K.  Genel: The effect of pitting on the bending fatigue performance of high-strength aluminum alloy, Scr. Mater. 57 (2007) 297–300. doi:10.1016/j.scriptamat.2007.04.045
• 8)   Y.  Xue,  H.  El Kadiri,  M.F.  Horstemeyer,  J.B.  Jordon and  H.  Weiland: Micromechanisms of multistage fatigue crack growth in a high-strength aluminum alloy, Acta Mater. 55 (2007) 1975–1984. doi:10.1016/j.actamat.2006.11.009
• 9)   J.  Payne,  G.  Welsh,  R.J.  Christ, Jr.,  J.  Nardiello and  J.M.  Papazian: Observations of fatigue crack initiation in 7075-T651, Int. J. Fatigue 32 (2010) 247–255. doi:10.1016/j.ijfatigue.2009.06.003
• 10)   L.  Li,  L.  Shen and  G.  Proust: Fatigue crack initiation life prediction for aluminium alloy 7075 using crystal plasticity finite element simulations, Mech. Mater. 81 (2015) 84–93. doi:10.1016/j.mechmat.2014.11.004
• 11)   R.  Yamada,  G.  Itoh,  A.  Kurumada and  M.  Nakai: Fatigue Crack Growth Behavior and Susceptibility to Hydrogen Embrittlement in 2000 and 7000 Series Aluminum Alloys, J. Japan Inst. Met. Mater. 80 (2016) 745–752. doi:10.2320/jinstmet.JD201606
• 12)   A.  Qian,  X.  Tan,  D.  Wang and  M.  You: Study on Initiation and Growth of Fatigue Crack of Pre-corroded 7B04 Aluminium Alloy Specimen, IOP Conf. Ser.: Mater. Sci. Eng. 493 (2019) 012022. doi:10.1088/1757-899X/493/1/012022
• 13)   Y.  Li,  D.  Retraint,  H.  Xue,  T.  Gao and  Z.  Sun: Fatigue properties and cracking mechanisms of a 7075 aluminum alloy under axial and torsional loadings, Procedia Struct. Integr. 19 (2019) 637–644. doi:10.1016/j.prostr.2019.12.069
• 14)   Z.  Chen,  Z.  Cai,  P.  Deng,  X.  Jiang,  S.  Chen and  H.  Sun: Effect of different surface conditions on fatigue properties of 7N01 aluminum alloy and the behavioral mechanism of crack of the alloy under alternating load, Mater. Res. Express 7 (2020) 026558. doi:10.1088/2053-1591/ab756b
• 15)   T.  Shiraiwa,  F.  Briffod and  M.  Enoki: Prediction of Fatigue Crack Initiation of 7075 Aluminum Alloy by Crystal Plasticity Simulation, Materials 16 (2023) 1595. doi:10.3390/ma16041595
• 16)   P.J.E.  Forsyth: Exudation of Material from Slip Bands at the Surface of Fatigued Crystals of an Aluminium–Copper Alloy, Nature 171 (1953) 172–173. doi:10.1038/171172a0
• 17)   N.  Thompson,  N.J.  Wadsworth and  N.  Louat: Xi. The origin of fatigue fracture in copper, Philos. Mag. 1 (1956) 113–126. doi:10.1080/14786435608238086
• 18)  P.J.E. Forsyth: Proceedings of the Crack Propagation Symposium, Cranfield, (1961) pp. 76–94.
• 19)   P.J.E.  Forsyth: Fatigue damage and crack growth in aluminium alloys, Acta Metall. 11 (1963) 703–715. doi:10.1016/0001-6160(63)90008-7
• 20)   E.E.  Laufer and  W.N.  Roberts: Dislocation structures in fatigued copper single crystals, Philos. Mag. 10 (1964) 883–885. doi:10.1080/14786436408225391
• 21)   E.E.  Laufer and  W.N.  Roberts: Dislocations and persistent slip bands in fatigued copper, Philos. Mag. 14 (1966) 65–78. doi:10.1080/14786436608218989
• 22)   C.  Laird: The Influence of Metallurgical Structure on the Mechanisms of Fatigue Crack Propagation, ASTM Spec. Tech. Publ. 415 (1967) 131–180. doi:10.1520/STP47230S
• 23)   S.J.  Basinski,  Z.Z.  Basinski and  A.  Howie: Early stages of fatigue in copper single crystals, Philos. Mag. 19 (1969) 899–924. doi:10.1080/14786436908225856
• 24)   J.M.  Finney and  C.  Laird: Strain localization in cyclic deformation of copper single crystals, Philos. Mag. 31 (1975) 339–366. doi:10.1080/14786437508228937
• 25)   J.G.  Antonopoulos,  L.M.  Brown and  A.T.  Winter: Vacancy dipoles in fatigued copper, Philos. Mag. 34 (1976) 549–563. doi:10.1080/14786437608223793
• 26)   U.  Essmann,  U.  Gösele and  H.  Mughrabi: A model of extrusions and intrusions in fatigued metals I. Point-defect production and the growth of extrusions, Philos. Mag. A 44 (1981) 405–426. doi:10.1080/01418618108239541
• 27)   J.C.  Figueroa and  C.  Laird: Crack initiation mechanisms in copper polycrystals cycled under constant strain amplitudes and in step tests, Mater. Sci. Eng. 60 (1983) 45–58. doi:10.1016/0025-5416(83)90076-9
• 28)   J.  Polák: On the role of point defects in fatigue crack initiation, Mater. Sci. Eng. 92 (1987) 71–80. doi:10.1016/0025-5416(87)90157-1
• 29)   Z.S.  Basinski and  S.J.  Basinski: Fundamental aspects of low amplitude cyclic deformation in face-centred cubic crystals, Prog. Mater. Sci. 36 (1992) 89–148. doi:10.1016/0079-6425(92)90006-S
• 30)   T.  Zhai,  J.W.  Martin and  G.A.D.  Briggs: Fatigue damage in aluminum single crystals—I. On the surface containing the slip burgers vector, Acta Metall. Mater. 43 (1995) 3813–3825. doi:10.1016/0956-7151(95)90165-5
• 31)   H.-J.  Burmeister and  R.  Richter: Investigations on the origin of grain boundary cracks in fatigued f.c.c. metals, Acta Mater. 45 (1997) 709–714. doi:10.1016/S1359-6454(96)00173-5
• 32)   T.  Zhai,  G.A.D.  Briggs and  J.W.  Martin: Acoustic microscopy of room-temperature fatigue damage in aluminium single crystals, Philos. Mag. 77 (1998) 957–980. doi:10.1080/01418619808221222
• 33)   P.J.  Golden,  A.F.  Grandt, Jr. and  G.H.  Bay: A comparison of fatigue crack formation at holes in 2024-T3 and 2524-T3 aluminum alloy specimens, Int. J. Fatigue 21 (1999) S211–S219. doi:10.1016/S0142-1123(99)00073-0
• 34)   T.S.  Srivatsan,  S.  Anand,  S.  Sriram and  V.K.  Vasudevan: The high-cycle fatigue and fracture behavior of aluminum alloy 7055, Mater. Sci. Eng. A 281 (2000) 292–304. doi:10.1016/S0921-5093(99)00716-9
• 35)   Z.F.  Zhang and  Z.G.  Wang: Dependence of intergranular fatigue cracking on the interactions of persistent slip bands with grain boundaries, Acta Mater. 51 (2003) 347–364. doi:10.1016/S1359-6454(02)00399-3
• 36)   A.J.  McEvily,  M.  Endo,  S.  Cho,  J.  Kasivitamunuay and  H.  Matsunaga: Fatigue Striations and Fissures in 2024-T3 Aluminum Alloy, Mater. Sci. Forum 567–568 (2007) 397–400. doi:10.4028/www.scientific.net/MSF.567-568.397
• 37)   S.  Kobayashi,  T.  Inomata,  H.  Kobayashi,  S.  Tsurekawa and  T.  Watanabe: Effects of grain boundary- and triple junction-character on intergranular fatigue crack nucleation in polycrystalline aluminum, J. Mater. Sci. 43 (2008) 3792–3799. doi:10.1007/s10853-007-2236-z
• 38)   Z.F.  Zhang and  Z.G.  Wang: Grain boundary effects on cyclic deformation and fatigue damage, Prog. Mater. Sci. 53 (2008) 1025–1099. doi:10.1016/j.pmatsci.2008.06.001
• 39)   J.  Man,  K.  Obrtlík and  J.  Polák: Extrusions and intrusions in fatigued metals. Part 1. State of the art and history, Philos. Mag. 89 (2009) 1295–1336. doi:10.1080/14786430902917616
• 40)   M.D.  Sangid,  T.  Ezaz,  H.  Sehitoglu and  I.M.  Robertson: Energy of slip transmission and nucleation at grain boundaries, Acta Mater. 59 (2011) 283–296. doi:10.1016/j.actamat.2010.09.032
• 41)   P.  Zhang,  S.  Qu,  Q.Q.  Duan,  S.D.  Wu,  S.X.  Li,  Z.G.  Wamg and  Z.F.  Zhang: Low-cycle fatigue-cracking mechanisms in fcc crystalline materials, Philos. Mag. 91 (2011) 229–249. doi:10.1080/14786435.2010.518169
• 42)   A.A.  Shanyavskiy and  L.M.  Burchenkova: Mechanism for fatigue striations as formed under variable negative R-ratio in Al-based structural alloys, Int. J. Fatigue 50 (2013) 47–56. doi:10.1016/j.ijfatigue.2012.04.006
• 43)   M.D.  Sangid: The physics of fatigue crack initiation, Int. J. Fatigue 57 (2013) 58–72. doi:10.1016/j.ijfatigue.2012.10.009
• 44)   L.L.  Li,  P.  Zhang,  Z.J.  Zhang and  Z.B.  Zhang: Effect of crystallographic orientation and grain boundary character on fatigue cracking behaviors of coaxial copper bicrystals, Acta Mater. 61 (2013) 425–438. doi:10.1016/j.actamat.2012.08.045
• 45)   J.  Polák and  J.  Man: Mechanisms of extrusion and intrusion formation in fatigued crystalline materials, Mater. Sci. Eng. 596 (2014) 15–24. doi:10.1016/j.msea.2013.12.005
• 46)   L.L.  Li,  Z.J.  Zhang,  P.  Zhang,  J.B.  Yang and  Z.F.  Zhang: Distinct fatigue cracking modes of grain boundaries with coplanar slip systems, Acta Mater. 120 (2016) 120–129. doi:10.1016/j.actamat.2016.06.032
• 47)   W.  Ye,  P.  Efthymiadis,  C.  Pinna,  A.  Ma,  B.  Shollock and  R.  Dashwood: Experimental and modelling study of fatigue crack initiation in an aluminium beam with a hole under 4-point bending, Int. J. Solids Struct. 138 (2018) 87–96. doi:10.1016/j.ijsolstr.2018.01.001
• 48)   W.  Liu,  Y.  Liu,  H.  Sui,  L.  Chen,  L.  Yu,  X.  Yi and  H.  Duan: Dislocation-grain boundary interaction in metallic materials: Competition between dislocation transmission and dislocation source activation, J. Mechan. Phys. Solids 145 (2020) 104158. doi:10.1016/j.jmps.2020.104158
• 49)   W.  Liu,  L.  Yu,  Y.  Liu,  H.  Sui,  H.  Fan and  H.  Duan: Dislocation pile-up polarization model for mechanical properties of polycrystalline metals based on grain boundary resistance variability, J. Mechan. Phys. Solids 160 (2022) 104793. doi:10.1016/j.jmps.2022.104793
• 50)   A.  Chlupová,  I.  Šulák,  T.  Babinský and  J.  Polák: Intergranular fatigue crack initiation in polycrystalline copper, Mater. Sci. Eng. A 848 (2022) 143357. doi:10.1016/j.msea.2022.143357
• 51)   V.  Mazánová,  M.  Heczko and  J.  Polák: On the mechanism of fatigue crack initiation in high-angle grain boundaries, Int. J. Fatigue 158 (2022) 106721. doi:10.1016/j.ijfatigue.2022.106721
• 52)   W.  Liu,  Y.  Chen,  H.  Sui,  J.  Fu and  H.  Duan: Microstructure-based intergranular fatigue crack nucleation model: Dislocation transmission versus grain boundary cracking, J. Mechan. Phys. Solids 173 (2023) 105233. doi:10.1016/j.jmps.2023.105233
• 53)   P.  Paris and  F.  Erdogan: A Critical Analysis of Crack Propagation Laws, J. Basic Eng. 85 (1963) 528–533. doi:10.1115/1.3656900
• 54)   A.  Kinnunen,  I.V.  Lomakin,  T.  Mäkinen,  K.  Widell,  J.  Koivisto and  M.J.  Alava: Striation lines in intermittent fatigue crack growth in an Al alloy, Phys. Rev. Mater. 7 (2023) 053602. doi:10.1103/PhysRevMaterials.7.053602
• 55)   K.  Tanaka and  T.  Mura: A Dislocation Model for Fatigue Crack Initiation, J. Appl. Mech. 48 (1981) 97–103. doi:10.1115/1.3157599
• 56)   K.  Tokuno,  T.  Yonezawa,  G.  Shigesato,  H.  Tsutamori and  M.  Hagino: Surface Relief Formation of Cyclic-Loaded Coarse-Grained Aluminum Polycrystals with Point Defect Clusters, Mater. Trans. 65 (2024) 502–511. doi:10.2320/matertrans.MT-M2023185
• 57)   R.  Fukuoka,  K.  Tokuno,  M.  Mitsuhara,  K.  Yamakoshi,  S.  Tsuchida,  J.  Miyamoto and  M.  Hagino: Dislocation Structures Formed inside Dislocation Channels of Rapid-Cooled and Tensile-Deformed Aluminum Single Crystals, Mater. Trans. 62 (2021) 221–228. doi:10.2320/matertrans.MT-M2020255
• 58)   S.  Tsuchida,  K.  Tokuno,  M.  Mitsuhara,  R.  Fukuoka,  R.  Tsuboi,  J.  Miyamoto and  M.  Hagino: Dislocation Cell Structures Formed inside Dislocation Channels of Rapid-Cooled and Tensile-Deformed Aluminum Single Crystals, Mater. Trans. 62 (2021) 1109–1117. doi:10.2320/matertrans.MT-M2021021
• 59)   K.  Tokuno,  M.  Mitsuhara,  S.  Tsuchida,  R.  Tsuboi,  J.  Miyamoto,  M.  Hagino,  T.  Inoue and  K.  Nishidate: Tangled Dislocation Structures inside Dislocation Channels of Rapid-Cooled and Tensile-Deformed Aluminum Single Crystals, Mater. Trans. 63 (2022) 562–569. doi:10.2320/matertrans.MT-M2021230
• 60)   K.  Tokuno,  M.  Mitsuhara and  M.  Hagino: Early-Stage Dislocation Structures inside the Dislocation Channels of Face-Centered-Cubic Metals with Point Defect Clusters, J. Japan Inst. Met. Mater. 87 (2023) 67–80. doi:10.2320/jinstmet.J2022030
• 61)   K.  Tokuno,  M.  Mitsuhara and  M.  Hagino: Early-Stage Dislocation Structures inside the Dislocation Channels of Face-Centered-Cubic Metals with Point Defect Clusters, Mater. Trans. 64 (2023) 2027–2038. doi:10.2320/matertrans.MT-M2023032