*3.4. Fatigue Performance*

The results of the fatigue tests are shown in Figure 6. The fitted curves for each specimen were obtained using Stromeyer's expression, log(σ − *a*) = −*b* log *N* + *c*, where σ is the stress amplitude, *N* is the number of cycles to failure, and *a*, *b*, and *c* are the fitting parameters. The fatigue performances of the as-welded specimens with and without reinforcement were worse than that of the BM. Although the fatigue lives of these specimens at a stress amplitude of 180 MPa were almost the same, that of the reinforcement-removed welded specimen was shorter than that of the as-welded specimen at 120 MPa. After DryLP treatment, the fatigue performances of the specimens with and without reinforcement were enhanced to a similar degree. The fatigue life increased by a factor of almost two at a stress amplitude of 180 MPa and more than 50 times at 120 MPa, which indicates that the DryLP treatment is more effective at lower stress amplitudes.

**Figure 6.** Results of fatigue tests for the base material (BM) and as-welded specimens (with and without reinforcement) before and after DryLP treatment.

The fracture surfaces of samples broken at 120 MPa and 180 MPa are shown in Figure 7, where the red arrows indicate crack initiation sites. The crack initiation sites for any specimens, such as as-welded and reinforcement-removed specimens before and after DryLP treatment, are undercuts, not blowholes. The fractures initiated at the boundary between the WM and HAZ for all specimens, regardless of DryLP treatment or the existence of weld reinforcement. Cracks initiated at undercuts are shown in the magnified views of the surfaces in Figure 7g.

**Figure 7.** Fracture surfaces of the welded specimens. (**a**) As-welded specimen broken at 120 MPa, (**b**) reinforcement-removed welded specimen broken at 120 MPa, (**c**) as-welded specimen broken at 180 MPa, (**d**) reinforcement-removed welded specimen broken at 180 MPa, (**e**) DryLPed welded specimen broken at 180 MPa, (**f**) DryLPed reinforcement-removed welded specimen broken at 180 MPa. (**g**) Magnified image of a typical crack initiation site, as indicated by the yellow box in (**c**).

### *3.5. Microstructure in WM*

Bright-field TEM images of the region ~10 μm below the surface in the WM of as-welded and DryLPed specimens (with reinforcement) are shown in Figure 8. The incident electron beam direction was nearly parallel to the [110] direction of Al, where the {111} reflection of Al was excited. The dislocations were observed as the darker areas. The dislocation density was estimated using Keh's equation, ρ = (*n*1/*L*1 + *<sup>n</sup>*2/*L*2)/*t*, where ρ is the dislocation density, *n*1 and *n*2 is the number of intersection points between the dislocation lines and the vertical and horizontal grid lines drawn on the TEM image, respectively, *L*1 and *L*2 is the total length of the vertical and horizontal grid lines, respectively, and *t* is the thickness of the TEM sample [39]. The dislocation densities of these samples before and after DryLP treatment were estimated as 1.0 × 10<sup>14</sup> m<sup>−</sup><sup>2</sup> and 5.1 × 10<sup>14</sup> m<sup>−</sup>2, respectively. This indicates that DryLP plastically deformed the WM, resulting in hardening and inducing compressive residual stress.

**Figure 8.** TEM images of weld material (WM) microstructures in laser-welded specimens (with reinforcement) (**a**) before and (**b**) after DryLP treatment.

### *3.6. E*ff*ect of Welding Defects on Fatigue Performance*

During fatigue tests, cracks initiated from undercuts at the weld toe, owing to the reduced hardness and tensile residual stress which remained after welding. The fatigue performances of the as-welded specimens with and without reinforcement were comparable at a stress amplitude of 180 MPa, indicating that the stress concentration at undercuts has a greater influence on fatigue performance than stress concentration at the weld toe. The fatigue life of the specimen with the reinforcement removed was shorter than that of the as-welded specimen at a stress amplitude of 120 MPa. Small blowholes were observed in the fracture surface of the specimen without reinforcement, which were expected to influence the fatigue performance at lower stress amplitudes. Welding defects, such as undercuts and blowholes, in addition to softening or tensile residual stress, decreased the fatigue life of the welded specimens.

The fatigue performances of the specimens with and without reinforcement after DryLP treatment were improved compared to the equivalent specimens before DryLP treatment, attributed to hardening of the WM up to the value of the original BM and the introduction of compressive residual stress. The fatigue lives of both specimens after DryLP treatment were significantly increased compared to that of the unpeened welded specimens at lower stress amplitudes. Blowhole defects can lead to stress concentration. However, their contribution is very small, as these features are generally spherical. In addition, the stress concentration at undercuts is smaller at lower stress amplitudes. Therefore, the effect of positive factors induced by DryLP, such as hardening and compressive residual stress, was larger than that of the negative factors, such as stress concentration at undercuts and blowholes at lower stress amplitudes. In addition, for gas metal arc welding lap fillet joint in GA 590 MPa steel sheets, the blowholes in the WM did not significantly a ffect the fatigue life at relatively lower stress amplitudes, although the fatigue life decreased in the presence of blowholes and surface pores [40]. Overall, DryLP e ffectively improved the fatigue performance of laser-welded specimens containing welding defects at lower stress amplitudes.

### *3.7. Plastic Deformation Induced by Femtosecond Laser-Driven Shock Wave*

When a peak pressure of a shock wave exceeds a threshold that depends on a material, the pressure increases as a function of the time or the travel distance exhibits a single structure, where the plastic component overtakes the elastic component. The threshold stress for aluminum when the single structure of the shock front is clearly formed is 25 GPa [41]. It was reported that the single structure was observed in the surface layer of 500 nm in pure aluminum, which was irradiated using the intensity of 8.7 × 10<sup>12</sup> <sup>W</sup>/cm<sup>2</sup> with the pulse duration of 150 fs [42]. Therefore, the shock wave with a single structure over 25 GPa should be driven and propagated in the 2024 aluminum alloy, which was irradiated at the intensity of 1.2 × 10<sup>14</sup> <sup>W</sup>/cm<sup>2</sup> with the pulse duration of 130 fs in this research.

It was empirically observed that the strain rate η of the shock wave with the single structure was proportional to the fourth-power of the shock stress σ [43,44]. For the aluminum alloy, η = 9100σ<sup>4</sup> has been reported [41]. Therefore, the strain rate η of 3.5 × 10<sup>9</sup> s<sup>−</sup><sup>1</sup> was obtained for the shock stress of 25 GPa. The dimensionless Bland number B = 3*hs*η/8*c* was defined [43], where *h* is the sample thickness, *c* is the bulk sound velocity under normal pressure, and *s* is the slope of the *<sup>u</sup>*p-*u*s relation, *u*s = *c* + *su*p, where *<sup>u</sup>*p is the particle velocity and *u*s is the shock velocity. When *B* is greater than 1, steady-wave conditions are expected [41]. The thickness *h* was estimated to be 3.0 μm for *B* = 1, η = 3.5 × 10<sup>9</sup> s<sup>−</sup>1, *s* = 1.338, *c* = 5.328 km/s [45]. Therefore, the shock wave with the single structure propagates in the surface layer of 3.0 μm. The single structure splits into two structures, elastic and plastic waves, at the depth of 3.0 μm, and the shock wave with the two-wave structure propagates into the deeper region.

Figure 9 shows the hardness in the BM region in the DryLPed 2024 aluminum alloy as a function of the depth measured using nanoindentation (ELIONIX, ENT-1100a, Japan) with an applied load of 1 mN and loading time of 2 s. The increase in hardness is more significant at a depth of 3 μm from the surface rather than depths of 3–20 μm, although the hardness increased in the surface layer with 20 μm thickness. The thickness of the significantly hardened layer of 3 μm corresponds to the thickness of 3.0 μm where the shock wave with the single structure propagates. This implies that the single structure induces plastic deformation more effectively, thereby increasing the hardness. A high-density-dislocation structure, shown in Figure 8a, was formed in a layer where the shock wave with the single structure propagates, because dislocation generation, rather than dislocation multiplication, was dominant [9,10,46].

**Figure 9.** Depth profile of hardness in the BM region in the DryLPed 2024 aluminum measured using nanoindentation.
