*3.2. Behavior of Mechanical Properties during PWHT*

Figure 9 presents the outcomes of tests of the Vickers hardness and Charpy impact absorption energy in the stir zones of FSW joints subjected to a post-weld heat treatment. When the FSW was performed, the hardness decreased significantly to 162 HV from 380 HV. This provides evidence that the hardness was decreased in the SZ due to the dissolution of the strengthening precipitates, as shown in Figure 8. When the joint was subjected to PWHT at 315 ◦C, the microhardness increased steeply during the under-aging step (US) of the PWHT process. Furthermore, as PWHT lasted for approximately three hours, the results for the joint showed that the hardness values recovered similarly to the BM outcome. In other words, this was the peak-aging step (PS). When PWHT lasted for more than four hours, the hardness decreased slightly. Thus, this was the over-aging step (OS). The Charpy impact absorption energy increased up to 92.2 J from 4.6 J, as listed in Table 2, as the FSW was conducted. However, it decreased severely during the under-aging step. Subsequently, the decrease lasted for four

hours. However, the Charpy impact absorption energy value increased slightly during the over-aging step. Figure 10 shows the behavior of the tensile properties of the FSW joints after PWHT. All of the tensile test specimens were fractured at the BM. Both the tensile and yield strength increased early during the under-aging step. Subsequently, these increases lasted for four hours of PWHT. After the peak-aging step, these values decreased gradually. On the other hand, the strain decreased abruptly early in the under-aging step. Additionally, it showed little change during the peak-aging and over-aging steps. Hence, PWHT has a decisive effect on the recovery of mechanical properties such as the hardness, yield strength and tensile strength of FSW joints. These results also indicate that microstructural evolution such as precipitation can occur during PWHT.

**Figure 9.** Behavior of the Vickers hardness and Charpy impact absorption energy in the stir zone of FSW joints subjected to a post-weld heat treatment. PWHT: post-weld heat treatment.

**Figure 10.** Behavior of the tensile properties of FSW joints after a post-weld heat treatment.

#### *3.3. Microstructural Evolution during PWHT*

Figure 11 shows SEM images of DP cells formed at the grain boundaries of the stir zone after the samples were subjected to PWHT for 30 min, three hours and eight hours. In the early stage of the under-aging step, very few and relatively small (Figure 11a) DP cells were observed. When the PWHT process was extended to three hours, the number of DP cells increased (Figure 11b). After eight hours, many colonies of DP cells formed (Figure 11c). The DP cells are involved in the formation of a solute-depleted matrix phase (α ) and a precipitate phase (γ) as a duplex transformation product, typically with nucleation at the grain boundaries with growth into one side of the supersaturated matrix (α phase) [22–25]. The DP cells consumed numerous solute atoms and diminished the precipitation hardenability; in this case, the γ phase readily formed instead of the finer metastable γ and γ phases. TEM observations were used to characterize the precipitates which formed during PWHT, and these results are shown in Figure 12. As shown in Figure 8, no precipitates were found at the foil prepared from the FSW joints. However, numerous nano-scale sphere-type precipitates were observed in the foil prepared from the sample which underwent PWHT for 30 min (Figure 12a). Many γ phase precipitates arose depending on the duration of PWHT (Figure 12b). When the PWHT process lasted for eight hours, coarsened γ phases were observed (Figure 12c).

**Figure 11.** SEM images of discontinuous precipitation (DP) cells formed at the grain boundaries of the stir zone subject to PWHT: (**a**) 0.5 hour, (**b**) three hours and (**c**) eight hours.

**Figure 12.** High-resolution transmission electron microscope (HR-TEM) images and selected area diffraction pattern (SADP) outcomes of the post-weld heat treated foils: (**a**) 0.5 hour, (**b**) three hours and (**c**) eight hours.

Figure 13 shows high-magnification CS-TEM bright-field images and corresponding SAD patterns of samples after PWHT for 30 min (a) and three hours (b) with the axis parallel to [001]α. In Figure 13a, the γ precipitates were nucleated and grew to a length of 5~15 nm along the (−100)<sup>α</sup> direction with three to six layers of Be. With a PWHT duration up to three hours, the γ precipitates grew and coarsened to γ in the same direction and more than eight Be layers formed. The images indicate the distance between the atoms with a straight line in two directions, showing results of a = 0.236 nm, b = 0.232 nm (Figure 13a), and a = 0.270 nm, b = 0.279 nm (Figure 13b). These outcomes were fairly close to the phase parameter of γ as reported by Guoliang et al. [21].

**Figure 13.** *Cont.*

**Figure 13.** CS-TEM images and SADP outcomes of precipitates consisting of (**a**) γ and (**b**) γ phases in PWHTed foils.

#### **4. Discussion**

#### *Relationship between the Mechanical Properties and Microstructure during FSW and Consequent PWHT*

In case of laser beam welding butt joint of beryllium-copper alloy, the joint shows liquation crack in HAZ involved in fracture. Also it is possible to get sound joint properties of 0.2 mm of thickness plate after welding; however, there is currently limited research on more than 3 mm of thickness [9].

In the case of solid-state joining process, which use not only thermal factor but secondary factors such as mechanical or chemical factor, the problems caused by solidification from melting state is expected to be suppressed.

Esmati et al. attempted diffusion brazing lap joint of beryllium copper alloy with filler metal of Ag content. The brazing joint was successfully fabricated without defect at 750 ◦C for 1200 s. However, maximum tensile strength was comparatively low value as 173 MPa [10].

Especially, this study was conducted to get a sound thick plate joint that has good joint properties without solidification defects by proposing the properly controlled friction stir welding condition.

Based on preliminary welding test in several friction stir welding conditions, the authors were able to figure out optimal condition to control the heat input properly. When the welding condition accompanied higher heat input, since the microstructure of the joint is expected to exhibit greater dissolution of the γ and γ precipitates, the faster and the stronger hardening effect is expected in the same post-weld heat treatment condition due to the higher driving force. Meanwhile, when the heat input during the welding process was lower, since the less γ and γ precipitates are dissolved, it is assumed to show the slower and the weaker hardening effect due to the less driving force to form precipitates.

The effect of a post-weld heat treatment on the mechanical properties and microstructures of friction-stir-welded beryllium-copper alloy were investigated, as presented in the results section. In this section, the relationship between the mechanical properties and precipitation behavior during FSW and consequent PWHT is discussed. The process flow from BM to PWHT has five steps based on the alternation of the mechanical properties and microstructures. This process is shown in Figure 14. Initially, the base metal has acceptable hardness and tensile strength, as a considerable amount of γ precipitates, which are the primary strengthening mechanism of the present alloy, were present in the BM, as shown in Figure 5b. Interestingly, in the second step just after FSW, the tensile strength and

hardness decreased abruptly, while the toughness and ductility increased sharply. The non-existing presence of strengthening γ precipitates is considered to be the primary cause of the fall in both the hardness and tensile strength. The authors believe that the γ precipitates dissolved into the base metal owing to the frictional heat generated by the rotation of the tool. A significantly refined grain microstructure was formed in the stir zone. Both the refined grains and softening through the dissolution of the γ precipitates had an effect on the upsurge of the toughness. The PWHT period includes steps 3, 4 and 5. During the third step, called the under-aging stage of PWHT within a half hour (PWHT, US), a recovery of the hardness and tensile strength was obtained. In contrast, the toughness and ductility expressed a large decline. This outcome appears to be related mainly to the formation of globular γ precipitates, as shown in Figures 12a and 13a. During the fourth stage of the peak-aging step of PWHT from two hours to four hours (PWHT, PS), the gradual strengthening tendency continues as the aging time is extended. The peak tensile strength and hardness could be determined from microstructures mainly containing γ precipitates which grew from γ precipitates. Finally, during the fifth stage of the over-aging step of PWHT from four hours to eight hours, a slight decrease in hardness and tensile strength occurred. The growth and coarsening of γ precipitates are considered to be the dominant causes of this finding. If PWHT involves longer aging times up to a few days, the low mechanical properties through the softening of the microstructure are suggested to appear. It was clearly revealed that when the Giga-grade high-strength beryllium-copper alloy, which shows mechanical properties mainly due to precipitation of the γ phase is subjected to FSW, the hardness and tensile strength decrease sharply because the γ precipitates dissolve into the base metal due to the frictional heat generated during the FSW process. It was also found that PWHT is indispensable to recover the hardness and tensile strength of FSW joints. However, excess aging times exceeding three to four hours at 315 ◦C bring about a decline in the hardness and tensile strength. In conclusion, it is crucial to control metastable precipitates such as γ and γ to secure reliable FSW joints with beryllium-copper alloys.

**Figure 14.** Schematic diagram of the evolution of the microstructure during the FSW and PWHT processes.
