Fig. 6 — Macrographs of the FSWs in this study with horizontal line indicating location of nanoindentations. plug of metal around the tool. By assuming that 100% of the weld torque was converted JANUARY 2013, VOL. 92 14-s WELDING RESEARCH into thermal energy and contact conditions remain constant, an energy balance was used to equate the heat input (Qg) with the heat loss terms as given in Equation 1. The heat loss terms included conduction, through the weldment (Qw), anvil (Qa), and spindle (Qsp), in addition to convection, which captured the preheating of metal (Qv) passing through the shear surface in advance of the weld. Qg = Qw + Qa + Qsp+Qv (1) The resulting relationship given in Equation 2 was used to determine a FSW temperature from the actual weld torque (Mt) (Refs. 27, 49) where ω was the tool rotation, τ was the flow stress, Rs was the radius of shear surface, R was the radius of the tool pin, and H was the length of the tool pin. (2) ⎛ The flow stress (τ) was approximated by Mt* ΔT such that as the shear zone temperature approached Tm, the flow stress approaches zero (Refs. 27, 49). This linear approximation was based on Fig. 3, which plots the flow stress vs. temperature for AA2219-T87 7 and shows a precipitous drop at around 0.5 Tm reaching a constant, linear plateau at approximately 0.7 Tm (Ref. 50). The value of flow stress at Tm was assumed to be zero. Thus, considering the range of published temperature measurements for FSW Al alloys of 0.8 to 0.9 Tm, the corresponding flow stress was relatively unaffected by temperature and was considered linear just prior to reaching Tm. Results and Discussion AA2219 is an Al-Cu alloy whose nominal composition is slightly above the maximum solid solubility as shown in the equilibrium diagram in Fig. 4 (Ref. 51). This yields a microstructure composed primarily of the saturated α-aluminum matrix plus a small amount of excess θ phase. The T87 temper used in this study refers to a heat treatment that artificially ages the Cu-rich precipitates in the α-matrix through a well-accepted sequence of equilibrium transformation given in Equation 3, where αss refers to a solid solutionized Al matrix. αss→GPI→GPII→θ′→θ (3) The T8 temper refers to a solid-solution heat treatment of the α phase at 535°C, followed by cold work and artificial aging at 175°C for 18 h (Ref. 52). This results in a base metal with the main strengthening metastable phase of θ′phase as shown in Fig. 5. Figure 5A is a low-magnification image of the base material microstructure, which shows a few large Cu-rich particles around 200–500 nm, corresponding to the excess θ phase. Figure 5B is a higher magnification image that shows the θ′strengthening metastable phase with a reported morphology of tetragonal discs that are semicoherent with the α-aluminum matrix (Ref. 53). The FSW process is considered to occur at high strain rates and impart a high strain to the metal surrounding the weld tool (Refs. 2, 8–12). Thus, the kinetics of the dynamic microstructural evolution would be expected to differ from the static equilibrium conditions (Refs. 13, 19). The occurrence of a high strain rate acting on the metal as it moves around the weld tool implies very rapid deformational heating and associated up-quenching followed by slow cooling. The strengthening precipitates in the base metal undergo coarsening during the FSW process and eventually lose their strengthening effectiveness due to elevated temperatures and/or longer times at elevated temperatures. Near the workpiece/weld tool interface, where the rate of heating was the highest due to the high shear strain rates, the Curich phases underwent dissolution. During the rapid up-quenching, if the eutectic temperature was exceeded at the workpiece/ weld tool interface, the remaining θ phase may have liquated (Refs. 13, 54). However, if the temperature remained below the eutectic, an increasing degree of dissolution of the Cu-rich phases was expected as the rate of temperature rise increased at the workpiece/weld tool interface, thereby increasing the solute concentration. At lower strain rate regions away from the shear zone, the Cu-rich phases would have continued to coarsen, depleting the solute from the α matrix (Ref. 19). Estimations of the strain rate associated with FSW have been based on various analytical or numerical models that rely on material property databases (Refs. 8–11) in addition to use of the Zener- Holloman parameter, which relates grain size to strain rate (Ref. 12). These methods have provided estimates in the range of 104 to 101 s–1, respectively, with lower values calculated from the Zener-Holloman method. Studies have indicated that the grain size at Q R R R s H g = R ⎝ ⎜ ⎞ ⎠ ⎟ + ⎡ ⎣ ⎢⎢ ⎤ ⎦ ⎥⎥ ωτ 2π 1 3 3 3 Table 3 — Tensile Strength of the FSW SZ Specimens Specimen Yield Strength Ultimate Tensile Strength (MPa) (MPa) 150 151 ± 2 269 ± 11 200 163 ± 11 295 ±11 300 190 ± 2 332 ± 9 PM 396 469 Table 4 — Bulk Eddy Current Measurements Specimen Eddy Current (% IACS) 150 26.2 ± 0.1 200 26.2 ± 0.7 300 22.7 ± 0.1
Welding Journal | January 2013
To see the actual publication please follow the link above