This was done in order to find out the effect of shoulder diameter, pin diameter and pin profile on the defect formation and tensile properties. The frustum shaped pin profile is chosen according to the self-optimised tool geometry suggested by R A Prado and L E Murr. The length of the pin was kept 0.2- 0.3 mm shorter than the plate thickness. The welded joints were produce with constant shoulder penetration 0.1 mm into the base material.
Samples were cut from the welded joints to study micro structural features and mechanical properties. The tensile testing samples were cut perpendicular to welding direction and the samples prepared according to the ASTM E8M sub size standard. The section was prepared by standard metallographic practice and Kellers reagent was used for etching. Then it was subjected to hardness testing and micro structural analysis. Samples for various testing were cut from welds at various positions. Vickers macro hardness testing was done with 3 kg load for 10 seconds. Tensile testing was carried out in tensometer of 2 tonne capacity with 3.2 mm cross head speed.
Results and discussion
Aluminium alloy AFNOR 7020 is widely used for structural application in aerospace industries like rocket liquid propulsion tank. This alloy is a precipitation hardenable Al-Zn-Mg (7XXX) alloy. It gains strength through the following complex precipitation sequence including metastable phases.
Slid solution -GP zone- ?- ? (Mg Zn2) [13,14]
Where GP zone and h are metastable. These precipitates exhibit limited temperature stability (~100 °C for GP zones, ~150 °C for h and ~200 °C for h). Thus their modification during welding leads to the reduction in strength.
Microstructural analysis: The base material microstructure of the 7020 used is shown in Figure 4. It has pancake shaped grains elongated in the rolling direction with size varying from 300 to 500 µm. A low magnification light optical macrograph of a defect free AFNOR 7020 weld cross section is shown in Figure 5 in which various regions; the weld nugget (WN), thermo mechanically affected zone (TMAZ) and heat affected zone (HAZ) can be clearly seen. The difference in the microstructures of these regions can be clearly seen when comparing the high magnification optical micrographs that are shown in Figure 6. The HAZ (Figure 6a) has slightly broadened grains when compared to the base material (Figure 4). This indicates that grain growth takes place in this region.
The grain structure has not changed significantly in the HAZ, which indicates that theres no plastic deformation in HAZ. In the TMAZ (Figure 6b) the grain structure has changed and flow lines are altered by the plastic deformation. Weld nugget (Figure 6c) grain structure, which consists of 5 to 20 µm grains, clearly indicates that a large plastic deformation has taken place in the region. The fine grain structure has been identified as dynamically re-crystallised structure. In this region the grain size, when compared to the base material, are almost two orders less. The amount of plastic deformation and temperature existing in TMAZ is less compared to the amount of plastic deformation and temperature existing in the WN. The strain, strain rate and temperature of plastic deformation, which control the micro structural evolution of the material, in the stirred zone is then expected to match the strain, strain rate and temperature required for dynamic re-crystallisation.
One of the other points to be noted is that the weld nugget is not symmetric about the weld line. Also, the micro structural evolution in the advancing interface (Figure 7a) and retreating side interface (Figure 7b) are different from each other. The micro structural evolution in the advancing side (Figure 7a) shows a much sharper interface and one can thus expect a much shaper strain, strain rate and temperature gradient to exist in this region. In the retreating side (Figure 7b) the sharp interface is not seen. This means that the gradients of strain, strain rate, and temperature would be lower in this region. Flow visualisation studies during friction stir welding have shown that material from the advancing side (closer to the leading edge) moves to the retreating side (closer to the trailing edge) and finally moves to the trailing edge of the advancing side. The material that deforms first would be at a lower temperature as the amount of plastic deformation is lower and sufficient time is not there for heat to get dissipated. This would mean that the gradients of temperatures and the concomitant gradients in strain and strain rates would be higher in the advancing side of the weld.
From Figure 8 the effect of tool diameter on the grain size in the weld nugget can be clearly seen. The micrographs are taken in the same region for both the welds. The weld produced using 15 mm shoulder diameter (Figure 8a) has a grain size in the range of 5 µm and the weld produced using 20 mm shoulder diameter (Figure 8b) has a grain size of around 15 µm. This indicates that the grain is strongly dependant on the shoulder diameter. Increase in shoulder diameter and pin diameter increases the heat input of the process. Due to increase in heat input, it is possible that the static grain growth mechanism takes place after the dynamic re-crystallisation, taking place during the plastic deformation. This type of grain growth has also been observed by Y S Sato et al, on AA 6063 with increase in rotation speed, where the higher rotation speed is expected to increase the heat input to the material.
Effect of tool geometry on defect formation
The different kinds of defects formed in the FSW welds are shown in Figures 9 and 10. Flash formation and improper bonding in this region (figure 9a), indicated by an arrow, is observed in the retreating side of the welds. Bonding takes place properly below the region where the flash forms. Unfilled regions are noticed only in the lower half of the advancing side of the welds (Figure 9b). These unfilled regions are formed closer to the horizontal portion of the pin near the bottom of the weld. The amount of the unfilled region is larger when the weld is done using cylindrical pin with flat end. Figure 10 shows similar defect with frustum shaped rounded end tools. While comparing the unfilled regions and shape of the defect (Figures 9b and Figures 10 a, b, c, d) it can be seen that the size and location are changing when the tool profiles were changed or the shoulder/pin diameter was changed. Within the range of the present experiments, a higher pin diameter (compare Figures 10 a and b) or a higher shoulder diameter (compare Figures 10 a and d) reduces the size of the unfilled region. The effect of shoulder diameter seems to be more significant in the present set of experiments. This is expected as most of the heat generated during the process is by the shoulder. It must be mentioned here that all the results presented till now is with 0o tilt angle.
As mentioned earlier, during the FSW process the rotating tool draws material from advancing side leading edge, flows around the tool and gets deposited near the trailing edge of the advancing side. Providing a backward tilt would help in consolidating the plastically deforming material near the trailing edge of the advancing side of the tool by creating a compressive stress in the retreating side. Thus, a backward inclination would help in reducing the tendency of cavity formation seen when the tile angle is zero. The backward tilt will also compensate for any loss of the material by flash formation. Experiments carried out using a tilt of 20o, with all other conditions remaining same, showed complete filling and almost no defect formation in the weld. The cross-section of the weld shown in figure 2 is from such an experiment with a tilt angle of 20o. It is also observed that the amount of flash formed drastically reduces when the tilt angle is given. Thus, one could say that it is beneficial to have a backward tilt angle during the welding process.
Mechanical properties
The base material hardness of aluminium alloy AFNOR 7020 used for the present experiments was 125 VHN3kg. The hardness profiles across weld cross section at the mid plane are shown in the Figure 11. The minimum hardness occurred around 7 mm and 16 mm away from both interfaces in the weld produced using 10 mm and 20 mm shoulder diameters, respectively. This means that the minimum hardness exists in a region beyond the diameter of the shoulder. From Figure 2 it can be seen that the WN and TMAZ is well within this distance. The possible reasons for this lower hardness in the HAZ could be due to grain coarsening (Figure 6a) and over aging taking place in the material in the HAZ. Clearly the kinetics of the process is also important in the HAZ. The minimum hardness seen is around 100 VHN3kg and 90 VHN3kg for weld produced using 10 mm and 20 mm shoulder diameter, respectively. The hardness of the material in the stirred zone and thermo mechanically affected zone is comparable with base material hardness. The hardness observed in the WN and TMAZ region is slightly lower than the BM. Precipitate dissolution and reforming due to post weld thermal cycle can be the reason for regaining the hardness in the stirred zone and TMAZ. A more detailed analysis is being carried out.
The base material tensile strength was 428.6 MPa and elongation was 16.6. Tensile properties of welds are shown in Table 3. Even though the softest zone occurred away from the weld, the tensile specimens fractured in the weld region for all three welds produced using un-chamfered shoulder with cylindrical flat-end pin. The only exception was the weld made using chamfered shoulder having a frustum shaped rounded endpin with 10 mm shoulder diameter and 6 mm diameter pin. The defects present in the weld make the weld weaker. From the data generated in Table 3, friction stir welding reduces both the tensile strength and ductility of the material. The maximum weld strength for the weld carried out using a tool with 20 mm shoulder diameter and 6mm pin diameter is 395.84 MPa with an elongation of 9.8 per cent. The maximum elongation for 15 mm shoulder diameter and 4mm pin diameter is 14 per cent with tensile strength of 355.41 MPa.
Conclusions
o A defect-free weld is produced in AFNOR 7020 aluminium alloy.
o In the stirred zone, fine equi-axed grains of size ranging from 5 - 20 µm are transformed from the initial pancake shaped 300 to 500µm size parent metal grain structure
o Grain size in the weld nugget varied with the tool geometry
o Two different tools geometries, un-chamfered shoulder with cylindrical flat end pin and chamfered shoulder with frustum shaped rounded end pin, with different dimensions were used. Chamfered shoulder with frustum shaped rounded end pin produces better quality weld than the other
o The experimental results showed that the tensile properties and fracture locations of the joints are significantly affected by tool geometry
o The reduction in strength of weld joint had been significantly attributed to the presence of the defects rather than micro structural changes
o The size of the unfilled region present in the advancing bottom side of the weld decreases as the shoulder diameter is increased or a tilt angle is given
o Friction stir welds of AFNOR 7020 do not show major reduction in hardness in the weld nugget and thermo-mechanically-affected zone. The softest region was formed away from the weld, in the HAZ and the minimum hardness was 90VHN3kg
o The maximum weld strength obtained in this study was 395.84 MPa (92.35 per cent) with 9.8 per cent (59 per cent) elongation is archived in the weld.
Detailed micro-structural analysis using transmission electron microscopy and precipitation kinetic analysis using differential scanning calorimetry is being carried out for further characterisation of weld and to find the possibility of increasing the weld strength.
|
