T-FSW of Dissimilar Aerospace Grade Aluminium Alloys: Inﬂuence of Second Pass on Weld Defects

: The restoration of numerous aircraft structures is achievable with e ﬀ ective repair of welded joints. T-joints are often utilized in these structures to provide structural stability, keeping minimal body weight. Multi-pass friction stir welding (FSW) has proved to be useful for improving the quality of aluminium alloy welds employed in the aerospace sector. However, FSW of these alloys in T-conﬁguration has not been su ﬃ ciently addressed yet. Even rarer is the discussion of e ﬃ cacy of second FSW pass, with altered process parameters for improving the weld quality in T-joints. Hence, two commonly used aerospace grade aluminium alloys, namely, AA2024 and AA7075, were friction stir welded in T-conﬁguration, varying three process parameters, i.e., tool rotational speed, welding speed and shoulder diameter. The e ﬀ ect of second FSW pass, performed at an optimum set of parameters, on kissing bond and tunnelling defect was studied in detail. A substantial reduction in the detrimental e ﬀ ect of these weld defects was discussed via tensile testing, micro-hardness and micro-structural observations.


Introduction
The safety and structural integrity of aircraft structures are largely dependent upon the skin-stringer joining [1][2][3]. T-joints in these structures enable a considerable increase in resistance against inertia and strength without substantial increase in weight, making them a vital component in the aerospace and transportation sectors [4]. The fabrication of skin-stringer welds using friction stir welding (FSW) not only eliminates the use of rivets [5], but also adds the benefits associated with this solid state joining technology, namely, zero porosity welds [6][7][8], low residual stresses [9], good dimensional stability [10], improved corrosion resistance [11], etc. Moreover, FSW has proved to be the most effective technology for joining aerospace grade aluminium alloys, mainly the 2xxx and 7xxx series in both similar and dissimilar arrangements [12][13][14][15][16][17][18][19]. Specifically, the joining of AA2024-T3 and AA7075-T6 alloys is commonly useful in the aerospace industry [20,21].
The main challenge faced during the T-section FSW is improper mixing between the skin and stringer metals. Previous studies have indicated that during the FSW of T-joints, mixing along the skin and stringer sections plays a key role in the determination of the weld quality [22][23][24]. Due to the same reason, various defects like kissing bond, hooking and tunnels occur in the weld zone. A kissing bond is formed due to oxide layer retention at faying interface, which is caused primarily because of lack of sufficient heat and plastic deformation. Whereas, a tunnel is formed either due to (1) inadequate heating of the base metal leading to insufficient material flow, or (2) excessive softening and lack of consolidation of the metal in the solid phase joint. Zhao et al. [22] demonstrated the disparity in the effects of kissing bond and tunnelling defect in deterioration of the weld strength along the skin and stringer, respectively. It was established that an increase in the traverse speed results in larger voids and the kissing bond shifts towards the stir zone (SZ). However, if the kissing bond or oxide bands are located near the edges of the fillet zones, the weld strength decreases significantly [25]. Sato et al. [26] showed that no metallic bonding takes place on either side of the kissing bond line. Investigations with variation in shape (flat and curved) of the die corners (corners at both sides of the weld throat) and T-joint configuration, i.e., joint elements in two pieces and three pieces, have also been studied in detail.
In the past, research studies have resulted in an increase in weld properties, by employing multiple FSW passes at the same parameters [1,[27][28][29]. Studies conducted by Sathari et al. [28] showed improved hardness, tensile strength and ductility for friction stir welds obtained via double pass, on comparison with the welds obtained through single pass. He et al. [1] concluded that the stability of the properties along the weld length improves after the double pass. A decrease in residual stress after multiple FSW passes has also been reported in the literature [27]. However, the potential for increase of friction stir weld properties by double pass has scarcely been utilized for dissimilar skin-stringer joining in general and 7xxx-2xxx combination in particular. Moreover, the effect of double pass in FSW, along with altered process parameters for the second pass has been accounted even more rarely. Further, most of the studies on FSW are focussed on welding in butt configuration. The nature of plasto-mechanics in the material required for successful FSW in T-configuration differs from that of butt configuration. Since a greater heat dissipation from the weld takes place due to clamping of the stringer, process parameters required to produce fine weldments are also different from those required in butt-welding [30,31]. Therefore, in the current study, two alloys vital for the aircraft industry, namely, AA2024-T3 and AA7075-T6, have been friction stir welded in T-joint configuration. The effect of second friction stir pass performed with changed process parameters upon the weld defects, micro-structural and mechanical properties has been addressed in detail.

Materials and Methods
FSW was performed in T-configuration, keeping AA2024-T3 alloy and AA7075-T6 alloy as skin and stringer materials, respectively. Since most applications of T-joints involve high strength material used as stringer [4], AA7075 alloy was kept as the stringer material. The dimensions of both the AA2024 and AA7075 alloy sheets were 200 mm × 50 mm × 2 mm. The chemical composition of these alloys is shown in Tables 1 and 2. Relevant mechanical and thermal properties are stated in Table 3. The incipient melting-liquidus temperature, thermal conductivity and specific heat capacity of the alloys has been obtained from the literature [32][33][34][35][36].  A robust vertical milling machine with an indigenously developed fixture and tool adopter were utilized for carrying out FSW. An experimental setup, including the fixture used for clamping the skin and stringer plates, is shown in Figure 1.  A robust vertical milling machine with an indigenously developed fixture and tool adopter were utilized for carrying out FSW. An experimental setup, including the fixture used for clamping the skin and stringer plates, is shown in Figure 1. A die corner radius of 2 mm was used to clamp the stringer. The fillet formed in the T-joint due to the die corners radius is necessary to avoid stress concentration at these sites [37]. High chromium high carbon steel was used for making the FSW tool. Tool pin failure due to process forces is another major challenge in the FSW of high strength aluminium alloys [38]. Continuous tapered pin profile with pin length 1.9 mm, a pin tip diameter 1.8 mm and pin root diameter 5.6 mm was selected after trial experimentation to induce effective material flow at the skin/stringer interface and prevent stress concentration at the root of the pin. The tool used for performing FSW and the transverse crosssection of weld geometry is schematically illustrated in Figure 2. A die corner radius of 2 mm was used to clamp the stringer. The fillet formed in the T-joint due to the die corners radius is necessary to avoid stress concentration at these sites [37]. High chromium high carbon steel was used for making the FSW tool. Tool pin failure due to process forces is another major challenge in the FSW of high strength aluminium alloys [38]. Continuous tapered pin profile with pin length 1.9 mm, a pin tip diameter 1.8 mm and pin root diameter 5.6 mm was selected after trial experimentation to induce effective material flow at the skin/stringer interface and prevent stress concentration at the root of the pin. The tool used for performing FSW and the transverse cross-section of weld geometry is schematically illustrated in Figure 2.
Three process parameters, namely, tool rotational speed, traverse speed and shoulder diameter, were varied across three levels. A Taguchi L9 orthogonal array was employed to design the experiments. Prior trial experimentation was also performed to identify the range of process parameters which yields sound welds.
Furthermore, to eliminate weld defects and enhance joint strength, second FSW pass was performed on welds obtained from Exp. Nos. 3 and 5. The type of defects found on advancing side (AS) and retreating side (RS) differ because of variation in material flow characteristics on the two sides. Similar flow kinematics during second pass can thus increase the size and detrimental effect of these defects. Therefore, the AS and RS were switched for the second FSW pass. These experiments have been termed as Exp. Nos. 3a and 5a, respectively. The parameters for double pass were kept same as those of Exp. No. 8. Peak tensile strength along stringer was selected as the criterion for selection of parameters for the second pass. The strategy behind this particular approach for second  Table 4 shows the experimental design for the current investigation. Three process parameters, namely, tool rotational speed, traverse speed and shoulder diameter, were varied across three levels. A Taguchi L9 orthogonal array was employed to design the experiments. Prior trial experimentation was also performed to identify the range of process parameters which yields sound welds.
Furthermore, to eliminate weld defects and enhance joint strength, second FSW pass was performed on welds obtained from Exp. Nos. 3 and 5. The type of defects found on advancing side (AS) and retreating side (RS) differ because of variation in material flow characteristics on the two sides. Similar flow kinematics during second pass can thus increase the size and detrimental effect of these defects. Therefore, the AS and RS were switched for the second FSW pass. These experiments have been termed as Exp. Nos. 3a and 5a, respectively. The parameters for double pass were kept same as those of Exp. No. 8. Peak tensile strength along stringer was selected as the criterion for selection of parameters for the second pass. The strategy behind this particular approach for second pass has been addressed in detail in Section 3.1. Tool tilt angle was kept constant at 2°. Table 4 shows the experimental design for the current investigation.  Taguchi L9   1  14  710  40  2  16  710  50  3  12  900  50  4  12  560  40  5  16  560  63  6  14  900  63  7  14  560  50  8  12  710  63  9  16  900  40   Second Pass  3a  12  710  63  5a 12 710 63
Metals 2020, 10, x FOR PEER REVIEW 5 of 15 A sample friction stir welded T-joint is shown in Figure 3. The test specimens were cut using the CNC wire electric discharge machine (WEDM). The standard metallographic procedure was followed for micro-structural examination. The polished specimens were etched using 1.5 mL HNO3 + 1 mL HF + 97.5 mL H2O and observed under an optical microscope (OM). Vickers micro-hardness was measured on the transverse cross section, along the The test specimens were cut using the CNC wire electric discharge machine (WEDM). The standard metallographic procedure was followed for micro-structural examination. The polished specimens were etched using 1.5 mL HNO 3 + 1 mL HF + 97.5 mL H 2 O and observed under an optical microscope (OM). Vickers micro-hardness was measured on the transverse cross section, along the skin and stringer of the welded joint. Each indentation was made at a distance of 0.5 mm with axial load and dwell time of 100 gram and 15 s, respectively. Tensile strength of the T-joints was measured along the skin and stringer using a computer controlled Tensometer with 20 kN capacity. The parameter combination which yielded optimum tensile strength along the stringer was selected for the second pass. Tests were performed at room temperature and at a cross head speed of 2 mm/min. The fixtures used for tensile testing with the firmly fixed samples are shown in Figure 4. The test specimens were cut using the CNC wire electric discharge machine (WEDM). The standard metallographic procedure was followed for micro-structural examination. The polished specimens were etched using 1.5 mL HNO3 + 1 mL HF + 97.5 mL H2O and observed under an optical microscope (OM). Vickers micro-hardness was measured on the transverse cross section, along the skin and stringer of the welded joint. Each indentation was made at a distance of 0.5 mm with axial load and dwell time of 100 gram and 15 s, respectively. Tensile strength of the T-joints was measured along the skin and stringer using a computer controlled Tensometer with 20 kN capacity. The parameter combination which yielded optimum tensile strength along the stringer was selected for the second pass. Tests were performed at room temperature and at a cross head speed of 2 mm/min. The fixtures used for tensile testing with the firmly fixed samples are shown in Figure 4.

Tensile Strength
The tensile strength, percentage elongation and fracture location for all the welds along the skin and stringer have been shown in Table 5. The experiments have been classified according to their respective tool rotational speeds. The failed tensile specimen tested along the stringer and skin are shown in Figures 5 and 6, respectively. Apart from SZ, the specimen also failed from the thermo-mechanically affected zone (TMAZ), tunnel and the kissing bond (KB) defect. It can be clearly observed from Table 5 that a tool rotation speed of 710 rpm yields welds with relatively higher tensile strength and elongation along the skin and stringer. However, at a constant rotational speed, increase in tensile strength along the stringer is discernible with increasing welding speed, whereas no such pattern is visible for tensile strength along the skin. For single pass, a peak tensile strength (along skin) of 330 MPa was observed for Exp. No. 1 while maximum tensile strength along stringer (155 MPa) was observed for Exp. No. 8.   The second pass was strategically performed on the parameters which yielded optimum tensile strength along the stringer, i.e., Exp. No. 8. The effect of tool rotational speed on tensile strength was found to be dominant amongst the three parameters. Thus, to study the effect of second pass with altered process parameters from the scheme of first pass, one weld with lower and one weld with higher initial rotational speed than that of Exp. No. 8 were selected for second pass. At the rotational speeds of 560 rpm and 900 rpm, the welds with least tensile strength along the skin, i.e., Exp. Nos. 3 and 5, were chosen to achieve an overall improvement in the tensile properties of the welds. For the specimen incurring a tunnel defect, i.e., Exp. No. 3 (as shown in Section 3.2), an increase of more than 100% and 200% after the second pass was witnessed in the tensile strength along skin and stringer, respectively. Since Exp. No. 5 at the rotational speed of 560 rpm showed minimum strength along the skin, the second pass was performed on this weld. In this case, the tensile strength along skin improved from 177 MPa to 273 MPa, whereas along the stringer, the strength improved from 85 MPa to 165 MPa. The increase in tensile strength along the stringer for Exp. Nos. 3a and 5a demonstrates that the effect of kissing bond defect has decreased to a considerable degree after the second pass. A minimum rotational speed of 560 rpm does not generate sufficient heat to induce effective metallic bonding between the two base alloys. On the other hand, the highest rotational speed of 900 rpm excessively softens the base alloy. As a result, the tool is not able to consolidate the material effectively, which ultimately lowers the tensile strength. Superimposing the optimum tool rotational speed of 710 rpm on these welds successfully eliminates these shortcomings, leading to the substantial enhancement of tensile properties. The second pass was strategically performed on the parameters which yielded optimum tensile strength along the stringer, i.e., Exp. No. 8. The effect of tool rotational speed on tensile strength was found to be dominant amongst the three parameters. Thus, to study the effect of second pass with altered process parameters from the scheme of first pass, one weld with lower and one weld with higher initial rotational speed than that of Exp. No. 8 were selected for second pass. At the rotational speeds of 560 rpm and 900 rpm, the welds with least tensile strength along the skin, i.e., Exp. Nos. 3 and 5, were chosen to achieve an overall improvement in the tensile properties of the welds. For the specimen incurring a tunnel defect, i.e., Exp. No. 3 (as shown in Section 3.2), an increase of more than 100% and 200% after the second pass was witnessed in the tensile strength along skin and stringer, respectively. Since Exp. No. 5 at the rotational speed of 560 rpm showed minimum strength along the skin, the second pass was performed on this weld. In this case, the tensile strength along skin improved from 177 MPa to 273 MPa, whereas along the stringer, the strength improved from 85 MPa to 165 MPa. The increase in tensile strength along the stringer for Exp. Nos. 3a and 5a demonstrates that the effect of kissing bond defect has decreased to a considerable degree after the second pass. A minimum rotational speed of 560 rpm does not generate sufficient heat to induce effective metallic bonding between the two base alloys. On the other hand, the highest rotational speed of 900 rpm excessively softens the base alloy. As a result, the tool is not able to consolidate the material effectively, which ultimately lowers the tensile strength. Superimposing the optimum tool rotational speed of 710 rpm on these welds successfully eliminates these shortcomings, leading to the substantial enhancement of tensile properties.
Along the stringer, all the welds failed from the interface of the two base alloys in SZ, as shown in Figure 5. Whereas, along the skin, 7 out of 9 welds from Taguchi L9 experimental design failed from SZ, in such a manner that the stringer remained intact with the AS of skin section, as shown in Figure 6. This is principally due to better intermixing between the two alloys at the AS as compared to the RS. Furthermore, Exp. Nos. 3a and 5a can also be observed to adopt the failure location of Exp. No. 8, failing from the RS-KB. This is evident of the effective superimposition of material flow characteristics which improved the weld structure of Exp. Nos. 3 and 5, in coherence with the weld structure of Exp. No. 8. Along the stringer, all the welds failed from the interface of the two base alloys in SZ, as shown in Figure 5. Whereas, along the skin, 7 out of 9 welds from Taguchi L 9 experimental design failed from SZ, in such a manner that the stringer remained intact with the AS of skin section, as shown in Figure 6. This is principally due to better intermixing between the two alloys at the AS as compared to the RS. Furthermore, Exp. Nos. 3a and 5a can also be observed to adopt the failure location of Exp. No. 8, failing from the RS-KB. This is evident of the effective superimposition of material flow characteristics which improved the weld structure of Exp. Nos. 3 and 5, in coherence with the weld structure of Exp. No. 8.

Microstructure
The macrostructure of the weld obtained from Exp. No. 8 is shown in Figure 7. Figure 8 represents the corresponding micro-structures as observed under optical microscope. As it is well known for FSW, the weld structure consists of three different zones, namely, SZ, TMAZ and the heat affected zone (HAZ), which possess different mechanical properties and grain sizes. These zones can be distinctly demarcated in both the skin and stringer material of the welded joint, as shown in Figure 7. The interface of skin and stringer metals in the SZ can be observed in Figure 8a. The SZ experiences highest temperature and degree of plastic deformation, because of the stirring action of the tool, due to which the material undergoes dynamic recrystallization resulting in very fine, equi-axed grains. The TMAZ experiences comparatively smaller heat input and mainly undergoes induced plastic deformation. Since partial recrystallization occurs in the TMAZ, grains in this zone can be observed to be coarser than SZ. For both the materials, the HAZ consists of grains which are slightly larger than the base metal, because of the prevalent thermal cycle. It can be seen from Figure 7 that a broader SZ and thinner TMAZ exists for the stringer material as on the AS, in contrast to the steeper SZ and broader TMAZ in the RS. This elucidates the fact that a much higher degree of plastic deformation due to stirring occurs on the AS, in comparison with the RS. deformation. Since partial recrystallization occurs in the TMAZ, grains in this zone can be observed to be coarser than SZ. For both the materials, the HAZ consists of grains which are slightly larger than the base metal, because of the prevalent thermal cycle. It can be seen from Figure 7 that a broader SZ and thinner TMAZ exists for the stringer material as on the AS, in contrast to the steeper SZ and broader TMAZ in the RS. This elucidates the fact that a much higher degree of plastic deformation due to stirring occurs on the AS, in comparison with the RS.    The TMAZ experiences comparatively smaller heat input and mainly undergoes induced plastic deformation. Since partial recrystallization occurs in the TMAZ, grains in this zone can be observed to be coarser than SZ. For both the materials, the HAZ consists of grains which are slightly larger than the base metal, because of the prevalent thermal cycle. It can be seen from Figure 7 that a broader SZ and thinner TMAZ exists for the stringer material as on the AS, in contrast to the steeper SZ and broader TMAZ in the RS. This elucidates the fact that a much higher degree of plastic deformation due to stirring occurs on the AS, in comparison with the RS.     Figure 9 represents the presence of kissing bond (KB) defect on both AS and RS of the interface between the two alloys. A kissing bond can be seen where the base metals meet in the TMAZ regions. However, in the SZ, efficient intermixing in AS diminishes the kissing bond as explicit from Figure 9a. Whereas, the kissing bond remains obtrusive even in the SZ on the RS, due to a lower degree of plastic deformation.
Weld macro-structures for Exp. Nos. 3, 3a and 5, 5a are shown in Figures 10 and 11, respectively. A tunnel or void can be clearly seen on the AS of the skin/stringer interface in Figure 10a, which disappears after second pass. A very high heat input, due to the maximum rotational speed of 900 rpm in Exp. 3, plasticises a greater amount of material in the SZ. However, the smallest shoulder diameter of 12 mm appears unable to forge/consolidate this material effectively in the wake of the weld. As a result, the softened material would flow out of the weld in the form of flash, leading to metal deficiency and void in the SZ. Pishevar et al. [39] and Chen et al. [40] showed that the heat input during FSW is dominated by the rotational speed, in comparison with the other parameters, which also supports the above inference. However, in the SZ, efficient intermixing in AS diminishes the kissing bond as explicit from Figure  9a. Whereas, the kissing bond remains obtrusive even in the SZ on the RS, due to a lower degree of plastic deformation. Weld macro-structures for Exp. Nos. 3, 3a and 5, 5a are shown in Figures 10 and 11, respectively. A tunnel or void can be clearly seen on the AS of the skin/stringer interface in Figure 10a, which disappears after second pass. A very high heat input, due to the maximum rotational speed of 900 rpm in Exp. 3, plasticises a greater amount of material in the SZ. However, the smallest shoulder diameter of 12 mm appears unable to forge/consolidate this material effectively in the wake of the weld. As a result, the softened material would flow out of the weld in the form of flash, leading to metal deficiency and void in the SZ. Pishevar et al. [39] and Chen et al. [40] showed that the heat input during FSW is dominated by the rotational speed, in comparison with the other parameters, which also supports the above inference.   Weld macro-structures for Exp. Nos. 3, 3a and 5, 5a are shown in Figures 10 and 11, respectively. A tunnel or void can be clearly seen on the AS of the skin/stringer interface in Figure 10a, which disappears after second pass. A very high heat input, due to the maximum rotational speed of 900 rpm in Exp. 3, plasticises a greater amount of material in the SZ. However, the smallest shoulder diameter of 12 mm appears unable to forge/consolidate this material effectively in the wake of the weld. As a result, the softened material would flow out of the weld in the form of flash, leading to metal deficiency and void in the SZ. Pishevar et al. [39] and Chen et al. [40] showed that the heat input during FSW is dominated by the rotational speed, in comparison with the other parameters, which also supports the above inference.  Furthermore, a minimum rotational speed of 560 rpm was employed for Exp. No. 5, which results in lesser softening and the movement of material becomes difficult. As a result, higher flow stress prevails during welding. Notably and understandably, a tool tilt angle of 2 • also influences the material flow dynamics in such a manner that upward extrusion and downward forging of base materials occurs on the AS and RS, respectively. Thus, the stringer material can be seen to be stretched upwards and skin material can be seen to be pulled downwards for Exp. No. 5 in Figure 11a. This also elucidates insufficient metallic bonding between the two materials, due to inadequate heat input. Importantly, an improvement in the homogenisation of AA2024-AA7075 intermixing is evident after the second pass for Exp. Nos. 3a and 5a. This homogenisation occurs due to the superimposition of more suitable material flow kinematics at optimum value of heat input. Since smaller values of flow stress exist for AA2024 as compared to AA7075 at identical temperatures, weld fillets have been formed by AA2024 in Exp. No. 8. However, due to the insufficient softening of AA2024, the AA7075 occupies the weld fillet in Exp. No. 5. Metals 2020, 10, x FOR PEER REVIEW 10 of 15 Furthermore, a minimum rotational speed of 560 rpm was employed for Exp. No. 5, which results in lesser softening and the movement of material becomes difficult. As a result, higher flow stress prevails during welding. Notably and understandably, a tool tilt angle of 2° also influences the material flow dynamics in such a manner that upward extrusion and downward forging of base materials occurs on the AS and RS, respectively. Thus, the stringer material can be seen to be stretched upwards and skin material can be seen to be pulled downwards for Exp. No. 5 in Figure 11a. This also elucidates insufficient metallic bonding between the two materials, due to inadequate heat input. Importantly, an improvement in the homogenisation of AA2024-AA7075 intermixing is evident after the second pass for Exp. Nos. 3a and 5a. This homogenisation occurs due to the superimposition of more suitable material flow kinematics at optimum value of heat input. Since smaller values of flow stress exist for AA2024 as compared to AA7075 at identical temperatures, weld fillets have been formed by AA2024 in Exp. No. 8. However, due to the insufficient softening of AA2024, the AA7075 occupies the weld fillet in Exp. No. 5.

Micro-Hardness
Micro-hardness distribution along the centre line of skin and stringer on the transverse crosssection for Exp. No. 8 is shown in Figures 12 and 13, respectively. Along the skin, the characteristic "W" plot for friction stir welded precipitate hardenable alloys can be clearly observed. As found out by various other researchers, the dissolution of strengthening precipitates causes reduction in hardness in the HAZ [41,42] and the grain refinement causes its relative increase in SZ [43]. It is worth noticing that the minimum micro-hardness of 103.4 HV exists at a distance of 8 mm from the weld centre on the AS, whereas in the stringer section, a minimum micro-hardness of 124.2 HV occurs at a distance of 9 mm from the weld centre. For AA2024, a maximum micro-hardness of 150.4 HV was observed in the SZ, which is comparable to that of the base alloy, i.e., 168 HV. However, the microhardness values in the welded AA7075 section range from 168.9 HV to 124.2 HV, which is considerably less than the base alloy micro-hardness of 225 HV. The above observation infers that amount of softening incurred by AA7075 alloy as stringer is much larger than AA2024 alloy as skin section.

Micro-Hardness
Micro-hardness distribution along the centre line of skin and stringer on the transverse cross-section for Exp. No. 8 is shown in Figures 12 and 13, respectively. Along the skin, the characteristic "W" plot for friction stir welded precipitate hardenable alloys can be clearly observed. As found out by various other researchers, the dissolution of strengthening precipitates causes reduction in hardness in the HAZ [41,42] and the grain refinement causes its relative increase in SZ [43]. It is worth noticing that the minimum micro-hardness of 103.4 HV exists at a distance of 8 mm from the weld centre on the AS, whereas in the stringer section, a minimum micro-hardness of 124.2 HV occurs at a distance of 9 mm from the weld centre. For AA2024, a maximum micro-hardness of 150.4 HV was observed in the SZ, which is comparable to that of the base alloy, i.e., 168 HV. However, the micro-hardness values in the welded AA7075 section range from 168.9 HV to 124.2 HV, which is considerably less than the base alloy micro-hardness of 225 HV. The above observation infers that amount of softening incurred by AA7075 alloy as stringer is much larger than AA2024 alloy as skin section.    Micro-hardness distribution along the skin section for Exp. Nos. 3 and 5, before and after the second pass is plotted in Figure 14 and Figure 15, respectively. For Exp. No. 3, the hardness values in SZ after the second pass range from 117-126 HV, which is similar to the micro-hardness in SZ after the first pass. On the contrary, a significant decrease in micro-hardness values of SZ after the second pass for Exp. No. 5 is evident from Figure 15. This can be explained with reference to the rotational speed to which the welds were subjected in the first FSW pass. For Exp. No. 5, an increase in rotational speed occurs, from 560 rpm in the first pass to 710 rpm in the second pass. As discussed in Section 3.2, such a change in rotational speed leads to greater heat input in the second pass, which results in the better softening and higher degree of plastic deformation of SZ. However, this is not the case for Exp. Nos. 3 and 3a. Thus, such an observation has not been encountered for Exp. No. 3. The aberrant peaks in micro-hardness values in the SZ for Exp. No. 5a likely occur due to heterogeneous mixing with the harder stringer material, which can be observed in Figure 11. Notably, a considerable improvement in micro-hardness values occurs beyond the SZ after the second pass for both Exp. Nos. 3 and 5.
The presence of a void in the SZ has also been reflected as a sudden fall of micro-hardness for Exp. No. 3. Micro-hardness distribution along the skin section for Exp. Nos. 3 and 5, before and after the second pass is plotted in Figures 14 and 15, respectively. For Exp. No. 3, the hardness values in SZ after the second pass range from 117-126 HV, which is similar to the micro-hardness in SZ after the first pass. On the contrary, a significant decrease in micro-hardness values of SZ after the second pass for Exp. No. 5 is evident from Figure 15. This can be explained with reference to the rotational speed to which the welds were subjected in the first FSW pass. For Exp. No. 5, an increase in rotational speed occurs, from 560 rpm in the first pass to 710 rpm in the second pass. As discussed in Section 3.2, such a change in rotational speed leads to greater heat input in the second pass, which results in the better softening and higher degree of plastic deformation of SZ. However, this is not the case for Exp. Nos. 3 and 3a. Thus, such an observation has not been encountered for Exp. No. 3. The aberrant peaks in micro-hardness values in the SZ for Exp. No. 5a likely occur due to heterogeneous mixing with the harder stringer material, which can be observed in Figure 11. Notably, a considerable improvement in micro-hardness values occurs beyond the SZ after the second pass for both Exp. Nos. 3 and 5.

Conclusions
In the present study, the effect of the second pass with altered process parameters on defects occurring in dissimilar T-friction stir welds was addressed, with reference to the mechanical and micro-structural features. The following conclusions can be drawn based upon the outcomes of this study: (1) The tunnelling defect was eliminated and a significant decrement in the obtrusive effect of KB defect was achieved by applying second welding pass at optimum parameters which induced increased inter-mixing between the AA2024 and AA7075 alloys. A greater degree of plastic deformation at AS helps in overcoming KB defect at this side of the weld.
(2) A considerable decrease in micro-hardness of the SZ occurs when a higher tool rotational speed of 710 rpm was applied in the second pass. Whereas, when a lesser rotational speed than that of first pass is applied, a negligible change in the micro-hardness of SZ was encountered. The decrease in micro-hardness of the stringer alloy, i.e., AA7075 alloy, was much larger than the decrement in the The presence of a void in the SZ has also been reflected as a sudden fall of micro-hardness for Exp. No. 3.

Conclusions
In the present study, the effect of the second pass with altered process parameters on defects occurring in dissimilar T-friction stir welds was addressed, with reference to the mechanical and micro-structural features. The following conclusions can be drawn based upon the outcomes of this study: (1) The tunnelling defect was eliminated and a significant decrement in the obtrusive effect of KB defect was achieved by applying second welding pass at optimum parameters which induced increased inter-mixing between the AA2024 and AA7075 alloys. A greater degree of plastic deformation at AS helps in overcoming KB defect at this side of the weld.
(2) A considerable decrease in micro-hardness of the SZ occurs when a higher tool rotational speed of 710 rpm was applied in the second pass. Whereas, when a lesser rotational speed than that of first pass is applied, a negligible change in the micro-hardness of SZ was encountered. The decrease in micro-hardness of the stringer alloy, i.e., AA7075 alloy, was much larger than the decrement in the hardness values of the AA2024 alloy. The appearance of aberrant peaks in micro-hardness distribution in SZ after the second pass indicate improved skin-stringer intermixing.
(3) For the weld exhibiting a void in the SZ in first pass, the ultimate tensile strength along skin increased from 150 MPa to 307 MPa after the second pass. Tensile strength along stringer also improved over 100% in comparison with the welds before the second pass. This positive restoration of weld structures occurs due to enhanced material flow dynamics at ideal heat input.