**5. Experimental findings and evidence for sound and better FSW joints**

## **5.1. The DS-FSW method**

**Figure 2.** Comparison between conventional (R-type) and T-type FSW configurations.

after (post-weld annealing, PWA), followed by furnace cooling.

treatable AA5000-series alloy (AA5754) was welded.

1 mm);

12 Joining Technologies

**4. The material**

The welding motion combines two different plate-to-pin mutual motion setups:

(2) a pin translation along a direction parallel to the welding centerline line.

(1) a pin axial spin rotation sets perpendicular to the sheet blanks, changing the rotation along the plate centerline by a radius equal to R (=0, corresponding to the conventional FSW, 0.5, and

The RT-type FSW innovative approach was compared with the conventional T-type (linear welding motion, i.e., for *R* = 0). In both the RT-type and T-type FSW processes, the stirring action was exerted by the pin tool rotation around its axis; the pin tilt angle was set at 2°, with respect to the normal direction to the plate surface. The RT-type and T-type FSW were performed using a pin rotational speed, *ω* = 2000 rpm, and a transverse speed, *v* = 30 mm/min. All experiments were carried out with a tool plunging speed of 1.5 mm/min. The above reported setting parameters were chosen by an optimization FSW processing study reported in [22], where the effect of the welding parameters and tool configuration on micromechanical and macromechanical properties of FSW joints in AA5754 sheets were investigated. The AA5754 was subjected to an annealing treatment at 415°C/3 h, both prior (AA5754-O), and

In both the cases, aluminum alloys were tested. In the first methodology, a heat-treatable AA6000-series alloy (AA6082) was used; whereas in the second methodology, a nonheat-

The AA5000-series alloys, such as the AA5754, are widely used in automotive, aerospace, marine, and military applications. They are characterized by a good strength-to-weight ratio, an appropriate weldability, and a good corrosion resistance. This class of aluminum alloy is difficult to join by conventional fusion welding techniques. This is mainly due to a dendritic structure, which typically forms in the melted zone and it seriously weakens the mechanical properties of the joint AA5000 series alloys. In this context, FSW has emerged as a promising solid-state process with the successfully overcome the fusion welding problems, making the welding process of AA5000 series alloys a sound one [23]. FSW of AA5754 is thus a promising

There is a strong need for an improvement in ductility and formability of FSWed joints. Some previous studies reported significant mechanical improvements by carrying out multipass [29], double lap [30], reverse dual rotation [31] FSW, and FS spot welding [32]. With this respect, the DS-FSW showed better strength, elongation, and formability of FSWed aluminum joints. The DS-FSW was proven to induce the serration of the geometric discontinuities, thus promoting a significant microstructure homogeneity at the NZ.

### *5.1.1. Mechanical properties*

**Figure 3** shows typical nominal stress versus nominal strain curves of FSWed joints in AA6082 obtained under different values of the rotational speed and welding speed. The joints ductility is shown to be lower in the NZ, with respect to the base metal (BM), irrespective of the welding parameters and process methodology [22]. In general, in terms of both the ultimate values of tensile strength and elongation, the conventional FSWed joints show a tensile behavior better than the one exhibited by the DS-FSWed joints. Actually, the conventional FSW process requires a high sinking value in order to generate the frictional heating allowing the material flow necessary to obtain sound joints, according to Mishra and Ma [7]. Thus, in the first pass, by using the same tool sinking as of conventional FSW produces a step in the blank surface that acts as a notch during the second pass. Therefore, the tool sinking value imposed in the second pass had to be further decreased in order to reduce the formation of surface defects. The pinless-pinless configuration has provided the worst tensile properties. In particular, the AS-AS configuration showed low mechanical properties of the joint, while the AS-RS config‐ uration did not reach a sound weldment.

The mechanical behavior is strongly improved when welding is performed using the pinpinless configuration. In this case, ductility levels similar to the ones showed by the conven‐ tional FSWed samples were obtained. In this case, the tensile fracture occurred at the HAZ, in the RS zone. The tensile properties of the joints are slightly affected by the rotational and welding speeds. This is not the case in the DS-FSW pinless-pinless tool configuration, which exhibits ultimate tensile strength (UTS) and UE values strongly dependent on the process parameters (**Figure 3**). As a matter of fact, as the thickness increases, the shoulder influence becomes ever more localized near the top surface of the sheet and, consequently, the stirring action becomes less and less effective. With this respect, in published work by these authors, the FSW capability to obtain sound joints in 1- and 1.5-mm-thick sheets using a pinless tool was widely documented [20, 22, 33].

**Figure 4.** Hemispherical punch test for different testing conditions (LDH is the Limit Dome Height).

T arrangement surface 1 is opposite.

arrangements, respectively (**Figure 4**).

**Figure 5.** Hemispherical punch test configurations. In the B arrangement, surface 1 is in contact with the punch; in the

New Approaches to the Friction Stir Welding of Aluminum Alloys

http://dx.doi.org/10.5772/64523

15

In the B arrangement, the local stress field intensity rise, caused by the notch, is responsible of the FSW sample failure at the geometric discontinuity. In the T arrangement, that is, FSWed blank having the notch in contact with the punch, the failure of the deformed joint occurs at the step produced by the sinking action applied by the shoulder [19]. This is mainly due to the biaxial tensile stress state to which the notch is subjected. This appeared to be less severe in

The DS-FSW joints showed LDH values higher than those measured on the conventional FSWed joints. This is likely to be attributed to the beneficial effect of the second pass of the DS-FSW. This second welding induces a dual beneficial effect: it allows both the closure of the geometric discontinuity, and the reduction in the height of the step produced by the first welding on the opposite plate surface. Furthermore, the DS-FSW is characterized by more uniform recrystallized grains across the NZ, and also partially across the thermo-mechanical affected zone (TMAZ), than in the case of the conventional FSW [19]. Finally, the joints obtained using the pin-pinless tool configuration lead to LDH values higher than the ones obtained by using the pin-pin configuration, irrespective of the sheets arrangement. Thus, in the pin-pinless configuration, the LDH value was only ~12% lower than that of the BM. This result appears to be virtually independent of the sheet arrangement. On the other hand, in the pin-pin tool configuration, the LDH reaches values of ~19 and 25% lower than that of BM in the T and B

A more accurate evaluation of formability is obtained by means of the forming limit curves (FLCs). This is obtained by plotting the major strain versus minor strain data (**Figure 6**). It

the T arrangements, with respect to that in the B arrangement.

**Figure 3.** Tensile stress-strain curves of the DS-FSW, with different welding parameters and tool configurations.

#### *5.1.2. Post-welding formability*

In comparison with conventional fusion welding techniques, one of the most important advantages offered by FSW is the relatively high post-welding formability. In this sense, the conventional FSW and the DS-FSW formability, obtained under the same process conditions, was detected. For this purpose, limit dome height (LDH) analyses were carried out. These values represent the punch stroke at the peak of the load versus the stroke curves. It actually represents the dome height of the deformed samples at the onset of necking, and the results are reported in the plot of **Figure 4**. These data agree with the joint ductility obtained by tensile tests (**Figure 3**). The LDH values were lower than those obtained on the BM, no matter what welding methodology was used. Such results reveal that a noticeable formability reduction along the welding zone [18, 20, 22, 33–35]. More specifically, the B arrangement leads to a LDH value lower than the T arrangement (as reported by the letter B and T in **Figure 4**, and according to the configuration reported in **Figure 5**).

**Figure 4.** Hemispherical punch test for different testing conditions (LDH is the Limit Dome Height).

exhibits ultimate tensile strength (UTS) and UE values strongly dependent on the process parameters (**Figure 3**). As a matter of fact, as the thickness increases, the shoulder influence becomes ever more localized near the top surface of the sheet and, consequently, the stirring action becomes less and less effective. With this respect, in published work by these authors, the FSW capability to obtain sound joints in 1- and 1.5-mm-thick sheets using a pinless tool

**Figure 3.** Tensile stress-strain curves of the DS-FSW, with different welding parameters and tool configurations.

In comparison with conventional fusion welding techniques, one of the most important advantages offered by FSW is the relatively high post-welding formability. In this sense, the conventional FSW and the DS-FSW formability, obtained under the same process conditions, was detected. For this purpose, limit dome height (LDH) analyses were carried out. These values represent the punch stroke at the peak of the load versus the stroke curves. It actually represents the dome height of the deformed samples at the onset of necking, and the results are reported in the plot of **Figure 4**. These data agree with the joint ductility obtained by tensile tests (**Figure 3**). The LDH values were lower than those obtained on the BM, no matter what welding methodology was used. Such results reveal that a noticeable formability reduction along the welding zone [18, 20, 22, 33–35]. More specifically, the B arrangement leads to a LDH value lower than the T arrangement (as reported by the letter B and T in **Figure 4**, and according

was widely documented [20, 22, 33].

14 Joining Technologies

*5.1.2. Post-welding formability*

to the configuration reported in **Figure 5**).

**Figure 5.** Hemispherical punch test configurations. In the B arrangement, surface 1 is in contact with the punch; in the T arrangement surface 1 is opposite.

In the B arrangement, the local stress field intensity rise, caused by the notch, is responsible of the FSW sample failure at the geometric discontinuity. In the T arrangement, that is, FSWed blank having the notch in contact with the punch, the failure of the deformed joint occurs at the step produced by the sinking action applied by the shoulder [19]. This is mainly due to the biaxial tensile stress state to which the notch is subjected. This appeared to be less severe in the T arrangements, with respect to that in the B arrangement.

The DS-FSW joints showed LDH values higher than those measured on the conventional FSWed joints. This is likely to be attributed to the beneficial effect of the second pass of the DS-FSW. This second welding induces a dual beneficial effect: it allows both the closure of the geometric discontinuity, and the reduction in the height of the step produced by the first welding on the opposite plate surface. Furthermore, the DS-FSW is characterized by more uniform recrystallized grains across the NZ, and also partially across the thermo-mechanical affected zone (TMAZ), than in the case of the conventional FSW [19]. Finally, the joints obtained using the pin-pinless tool configuration lead to LDH values higher than the ones obtained by using the pin-pin configuration, irrespective of the sheets arrangement. Thus, in the pin-pinless configuration, the LDH value was only ~12% lower than that of the BM. This result appears to be virtually independent of the sheet arrangement. On the other hand, in the pin-pin tool configuration, the LDH reaches values of ~19 and 25% lower than that of BM in the T and B arrangements, respectively (**Figure 4**).

A more accurate evaluation of formability is obtained by means of the forming limit curves (FLCs). This is obtained by plotting the major strain versus minor strain data (**Figure 6**). It resulted that the formability of the BM is always higher than that of the welded joints. In the stretching side of the FLD, for a given minor strain, the major strain measured on the DS-FSWed joints appeared systematically higher than that provided by the conventional FSWed ones. This result agrees with the behavior exhibited by the LDH (**Figure 4**). The higher vertical position of the FLCs confirms that formability is strongly improved when the DS-FSW technique is used. The comparison among the different FLCs obtained in the DS-FSW, using both the pin-pinless and the pin-pin configurations, shows that the FLCs are scarcely affected by the tool configuration used in the second pass. Finally, the B arrangement is characterized by the lowest major strain values, and this agrees with the LDH results shown in **Figure 4**. This differentiation tends to vanish when DS-FSW is used. It is noteworthy to observe that the process methodology and sample arrangement also affect the extension of the FLCs. In fact, **Figure 6** shows FLCs smaller extension in the welded joints, compared to the BM. Such behavior is almost negligible in the drawing zone of the FLD. This becomes significant in the stretching region, as confirmed by LDH values shown in **Figure 6**.

**Figure 7.** Nanoindentation hardness profiles for the different configurations used.

**Figure 8.** Nanoindentation Young's modulus profiles for the different configurations used.

Moreover, in both the pin and pinless conventional FSW, the lower *H* and *Er* values, recorded at the surface in contact with the shoulder, also pertain to the outer part of the SZ. On the contrary, in both pin and pinless FSW, the profile along the centerline of the sheet section did not show any reduction of *H* and *Er* across the NZ, TMAZ, and HAZ. The hardness, at the section centerline of the NZ, is constantly higher than the BM, with values that peak at 2.05 GPa. In particular, the two AS-TAMZ and RS-TMAZ showed values ranging 1.40–1.55 GPa, while the hardness in the NZ ranged from 1.65 to 2.05. Along the centerline of the pinless FSWed

New Approaches to the Friction Stir Welding of Aluminum Alloys

http://dx.doi.org/10.5772/64523

17

**Figure 6.** Forming limit curves (FLC) obtained using different welding processes and sample arrangements.

**Figures 7** and **8** show the nanoindentation experimental results, in terms of harness and Young's modulus. The main difference between the pin and the pinless FSW consists of the low hardness (*H*) and elastic modulus (*Er*) values obtained at the TMAZ. In the conventional pin FSW, low *H* and *Er* are in the retreating TMAZ, whereas in the pinless FSW, both the advancing and the retreating TMAZ experienced such low *H* and *Er* values. In both the cases, these lower values accounted for a drastic hardness reduction, being three times less, and an elastic modulus reduction of almost ten times, with respect to the BM values.

New Approaches to the Friction Stir Welding of Aluminum Alloys http://dx.doi.org/10.5772/64523 17

**Figure 7.** Nanoindentation hardness profiles for the different configurations used.

resulted that the formability of the BM is always higher than that of the welded joints. In the stretching side of the FLD, for a given minor strain, the major strain measured on the DS-FSWed joints appeared systematically higher than that provided by the conventional FSWed ones. This result agrees with the behavior exhibited by the LDH (**Figure 4**). The higher vertical position of the FLCs confirms that formability is strongly improved when the DS-FSW technique is used. The comparison among the different FLCs obtained in the DS-FSW, using both the pin-pinless and the pin-pin configurations, shows that the FLCs are scarcely affected by the tool configuration used in the second pass. Finally, the B arrangement is characterized by the lowest major strain values, and this agrees with the LDH results shown in **Figure 4**. This differentiation tends to vanish when DS-FSW is used. It is noteworthy to observe that the process methodology and sample arrangement also affect the extension of the FLCs. In fact, **Figure 6** shows FLCs smaller extension in the welded joints, compared to the BM. Such behavior is almost negligible in the drawing zone of the FLD. This becomes significant in the

stretching region, as confirmed by LDH values shown in **Figure 6**.

16 Joining Technologies

**Figure 6.** Forming limit curves (FLC) obtained using different welding processes and sample arrangements.

elastic modulus reduction of almost ten times, with respect to the BM values.

**Figures 7** and **8** show the nanoindentation experimental results, in terms of harness and Young's modulus. The main difference between the pin and the pinless FSW consists of the low hardness (*H*) and elastic modulus (*Er*) values obtained at the TMAZ. In the conventional pin FSW, low *H* and *Er* are in the retreating TMAZ, whereas in the pinless FSW, both the advancing and the retreating TMAZ experienced such low *H* and *Er* values. In both the cases, these lower values accounted for a drastic hardness reduction, being three times less, and an

**Figure 8.** Nanoindentation Young's modulus profiles for the different configurations used.

Moreover, in both the pin and pinless conventional FSW, the lower *H* and *Er* values, recorded at the surface in contact with the shoulder, also pertain to the outer part of the SZ. On the contrary, in both pin and pinless FSW, the profile along the centerline of the sheet section did not show any reduction of *H* and *Er* across the NZ, TMAZ, and HAZ. The hardness, at the section centerline of the NZ, is constantly higher than the BM, with values that peak at 2.05 GPa. In particular, the two AS-TAMZ and RS-TMAZ showed values ranging 1.40–1.55 GPa, while the hardness in the NZ ranged from 1.65 to 2.05. Along the centerline of the pinless FSWed section, the hardness profile appeared considerably more uniform than the one obtained in the pin FSW. In the former case, the hardness ranged from 1.75 to 2.10 GPa, across the characteristic regions of the FSW joint.

morphology across the SZ, from surface to surface, greatly favored the soundness and better

New Approaches to the Friction Stir Welding of Aluminum Alloys

http://dx.doi.org/10.5772/64523

19

**5.2. The friction stir welding method by pin rotation deviation from centerline: RT-FSW**

Rotation of the pin from its centerline progression, during the FSW process, is intended as a

**Figure 9** shows typical stress-strain tensile curves of both as-received and FSW AA5754 sheets. The ultimate tensile strength (UTS) and plastic elongation (El) values are reported in **Figure 10**. The mechanical response is quite different as the pin moves in RT-type configu‐ ration with *R* = 0.5 mm. In this case, the UTS reduction respect to the unwelded sheet was 25%, that is double that for *R* = 0 mm (T-type FSW). At *R* = 0.5 mm, ductility (El) was 78%, respect to the unwelded sheet, and thus it was three times lower with respect to the ductili‐ ty obtained at *R* = 0 mm (in the T-type FSW). UTS reduction, compared to the unwelded sheet, accounted for 20 and 22%, respectively; the ductility reduction was 48 and 74%, ac‐

further possible improvement in the welded soundness of aluminum plates.

**Figure 9.** Tensile stress-strain curves for RT-FSW at *R* = 0 (conventional FSW), and 0.5 mm.

The tensile curves, irrespective of the specific FSW setup (T- and RT-type FSW), clearly showed the occurrence of serrated yielding, also termed the Portevin-Le Chatelier (PLC) effect, that is a common phenomenon in 5xxx aluminum alloys [36, 37]. The PLC phenomenon is driven by Mg solute atom cloud formation. This microstructure atomic-level evolution is actually

**Figure 10.** UTS and El of the AA5754 RT-FSW.

post-welding response of the welded Al-sheets.

*5.2.1. Mechanical properties*

cordingly.

Uniform *H* and *Er* values were obtained across the welded zone in all the three DS-FSW configurations. This was not the case in the two conventional, where hardness appeared far from being uniform across the FSW joint regions. In particular, in the pin-pin AS-AS DS-FSW, the hardness decreased in the TMAZ AS, while the rest of the welded zone (i.e., NZ and RS-TMAZ) showed hardness values significantly higher than the ones of the BM. This trend was common to all the three profiles (upper surface, centerline, lower surface). The elastic modulus increased up to 50%, in the NZ. The top values were reached in the surface where the second FSW took place (Surface 2). Quite similar hardness trends were found in the pin-pin AS-RS DS-FSW. In particular, in the TMAZ RS (at the second welding), the elastic modulus steadily increased from values of almost half respect the BM, to reach values of some 30–35% higher than those of BM, in the TMAZ AS. Finally, in the pin-pinless AS-AS DS-FSW, *H* and *Er* profiles, taken along the upper-surface, centerline, and lower-surface, were essentially similar to those observed in the pin-pin AS-AS DS-FSW. The only significant difference was the rather fuzzy and wavy hardness trend obtained in this case, at the AS-TMAZ, SZ, and RS-TMAZ.

The better formability of the DS-FSW, with respect to the conventional FSW, is most likely related to the local elastic modulus uniformity (i.e., the reduced Young's modulus) across the weld, and to the less dramatic hardness variation, from top to bottom of the sheet section.

#### *5.1.3. Microstructure of joints*

**Table 2** reports the mean grain size as a function of welding methodology and tool configu‐ ration, measured at different zones of the welded joints. The considerably small grain size in the SZ, combined with the equiaxed grain shape, implies the occurrence of dynamic recrys‐ tallization because of the very high levels of deformation and temperature reached in such zone during FSW [35].


**Table 2.** AA6082 DS-FSW mean grain size; the BM SZ had a mean grain size of 20 ± 2 μm.

In the conventional FSW, the grain size within the SZ tended to increase near the top of the weld zone, and this is chiefly due to the temperature variation within the weld zone [36]. In all the three DS-FSW configurations described here, the observed grain size uniformity and morphology across the SZ, from surface to surface, greatly favored the soundness and better post-welding response of the welded Al-sheets.

#### **5.2. The friction stir welding method by pin rotation deviation from centerline: RT-FSW**

Rotation of the pin from its centerline progression, during the FSW process, is intended as a further possible improvement in the welded soundness of aluminum plates.

#### *5.2.1. Mechanical properties*

section, the hardness profile appeared considerably more uniform than the one obtained in the pin FSW. In the former case, the hardness ranged from 1.75 to 2.10 GPa, across the

Uniform *H* and *Er* values were obtained across the welded zone in all the three DS-FSW configurations. This was not the case in the two conventional, where hardness appeared far from being uniform across the FSW joint regions. In particular, in the pin-pin AS-AS DS-FSW, the hardness decreased in the TMAZ AS, while the rest of the welded zone (i.e., NZ and RS-TMAZ) showed hardness values significantly higher than the ones of the BM. This trend was common to all the three profiles (upper surface, centerline, lower surface). The elastic modulus increased up to 50%, in the NZ. The top values were reached in the surface where the second FSW took place (Surface 2). Quite similar hardness trends were found in the pin-pin AS-RS DS-FSW. In particular, in the TMAZ RS (at the second welding), the elastic modulus steadily increased from values of almost half respect the BM, to reach values of some 30–35% higher than those of BM, in the TMAZ AS. Finally, in the pin-pinless AS-AS DS-FSW, *H* and *Er* profiles, taken along the upper-surface, centerline, and lower-surface, were essentially similar to those observed in the pin-pin AS-AS DS-FSW. The only significant difference was the rather fuzzy

and wavy hardness trend obtained in this case, at the AS-TMAZ, SZ, and RS-TMAZ.

The better formability of the DS-FSW, with respect to the conventional FSW, is most likely related to the local elastic modulus uniformity (i.e., the reduced Young's modulus) across the weld, and to the less dramatic hardness variation, from top to bottom of the sheet section.

**Table 2** reports the mean grain size as a function of welding methodology and tool configu‐ ration, measured at different zones of the welded joints. The considerably small grain size in the SZ, combined with the equiaxed grain shape, implies the occurrence of dynamic recrys‐ tallization because of the very high levels of deformation and temperature reached in such

Conventional 14 ± 2 7.8 ± 0.3 7.6 ± 0.3 6.9 ± 0.3 14 ± 2

In the conventional FSW, the grain size within the SZ tended to increase near the top of the weld zone, and this is chiefly due to the temperature variation within the weld zone [36]. In all the three DS-FSW configurations described here, the observed grain size uniformity and

**Table 2.** AA6082 DS-FSW mean grain size; the BM SZ had a mean grain size of 20 ± 2 μm.

**AS/TMAZ Surface 1/SZ Middle/SZ Surface 2/SZ RS/TMAZ**

12 ± 2 6.3 ± 0.2 6.2 ± 0.2 6.4 ± 0.2 13 ± 2

12 ± 2 5.7 ± 0.2 5.7 ± 0.2 5.8 ± 0.2 13 ± 2

characteristic regions of the FSW joint.

18 Joining Technologies

*5.1.3. Microstructure of joints*

zone during FSW [35].

DS-FSW pin-pin

DS-FSW pin-pinless

**FSW configuration Mean grain size, μm**

**Figure 9** shows typical stress-strain tensile curves of both as-received and FSW AA5754 sheets. The ultimate tensile strength (UTS) and plastic elongation (El) values are reported in **Figure 10**. The mechanical response is quite different as the pin moves in RT-type configu‐ ration with *R* = 0.5 mm. In this case, the UTS reduction respect to the unwelded sheet was 25%, that is double that for *R* = 0 mm (T-type FSW). At *R* = 0.5 mm, ductility (El) was 78%, respect to the unwelded sheet, and thus it was three times lower with respect to the ductili‐ ty obtained at *R* = 0 mm (in the T-type FSW). UTS reduction, compared to the unwelded sheet, accounted for 20 and 22%, respectively; the ductility reduction was 48 and 74%, ac‐ cordingly.

**Figure 9.** Tensile stress-strain curves for RT-FSW at *R* = 0 (conventional FSW), and 0.5 mm.

**Figure 10.** UTS and El of the AA5754 RT-FSW.

The tensile curves, irrespective of the specific FSW setup (T- and RT-type FSW), clearly showed the occurrence of serrated yielding, also termed the Portevin-Le Chatelier (PLC) effect, that is a common phenomenon in 5xxx aluminum alloys [36, 37]. The PLC phenomenon is driven by Mg solute atom cloud formation. This microstructure atomic-level evolution is actually responsible for the strain rate dependency of the observed serrated yielding phenomenon. The Mg solid solution, induced in the grains of the cold-rolled AA5754 sheets, effectively pins the dislocation sliding motion induced by the tensile test. This, in turns, generates the yielding phenomenon.

welding process. The NZ, obtained with *R* = 1 mm, still presented some oxide layers both in

New Approaches to the Friction Stir Welding of Aluminum Alloys

http://dx.doi.org/10.5772/64523

21

The microstructure of the post-weld annealed (PWA) sheets, for the different rotational radii investigated, is shown in **Figure 13**. For *R* = 0 mm, fine equiaxed grains characterize the whole extension of the FSW sheet. These equiaxed grains had a mean size substantially same as the ones in the BM, and this was found in the HAZ and the TMAZ of the AS and RS. The only exception consisted in the grain size and morphology in upper welded zone, that is, the surface directly in contact with the shoulder, during the FSW. In this zone, very coarse irregular grains were induced to form by the stirring effect, and by the heat flow introduced in the aluminum plate by the tool shoulder. The depth extension of this coarse-grain region accounted for a minimum of one-third to a maximum of half of the whole sheet thickness. In particular, at *R* = 0.5 mm, fine recrystallized equiaxed grains characterized the whole extension of the BM, HAZ, and the TMAZ, in both the AS and the RS. More specifically, in this case, the NZ was characterized by the occurrence of very coarse irregular grains, mixed with fine recrystallized grains strips located throughout across the NZ, from the upper to the lower surface. The primary factors leading to the occurrence of abnormal grain growth process, during FSW, are

**Figure 12.** Montage of POM RT-FSW AA5754-O at *R* = 0 (conventional FSW), 0.5, and 1 mm.

associated with inhomogeneous gran deformation.

**Figure 13.** Montage of POM RT-FSW PWA at *R* = 0 (conventional FSW), 0.5, and 1 mm.

its center and near the TMAZ/NZ boundary.

### *5.2.2. Microstructure of the joints at the different pin rotation radii*

**Figure 11** shows an overview of the FSW plate microstructure, in which the occurrence of a grain dynamic recrystallization process in the NZ is evident.

**Figure 11.** Montage of polarized optical micrographs (POM) RT-FSW at *R* = 0 (conventional FSW), 0.5, and 1 mm.

#### *5.2.3. Microstructure modifications induced by pre- and post-welding annealing*

The AA5754 was subjected to an annealing treatment at 415°C/3 h followed by furnace cooling, in one case prior FSW (AA5754-O state), and in another case, after FSW (post-weld annealing: PWA).

The microstructure of the annealed FSW AA5754-O sheets, obtained both under T-type and RT-type FSW configurations, is shown in **Figure 12**. As expected, the base material (BM) is fully recrystallized. It appeared that the equiaxed recrystallized mean grains did not change significantly in the BM, HAZ, and TMAZ, on either AS and RS of joint. This was found irrespective of the pin deviation extent. The NZ-grained structure, in the conventional (*R* = 0), and for 0.5 mm pin rotation deviation, appeared to be mixed, and characterized by the coexistence of fine equiaxed grains and stirred elongated grains (still remaining of the stirring effect induced by the FSW). It is actually a microstructure modification induced by concurring effect driven by the first recrystallization stage (due to the annealing treatment at 415°C/3 h), and by the following mechanical heat flow during the FSW. This latter is known to rise the temperature in the NZ aluminum alloys typically by 350–500°C [20, 27, 37]. The tool shoulder rotation, during welding, induces a large heat transfer and a high strain level at the top surface, which is considerably higher than that induced at the bottom surface. The bottom surface was in contact with a back-plate support, during the welding process. This indeed acted as a heat sink lowering the peak temperature and reducing the time to the peak temperature. Thus, in turns, grain growth in the bottom surface of the NZ was effectively slowed down during the welding process. The NZ, obtained with *R* = 1 mm, still presented some oxide layers both in its center and near the TMAZ/NZ boundary.

**Figure 12.** Montage of POM RT-FSW AA5754-O at *R* = 0 (conventional FSW), 0.5, and 1 mm.

responsible for the strain rate dependency of the observed serrated yielding phenomenon. The Mg solid solution, induced in the grains of the cold-rolled AA5754 sheets, effectively pins the dislocation sliding motion induced by the tensile test. This, in turns, generates the yielding

**Figure 11** shows an overview of the FSW plate microstructure, in which the occurrence of a

**Figure 11.** Montage of polarized optical micrographs (POM) RT-FSW at *R* = 0 (conventional FSW), 0.5, and 1 mm.

The AA5754 was subjected to an annealing treatment at 415°C/3 h followed by furnace cooling, in one case prior FSW (AA5754-O state), and in another case, after FSW (post-weld annealing:

The microstructure of the annealed FSW AA5754-O sheets, obtained both under T-type and RT-type FSW configurations, is shown in **Figure 12**. As expected, the base material (BM) is fully recrystallized. It appeared that the equiaxed recrystallized mean grains did not change significantly in the BM, HAZ, and TMAZ, on either AS and RS of joint. This was found irrespective of the pin deviation extent. The NZ-grained structure, in the conventional (*R* = 0), and for 0.5 mm pin rotation deviation, appeared to be mixed, and characterized by the coexistence of fine equiaxed grains and stirred elongated grains (still remaining of the stirring effect induced by the FSW). It is actually a microstructure modification induced by concurring effect driven by the first recrystallization stage (due to the annealing treatment at 415°C/3 h), and by the following mechanical heat flow during the FSW. This latter is known to rise the temperature in the NZ aluminum alloys typically by 350–500°C [20, 27, 37]. The tool shoulder rotation, during welding, induces a large heat transfer and a high strain level at the top surface, which is considerably higher than that induced at the bottom surface. The bottom surface was in contact with a back-plate support, during the welding process. This indeed acted as a heat sink lowering the peak temperature and reducing the time to the peak temperature. Thus, in turns, grain growth in the bottom surface of the NZ was effectively slowed down during the

*5.2.3. Microstructure modifications induced by pre- and post-welding annealing*

*5.2.2. Microstructure of the joints at the different pin rotation radii*

grain dynamic recrystallization process in the NZ is evident.

phenomenon.

20 Joining Technologies

PWA).

The microstructure of the post-weld annealed (PWA) sheets, for the different rotational radii investigated, is shown in **Figure 13**. For *R* = 0 mm, fine equiaxed grains characterize the whole extension of the FSW sheet. These equiaxed grains had a mean size substantially same as the ones in the BM, and this was found in the HAZ and the TMAZ of the AS and RS. The only exception consisted in the grain size and morphology in upper welded zone, that is, the surface directly in contact with the shoulder, during the FSW. In this zone, very coarse irregular grains were induced to form by the stirring effect, and by the heat flow introduced in the aluminum plate by the tool shoulder. The depth extension of this coarse-grain region accounted for a minimum of one-third to a maximum of half of the whole sheet thickness. In particular, at *R* = 0.5 mm, fine recrystallized equiaxed grains characterized the whole extension of the BM, HAZ, and the TMAZ, in both the AS and the RS. More specifically, in this case, the NZ was characterized by the occurrence of very coarse irregular grains, mixed with fine recrystallized grains strips located throughout across the NZ, from the upper to the lower surface. The primary factors leading to the occurrence of abnormal grain growth process, during FSW, are associated with inhomogeneous gran deformation.

**Figure 13.** Montage of POM RT-FSW PWA at *R* = 0 (conventional FSW), 0.5, and 1 mm.

## *5.2.4. Defect and void formation during FSW*

Traces of the presence of oxide layers (the lazy S-lines) are evident in the NZ microstructure (**Figures 12** and **13**). These, actually, follow the location of the fine grain strips. It thus appeared that the fine grains are formed where the oxide layers, the lazy S-line oxides, are present, and were formed at *R* = 1 mm of RT-type FSW. Thus, it appeared that the fine grain strips, at *R* = 0.5 mm, are being formed along already existing lazy S-line oxide, which formed during FSW.

**6. Concluding remarks**

with respect to the FSW;

measured on conventional FSW.

DS-FSW:

was obtained;

RT-FSW:

energy.

**Author details**

University, Ancona, Italy

view, were obtained and are here summarized.

the characteristic welded zone were quite similar;

Marcello Cabibbo1\*, Archimede Forcellese1

\*Address all correspondence to: m.cabibbo@univpm.it

2 Università degli Studi e-Campus, Novedrate, Italy

In this contribution, two novel approaches and methodologies of friction stir welding on aluminum alloys were presented. The first approach consists of a double-side FSW (DS-FSW). The second approach is represented by a radial deviation of the rotating pin from its centerline, during FSW (RT-FSW). Both new methods were tested in a conventional pin and nonconven‐ tional pinless configuration. Several interesting achievements, from a technological point of

New Approaches to the Friction Stir Welding of Aluminum Alloys

http://dx.doi.org/10.5772/64523

23

DS-i: the elastic modulus and the hardness showed a larger uniformity across the sheet section,

DS-ii: A better formability of the DS-FSW, compared to the conventional pin and pinless FSW,

DS-iii: The DS-FSWed joints are characterized by LDH, and FLC values higher than those

RT-i: The RT setup, for a pin rotation radius of 0.5 mm, induced a low reduction of the mechanical response, compared to the conventional T setup FSW (i.e., with no pin deviation from the welding line). Accordingly, both the microstructure and the hardness profiles of all

RT-ii: The post-weld annealing (PWA) showed the best mechanical response respect to the unwelded annealed AA5754 sheet. The best experimental setups were obtained setting a pin rotation radius *R* = 0.5 mm. In this configuration, UTS was 15% higher, and a ductility reduction of up to 30%, respect to the unwelded annealed sheet. In this condition, the micro‐ structure of the NZ appeared to be characterized by very coarse grains. These coarse grains were generated by geometric dynamic recrystallization (GDR), which is induced by the combined effect of shoulder pressure (heat input), and post-welding annealing (PWA) thermal

1 DIISM – Department of Industrial Engineering and Mathematics, Marche Polytechnic

and Michela Simoncini2

#### *5.2.5. Mechanical properties and hardness modifications induced by pre- and post-welding annealing*

Typical stress-strain curves are shown in **Figure 14**. It appeared that the closest mechanical response to the unwelded annealed AA5754 sheet is obtained by welding with *R* = 0.5 mm in the PWA condition, where UTS differed only by 5%, and ductility differed by 30% with respect to the ductility of the unwelded annealed condition. In the other conditions, the UTS remained within a range of 14% of difference, with respect to the annealed sheet, with a ductility reduction ranging from 76 to 30%.

**Figure 14.** Tensile stress-strain curves for RT-FSW, in the AA5754-O stare and in the PWA condition, at *R* = 0 (conven‐ tional FSW), 0.5, and 1 mm.

Therefore, based on the microstructure evidence, and the obtained hardness and mechanical response, the use of a RT-type welding motion is justified when the plate is homogenized prior, or, even better, after FSW. Conversely, there is no need to deviate the pin, from its welding centerline, in the case of non-annealed AA5000 FSW.
