Effect of faying surfaces and characterization of aluminium AA6063–steel AISI304L dissimilar joints fabricated by friction welding with hemispherical bowl and threaded faying surfaces

This work describes the effect of newly introduced faying surfaces on the microstructure and the mechanical properties of dissimilar weld joints of AA6063 and AISI304L alloys that are fabricated through the rotary friction welding process (RFW). The experiments were done as six different experimental methods (‘A’ to ‘F’) at 1300 rpm rotation, 18 MPa friction pressure (FP), 24 MPa upset pressure (UP) and 5 s friction time (FT) with the faying surfaces of hemispherical bowl and thread of 1 mm pitch on the weld specimens. The fabricated joints and the weld zones were characterized by macro and microstudy, energy dispersive X-ray spectroscopy (EDS) spectrums, tensile properties, Vickers microhardness, impact toughness and fractography. The results showed that these faying surface modifications strengthen the bonding between the weld specimens and influences the performance of the joints. The hemispherical bowl showed better results than the threaded surfaces. Axial shortenings were within the acceptable limit in the range of 20–27 mm. Macro and microstructural studies showed the defect-free weld joints and the strong bonding between AA6063 and AISI304L alloys. The hemispherical faying surface on AISI304L alloy formed a U-shaped weld interface (WI) in the dissimilar joints. EDS proved the formation of the Fe–Al intermetallic and the element ‘O’ at weld zone. The joint efficiency for all the methods was around ≥100%. Maximum tensile strength was recorded as 238 MPa for method F. The threaded surface showed good hardness property nearby WI, and method A yielded maximum impact toughness for the joint.


Introduction
As the different chemical compositions and material properties, joining of stainless steel with aluminium is difficult by fusion welding processes [1]. Friction welding (FW), a nonfusion joining, can be used for joining ferrous-nonferrous dissimilar metals. In this study, RFW was used for joining AA6063 and AISI304L dissimilar alloys with the contacting surface modifications on the weld specimen. The bonding mechanism between steel to aluminium relies on the diffusion of metal atoms at the interface. FW is an eco-friendly solid-state joining process, which means the welding takes place below its melting point [2]. In FW, a joint is made through frictional heating by the sliding action between the two faying surfaces of the weld specimens held together under pressure. In FW, the frictional force is the mechanical energy and the heat generated is the thermal energy. The soft material is expelled out as a flash under heat and pressure [3]. The FP is required to give the axial movement which is one of the main factors affecting the weld strength. The UP is for the consolidation of the weld [4]. The contact pressure of the interface centre is higher than that of the edge and the temperature in the core becomes greater than the perimeter of the weld in FW. The dissimilar metal welds are widely used in boiler, oil refining industries. Many dissimilar combinations have been successfully welded, including steel to aluminium, steel to copper base alloys, stainless steel to nickel-base alloys, etc. There are numerous applications in which welds are made from metals of different compositions considering different properties are required from different parts of the same weldment. This fetches the need for joining dissimilar metals in industries. The dissimilarity in the melting point of dissimilar metals to be joined must be considered in the FW since one metal will be in the molten stage long before the other when subjected to the same heat amount. Zinda et al. [5] researched the frictional joining between aluminium (5A33) and magnesium (AZ31B) alloys and reported that the tensile strength increased with the increase of the FT (5 s) and the formed intermetallics improved the hardness of joints. The effect of FT was identified through their research. Kimura et al. [6] joined Al.6063 and SS304 steel dissimilar alloys through the FW process and achieved 100% efficiency with sound weld joints. Radosław Winiczenko [7] achieved 87% efficiency when joining AISI1020 with A536 dissimilar alloys and found the diffusion of carbon from iron to steel through EDS. It is informed through their research that a small amount of the base metal was heated and that molten metal was thrown from the joint during the FW, therefore, the intermetallic material was kept to a very minimum. The heat-affected zones (HAZ) are also minimal and narrow [8]. A successful weld between dissimilar metals is one, that is as strong as the weaker of the two metals being joined and research stated that the quality of weld surface joint can be checked and the crack can be found with image segmentation technique as the coefficient of friction between the metals during friction welding is prominent [9]. Also, the problem of joining dissimilar metals depends on the quality of the transition zone in the joint and the intermetallic phases that occurred in this transition zone. For example, a discontinuous layer that formed at the joint interface during the friction welding between the pure Al/ Fe metals caused the failure almost at their joints when the load is applied [10]. Similarly, Gawhar and Sherko [11] joined the SS316L steel to AISI1045 steel and found the failure nearby the thermo-mechanically affected zone (TMAZ), also the author informed that the forge pressure is responsible to increase the hardness. The dissimilar joints are made successful only if there is mutual solubility of the two metals. If there is little or no solubility between the two different metals to be joined then the weld joint will not be successful. Sometimes, it is necessary to use a third metal that is soluble with each metal to produce a successful joint. The transition zone has to be examined to determine its crack sensitivity, strength and susceptibility to corrosion [12]. Thermal expansion of metals plays a vital role in the failures of the dissimilar weld by developing high stress at the weld interface [13]. If the coefficient of thermal expansion of both materials is widely different, there would be internal stresses set up in the intermetallic zone during any temperature change of the weldment. If the intermetallic zone is brittle, it may lead to having weld failure soon. Prasanthi et al. [14] successfully welded mild steel with titanium rod of ϕ10 mm × 100 mm size using the FW process. Through their study, they confirmed the formation of fine Fe-Ti intermetallic phase at the weld interface and the hardness of 350 VHN was obtained at the interface. Prashanth et al. [15] did FW of Al-12Si metal generated by selective laser melting process. The results showed the improvement in ductility whereas the drop in hardness in weld interface compared to the base metal. Marlon and Alexandre [16] studied the influence of the surface contact geometry (faying modifications) of joining Al.6351/1020 steel through FW and accepted that the geometry on the contact pin would influence the mechanical properties. Senthil et al. [17] did the experiments of joining AISI304 and AA6063 with tapered faying modification on the AISI304 side and proved the effect of faying surface on the efficiency of the weld joint and the improvement in their joint properties through their research. A novelty of this work is to newly introduce the hemispherical bowl and threaded faying surfaces on the weld metals for the friction welding of dissimilar alloys. In this work, dissimilar joints 2 Materials and methods

Materials
The materials used in this study consist of aluminium AA6063 (T6 tempered) and stainless steel AISI304L alloys. These dissimilar combinations are used in cryogenic applications [18]. The chemical elements present in the materials and their  values in the mass fraction are given in Table 1. The chemical compositions of the materials were confirmed with optical emission spectrography (OES) as per the standard ASTM E1251. The materials were in the form of a cylindrical rod for the welding experiments. AA6063 is a soft material whereas AISI304L having a low percentage of carbon is a hard material. The joining of both materials is a mandatory one in structural applications. The melting point of AISI304L alloy is almost two times greater than AA6063 alloy, so a huge variation is there in the performance while welding. AISI304L is widely used in many applications due to excellent properties [19], and it is also becoming popular for the combination with other material. The properties of weld metals are reported in Table 2.

Methods
The experiments were carried out by a direct drive rotary friction welding machine (model: KAKA, German). Initially, the faying surfaces of the weld specimen of having ϕ12 mm diameter and 100 mm length rods were modified with three different kinds of faying surfaces on weld specimens like a hemispherical bowl (Fig. 1a, d), flat surface ( Fig.  1b, e) and thread of 1 mm pitch for 10 mm long along with a length of faying surfaces (Fig. 1c, f). The cleanliness of the weld specimen is necessary since the contaminants like grease, dust and oil may reduce the quality of joints. According to the experimental description given in Fig. 2, the joints were made with different surface modifications. A schematic view of the welding process followed in this study is shown in Fig. 3. The joints were prepared at 1300 rpm, 18 FP, 24 UP, 5 s. FT, 3-s upset time (UT) and 3 mm/s feed rate. The welding parameters were selected based on the trial and error methods and the literature survey. The recommended parameter values were chosen for this study. The upset pressure is higher than that of friction pressure for its better efficiency. After the fabrication of dissimilar welds, the samples were prepared from the welds as per ASTM standard for characterisation and mechanical testing. The ASTM E8 standard ( Fig. 4a) and ASTM E23 standard (Fig. 4b) were followed for tensile and impact testing, respectively. The macrostructure and microstructure of joints were taken with the help of an optical microscope (OM) and scanning electron microscope (SEM). The Energy dispersive spectroscopy (EDS) analysis was also done on the weld zones. SEM (Model: Geminis SEM 300, Carlzeiss) affords the images at 15 kV with the apertures range: of 10 to 300 μm. The hardness of the weld zone of each experimental methods was measured using a Vickers hardness angle. An indenter will move on this scale when a pendulum is allowed from its horizontal static position to impact the V-notched specimen. There is a stand at the bottom of the machine where the V-notched specimen is supported as a beam in a horizontal position. In addition to that, the fracture analysis was also done on the tensile and impact tested samples using SEM equipment.

Observation and axial shortening
The friction welded dissimilar welds are shown in Fig. 5a along with its weld joint images ( Fig. 5b- it is observed that the faying surface modifications may also show their effect on the metal flash expelled during the application of frictional force during the friction welding. The experiment with a hemispherical bowl shape on AISI304L (methods 'E' and 'F') shows the maximum volume of flash expelled during the joining. But, the hemispherical bowl shape on AA6063 (method 'D') gives not much flash volume in the weld joint as compared to methods E and F. From Fig.  5a, it is observed that the axial shortening in the AA6063 side is much more than that of the AISI304L side. As the melting point of AA6063 is lower than that of AISI304L, the flash may contain aluminium at the interface. The experimental method A showed a fine weld interface (Fig. 5b). In the experimental method B, the penetration of threaded AISI304L alloy is about 3 pitches into AA6063 alloy (Fig. 5c). But, in method 'C', the penetration of threaded AISI304L alloy is about 5 pitches into AA6063 alloy as the threaded faying surface of AA6063 stimulates the bonding between AISI304L and AA6063 specimens during FW (Fig. 5d). The weld joints of methods D, E and F are shown in Fig. 5e-g, respectively. The specimens were bisected with an EDM machine to analyse the weld interface. The weld interfaces of all joints are shown in Fig. 6a-f. The EDM cutting efficiency was measured in terms of the surface roughness between 0.18 μm -0.44 μm. In Fig 6a, a threaded portion of the AA6063 side is shown. Figure 6b shows the penetration of threaded stainless steel into the aluminium alloy. The thread on the AA6063 side (joining method C) caused to have more insertion of threaded AISI304L alloy into the AA6063 specimen as shown in Fig.  6c. The joint with method D had a straight interface as in Fig.  6d even though the AA6063 side has a hemispherical bowl during FW. The U-shaped weld interfaces were seen in Fig.  6e, f for methods E and F, respectively. But the AISI304L alloy with hemispherical bowl only creates the U-shape weld interface. Figure 7 shows the axial shortening of different weld joints observed during FW. The maximum shortening of 26.5 mm was regarded for method E as the effect of the hemispherical surface of the AISI304L side.

Macrographs of weld zones
The weld zones of friction welded dissimilar joints were characterized by macrographs, optical microstructures (OM) and scanning electron microscope (SEM) images, and point and line EDS spectrum with colour mapping. The characterization results are presented here. Figure 8af shows the macro-images of welded joints with different faying surface modifications under ×10 magnification. The images were taken using a stero microscope following the ASTM E340-15 standard. But in some cases, the etchant "Aqua regia" was used for identifying the defects if any and the clear weld flow on the specimen. The images reveal the complete fusion between the weld specimen and the base metal. No defects were found in the weld zone. In Fig. 8a, the edges of the threaded AA6063 and the penetration of AISI304L into AA6063 were seen. The weld interface was showing a good bonding between the dissimilar alloys. Figure 8b shows the two pitches of AISI304L penetration into the AA6063 alloy and incomplete bonding on one of its edges between AA6063 and AISI304L. In Figure 8c, in one of the weld joint edges, the three pitches of the AISI304L specimen were having unsatisfied contact with the pitches of AA6063 alloy but the other edge shows the good bonding between base alloys. From the images, it was understood that the threading was stimulating the bonding between the base metals. Image 8d shows the macroimage of the weld joint fabricated with method D and the weld interface is good with no defects. Though the AA6063 specimen was having hemispherical bowl faying surfaces, it was not seen in Fig. 8d. This is due to the rotational impact of the AISI304L specimen with 18 MPa FP. Though there was an incomplete bonding found on one of the weld joint edges of the welded sample joint fabricated by method E (Fig. 8e), the WI is strong and shows good results.
Here, both edges of the weld interface were not in a straight line. Figure 8f shows the good weld joint and the perfect complete bonding with no defects

Optical microstructures and width measurement of weld zones
The weld zones were characterized by an optical microscope. Figure 9a-r shows the optical microstructures of the weld joints (base metals and weld regions) of all experimental methods taken using a metallurgical microscope, Dewinter optical tech following ASTM E3-11 (RA 17) standard. Under the ×200 magnification, microstructures show very good bonding between the base metals. The weld zone of joints consists of a WI, TMAZ [20] and HAZ. The widths of these zones were measured and the graph was plotted with their values as shown in Fig. 10. To find the clear microstructures, the etchants used for AA6063 and AISI304L were Kroll's and Glyceregia, respectively. From the microstructures' images, the AA6063 base metal shows fine Al-Si eutectic particles in a matrix of Al. solid solution. Whereas, the AISI304L structure shows fine-twinned austenite grains with carbide precipitation. For method A (Fig. 9a-c), the width of WI was about 33 μm and the HAZ and TMAZ were observed nearby WI. The WI was narrow and grains nearby WI of base metals were modified due to the FP. Figure 9d-f, the width of WI was very narrow at 24 μm. The WI shows the wavy-like format which might be due to the threaded faying surfaces on the AISI304L specimen. The HAZ was also regarded as 57 μm and comparatively lower than method A. From Fig.  9g-i, a very narrow WI of 14 μm is observed. The elongated and the refined grains were seen nearby WI. The narrow HAZ of 20 μm size is also found here. Figure 9j-l shows the structure of method D. From the images, the widths of WI, TMAZ and HAZ were measured. A very narrow TMAZ of around 25 μm size was identified. Figure 9m-o is of the weld joint with the joining method E and the WI is seen. The narrow WI and TMAZ were identified with 22 and 11-μm size, respectively. The deformed and elongated grains were also seen in the images nearby WI. Figure 9p-r shows the AA6063 bases, WI and AISI304L base structures, respectively. Eutectic particles are identified on the AA6063 base side and austenitic grains with carbide precipitates are identified on the AISI304L side. The HAZ width maximum of 140 μm was measured and the WI of the strong joint was observed as 37 microns. In Fig.  10, it is shown that the width of the TMAZ is smaller than HAZ and the joining method C shows the minimum width compared to others.

SEM images of weld joints
The SEM images of the weld interface of the various dissimilar joints for the experimental methods A to F are given in Fig.  11a-f, respectively. The structures (Fig. 11) were taken at the central core and middle of the weld interface ( Fig. 8) with 200 μm scale. In the SEM images, the weld joint lines were mentioned as 'WI' which means dissimilar weld interface. The SEM images show no defects along the WI line. In Fig. 11a, the diffused aluminium is seen nearby WI, which is due to the hard force given by AISI304L alloy during FW. So it is decided that the threaded face of AA6063 alloy stimulate the bonding between stainless steel and aluminium alloys weld specimens. In Fig. 11b, the thread-like WI is shown; it may be due to the threaded effect of the AISI304L specimen. But here no diffused AA6063 alloy is seen nearby WI. Figure 11c has some elongated grains nearby WI with no AA6063 alloy debris in WI and  Figure 11d has micron-size elements in the weld zone with no defects. But in Fig. 11e, the AA6063 side nearby WI got forced by the hemispherical bowl faying surface of AISI304L alloy during FW and then bonding was created between the weld specimens. This hemispherical faying surface created a thicker WI than that of the threaded faying surface. The diffusion of elements nearby WI on both sides is seen. Figure 11f shows the bonding between both hemispherical bowl faying surfaces of AA6063 and AISI304L weld specimens. Though the faying surface of AISI304L in method F  Fig. 11f seems to be different compared to Fig. 11e belonging to method E. This difference takes place mainly due to the faying surface of AA6063 part as in method F. From the SEM resuts, the thickness and the shape of the WI and the weld zone property are changine depending on the faying surfaces of the weld specimens. It is inferred that the faying surface modifications can influence the bonding in weld boundary of the dissimilar joint.

Line EDS analysis on weld zone of different joints
Energy-dispersive X-ray spectroscopy analysis like line elemental analysis, point analysis and colour mapping analysis  was performed on the friction welded dissimilar joints to investigate the weight per cent of elements present and their distribution along the weld zone of the dissimilar joint on both AA6063 and AISI304L sides; besides, to find the phases that occur at the weld interface during the friction welding. Figure 12 (a-d6) shows the description of the line EDS analysis along the weld zone area of the fabricated joints. In the figure, the FE-SEM images of weld regions where the EDS analysis was done on the weld joint along weld zone are shown as images in Fig. 12 like (a, a1, a2), (b, b1, b2), (c, c1, c2, c5, c6) and (d, d1, d2) for the experimental methods A, C, E and F respectively. For methods 'B' and D, the focusing was not good on the weld joint by the equipment during scanning so it could not be able to take line EDS report for the joints fabricated through methods B and D and the results are not presented in this paper. In Fig. 12 (a), the line EDS was analysed along the regions 'AB' and 'CD' and the results are presented. Figure 12 (a3 and a4) is the element distribution along AB and CD regions, respectively. Similarly, the spectra of elements that present at AB and CD are reported in Fig. 12  In Fig. 12 (b, b1, b2), the expelled AA6063 content during FW is seen. The line EDS analysis (Fig. 12 (b3 and b4)) shows the element distributions along the regions 'EF' and 'GH' is given in different colours. Very few traces of 'Ag' and 'Mg' elements were seen, and according to line spectrum images ( Fig. 12 (b5 and b6)), the 'Al' content is a little bit higher than 'Fe' content. Figure 12 (c) belongs to method E which was having a hemispherical bowl shape on its AISI304L specimen during FW. Due to the hemispherical shape of AISI304L, the U-shaped strong weld interface is shown in Fig. 12 (c). Here, the weld boundaries (IJ, KL- Fig. 12 (c1 and c2)) and HAZs of AISI304L (MN- Fig. 12 (c7)) and AA6063 (OP- Fig. 12 (c8)) were analysed. The line spectra by the analysis are present in Fig. 12 (c5, c6, c11 and c12). The presence of iron (Fe) content is much higher in the IJ region (c2, c4) than in the KL region (c1, c3). The element distribution phases present along HAZ are shown in Fig. 12 (c9-c12). Figure 12 (d, d1 and    interface (point A in Fig. 13 (a)) while the HAZ of AA6063 side (point C in the Fig. 13 (a)) showed the presence of Al, O and C only. But, the HAZ of AISI304L (point B in Fig. 13 (a)) has Fe, Cr, C etc. as shown in Fig. 13 (a2). Figure 13 (b) is the SEM image of the joint by method B. From Fig. 13 (b1), it is observed that method B has majorly Fe, Cr, Ni and C; but in the HAZ of AA6063 side (point B in Fig. 13 (b)), the Fe presence of 5.3% is shown, which proves the element transfer of iron from AISI304L side to AA6063 side crossing the weld boundary during FW. Similarly, the presence of O is also recorded in Fig. 13 (b3). From Fig. 13 (b2), the chromium content of around 18% is proved along with 7.5% Ni in the weld interface of the joint. It showed the improvement of corrosion resistance of the joint. Figure 13 (c) is the image of the experimental method C. The HAZ of AA6063 and AISI304L sides are shown in Fig. 13 (c1) and (c2), respectively, and the weld interface elements are shown in Fig. 13 (c3). Through the point EDS analysis, it is known that the joint for the methods E and F were analysed and their elements in wt. % are shown in Fig. 13 (e1-e3) and (f1-f4), respectively.

EDS colour mapping on weld zone
The colour mapping and their EDS analysis for the joints fabricated by the methods A, B, C, E and F are denoted by Fig. 14 (a-c), The good bonding between the alloys is confirmed.

Tensile properties and tensile fracture analysis
The fabricated weld joints through the new experimental methods A to F were initially dropped from 1-m height for the preliminary qualifications to check their strength using the drop test [21]. Once the joints pass the drop test, then they have undergone mechanical testing. The welded dissimilar joints after the tensile testing are shown in Fig. 15. The friction welding parameters determine the tensile strength of the joint [22]. In the figure, the plastic deformation and neck formation can be seen in the metal parts. All the breakage happed to the outside of the weld during the tensile testing. While welding, the aluminium at semi-solid state extruded at the interface and tied up with frictionally heated AISI304L metal faying surfaces due to the metal affinity property of both Al and Fe. Based on the mechanism, the metal bonding was formed with

Map spectrum for the method 'B'
Map spectrum for the method 'C'

AA6063 AISI304L
Joining method 'C' Fig. 14 (continued) AA6063 specimen. Like this, the methods with the threaded faying surfaces on the AISI304L side gave good results. Though the faying surface modifications on the AA6063 side did not give the results equal to the results of the joints with the faying surface modification on the AISI304L side, the threaded and hemispherical faying surfaces on the AA6063 side stimulate the bonding between AA6063 and AISI304L during FW. The tensile test results are reported in Fig. 16 (ag). The maximum 238 MPa tensile strength was recorded for the joint fabricated by method F. But the method E also produced almost equal to the strength of method F's joint. Similarly, the yield strength was also noted and given in Fig.  16  hemispherical faying surface produced maximum strength than the threaded faying surface modifications. Figure 16 (b) shows the elongation that occurred during tensile testing of the welded joints. Even though the strength is good, it showed low elongation on the joints except for methods E and F. Joining method E had the maximum elongation (%) with the value of 13% followed by method F with 11%. This elongation is higher than that of AA6063-T6 base metal ( Table 2).
The methods A to D may reduce the elongation and the plastic deformation on the joints in this work. Figure 16 (c) and (d) has the relationship of tensile and yield strength because this relationship decides the elongation and the strain hardening behaviour of the joints. From Fig. 16 (c), the hemispherical faying modifications have a much better tensile-to-yield strength (TS/YS) ratio and a maximum of 1.4 for method E. If this ratio is higher, it means that the ductility and strain hardening nature [23,24] before the ultimate strength is higher for the joints. The TS and YS differences are also given in Fig.   16 (d). Methods A to C show almost the same TS~YS difference, but method E has a much better value. Figure 16 (f) is the peak value during testing. The maximum value of 15 kN was recorded for methods E and F, which means that a hard effort was given by the joints of methods E and F against the pulling load during the tensile test. It is mandatory to measure the weld joint efficiency to identify the effectiveness of the experimental welding methods and the quality of the joint considering the tensile strengths of both the joint and the soft base metal in the weld joint [25,26]. Here, the joint efficiency of the methods is calculated based on the ratio of tensile strength of the fabricated dissimilar joint to the tensile strength of the AA6063 base metal ( Table 2). The experimental methods own almost equal or greater than 100% with a maximum of 116% for method F followed by method E with 115%. It is a good sign of using these kind of experimental methods in industries. In this study, the shear yield strength is mathematically calculated from the yield strength of the joints Map spectrum for the method 'F'

Ts~Ys difference (d)
as per Von Mises yield criterion following Eq. (1) [27] and the calculated shear values are in Fig. 16 (g).
where 'τ' is the shear yield stress and 'σ Y ' is the yield strength.
After the tensile testing, the fractography was studied on the tensile fractured specimens using SEM equipment. Figure 17 shows the fractured images of both AA6063 and AISI304L sides of the fractured weld joints; those were fabricated through methods A to F. The specimens with A to C methods show the mixed fracture and sliding behaviour of the metal. But, the specimens with methods D to F show the dimple rupture and also have the AA6063 debris on the AISI304L side. The ring pattern that formed during the welding can be seen in the images particularly in methods A to C. The plastic deformation and neck formation can be seen in the images of methods E and F. The chance of having dimple rupture on tensile fracture was more by the hemispherical faying surface rather than the welding with threaded surfaces.

Microhardness variations
The sample of dissimilar joint shown in Fig. 18 is suitable for determining the hardness across dissimilar interface [28]. The hardness was measured on both AA6063 and AISI304L sides of the joint along with the regions like WI, TMAZ and HAZ and base metal (BM) from the mid-axis of the weld joint under 300 g loads. The values were obtained as shown in Fig. 18 along the x and y horizontal directions from the weld interface. The determined microhardness distribution values of all the experimental methods A to F are given in Fig. 19. In the figures, the weld interface line is considered the reference line. AA6063 side has the hardness value in the range of 50-85 Hv; whereas, AISI304L side has 250-350 Hv. The reason for the low hardness on the aluminium side is the soft nature of the metal. When analyzing the weld zone of the AA6063 side, the hardness nearby WI is better than that of AA6063 alloy BM except for method B, where the BM has a little higher value. Method A (of the threaded faying surface) has a maximum hardness value nearby the WI of both AA6063 and AISI304L AA6063 side AA6063 side AISI304L side From this research, it is considered, apart from the welding parameters, the faying surface modifications on the weld specimens can control the hardness property of the joints. Figure 20a, b is the hardness analysis for joining methods along with BM, HAZ, TMAZ and WI on AA6063 and AISI304L sides, respectively. From Fig. 20a, method D has a maximum hardness nearby WI; this is because of the hard force given by the AISI304L part during the frictional rotation as the hemispherical faying modification on the AA6063 side improved the hardening property of joining nearby WI by increasing the fine-grain boundaries. From Fig. 20b, method A showed a maximum hardness nearby WI followed by B and E methods. It is hereby identified that the hardness was not uniformly changing from WI to BM; it changes depending on the faying surface modifications. Table 3 is the overall comparison of the performance of different experimental methods in terms of microhardness from the BM of the AISI304L region to the BM of the AA6063 region. Considering the AA6063 side, the hardness is decreasing from WI to BM of the joint. But, AISI304L side hardness is decreasing from WI region to BM region for the methods A and B, but for the other methods, it is vice versa. On the WI side, the maximum and minimum hardness values were recorded as 348 Hv and 295 Hv for methods A and E respectively. While, in WI of AA6063, a maximum of 88 Hv was recorded for method D and a minimum of 67 Hv for method B. Thermo-mechanically affected zone shows better   Fig. 19 Vickers microhardness distribution on weld specimens for experimental methods 'A' to 'F' values than the heat-affected zone and the values 241 Hv and 56 Hv were recorded as the minimum for the heat-affected zone region of AISI304L side and AA6063 side, respectively.

Charpy V-notch impact test and impact fracture analysis
The impact test was done on the weld joints that were produced by the different joining methods. This test is to find out the impact toughness of the weld joints by checking their capability to observe the sudden impact loads. As told, the experimental methods were classified based on the faying surface modifications on the weld specimens. The impact energy (in Joules) recorded during Charpy testing is given in Fig.  21, and the fractured images of impact tested specimens are shown in Fig. 22a-f. Method A showed a maximum energy of 30 J. Whereas, the B and C methods threaded on the AISI304L side showed a minimum of 10 J and 12 J, respectively. This means that the threaded is not giving ductile formation on the joints during testing. But fortunately, all the hemispherical bowled faying modifications showed good impact energy absorption especially method E which gained 26 J energy. From Fig. 22, method A (Fig. 22a) showed a mixed brittle and ductile fracture, but B (Fig. 22b) and C (Fig. 22c)  Fig. 20 (a, b) Microhardness of different regions at dissimilar joint by methods 'A' to 'F': (a) AAz063 side and (b) AISI304L sides  Fig. 22b, which showed insufficient bonding between AA6063 and AISI304L specimens by method B. Figure 22d-f belongs to the hemispherical bowl faying surfaces on the specimens. The shape of the hemispherical bowl faying surface is seen in Figure 22d, e, but it is not seen in Fig.  22f in which AA6063 and AISI304L both faying surfaces had a hemispherical bowl. From Fig. 22e, method E has a ring pattern in its core as it has the hemispherical bowl surface on its AISI304L side. It is mentioned that the faying modifications on the weld specimens can influence the toughness property of the weld joints. The impact-tested specimens were further studied for their fracture behaviour. The fractured SEM images are shown separately for AA6063 and AISI304L sides for different experimental methods in Fig. 23. Impact pendulum struck on the Vnotch. Though method A (Fig. 22a) has a sliding nature on the AA6063 side, the aluminium 6063 debris was sticking on the AISI304L side. Methods B and C show the brittle fracture on both AA6063 and AISI304L sides. In method D, the ductile formation was seen on both sides and the mixed fracture has also been seen in the images. But the methods E and F show the dimple fracture since it tried hard against the fracture with the help of formed ductility. It improved the toughness property of the joint. Thus, the excellence of different faying surfaces on the impact toughness is described. The nature of the metals on which the faying surfaces made in the dissimilar welds is also to be considered during the specimen desing in FW.

Conclusions
In this study, AISI304L and AA6063 materials were successfully joined through the RFW machine with the hemispherical bowl and threaded faying surface modifications on the weld specimens. The effectiveness of the experimental methods and the weld joints were investigated through various characterisation methods, and the following conclusions may be drawn: & The axial shortening of the dissimilar joints by the methods was less than 30 mm, which is within the limit