Journal Pre-proof FRACTURE MECHANICS AND MECHANICAL BEHAVIOUR IN AA5083-H111 FRICTION STIR WELDS (FSW) Oluwaseun John Dada PII: DOI: Reference: S2468-2276(20)30003-X https://doi.org/10.1016/j.sciaf.2020.e00265 SCIAF 265 To appear in: Scientific African Received date: Revised date: Accepted date: 23 January 2019 28 November 2019 13 January 2020 Please cite this article as: Oluwaseun John Dada , FRACTURE MECHANICS AND MECHANICAL BEHAVIOUR IN AA5083-H111 FRICTION STIR WELDS (FSW), Scientific African (2020), doi: https://doi.org/10.1016/j.sciaf.2020.e00265 This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2020 The Author(s). Published by Elsevier B.V. on behalf of African Institute of Mathematical Sciences / Next Einstein Initiative. This is an open access article under the CC BY license. (http://creativecommons.org/licenses/by/4.0/) ฀฀฀฀฀฀฀฀฀ Prof. Oluwaseun John Dada FRACTURE MECHANICS AND MECHANICAL BEHAVIOUR IN AA5083-H111 FRICTION STIR WELDS (FSW) 1,2 1 Oluwaseun John Dada Signature EcoSystems Technologies, DLF Center, 12A Pok Man Street, Tai Kok Tsui, Hong Kong. 2 Department of Metallurgical and Materials Engineering, Federal University of Technology, Akure, Ondo State, Nigeria. oluwaseundd@yahoo.com; john.dada@signecotech.org; ojdada@futa.edu.ng ABSTRACT: Friction Stir Welding (FSW) is one of the fastest developing welding technologies being implemented in the transport industry today. Fracture surfaces of AA5083-H111 specimens, which were friction-stir-welded using different parameters, were studied with a scanning electron microscope (SEM). The changes in the microstructure of the weld, the location and type of defects that accompanied each stirring process determined the magnitude of reduction in strength and soundness of joints for flawed samples. The paths of crack propagation and modes of fast fracture under static loads were also determined. The sharp edges of the defects were stress concentrators which aided crack initiation. Understanding of the effect of microstructure and defect on the failure of FSW joints will aid optimization of the process variables, tool design, weld quality assurance and decision making. On the long run, it will boost the confidence of medium-scale manufacturers in the adoption and implementation of this fairly new technology. KEYWORDS: Friction Stir Welds (FSW); Aluminium; Microstructure; Failure mechanism; Defects. ฀฀฀฀฀฀฀฀฀ Prof. Oluwaseun John Dada Bio: Dr Oluwaseun John Dada is Professorial Chair of Signature EcoSystems Technologies. He has worked over over the past 8 years on NanoScale materials, advanced materials manufacturing and applications, and over 10 years in Metallurgy and Aluminium Manufacturing. I have achieved Electrolytic syntheses, characterizations, and applications of novel graphene/nano-Carbons, resulting in Superior Electrical, Electronic, Thermal and Dielectric characteristics; also resulting in the Highest Capacities when applied to Lead Acid and Lithium-ion Batteries, resulting from higher Utilization of Electrode Materials, 1st to demonstrate electrode precipitation and dissolution mechanisms in the gel-crystal electroactive zone of the lead-acid battery; synthesized graphene also resulted in higher viscoelastic mechanical behaviour in graphene-enhanced epoxy nanocomposites prepared by in-situ reduction of GO by solvent evaporation and hot stirring. Synthesized highly scalable and diversely applicable crystalline pristine graphene directly from graphite, with high basal planarity using solid-state processes, or alternative colloidal sonication process. This has resulted in the high-performance graphene anode for lithium-ion battery. On my project on FSW, I have achieved parametric analysis, 1st SEM demonstration of material flow pattern, and failure characteristics of Friction Stir Welds of industrial and aeronautical alloys. Experience dates back to Alussuisse owned Nigerian Aluminium Extrusion, a Laboratory Extrusion project, and AIRBUS FSW project at Wits University. Dr Dada received PhD in Mechanical Engineering Nanotechnology Concentration at Hong Kong University of Science and Technology, Hong Kong. Dr Oluwaseun John Dada is the CEO/ Director of Research and Engineering at Signature EcoSystems Technologies, Hong Kong. Dr. Dada was ORTg, Hong Kong PhD Fellow (Mechanical Engineering, Nanotechnology Concentration), Senate Representative, Entrepreneurship Centre (EC) Venture Entrepreneurship (VEN) Program Manager, and Engineering School Board member at The Hong Kong University of Science and Technology (HKUST). He was on AMSEN (Carnegie-RISE network) MSc scholarship at University of Witwatersrand, and won the AMSEN conference 3rd presentation prize in Nairobi. Professor Dada was Lecturer, 1st Class and Best Engineering School Graduate at the Federal University of Technology FUT, Akure. I am currently an Chief\Editor for 3 materials science journal papers and published 3 books including: 1. Interfacial Performance of Applied Lead/Graphene; 2. Innovative Friction Stir Welds of Aluminium Alloys Characterizations), and contributed to more than 60 scholarly journal publications. ฀฀฀฀฀฀฀฀฀ Prof. Oluwaseun John Dada 1. INTRODUCTION Friction Stir Welding (FSW) is one of the fastest developing welding technologies being implemented in the transport industry today. It utilizes a robust rotating and non-consumable tool comprising a shoulder and a probe, which is pushed into the adjoining surfaces of the plates to stir, plasticise, forge and consolidate the material as it moves along the joint-line (Figure 1) [1–5]. Figure 1. Schematic diagram of the FSW process and forces on the tool [6]. The application of FSW to aluminium structures, especially those made of 5083-H111 aluminium alloy, is particularly significant for body structures of road vehicles and marine automobiles such as fast ferries and catamarans where higher speed, lower weight and energy savings is essential. AA5083-H111 has been notable for these applications because of its high strength, formability, weldability and corrosion resistance. Cost-effectiveness, ฀฀฀฀฀฀฀฀฀ Prof. Oluwaseun John Dada environmentally friendliness, enhanced superplasticity and superior mechanical properties have made FSW preferred to other techniques like metal inert gas (MIG) and tungsten inert gas welding (TIG). Also, FSW’s application scope has been extended beyond aluminium, to other materials such as steels, titanium, and polymers. FSW results in the creation of the following microstructural zones:  Weld nugget, which has undergone mechanical stirring, resulting in severe plastic deformation and recrystallization caused by frictional heat generated by the tool in the material workpiece [7, 8]. It has fine equiaxed grains in a banded structure. The flow arm zone is at the top of the weld nugget, where the grains are finer as a result of more intense stirring from the tool shoulder.  Thermo-mechanically affected zone (TMAZ) is located between the weld nugget and HAZ. It has undergone some plastic strain without recrystallisation. Visible elongation and rotation of the parent material grain structure occur in this zone during welding [9].  Heat affected zone (HAZ), which is typical in all welds; the material has undergone some thermal cycling, but no deformation and has a different microstructure from the base material. The difference in local properties of these microstructural zones and the presence of defects makes mechanical behaviour and failure of FSWs complex. Rotational and transverse speeds of the tool have been identified as key welding parameters that influence microstructure, mechanical behaviour and defect formation. The industrial viability of FSWs depends on the mechanical properties and the effect of defects. FSW AA5083-H111 of less than 5mm thickness, which is less studied compared to other alloys in the aluminium series such as 2XXX, 6XXX and 7XXX, was used in this investigation. The mechanical and microstructural ฀฀฀฀฀฀฀฀฀ Prof. Oluwaseun John Dada characteristics of FSWs of 2.8 mm thick AA5083-H111 and their effect on tensile fracture and failure mechanisms were ascertained. 2. EXPERIMENTAL PROCEDURE AA5083-H111 plates of 360mm × 120mm × 2.8mm, supplied by Hulamin Ltd, were welded using a cylindrical shoulder of 20mm diameter and tapered probe of 6mm diameter on a converted CNC milling machine along the direction of rolling in the plates. The FSW tool was made from H13 tool steel. The chemical composition of AA5083-H111 is given in Table 1. The surfaces of plates were cleaned with acetone to degrease them prior to welding. Combinations of rotational speeds between 400, 500 and 600 rpm and welding speeds between 50, 60 and 70 mm/min were used to produce six samples in Table 2. Table 1. Chemical composition of AA5083-H111 (wt%). Elements Si Fe Cu Mn Mg Cr Zn Ti Al wt % 0.12 0.33 0.03 0.51 4.39 0.08 0.01 0.02 Balance Table 2. Specimens and their welding parameters. FSW Samples A B C D E F Rotational speeds (rpm) 400 400 500 630 630 630 Welding speeds (mm/min) 50 60 50 50 70 60 Pin loaded dog-bone specimens were prepared according to ASTM standard E8/E8M. Transverse static tests were performed on an ESH hydraulic testing machine at a strain rate of 0.007mm/mm/s. Fracture surfaces were analysed using the Zeiss LEO 1525 Field Emission Gun Scanning Electron Microscope (FEGSEM) in mainly secondary electron mode. ฀฀฀฀฀฀฀฀฀ Prof. Oluwaseun John Dada Metallographic preparation FEG-SEM according to Struer’s guide [10] and etching was done using Keller’s reagent. A Leica microscope was used to observe the microstructure. The sizes of defects in the flawed samples were measured across the weld. Microhardness profiles across the welds were obtained using FM-700 Future-Tech Vickers hardness tester with a 300kgf load and 15s dwell time. 3. RESULTS AND DISCUSSION 3.1. Microstructural characterisation All expected microstructural zones such as the weld nugget (WN), thermo-mechanically affected zone (TMAZ) and heat-affected zone (HAZ) were present. There were no clear boundaries between the zones at the retreating side of the weld tool (LHS of Figure 2), although the size of the grains increased from the weld nugget outwards. On the advancing side of the weld tool, there was a clear boundary between the nugget and the HAZ (RHS of Figure 2). A higher shear rate in a narrower band, which resulted in a sharp change in microstructure was reported to be responsible for the advancing weld nugget/TMAZ interface [11]. A joint-line remnant, kissing bond, or “lazy-S”, which appeared as a thin dark line more towards the advancing side, was found in the weld nugget. It has been identified to comprise amorphous clusters of alumina both within grains and at grain boundaries [12], because it comprises the surface of aluminium which is nature of surface alumina, which are grains attacked by atmospheric oxygen. The finest grains were found at the crown of the weld in the flow arm zone due to additional torque and frictional heating from the tool shoulder. Tunnel defects were found at the base of the weld nugget/TMAZ interface on the advancing side (Figure 3) in samples B-F and their sizes are shown in Table 3. Pinholes are in samples D-F (Figure 4). Root flaws were also found at the root of all the welds. ฀฀฀฀฀฀฀฀฀ Prof. Oluwaseun John Dada Figure 2. Light microscope image through the FSW AA5083-H111, showing the microstructural zones in Sample A. Figure 3. Light microscope image through the FSW AA5083-H111, showing tunnel defect at the base of the weld nugget/TMAZ interface (Sample A). ฀฀฀฀฀฀฀฀฀ Prof. Oluwaseun John Dada Figure 4. The microstructure of Samples B-F showing positions of tunnel defects and pinholes. Table 3. Average sizes of tunnel defects in Sample B-F. Specimen Tunnel Size (µm2) B C D E F 17.8 29.3 63.3 35.8 37.7 3.2. Micro-hardness profiles There was an increase in hardness in the weld zone, more towards the retreating side, due to grain refinement and better weld consolidation (Figure 5). There was no discernible ฀฀฀฀฀฀฀฀฀ Prof. Oluwaseun John Dada relationship between weld parameters and hardness profiles. The hardnesses towards the retreating side are higher, while hardnesses are lower towards the advancing side due to greater frictional heat and grain enlargement, while there is reduced frictional heat in the retreating side. The grain size differences follow hall-perch relationship resulting in the slope in hardness profile inverse to the direction of tool flow. Figure 5. Microhardness profiles of Samples A-F across the weld nugget. 3.3. Static testing and failure The average static properties of the welds are given in Table 4. The fracture paths in Samples A-E were similar. The welds were expected to fail in the base material since the hardnesses there were lower than in the weld. The presence of tunnel defects changed the fracture location under tensile loading. In Samples A-E, fast fracture initiated due to the stress concentration from the sharp corners of the triangular weld tunnel. Fast fracture propagated in ฀฀฀฀฀฀฀฀฀ Prof. Oluwaseun John Dada two directions: first, upwards at 45°, slightly bending through the TMAZ and HAZ, and finally vertically upward in the flow arm zone. Secondly, fracture initiated from the tool marks and propagated vertically downwards from the right angle of the tunnel through the HAZ. Sample F fractured almost along the joint line, despite the tunnel defect at the base of the weld nugget/TMAZ being larger than the pin-hole at the centre of the weld nugget (Figures 5-6). This showed that the pinholes in the nugget determined the fracture location, and were more deleterious to the ductility and UTS (Tables 3-4, Figures 4,6-7). Sharp corners of the pinholes were stress concentrators for fast fracture within the weld nugget. Liu et al. [13] deduced that the tensile failure of FSWs near the weld nugget/TMAZ interface was due to significant differences in their internal structures which made the interface a weaker region. This was not found in the current study. The fracture path was bent away from this interface, despite the presence of tunnels defect with sharp corners in Specimen A, as shown in Figure 3. There was no evidence of weakness at the weld nugget/TMAZ interface, despite the microstructural differences. Fracture of welds at locations where microhardness was higher, as in this case, was an indication of the presence of defects. In this case, the fracture path approached the areas of lowest hardnesses in the weld, in agreement with Liu et al. [14], Biallas et al. [15] and Okamura et al. [16]. ฀฀฀฀฀฀฀฀฀ Prof. Oluwaseun John Dada Table 4. Static testing results of the welds and parent material. SPECIMEN PARENT A B C D E F UTS (MPA) 295 265 246 230 245 172 300 0.2% PROOF STRESS 140 135 139 138 139 135 145 STRAIN-TO-FAILURE 0.3 0.12 0.08 0.08 0.07 0.02 0.23 WELD EFFICIENCY (%) 98.3 88.3 82.0 76.7 81.0 57.0 100 MATERIAL PROPERTIES (MPA) Figure 6. Pictures of fractured tensile specimens A-F of Al Friction Stir Welded (FSWs) with different parameters. ฀฀฀฀฀฀฀฀฀ Prof. Oluwaseun John Dada (a) (b) Figure 7. Schematic diagrams of the fracture path for (a) Samples A, and (b) Sample F. 3.4. Fracture surface analysis Weld tunnels at the roots in Samples A-E ฀฀฀฀฀฀฀฀฀ Prof. Oluwaseun John Dada The overall fracture surface (Sample A, inverted) is shown in Figure 8. Features such as the root of the weld (root is at the top, Figure 9), tool marks (e.g. 1, Figure 9), tunnel defect (indicated by the arrow), part of the onion ring structure (e.g. 4, Figure 9) and the region of fast fracture (e.g. 5, Figure 9) were observed on the fracture surface towards the lower part of the weld (which is inverted). Tool marks and bands of unconsolidated material (lap pattern) that made up the weld tunnels resulted from the reduced bonding between the onion structure and the weld root. The tool mark was an incomplete semi-circular feature created by the motion of the tool on the material at the root of the weld, while the onion ring was created by successive deposition of material, rotated and pushed down around the tool pin, but which failed to be consolidated with the root of the weld at the advancing side of the tool. Similar features were found by Gratecap et al. [17] and Chen et al. [18]. Poor consolidation that produced the weld tunnel resulted from the reduction in fluidity and flow of material to the end of the pin on the advancing side [18]. Also, complete weld consolidation did not take place, because the vertical forging pressure was too low. Tilting the tool, which could have compensated for the low vertical pressure and improved downward flow of material [18], as the tool translated along the joint, could not be done on the current CNC milling machine. The unevenness in the depth of the unconsolidated material in Specimen 1 indicated that the tool lost vertical pressure occasionally while rotating and translating. This showed that position controlled welding was not optimal, if defects were to be avoided. Greater vertical pressure in force controlled welding mode should be necessary to avoid tunnel defects. ฀฀฀฀฀฀฀฀฀ Prof. Oluwaseun John Dada (a) (b) Figure 8.(a) SEM-SE image of the overall fracture surface of Specimen 1 along the weld, (b) SEM-SE image of the lower part of the fracture surface taken along the weld in Sample A. ฀฀฀฀฀฀฀฀฀ Prof. Oluwaseun John Dada The river marks diverging downwards on the surface of the tool marks (Figure 10a) and upwards at the ending of the unconsolidated onion rings (Figure 10b), indicated that they were the primary regions of crack initiation in the weld tunnel. (a) (b) Figure 9. SEM-SE image (Sample A) of (a) one of the tool marks with arrows showing lines indicating the initiation of fast fracture at the root of the weld, and (b) the unconsolidated materials that formed the hypotenuse of the triangle weld tunnel with arrows showing the direction of fracture in the TMAZ. Failure mechanisms and stress states in Sample A The weld root (below the tunnel) was characterized by a large number of small parabolic dimples with large lips on the cleavages, showing a failure by mixed micro-void coalescence and quasi-cleavage (e.g. 2-3, Figure 8). Closer to the base of the weld, there were areas both with and without dimples, indicating the quasi-cleavage mechanism. This is consistent with the results obtained by Elangovan and Balasubramanian [19], who identified poor weld consolidation in the root was being responsible. The grains in this region were fine, and high stresses were needed for failure, due to the increased hardness. Quasi-cleavage fractures are brought about by crack initiation by microvoid coalescence and propagation by tearing of cleavage facets ahead of the moving crack front [14]. ฀฀฀฀฀฀฀฀฀ Prof. Oluwaseun John Dada Above the unconsolidated material region (in the TMAZ/HAZ, Figure 10a), there were more equiaxed and small parabolic dimples. In the middle, there were larger parabolic dimples than those near the root, which indicated larger grains, larger precipitates and higher shear ductility. A large number of larger dimples, which sometimes had precipitated, indicated that the precipitates were active originators of microvoids under applied stresses. In the flow arm zone at the crown (Figure 10b), cleavages carrying a few tiny dimples were found, which indicated failure by quasi-cleavage mode. The precipitates that caused dimples were regions of local crack initiation, but the matrix failed in a brittle manner. Smaller dimples were from the smaller precipitates which were reduced by the extra torque from the tool shoulder. The lower ductility was due to the reduction of grain size by work-hardening induced by tool shoulder. There was a transition from the ductile tear to quasi-cleavage from the middle to the top of the weld fracture surface [20]. This is consistent with Zadpoor et al. [20] where brittle regions were formed in the top of the sheet which underwent severe plastic deformation when in contact with the rotating tool shoulder. Cleavage-like fracture is not common in FCC metals where there are many active slip systems [21]. Zadpoor et al. [20] suggested that high rotational speed could have resulted in the non-optimal grain structure and strain rate that weakened the recrystallized microstructure, and the mixture of ductile and brittle fracture could be the result of the blocking of dislocation movement by the precipitates. Quasi-cleavage fracture is different from cleavage, which is a low energy fracture mechanism that moves through well-defined low index crystallographic planes [21]. Quasi-cleavage is used to describe cleavage that has signs of plastic deformation, although limited [20, 21]. ฀฀฀฀฀฀฀฀฀ Prof. Oluwaseun John Dada (a) (b) Figure 10. SEM-SE image (Sample A) of (a) the middle of fracture surface along the weld HAZ showing dimples and precipitates, and (b) the crown (flow arm zone) of the weld (inverted) showing the fine dimples on cleavages. The shape of dimples on fracture surfaces indicates the responsible stress states [22]. Equiaxed dimples are associated with tensile stresses (Mode I), while shear dimples are from shear or tear stress systems (Mode II and III) and point in the direction of shear [14]. The lips of the shear and elongated dimples at the middle and near the top of the fracture surface pointed upwards (in the direction of the crown of the weld), while those at the root pointed downwards, and this was not associated with normal tensile loading. The presence of tunnels created a response of both tensile and tearing stresses when the tensile loads were applied. The sizes and shapes of the dimples were directly proportional to the sizes of the grains and precipitates in the different zones and ductility [23]. Zones, where the precipitates were large, had larger and deeper dimples (e.g. at the middle of the fracture surface). Where the grains were larger, the lips of the dimples were wider, indicating that the material was more ductile. Fine dimples, which are associated with work-hardening [23], were found in the flow arm zone and in the root, which indicates that more plastic deformation took place in these two ฀฀฀฀฀฀฀฀฀ Prof. Oluwaseun John Dada zones. The different fracture mechanisms observed in the SEM showed that the cracks propagated through different microstructural zones in Samples A-E. (a) (b) Figure 11. Schematic diagrams fracture path with arrows indicating dominant stress systems responsible for fracture in (a) Specimen A, and (b) Specimen F. ฀฀฀฀฀฀฀฀฀ Prof. Oluwaseun John Dada Failure mechanisms and stress states in Sample F The overall fracture surface is shown in Figure 12a. The lips of the dimples at the upper part of the fracture surface in the flow arm zone were elongated downwards (Figure 12b), opposite to the direction of failure. Where the direction of dimple elongation is opposite to that of fast fracture, as in this case, it is an indication of ductile shear fracture [22]. Towards the middle to the root of the fracture surface, the dimples were elongated downwards in the direction of the root, and in the direction of fast fracture in the root of the weld. This indicates that the mechanism of fast fracture here was a ductile tear. At the top in the flow arm zone and at the root, there was a gradual transition from ductile shear and tear to quasi-static fracture. The middle of the fracture surface contained shallow dimples which showed that the precipitates involved in dimple rupture were finer and the linear pattern on the surface was from the weld nugget bands, indicating that fracture occurred across the onion ring structure (Figure 12c). The failure mechanism here was ductile. Smaller and shallower dimples showed that failure in the weld nugget was less ductile than in Specimens A-E, which failed in the TMAZ/HAZ. When the tensile load was applied on Specimen 6, the presence of defects created a system of shear stresses at the upper part of the weld and a tear system at the lower part, which resulted in the fast fracture through the weld nugget (Figure 12b). ฀฀฀฀฀฀฀฀฀ Prof. Oluwaseun John Dada (a) (b) (c) Figure 12. SEM-SE image (Sample F) showing (a) Overall fracture surface, (b) weld crown (flow-arm zone) showing shear dimples on cleavages, and (c) Middle of the weld (in the weld nugget) showing a linear pattern from onion rings. 3.5. Effect of welding parameters on defect formation The defect size had no relationship with the advance of the tool per revolution. Increased welding speed led to decreased vertical force, thereby reducing weld consolidation and causing widening of the tunnel defects. Leonard and Lockyer [24] stated that the material is ฀฀฀฀฀฀฀฀฀ Prof. Oluwaseun John Dada given less work per unit of the weld length when welding speed is increased at a constant rotational speed. Thus, materials are cooler and more difficult to consolidate at higher welding speeds. Also, increased rotational speed at constant welding power of the CNC milling machine led to the formation of a pinhole at the advancing side and another pinhole in the nugget due to poor stirring (caused by sliding) in Specimens E and F. Increased rotational and welding speed has been linked to higher heat input and more intense stirring [25]. However, this is only obtainable when the welding power of the FSW machine is high enough to accommodate the increases, by ensuring other parameters such as the downward and transverse forces are not significantly reduced. Also, increased rotational speed to an optimum, without a constant or reducing downward force (pressure) leads to a reduction in the frictional forces that provides intense heating and stirring [26]. Depending on tool geometry, increased rotational speed causes sliding of the material on the tool. In this case, the tool used was a disadvantage, because it was unthreaded. The absence of a tilt angle on the tool was a major disadvantage. Tilting the tool (2-3°) causes the rear of the shoulder and pin to be lower than the front, which aids the downward flow of the material around the rotating pin. Also, the rear of the shoulder forges the material as the tool moves along the joint line [27]. A deeper plunge depth in position control would have ensured greater tool penetration, reducing or avoiding the root flaw and also achieving forging within the weld. Root flaws could have been avoided or reduced with increased rotational speed, which causes increased stirring at the root and improved weld penetration at constant downward pressure [25]. 3.6. Effect of defects on weld integrity Hardness results showed that AA5083-H111 FS welds were stronger than the base material, but tunnel defects reduced tensile properties. Jamshidi-Aval et al. [28] and Squillace et al. [29] found tunnel defects in joints produced using conical tools; reduced static mechanical ฀฀฀฀฀฀฀฀฀ Prof. Oluwaseun John Dada properties [28], or not [29]. In both cases, the sizes of the tunnels across the welds were not measured. In this study, the size of the tunnel defects affected the static properties. There is an optimum size of tunnel defect below which the overall strength of the weld was unaffected [29, 30]. In Sample A where the pin-hole (discontinuous tunnel) was 2µm2 across, the tensile properties were similar to that of the base material (Table 4). UTS, strain-to-failure and joint efficiency reduced with increasing tunnel size, while yield strength was almost unaffected (Figure 13), in agreement with Chen et al. [18]. The closeness of defects to the centre of the weld (in Sample F, FSW at RS 630rpm; WS 60mm/min) was more deleterious to UTS and ductility. The presence of root flaws, lazy-S’s (A-F) and pinhole at the upper part of the advancing side (E-F) had no noticeable effect on fracture or overall strength of FSW AA5083-H111. (a) (b) (c) Figure 13. Effect of tunnel defect size on UTS, proof stress, strain-to-failure and weld efficiency for FSWs AA5083-H111 (Samples A-E). ฀฀฀฀฀฀฀฀฀ Prof. Oluwaseun John Dada 4. CONCLUSIONS The difference in failure mechanisms in the welds were due to different microstructural zones which differed in grain structure. Tunnel defects were caused by a lack of consolidation of the onion ring structure at the lower part of the advancing side of the tool. This resulted from differences in the rate of material flow away and from the bottom of the advancing side. This study confirmed that fact that low vertical pressure and non-tilting of the tool are the overall causes of tunnels. Tool penetration was inadequate as the height of the root flaw was significantly high. The degree of severity of weld defects: Tunnels, pinholes and lazy-S’s determined the strength. The presence of a pinhole at the weld nugget significantly reduced tensile properties and caused a fracture, even when the tunnel at the weld nugget/TMAZ interface was larger. Thus, the tunnels were more severe than the pinholes, while the lazy-S’s had no effect on quasi-static fracture. Tensile fracture in regions of the weld where hardness values were higher indicated the likely presence of weld defects such as tunnels and pinholes. The presence of sharp transitions as a result of differences in microstructure between the weld nugget and TMAZ at the advancing side had no effect on hardness or fracture path. The weld nugget/TMAZ interface was not a weak region in FSW AA5083-H111, as previously reported, because cracks propagated away from tunnels located along which initiated the failure. At constant welding power, increased rotational and welding speeds resulted in a reduction in vertical pressure that caused increased size, number and severity of weld defects. In all the specimens, the presence of defects transformed the applied tensile loads into tear and shear stresses. Also, there was a transition from ductile shear and tear in the center of the fracture surface to mixed modes and quasi-cleavage at the top and root of the weld. ฀฀฀฀฀฀฀฀฀ Prof. Oluwaseun John Dada ACKNOWLEDGEMENT The author wishes to acknowledge Hulamin South Africa for supplying the materials. Professors Lesley Cornish, Claudia Polese, Tony Paterson of the University of the Witwatersrand, South Africa, and George F. Van der Voort for their contributions. REFERENCES 1. Dada OJ, Polese C, Cornish LA, Adewuyi BO, Borode JO (2012) CHARACTERISATION OF INNOVATIVE FRICTION STIR WELDING JOINTS FOR AIRFRAME STRUCTURES. AMSEN 2. Dada OJ (2012) Characterisation of Innovative Friction Stir Welding for Aeronautical Structures: M.Sc Eng Dissertation Submitted to University of The Witwatersrand, Johannesburg. 3. Dada OJ (2012) SEM Characterisation of Microstructure, Tunnels & Quasi-Static Failure in AA5083-H111 Friction Stir Welds. SAIW/IIW Reg. Congr. 2012 Available SSRN e-Journal 4. Dada OJ (2019) Effect of Microstructure and Tunnel Defect on Quasi-Static Fracture of Friction Stir Welded (FSW) AA5083-H111. 50th Annu. Conf. Microsc. Soc. South. Africa (MSSA 2012) B. Abstr. available SSRN E-Journal 5. Dada OJ (2019) Characterisation of Innovative Friction Stir Welding Joints for Airframe Structures: Research Overview. SSRN e-Journal 6. WEBSTER H. (2011) About FSW. HFW 4–6 7. Mishra RS, Mahoney MW (2007) Friction Stir Welding and Processing. Mater Sci Eng R Reports. https://doi.org/10.1361/fswp2007p001 8. Sato YS, Urata M, Kokawa H (2002) Parameters Controlling Microstructure and Hardness during Friction-Stir Welding of Precipitation-Hardenable Aluminum Alloy ฀฀฀฀฀฀฀฀฀ Prof. Oluwaseun John Dada 6063. Metall Mater Trans A 33:625–635 9. Genevois C, Deschamps a, Denquin a, Doisneaucottignies B (2005) Quantitative investigation of precipitation and mechanical behaviour for AA2024 friction stir welds. Acta Mater 53:2447–2458 10. Weidmann E, Anne G (2008) Struer’s Application Notes: Metallographic Preparation of Aluminium and Aluminium Alloys. Notes 1–6 11. Leitao C, Leal RM, Rodrigues DM, Loureiro a., Vilaça P (2009) Mechanical behaviour of similar and dissimilar AA5182-H111 and AA6016-T4 thin friction stir welds. Mater Des 30:101–108 12. Zhou N, Song D, Qi W, Li X, Zou J, Attallah MM (2018) Influence of the kissing bond on the mechanical properties and fracture behaviour of AA5083-H112 friction stir welds. Mater Sci Eng A. https://doi.org/10.1016/j.msea.2018.02.011 13. Liu HJ, Fujii H, Maeda M, Nogi K (2003) Tensile properties and fracture locations of friction-stir-welded joints of 2017-T351 aluminum alloy. J Mater Process Technol 142:692–696 14. Liu AF (2005) Mechanics and Mechanisms of Fracture: An Introduction. ASM International 15. Biallas G, Braun R, Donne CD, Staniek G, Kaysser WA (1999) Mechanical properties and corrosion behavior of friction stir welded 2024-T3 G Biallas, R Braun, C Dalle Donne, G Staniek and W A Kaysser. Proc First Int Conf Frict Stir Weld 2–11 16. Okamura H, Aota K, Sakamoto M, Ezumi M, Ikeuchi K (2002) Behaviour of oxides during friction stir welding of aluminium alloy and their effect on its mechanical properties. Weld Int 16:266–275 17. Gratecap F, Girard M, Marya S, Racineux G (2011) Exploring material flow in friction stir welding: Tool eccentricity and formation of banded structures. Int J Mater Form ฀฀฀฀฀฀฀฀฀ Prof. Oluwaseun John Dada 5:99–107 18. Chen H-B, Yan K, Lin T, Chen S-B, Jiang C-Y, Zhao Y (2006) The investigation of typical welding defects for 5456 aluminum alloy friction stir welds. Mater Sci Eng A 433:64–69 19. Elangovan K, Balasubramanian V (2008) Influences of post-weld heat treatment on tensile properties of friction stir-welded AA6061 aluminum alloy joints. Mater Charact 59:1168–1177 20. Zadpoor AA, Sinke J, Benedictus R (2009) Fracture mechanism of aluminium friction stir welded blanks. Int J Mater Form 2:319–322 21. ASM International Handbook Committee (1987) ASM Handbook: Volume 12: Fractography. ASM Int Mater Park 91–146 22. Charles JA, Crane FAA, Furness JAG (1997) Selection and Use of Engineering Materials (Google eBook). Butterworth-Heinemann 23. Sattarifar I, Abachi P, Taheri N (2008) INFLUENCE OF MECHANICAL TREATMENTS ON THE MECHANICAL PROPERTIES OF AA5083 ALUMINUM ALLOY. In: 17th Int. Conf. Metall. Mater. 13th to 15th 2008 Hradec Moravicí. pp 1–8 24. Leonard A., Lockyer S. (2003) FLAWS IN FRICTION STIR WELDS. 4th Int. Symp. Frict. stir Welding, Park City Utah, USA, May 14-16 25. Barlas Z, Ozsarac U (2012) Effects of FSW Parameters on Joint Properties of AlMg3 Alloy. Weld J 91:16–22 26. Hussain AK (2010) EVALUATION OF PARAMETERS OF FRICTION STIR WELDING FOR ALUMINIUM AA6351 ALLOY. Int J Sci Technol 2:5977–5984 27. Hattingh DG (2012) Private Communication. Inst. Adv. Manuf. Eng. Res. (IAMER), Fac. Eng. NMMU, Port Elizab. South Africa 28. Jamshidi Aval H, Serajzadeh S, Kokabi a H (2010) The influence of tool geometry on ฀฀฀฀฀฀฀฀฀ Prof. Oluwaseun John Dada thermo-mechanical and microstructural behaviour in friction stir welding of AA5086. In: Proc. Inst. Mech. Eng. Part C J. Mech. Eng. Sci. pp 1–16 29. Squillace A, Segreto T, Prisco U, Teti R, Campanile G, Engineering P, Federico N, Tecchio P (2006) Optimization of friction stir welds of aluminium alloys. Intell Prod Mach Syst 247–252 30. Jonckheere C, Meester B De, Simar A (2010) Flow pattern and mechanical properties of dissimilar friction stir welds joining 6061-T6 to 2014-T6 aluminium alloys. In: 8th Int. Symp. Frict. Stir Welding, May 2010, Timmendorfer Strand, Ger. pp 1–7