De auteur(s) geeft (geven) de toelating deze masterproef voor consultatie beschikbaar te stellen en delen van de masterproef te kopiëren voor persoonlijk gebruik. Elk ander gebruik valt onder de bepalingen van het auteursrecht, in het bijzonder met betrekking tot de verplichting de bron uitdrukkelijk te vermelden bij het aanhalen van resultaten uit deze masterproef.

The author(s) gives (give) permission to make this master dissertation available for consultation and to copy parts of this master dissertation for personal use. In the case of any other use, the copyright terms have to be respected, in particular with regard to the obligation to state expressly the source when quoting results from this master dissertation.

Ghent, June 2016 Preface

In this preface, I would like to take the opportunity to express my gratitude to the following people, who have helped and supported me during the course of this master dissertation.

Ir. Koen Faes of the Belgian Welding Institute who granted me the opportunity to do this dissertation. Thank you for your help, support and advice during the experiments.

Prof. dr. ir. Wim De Waele for your guidance, involvement and critical point of view.

Kevin Deplus of the Belgian Welding Institute for his practical support on the welding machine.

Gert Oost of the Belgian Welding Institute for his assistance during the preparations of the metallographic samples.

The Belgian Welding Institute for supplying materials and means.

Finally, I would like to thank my parents, for allowing me to continue studying and supporting me throughout my academic career.

Tom Kolba, June 2016

I

Abstract

Friction spot welding is a technique for joining lightweight aluminium and magnesium alloy sheets in the overlap configuration by means of frictional heat and mechanical work and has a high potential for industrial applications. As this is a very recent technique, little information is available regarding the evaluation and optimisation of process parameters for specific material combinations. The objective of this master dissertation is to investigate the weldability of high strength aluminium alloys EN AW-7475-T761, EN AW-6082-T6 and EN AW-7075-T6, aiming to produce high quality joints in terms of mechanical performance and microstructure. More specific, the influence of the plunge depth, rotation speed and welding time was investigated by metallographic examination as well as microhardness and lap shear strength tests. This master dissertation contains a literature study and continues with its own experimental program.

Keywords: EN AW-7475-T761; EN AW-6082-T6; EN AW-7075-T6; friction spot welding; refill friction stir spot welding

II

Experimental investigation of the weldability of aluminium alloys using friction spot welding

Tom Kolba

Supervisors: Wim De Waele and Koen Faes

Abstract Friction spot welding

Friction spot welding is a technique for joining sheets of Process description lightweight alloys in the overlap configuration by means of frictional heat and mechanical work and has a high potential for Friction spot welding (FSpW), also known as refill friction stir spot industrial applications. As this is a very recent technique, little welding is a technique for joining sheets in the overlap configuration. information is available regarding the evaluation and optimisation It was invented and patented by GKSS Research Centre GmbH [1]. of process parameters for specific material combinations. The The workpiece does not reach temperatures above the melting process has been used to investigate the weldability of the high temperature, making this a solid-state joining technique. This process strength aluminium alloys EN AW-7475-T761, EN AW-6082-T6 is suitable for spot welding lightweight materials with low melting and EN AW-7075-T6, aiming to produce high quality joints in point such as aluminium and magnesium alloys. The mechanical terms of mechanical performance and microstructure. More nature of this welding process makes it possible to weld any alloy specific, the influence of the plunge depth, rotation speed and which presents some degree of plasticity [2]. Friction spot welding welding time was investigated. The extended abstract first briefly evolved from friction stir welding (FSW), removing the traverse part describes the process and continues with the results of the optical of FSW and adding a refill stage. This is done by using a three piece microscopy, microhardness and static tensile test results. non-consumable tool, consisting of a concentric clamping ring, centre pin and sleeve. The welding tool during the different process steps is Keywords: EN AW-7475-T761; EN AW-6082-T6; EN AW-7075- schematically depicted in figure 1, courtesy of Rosendo et al. [3]. T6; friction spot welding; parameter study

Introduction

The increasing need for strong and lightweight materials in automotive and aerospace industries aiming for fuel savings, performance and safety have led to the development of new joining techniques for aluminium and magnesium alloys. Up to date, most spot joints are realised by mechanical fastening or resistance spot welding. Mechanical fastening methods as riveting, self-piercing riveting and Figure 1: Schematic representation of the FSpW process [3] bolting add considerable weight, when significant amounts are required, and are more difficult to automate. The pierced surface The process consists of four stages. In the first stage the tool is lowered resulting from these methods also give rises to sealant and corrosion to the zero level, which is equal to the top surface of the upper sheet. prevention challenges. Traditional fusion welding of some high Here the clamping ring, located at the outside of the tool, fixates the strength aluminium alloys, is difficult due to hot cracking and two plates between the weld head and the backing anvil by applying hydrogen void generation. The higher thermal and electrical pressure. At the same time, the pin and sleeve start to rotate at identical conductivity of aluminium also lead to a higher energy consumption rotational speeds, touching the upper sheet. The friction causes the in resistance spot welding or laser spot welding and thus to higher lightweight alloy to heat up and soften. In the following stage, the operating costs. sleeve is pushed into the workpiece whilst it is still rotating until it reaches the pre-set plunge depth. Simultaneously, the pin retracts, The new joining technique known as friction spot welding (FSpW), or creating a cavity of the same volume, where the material displaced by refill friction stir spot welding, is one of the alternatives overcoming the sleeve is accumulated. In stage three, the sleeve is retracted and the these disadvantages. It offers additional advantages such as: no filler pin pushes the softened material back down, filling the weld and material is required, environmentally friendly (no generation of fumes, creating a weld nugget. As a last stage the entire weld head is IR, UV and electromagnetic radiation), limited or no waste products, withdrawn from the workpiece, revealing a circular mark on the top fast processing speeds, no pre-cleaning needed and the good surface sheet, but without keyhole or significant material loss. This particular quality requires no post-processing. The goal of this research is the sequence is called sleeve-plunge. The pin-plunge variant also exists, investigation of the influence of the most important process parameters where the pin first plunges into the workpiece, but this would lead to a (plunge depth, rotation speed and welding time) on the microstructural smaller effective weld area and as stated by Suhudinn et al. [4], this and mechanical properties of friction spot welds in the alloys 7475- results in a lower joint strength. The fact that there is no keyhole leads T761, 6082-T6 and 7075-T6, aiming at strong connections with a high to less stress concentrations and corrosion problems. reproducibility.

III

The process parameters consist of the rotation speed (RT), the plunge strength. Partial bonding is a region where the bonding of the two depth (PD) and the joining time (JT). The joining time can be divided sheets is not that strong, located between the bonding ligament and the into three different parts, the plunge time (PT), the dwell time (DT) hooking [3]. It resembles a short jagged line underneath the sleeve. and the retraction time (RT). These time parameters are visualised in Together with hooking, it plays an important role in crack initiation as the figure below. cracks can easily propagate in the weakly bonded region to create a circumferential tear [3]. The bonding ligament (BL), found centrally below the stir zone, is a banded structure where the two sheets have a strong metallurgical bond. Depending on the alloy used, the evidence of the presence of the bonding ligament may be different.

Experimental data

Material

In the present research, sheets of aluminium alloys EN AW-7475- T761, EN AW-6082-T6 and EN AW-7075-T6 were welded in the overlap configuration. The sheets had a thickness of 1,6 mm, 2 mm and 1,6 mm respectively and were used in the bare condition, meaning Figure 2: Plunge depth as a function of the joining time with the plunge, dwell without any coating. The main mechanical properties following from and retraction time indicated their temper conditions are shown in table 1.

Weld properties Table 1: Typical mechanical properties of the alloys used [6]

Ultimate strength Yield strength Elongation A typical cross-section of a FSpW weld nugget is depicted in figure 3. It consists of 4 distinct zones, each with different microstructures and [MPa] [MPa] [%] axi-symmetrical to the tool axis [5]. The approximate locations of 7475-T761 524 448 12 these zones are also indicated on figure 3. The zones, from the middle of the weld outwards, are: the stir zone (SZ), the thermo-mechanically 6082-T6 290 250 10 affected zone (TMAZ), the heat affected zone (HAZ) and the base metal (BM). 7075-T6 572 503 11

Equipment

The welds were realised using a commercial RPS 100 friction spot welding machine manufactured by Harms & Wende Germany. The three piece welding tool, consisting of the clamping ring, the sleeve and the pin have the following dimensions: the outer diameters of the clamping ring and sleeve are 14,5 and 9 mm respectively while the pin has an outer diameter of 6,4 mm. The machine is capable of applying Figure 3: Typical cross-section by Rosendo et al. [3] axial forces and rotational speeds up to 15 kN and 3300 rpm respectively. There is a thread at the end of the pin and sleeve that acts The SZ has approximately the same width as the sleeve diameter and as a barrier, limiting the amount of plasticised material that is pressed is characterised by a fine equiaxed grain structure. This is the result of between the tool parts, but it also improves the material flow as dynamic recrystallisation caused by high strain and the intense thermal concluded by Zhao et al. [7]. cycle during welding, causing solubilisation of precipitates and the ageing afterwards [3]. The TMAZ is concentrated below the periphery Experiments of the sleeve and consists of elongated grain structures due to moderate deformation and frictional heating as concluded by Shen et al. [5]. The For each material, only the three most important process parameters HAZ undergoes, as the name implies, only a thermal cycle caused by were varied. This resulted in the following parameter windows. the frictional heating. This results in a non-deformed microstructure. However, the exact extension of the TMAZ and therefore, the boundary between the TMAZ and the HAZ is difficult to localise with Table 2: Process parameter windows for the alloys used an optical microscope. Rosendo et al. [3] defined the boundary Rotation speed [rpm] Plunge depth [mm] Dwell time [s] between the TMAZ and the HAZ as the location with the minimum 7475- 1500 - 2500 1,6 - 2,1 2,5 - 3,5 - 4,5 hardness between the stir zone and the base material. T761 6082-T6 1500 - 2250 - 3000 2 - 2,5 - 3 0 - 1 - 2 The cross-sections of the weld nugget also reveals some geometric patterns associated to the material flow and common to most welds as 7075-T6 1500 - 2000 - 2500 1,8 - 2,2 - 2,7 0 - 1 - 2 examined by Rosendo et al. [3]. These consist of hooking, partial bonding and the bonding ligament. Hooking is caused by plastic deformation of the lower sheet and the lack of mixing of the sheets at that location. It has an upside down V-shaped appearance. The tip of For the 7475 welds the plunge and retraction times were set at 2 and the hook acts as a crack initiation location and for that reason sharp 1,5 seconds respectively and for the other alloys both were set at 1,5 and large hooks are avoided in literature as these decrease the joint seconds. After welding, a longitudinal section through the centre of the weld nugget was ground, polished and etched using Keller’s reagent, IV to reveal the weld microstructure. Lap shear strength tests were performed on the welds, except the 7075 series, using an Instron testing machine model 1342 at 10mm/min displacement speed. Microhardness maps (HV0,3) were made using a Leco AMH43 system. Traverse microhardness lines (HV0,5) in the middle of the top sheet were also made with a Struers Duramin-A300. These tests were performed on weld samples resulting from low and high heat input parameter combinations.

Results Figure 5: Fine banded BL (7475 alloy, 1500 rpm RT, 2,1 mm PD and 2,5 s DT) Metallographic examination

The SZ of all alloys was characterised by fine equi-axed grains as described in literature. However, a brighter band of coarser grains is noticed down the middle of the SZ for welds made from the 7475 and 7075 aluminium alloys. This band, also visible in the middle of the top part of figure 4, was thinner for weld conditions with a lower rotational speed and increasing the dwell time results in the narrowing of this band. The space just below the interface of the pin and sleeve contained the finest grains. In all alloys these seem to align to form striations, like those seen in figure 4. These striations tend to become thinner and Figure 6: Un-bonded zone (7475 alloy, 2500 rpm RT, 1,6 mm PD and 4,5 s DT) fade away as the rotational speed increases. The striations of most of the 6082 welds were limited in length and confined closer to the The bonding ligament of the 6082 welds changed the most with plunge surface than those observed in the 7475 and 7075 welds. depth. Using a plunge depth of 2 mm resulted in a flat bonding ligament line, with partial bonding at both ends. In this case the BL length increased with increasing dwell time and rotational speed. The plunge depth of 3 mm resulted in a more upwards bended BL and no partial bonding is documented. This is due to the larger plunge depth causing a more upward material flow, dissolving the partial bonding. No correlation could be found between the appearance of the BL and the parameters used for the 7075 welds due to tool failure limiting the amount of metallographic samples.

A new geometrical pattern appeared in the form of a dark line in the

middle of the weld nugget, depicted in figure 7. It has a W-shape and Figure 4: Striations (6082 alloy, 2000 rpm RT, 2 mm PD and 2 s DT) connects the outer areas of the nugget surface. This line represents an axi-symmetrical plane around the nuggets centre, which consists of The TMAZ could be seen in all welds and has a distinct semi ellipse very fine and compact grains. It was only documented in some weld shape, with the flat side against the sleeve plunge path, which varies in conditions (RS 2500 rpm, 2,1 mm PD and 3,5 or 4,5 s DT) for the 7475 size. The grains inside the TMAZ are elongated and oriented alloy and (RS 2500 rpm, 2,7 mm PD and 2 s DT) for the 7075 alloy. It downwards and slightly inwards. The HAZ of the 7475 and the 7075 thus seems that the pattern is caused by a specific material flow welds were not visible with an optical microscope as opposed to the resulting from a rotational speed of 2500 rpm together with an average HAZ of the 6082 welds. In this alloy, the HAZ is recognised as a dark plunge depth and a sufficiently high dwell time. band between the base material and TMAZ. This zone becomes darker and extends further outward with increasing heat input (i.e. increasing RT and/or dwell time).

The bonding ligament behaviour changed with the different alloys. It is characterised by fine grains and was more clearly visible for the 6082 welds. The appearance of the bonding ligament of the 7475 welds changed the most with the rotational speed. All welds made with a rotational speed of 1500 rpm exhibit a very fine banded structure, centrally below the SZ. It is characterised by alternating finer and coarser grains and is limited in width, as seen in figure 5. Welds produced with the rotational speed of 2500 rpm showed an un-bonded Figure 7: W-shaped band (7075, 2500 rpm RT, 2,7 mm PD and 2 s DT) zone at the bonding ligament location. At the ends of this un-bonded zone, depicted in figure 6, normal bonding ligament is observed which Imperfections like lack of mixing and voids were consistently found in continues to the outside of the weld nugget. This debonding never welds produced with low rotational speed and short dwell time extends underneath the sleeve area, which suggests a stronger bond combinations. They can be attributed to a too low heat input. On the underneath the sleeve. Partial bonding was not visible in this alloy. other hand, for the 6082 alloy, one extreme weld condition resulted in an excessive heat input. This caused surface cracks to be formed at both free surfaces, as indicated in figure 8. Incomplete refill in all

V alloys could be eliminated by changing the zero level of the sleeve and causes an increase. These can be used to explain most of the changes. pin. The effect of the increased heat input changes with the alloy welded. An increased heat input for the 7475 alloy results in an increase in the hardness drop between TMAZ and HAZ, a more homogeneous SZ which extends slightly further and having a slightly reduced hardness. The location of the minima does not seem to change. Hardness peaks are observed for the low heat input combination indicating the fine grains under the sleeve/pin interface. The increased heat input for the 6082 alloy causes the HAZ (slope connecting base metal to minimum) and the TMAZ (slope connecting minimum to SZ) to significantly extend further while creating a plateau in the centre of the SZ. It also causes the minima to be slightly lower. However the SZ hardness is higher, due to more recrystallisation. The traverse line obtained for the 7075 alloy also revealed the Figure 8: Surface cracks (6082 alloy, 3000 rpm RT, 3 mm PD and 2 s DT) extension of the TMAZ and the HAZ and the slight lowering of the minima with increasing heat input. As opposed to the 7475 alloy from Microhardness the same family, where an increased SZ hardness is perceived, welding of the 7075 alloy causes a much lower SZ hardness relatively to the The traverse microhardness lines of the low heat input parameter base metal hardness. Slight peaks at the sleeve plunge path are noticed combinations can be seen in figure 9 while those of high heat input for the lower heat input together with a fairly flat SZ. The higher heat combinations are plotted in figure 10. The edge of the sleeve plunge input sample has a central hardness drop due to very coarse grains in path is indicated with black dotted lines, containing the SZ. The base the centre of the nugget, also seen in figure 7. metal hardness of each material is indicated with corresponding coloured dotted horizontal lines. The microhardness maps of the 7475 alloy lack well-defined structures, due to the natural ageing capability of the material and the long time between welding and testing. A higher overall SZ hardness is noticed for the low heat input, in line with the observations from the traverse hardness lines. The 6082 alloy exhibits well-defined structures and the two hardness maps can be seen in figure 11. This is because the natural ageing of the alloy is limited. The edge of the plunge path is indicated with a black dotted line together with the TMAZ and HAZ on one side, which are symmetrical to the nugget’s central axis.

Figure 9: Red: 7475 alloy, 1500 rpm, 1,6 mm PD, 2,5 s DT; Blue: 6082 alloy, 1000 rpm, 2 mm PD, 0 s DT; Green: 7075 alloy, 2500 rpm, 2,7 mm PD, 2 s DT

Figure 11: Low heat input (6082 alloy, 1500 rpm RT, 3 mm PD and 0 s DT) (top) and high heat input (6082 alloy, 3000 rpm RT, 3 mm PD and 2 s DT) (bottom) microhardness maps.

The extension of the TMAZ and HAZ can clearly be seen. The boundary between the two zones is defined by an area of lower hardness which widens with increasing heat input. The increased hardness in the SZ for higher heat input confirms the observations from the traverse hardness lines. The areas of lower hardness inside the SZ of the low heat input sample correspond to the darker areas that could be seen in the cross-section. The 7075 alloy hardness maps show the same behaviour, when looking Figure 10: Red: 7475 alloy, 2500 rpm, 1,6 mm PD, 4,5 s DT; Blue: 6082 alloy, at the TMAZ, HAZ and the boundary in between, as in the 6082 alloy. 3000 rpm, 2 mm PD, 2 s DT; Green:7075 alloy, 1500 rpm, 2,7 mm PD, 0 s DT However, as with the 7475 alloy, the SZ is fairly homogeneous due to natural ageing smoothing out the hardness distribution. In the high heat Two phenomena are important for changes in hardness level. input sample the bonding ligament appears as a zone of lower Dissolution of the precipitates causes a reduction and recrystallisation hardness. VI

Lap shear strength

The lap shear tests revealed the influences of the process parameters on the strength of the joints. For the 7475 alloy, the lap shear strength increased with both plunge depth and dwell time, without reaching a maximum. The increased plunge depth causes less sharp hooking and the increased dwell time results in a more homogeneous weld structure. A slight decrease is noticed when increasing the rotational speed, but this effect is statistically insignificant. This decrease can be attributed to the central debonding discussed before. The strongest joint, with a shear strength of 5,38 ± 0,56 kN, was obtained for RT 1500 rpm, 2,1 mm PD and 4,5 s DT. The main effects plots of the 6082 alloy are shown in figures 12 through 14. The maximum and minimum lap shear strengths are indicated with an upward and downward facing triangle respectively. The average value is indicated with a circle and the standard deviation is represented by an error bar. It is immediately clear from the lap shear strength range that very strong connections were made. Each Figure 14: Rotation speed influence on lap shear strength of the 6082 alloy parameter influence plot exhibits a maximum, meaning that the optimal welding parameter combination is contained in the suggested Acknowledgments parameter window. The highest average lap shear strength (10,52 ± 0,20 kN) is obtained when all these maxima are combined (i.e. RT 2250 rpm, PD 2,5 mm and DT 1 s).The main failure mode was plug This work has been carried out as part of a master dissertation program pull out on upper sheet, but through weld failure occurred at lower at Ghent University with the additional support of the Belgian Welding plunge depth (2 mm) and low heat input parameters. Institute for supplying materials and means. References

[1] C. Schilling and J. dos Santos, “Method and device for linking at least two adjoining work pieces by friction welding”. European patent number Patent EP 1230062 B1, 1999.

[2] C. Schilling, A. Strombeck and J. Santos, “Friction spot welding: new joining process for spot connections,” GKSS Research Center, Geesthacht, 2001.

[3] T. Rosendo, B. Parra, M. Tier, A. da Silva, J. dos Santos, T. Strohaecker and N. Alcantara, “Mechanical and microstructural investigation of friction spot welded AA6181-T4 aluminum alloy,” Materials & design, vol. 32, no. 3, pp. 1094-1100, 2011.

[4] U. Suhuddin, L. Campanelli, M. Bissolatti, H. Wang, R. Figure 12: Plunge depth influence on lap shear strength of the 6082 alloy Verastegui and J. dos Santos, “A review on microstructural and mechanical properties of friction spot welds in Al-based similar and dissimilar joints,” in Proceedings of the 1st international joint symposium on joining and welding, Osaka, Japan, 2013.

[5] Z. Shen, X. Yang, Z. Zhang, L. Cui and T. Li, “Microstructure and failure mechanisms of refill friction stir spot welded 7075-T6 aluminum alloy joints,” Materials & design, vol. 44, pp. 476-486, 2013.

[6] “MatWeb Material Property Data,” [Online]. Available: http://www.matweb.com. [Accessed 12 April 2016].

[7] Y. Zhao, S. Lin, L. Wu and F. Qu, “The influence of pin geometry on bonding and mechanical properties in friction stir welded 2014 Al alloy,” Materials letters, vol. 59, no. 23, pp. 2948-2952, 2005.

Figure 13: Dwell time influence on lap shear strength of the 6082 alloy

VII

Contents

Friction spot welding ...... 1 1.1 Introduction ...... 1 1.2 The process ...... 2 1.3 Process parameters ...... 4 Metallurgy of aluminium alloys ...... 7 2.1 Nomenclature and composition ...... 7 2.2 Weldability and precipitate behaviour ...... 9 Weld properties ...... 11 3.1 Microstructural and microhardness properties ...... 11 3.1.1 Weld zones ...... 11 3.1.2 Geometrical patterns and defects ...... 13 3.2 Mechanical properties ...... 15 3.2.1 Lap shear strength ...... 15 3.2.2 Failure modes ...... 17 3.2.3 Artificial and natural ageing ...... 19 Experimental conditions ...... 21 4.1 Experimental conditions ...... 21 4.2 Experimental test procedures ...... 26 4.2.1 Welding...... 26 4.2.2 Visual inspection ...... 27 4.2.3 Peel test ...... 28 4.2.4 Lap shear strength tests ...... 29 4.2.5 Metallographic examination ...... 30 4.2.6 Microhardness measurements ...... 31 Experimental results ...... 33 5.1 Weldability of EN AW-7475-T761 ...... 33 5.1.1 Metallographic examination ...... 33 5.1.2 Microhardness ...... 38 5.1.3 Lap shear strength ...... 40 5.1.4 Overview ...... 42 5.2 Weldability of EN AW-6082-T6 ...... 42 5.2.1 Metallographic examination ...... 43

VIII

5.2.2 Microhardness ...... 46 5.2.3 Lap shear strength ...... 50 5.2.4 Overview ...... 53 5.3 Weldability of EN AW-7075-T6 ...... 53 5.3.1 Metallographic examination ...... 54 5.3.2 Microhardness ...... 55 5.3.3 Overview ...... 57 Conclusion ...... 58 References ...... 60 Technical specifications FSpW machine ...... 63 Material certificate EN AW-7075-T6 ...... 67 Welding conditions ...... 69

IX

List of figures

Figure 1.1: Friction spot welded component of a BMW 5 series [1] ...... 1 Figure 1.2: Different stages of FSSW: a) plunging b) stirring c) retracting [4]...... 2 Figure 1.3: 3D visualisation of the weld head with extended pin and sleeve [5] ...... 3 Figure 1.4: Schematic cross section of sleeve plunge FSpW [7] ...... 3 Figure 1.5: Plunge depth in function of joining time with plunge, dwell and retraction time indicated 4 Figure 1.6: CAD image of the RPS 100 friction welding machine by Harms & Wende: 1) Friction welding head 2) Clamping mechanism 3) Control cabinet 4) Control panel 5) Trolley 6) Safety foot switch [11] ...... 5 Figure 1.7: From left to right: clamping ring, sleeve and pin ...... 6 Figure 2.1: Generalisation of the hydrogen solubility rate in function of temperature for aluminium [16] ...... 9 Figure 2.2: Strength recovery after welding for 7xxx series (left) and 6xxx series (right) [17] ...... 10 Figure 3.1: Typical cross-section by Rosendo et al. [5] ...... 11 Figure 3.2: Different weld zones for welded EN AW-6181-T4 [5] ...... 12 Figure 3.3: Typical W-shaped microhardness profile [5] ...... 12 Figure 3.4: Locations of partial bonding, bonding ligament and hooking in a cross section [5] ...... 13 Figure 3.5: Hook characteristics at different joining times and plunge depths [18] ...... 14 Figure 3.6: Close up of two typical weld defects, lack of mixing and incomplete refill [5] ...... 15 Figure 3.7: Shear strength as a function of the bonding ligament length [10] ...... 16 Figure 3.8: Indication of the bonding ligament length for a high strength (a) and low strength (b) weld [10] ...... 17 Figure 3.9: Illustrations of the three different failure modes due to shear loads [19] ...... 17 Figure 3.10: Schematic representation of the crack initiation sites and stress concentrations [5] ...... 18 Figure 3.11: Different failure modes observed by Shen et al. [18] for EN AW-7075-T6 welds under shear load ...... 18 Figure 3.12: Distribution of microhardness values in the SZ in samples a) aged at different temperatures for 16h and b) aged at 80°C for different durations [26] ...... 19 Figure 3.13: Microhardness profile of FSW 7075-T6 in as-welded and PWHTed conditions. 1000/150 means FSW at 1000 rpm and 150 mm/s [29] ...... 20 Figure 3.14: Graph showing average tensile properties of base material, as-welded and PWHTed conditions [29] ...... 20 Figure 4.1: New sleeve (left) and fractured sleeve (right)...... 24 Figure 4.2: Graphical representation of available samples in the material EN AW-7075-T6 from series 1 ...... 25 Figure 4.3: Graphical representation of available samples in the material EN AW-6082-T6 from series 1 ...... 25 Figure 4.4: Graphical representation of available samples in the material EN AW-7475-T761 from series 1 ...... 25 Figure 4.5: Class 1 weld surface (IJ-82-R1-17) ...... 28 Figure 4.6: Class 3 weld surface (IJ-82-R1-4) ...... 28 Figure 4.7: Plug pull-out on upper sheet from peel test on weld IJ-82-R1-13 ...... 28

X

Figure 4.8: Schematic representation of the lap shear strength samples ...... 29 Figure 4.9: Schematic representation of lap shear strength test ...... 29 Figure 4.10: Metallographic samples ...... 30 Figure 4.11: Embedded and etched sample ...... 31 Figure 4.12: Example of an indentation pattern ...... 32 Figure 4.13: Traverse hardness measurements on IJ-82-R1-4 ...... 32 Figure 5.1: Effect of dwell time and rotation speed on the central coarse grain band. From left to right: dwell time of 2,5; 3,5 and 4,5 seconds. Top to bottom: 1500 and 2500 rpm ...... 33 Figure 5.2: TMAZ of weld condition IJ-74-R1-1 ...... 34 Figure 5.3: Fine striations under pin/sleeve interface in IJ-74-R1-8 ...... 34 Figure 5.4: Broad striations under sleeve/pin interface in IJ-74-R1-1 ...... 35 Figure 5.5: The two types of bonding ligament...... 36 Figure 5.6: Fusion of BL and W-shaped fine grain structure in IJ-74-R1-12 ...... 36 Figure 5.7: Sharp hook (left) and M-shaped hook (right). H = hooking height ...... 37 Figure 5.8: Incomplete refill (left), lack of mixing (middle) and voids (right) ...... 37 Figure 5.9: Low heat input (left) and high heat input (right) microhardness profiles ...... 39 Figure 5.10: Low heat input microhardness map ...... 39 Figure 5.11: High heat input microhardness map ...... 40 Figure 5.12: Parameter influences on the lap shear strength ...... 41 Figure 5.13: Parameter interaction plots ...... 42 Figure 5.14: TMAZ of IJ-82-R1-14 ...... 44 Figure 5.15: Typical weld cross sections. White arrow indicates dark HAZ band. Left to right: 1000; 2000 and 3000 rpm ...... 44 Figure 5.16: Effect of rotation speed on the fine grain striations. Left to right: 1500; 2000; 3000 rpm ...... 44 Figure 5.17: Typical bonding ligament behaviour of welds with a plunge depth of 2 mm ...... 45 Figure 5.18: Typical bonding ligament behaviour of welds with a plunge depth of 3 mm ...... 45 Figure 5.19: Surface cracks of IJ-82-R1-14 ...... 46 Figure 5.20: Low heat input (left) and high heat input (right) microhardness profiles ...... 48 Figure 5.21: Low heat input microhardness map ...... 49 Figure 5.22: High heat input microhardness map ...... 49 Figure 5.23: Low heat input (left) and high heat input (right) metallographic cross-sections...... 49 Figure 5.24: Parameter influences on the lap shear strength ...... 51 Figure 5.25: Parameter interaction plots ...... 52 Figure 5.26: Failure modes: 1) trough weld 2) plug pull out on upper sheet 3) plug pull out on lower sheet 4) combination of through weld and plug pull out on upper sheet 5) combination of plug pull out on bottom and upper sheet ...... 53 Figure 5.27: Cross-section of IJ-75-R1-5 resulting from the unstable sleeve zero level ...... 54 Figure 5.28: Cross-section of IJ-75-R1-7 ...... 55 Figure 5.29: Low heat input (left) and high heat input (right) microhardness profiles ...... 56 Figure 5.30: Low heat input microhardness map ...... 57 Figure 5.31: High heat input microhardness map ...... 57

XI

List of tables

Table 2.1: Composition of the alloys used [12] [13]. *= max value ...... 8 Table 2.2: Typical mechanical properties of the alloys used [12] [13] ...... 8 Table 4.1: Common operating range of the FSpW welding machine (t = plate thickness) ...... 21 Table 4.2: Experimental parameter levels used for the material EN AW-7075-T6 ...... 22 Table 4.3: Experimental parameter levels used for the material EN AW-6082-T6 ...... 23 Table 4.4: Experimental parameter levels used for the material EN AW-7475-T761 ...... 23 Table 4.5: Visual inspection classes ...... 27 Table 4.6: Failure modes ...... 30 Table 4.7: Composition of Keller’s reagent ...... 31 Table 5.1: Overview of significant parameter correlations. 1) only for 1,6 mm PD ...... 42 Table 5.2: Overview of significant parameter correlations. 1) only at 2 mm PD ...... 53 Table 5.3: Overview of significant parameter correlations ...... 57

XII

List of abbreviations

ANOVA Analysis of variance BL Bonding ligament BWI Belgian Welding Institute DT Dwell time FSpW Friction spot welding FSSW Friction stir spot welding HAZ Heat affected zone IJ InnoJoin JT Joining time LSS Lap shear strength PD Plunge depth PT Plunge time PWHT Post weld heat treatment RT Retraction time RT Rotation speed SZ Stir zone TMAZ Thermo-mechanically affected zone

XIII

Chapter 1 Friction spot welding

1.1 Introduction

The increasing need for strong and lightweight components in automotive and aerospace industries aiming for fuel savings, performance and safety have led to the development of new joining techniques for aluminium and magnesium alloys. Up to date, most spot joints are made with mechanical joining or resistance spot welding. Mechanical fastening methods such as riveting, self- piercing riveting and bolting add considerable weight, when considering large amounts of welds, and are more difficult to automate. The pierced surface resulting from these methods also gives rise to sealing and corrosion prevention challenges. Traditional fusion welding, like resistance spot welding, is difficult due to the limited tool lifetime and the low weldability of aluminium and its alloys. The higher thermal and electrical conductivity of aluminium also lead to a higher energy consumption in resistance spot welding or laser spot welding and thus to higher operating costs. The new joining technique, known as friction spot welding (FSpW), is one of the alternatives overcoming these disadvantages and even offers other advantages. These include: no filler material is required, environmentally friendly and safe (no generation of fumes, IR, UV and electromagnetic radiation), limited or no waste products, fast processing speeds, no pre-cleaning needed and the good surface quality does not require post-processing. Figure 1.1 shows an example of a friction spot welded component. It is the objective of this master dissertation to investigate the weldability of high strength aluminium alloys using this technique but firstly, in this chapter, this new process is described in detail and the different process parameters are discussed to offer the reader the necessary insight.

Figure 1.1: Friction spot welded component of a BMW 5 series [1]

1

1.2 The process

Friction spot welding (FSpW), also known as refill friction stir spot welding is a technique for joining sheets in overlap configuration invented and patented by the GKSS Research Centre [2]. It is classified as a solid-state joining process, since the materials don’t melt during the weld cycle. This process is suitable for spot welding lightweight materials with a low melting point, such as aluminium and magnesium alloys, and a lower weldability when considering older fusion techniques. The mechanical nature of this welding process makes it possible to weld any alloy which presents some degree of plasticity [3]. Friction spot welding evolved from friction stir spot welding (FSSW), by adding a refill stage at the end. To explain the basic idea and to better comprehend the advantages compared to friction stir spot welding, the latter is explained first.

Friction stir spot welding works on the principle of plunging a single piece rotating non-consumable tool into the workpiece and heating the weld area due to friction. This frictional heat plasticises the material, making the realisation of a joint possible. Many different tool geometries exist to facilitate the material flow. The following picture by Kawasaki Heavy Industries Ltd. [4] can be used to explain the different process stages.

Figure 1.2: Different stages of FSSW: a) plunging b) stirring c) retracting [4]

In the first stage, the tool is brought up to the correct rotational speed and the machine plunges it into the workpiece. The second stage consists of stirring the material, adding more heat and plasticising the joint. In the third and last stage the tool is retracted from the workpiece, leaving a keyhole shaped cavity behind. Although a good quality weld can be made, the cavity causes some problems. The first problem is that the shape of the keyhole results in stress concentrations, thus reducing the strength of the joint. The second is that the keyhole can effectively contain water, giving rise to corrosion problems. Friction stir welding (FSW) is another variant of this process. FSW adds translational movement by the weld head, producing a seam weld.

These disadvantages have led to the development of refill friction stir spot welding. The process resembles to FSSW but uses a three piece non-consumable tool. A 3D visualisation of the tip of the weld head, indicating the three parts can be seen in figure 1.3.

2

Figure 1.3: 3D visualisation of the weld head with extended pin and sleeve [5]

The tool consists of a concentric clamping ring, centre pin and sleeve, each with an independent actuator for the axial movement. In addition, the pin and sleeve can also rotate independently. The primary function of the clamping ring is to apply pressure, but it also protects the operator from possible ejected plasticised material. The thread at the end of the pin and sleeve acts as a barrier, limiting the amount of plasticised material that is pressed between the tool parts, but it also improves the material flow as concluded by Zhao et al. [6]. The tool can operate in two operating modes, each consisting of four stages. The operator can choose between the sleeve-plunge variant and the pin-plunge variant. A schematic representation of the FSpW sleeve plunge variant is depicted below [7].

Figure 1.4: Schematic cross section of sleeve plunge FSpW [7]

The process can be described in four stages. In the first stage, the tool is lowered to the zero level, which is equal to the top of the upper sheet. Here the clamping ring, located at the outside of the tool, fixates the two plates between the weld head and the anvil, by applying pressure. At the same time, the pin and sleeve start to rotate at the same rotational speed, touching the upper sheet. The friction causes the lightweight alloy to heat up and plasticise. In the following stage, the sleeve is pushed into the workpiece whilst rotating. Simultaneously, the pin retracts, creating a void for the displaced material by the sleeve. In stage three, the sleeve is retracted and the pin pushes the material back down, filling the weld. As a last stage the weld head is withdrawn from the workpiece, revealing a circular mark on the top sheet, but no keyhole. Without the keyhole, the main problems of FSSW, stress concentration and

3 corrosion problems, are overcome. The pin-plunge variant, where the pin plunges into the workpiece and the sleeve is retracted is not often used although it reduces the required axial force. This is due to the fact that the effective weld area is smaller with this technique. Suhuddin et al. [8] reported that this results in a lower joint strength.

1.3 Process parameters

The FSpW welding machine used in this master dissertation is the RPS 100 by Harms & Wende and is depicted in figure 1.6, where the most important parts are indicated. The three piece weld tool, has following dimensions: the outer diameter of the clamping ring and sleeve is 14,5 mm and 9 mm respectively while the pin has a 6,4 mm outer diameter. The tool pieces used can be seen in figure 1.7. Further technical specifications of the FSpW machine can be found in appendix A. With this machine, the process is entirely PLC controlled, allowing to precisely change the process parameters to control the properties of the weld. This chapter is devoted to explaining the meaning of these parameters, not their influences on the weld properties as discussed further in this dissertation. These parameters are:

 Rotation speed (RS): the pin and sleeve rotate at this speed during the stirring stage.  Plunge depth (PD): the maximum axial displacement of the sleeve during the plunging stage, measured from the top surface of the upper sheet.  Joining time (JT): the overall duration of the actual friction phase of the FSpW process. This is the total welding time minus the time needed for clamping and releasing of the workpiece. This parameter comprises of three different parts: o Plunge time (PT) o Dwell time (DT) o Retraction time (RT) These time parameters are also visualized in figure 1.5. Note that the variation between two plunge depth positions happens linearly.

Figure 1.5: Plunge depth in function of joining time with plunge, dwell and retraction time indicated

4

The total heat input (Q) for the FSSW process according to Su et al. [9] can be calculated using the following discrete formula:

푁 푄 = ∑ 퐹(푛) ∗ (푥(푛) − 푥(푛 − 1)) + 푇(푛) ∗ 휔(푛) ∗ ∆푡 (1.1) 푛=1

This formula was derived by assuming that the mechanical work delivered by the weld machine is entirely transferred to the workpiece. Since some energy is used to heat up the tool, this formula is an approximation. The factors are the plunge force F, the axial displacement of the sleeve x, the torque T, the rotation speed ω (rad/s) and the joining time Δt. Su et al. [9] concluded that, for FSSW of EN AW-6061-T6, the torque term was around 200 times greater than the force term, resulting in the elimination of the force term. Tier et al. [10] concluded that the equation needs to be improved by taking the contact condition into account, meaning a slip or stick condition.

The following parameters do not affect the FSpW process directly, but a sensible choice of their settings needs to be made for the production of good welds:

 Clamping pressure (CP): the clamping ring applies this pressure to the workpiece. This should be set in such way that no indentation is made by the clamping ring.  Holding time (HT): duration of the applied CP after the retraction of the pin and sleeve, holding the sheets during the beginning of their cooldown.

Figure 1.6: CAD image of the RPS 100 friction welding machine by Harms & Wende: 1) Friction welding head 2) Clamping mechanism 3) Control cabinet 4) Control panel 5) Trolley 6) Safety foot switch [11]

5

Figure 1.7: From left to right: clamping ring, sleeve and pin

6

Chapter 2 Metallurgy of aluminium alloys

Three different materials were welded in overlap and investigated in this master dissertation. These are the EN AW-7075-T6, the EN AW-6082-T6 and the EN AW-7475-T761 aluminium alloys in sheets of 1,6 mm, 2 mm and 1,6 mm respectively. No dissimilar material combination joints were made. The 7075 and 7475 are cold finished aluminium wrought products and are classified as a high- strength aluminium alloys from the 7xxx series. Because of their low weight and superior strength, they are mostly used in highly stressed structural parts for aerospace applications, but can also be seen in high-performance transportation applications, where its high strength-to-weight ratio is beneficial. The alloy studied is in the bare condition, meaning without any alclad layer. An alclad layer is in fact a very thin layer of high-purity aluminium with the purpose of creating an oxide layer, protecting the base material aluminium from corrosion. The 6082 has a different nature, as it is a medium-strength alloy from the 6xxx series. The alloys is frequently used is automotive and nautical industries. It also can be welded by traditional fusion welding techniques like TIG and MIG but friction welding offers some extra benefits as mentioned before.

The following section will explain the nomenclature and what this implies for the composition. Later on in this chapter, the general properties of these materials are discussed and finally some information about the weldability of these alloys is given.

2.1 Nomenclature and composition

The alloys are named following the ANSI nomenclature. But due to the high use of the 7075 alloy in the aerospace industry, it is also referred to with terms as aircraft aluminium or aerospace aluminium. ANSI names can be divided into two parts, each giving different types of information. The first part gives insight into the composition. If the first part consists of 4 digits, it is a wrought aluminium. The alloys used here all fall under this category. Further classification into families can be made based on the main constituents. The specific family is designated by the first digit. In this dissertation, only two families are of importance:

 6xxx: Magnesium and silicon are the main constituents. The most commonly used alloys come from this family.  7xxx: Zinc is the main alloying element, but high amount of magnesium may also be present. The highest strength alloys come from this family.

Only the exact composition, as tested by the manufacturer, is available for the 7075 alloy and can be seen in table 2.1 below together with the typical compositions of the other two alloys, as defined by [12]. The complete material certificate of the 7075 can also be found in appendix B.

7

Table 2.1: Composition of the alloys used [12] [13]. *= max value

Chemistry Si Fe Cu Mn Mg Cr Zn Ti V Zr Other 7075 0,07 0,15 1,4 0,03 2,5 0,21 5,6 0,03 0,01 0,01 0,05 [wt%] 7475 1,2- 1,9- 0,18- 5,2- 0,10* 0,12* 0,06* 0,06* / / 0,15* [wt%] 1,9 2,6 0,25 6,2 6082 0,70- 0,40- 0,60- 0,50* 0,10* 0,25* 0,20* 0,10* / / 0,15* [wt%] 1,3 1 1,2

Both of the families used are heat treatable, meaning that the alloys and the forming of precipitates are significantly influenced by heat treatments. They are also named precipitation-hardened alloys. The given heat treatment is noted as the temper designation, which is given in the second part of the name. It consists of a letter and one or more digits. The two tempers seen in this dissertation have the following meaning:

 T6: This stands for a solution heat treatment followed by artificial ageing.  T761: The alloy is solution heat treated, followed by limited artificial over-ageing to increase corrosion resistance but with some reduction in strength.

The solution heat treatment process involves taking the constituents into solution by increasing the temperature and retaining them into place by rapid quenching. The second part of the heat treatment, ageing, also known as age hardening or precipitation hardening, can be best described as the diffusion and clustering of precipitates in the material with the result being an increase of the mechanical properties over time. The ageing can be done in a controlled environment or can happen at atmospheric conditions. When using a controlled environment, the process is called artificial ageing. Artificial ageing is usually carried out by putting the workpiece in a chamber at an elevated temperature for a defined period of time. Time is one of the critical factors here as over-ageing would lead to worse mechanical properties. On the other hand, ageing can also take place when the workpiece rests at room temperature. The material naturally converges to the final properties, giving it the name natural ageing. The mechanical properties, as a result of these heat treatments, of each alloy are listed in table 2.2. The exact properties for the 7075-T6 are given, alongside the theoretical values of the 6082-T6 and 7475- T761.

Table 2.2: Typical mechanical properties of the alloys used [12] [13]

Ultimate strength [MPa] Yield strength [MPa] Elongation [%] 7075-T6 578-580 506-507 13,3-13,7 7475-T761 524 448 12 6082-T6 290 250 10

8

2.2 Weldability and precipitate behaviour

ISO 581:2005 defines weldability as a property of metals (or combinations of metals) to generate connections with determined welding technology to meet the required construction conditions. There are many criterions listed in this standard for determining weldability but the most determining in the case of high strength aluminium alloys, such as the 7xxx series, and some medium-strength aluminium alloys, such as a the 6xxx series, are the resistibility against hot cracking and void generation. Certainly for the high strength alloys 7075 and 7475, traditional fusion welding techniques result in improper welds based on the above criteria. The EN AW-6082-T6 however can be fusion welded when using the proper filler material. The following is thus mostly relevant when considering the high strength alloys in this dissertation.

Cracks throughout the weld zone seriously compromise the integrity of the joint. These cracks can be located at the surface of the welds or internally, resulting in the need for more advanced inspection techniques. Hot cracking is the main cause of crack formation in most aluminium alloy welds [14]. This phenomenon is mainly driven by uneven thermal stresses caused by uneven shrinkage in the weld. The first and most obvious measure to reduce this effect is to reduce the welding temperature, which results in lowering the thermal stresses and the sensitivity to hot cracking. The alloying elements play an important role as well. The alloying elements (mainly magnesium and copper) in 7075 that provide its high strength properties also result in an increased crack sensitivity [15] [14]. Void generation in aluminium alloys due to the entrapment of hydrogen gas plays the second big role in the low weldability of high strength aluminium alloys when considering fusion welding. The temperatures reached by fusion welding give rise to a very high hydrogen solubility rate as seen in figure 2.1. The large increase in solubility when aluminium passes its solidus temperature is clearly visible. This causes hydrogen introduced trough contaminants (moisture, hydrocarbons, etc.) to diffuse into the molten aluminium and create hydrogen gas voids [15]. When the aluminium cools down, these voids are trapped due to the now low solubility. Voids also reduce the strength of the weld. When considering fusion welding, the workpiece thus needs to be very well pre-treated to reduce contaminants.

Figure 2.1: Generalisation of the hydrogen solubility rate in function of temperature for aluminium [16]

9

By using solid state welding techniques like friction spot welding, the weldability of these aluminium alloy improves significantly. This is achieved by not exceeding the solidus temperature of the alloys during the process. FSpW thus provides a solution by reducing the sensitivity to both cracking and void generation. The fact that the hydrogen solubility rate is very low during the entire process has the additional benefit that less measures have to be taken to reduce contamination.

Welding also causes a noticeable difference in the behaviour of the precipitates in the two different families. When 6xxx series aluminium is welded, the thermal weld cycle causes a coarsening of the precipitates in the heat affected zone. Whereas welding 7xxx series aluminium leads to a precipitate dissolution into the aluminium matrix [17]. This leads to less desirable mechanical properties, in the heat affected zone, like reduced strength because the movement of dislocations is less likely to be impeded. This reduction of the strength is partly or entirely recovered by the ageing of the weld afterwards. In this case, the two families exhibit different behaviour as well. Whereas the 7xxx series will recover its mechanical properties almost entirely by natural ageing, the 6xxx series will not [17]. As the peak temperature of the FSpW process is usually around 480°C, the much larger recovery of the 7xxx series and the almost irreversible strength loss of the 6xxx series strength can be seen in figure 2.2.

Figure 2.2: Strength recovery after welding for 7xxx series (left) and 6xxx series (right) [17]

10

Chapter 3 Weld properties

When taking a closer look at a completed FSpW weld, a couple of recurring elements can be seen. This chapter aims to identify those elements and their influence on the mechanical weld behaviour. It provides the basic insights necessary to interpret the experimental results later in this dissertation.

3.1 Microstructural and microhardness properties

This section discusses the weld morphology and the typical microstructures present in a FSpW weld. The microhardness profile is also discussed here, as it is closely related to the microstructures. The metallographic study is usually done by making cross sections followed by grounding, polishing, etching and finally photographing with an optical microscope. A typical cross section of a weld nugget is depicted below.

Figure 3.1: Typical cross-section by Rosendo et al. [5]

3.1.1 Weld zones Cross sections of FSpW welds (like the one in figure 3.1) reveal a bowl shaped weld appearance. The nugget thickness, which is the distance between interface and top surface, is said to increase with the plunge depth, but also with the welding time, as noted by Shen et al. [18]. The microstructural analysis reveals 4 distinct zones, each with different characteristics and all axi-symmetrical to the tool axis [19]. The approximate locations of these zones can also be seen in figure 3.1. The zones, named from the middle of the weld outwards, are: the stir zone (SZ), the thermo-mechanically affected zone (TMAZ), the heat affected zone (HAZ) and the base material (BM). Differences in microstructure between these zones for an EN AW-6181-T4 alloy are visible in figure 3.2.

11

Figure 3.2: Different weld zones for welded EN AW-6181-T4 [5]

Although the presence of the HAZ cannot be seen by optical microscopy, as it has the same grain size as the BM, it can be confirmed by a W shaped microhardness profile across the joint as illustrated by Rosendo et al. [5] for EN AW-6181-T4 in figure 3.3. This distinctive shape recurs for all heat treatable aluminium alloys in literature and is thus to be expected for the alloys studied in this dissertation.

Figure 3.3: Typical W-shaped microhardness profile [5]

The SZ has approximately the same width as the pin diameter and is characterised by a fine equiaxed grain structure. The result of the dynamic recrystallization caused by high strain and the intense thermal cycle during welding, causing solubilisation of the precipitates and the ageing afterwards [5]. This is the cause of the elevated hardness in the SZ [20]. Shen et al. [19] noticed that the intense stirring in the region below the space between the pin and the sleeve causes even finer grains than in the bulk of the SZ. The grain size relates to the microhardness according to the classic Hall-Petch relation:

퐾퐻 퐻 = 퐻0 + (3.1) √푑

12

With 퐾퐻, 퐻0 constants and 푑 the mean grain size. A similar relation exists for the yield strength. This together with the aforementioned fine microstructures explain the two hardness peaks and their location inside the SZ. The size of the grains is also influenced by the process parameters. Zhao et al. [7] reported that the grains in de SZ are coarsening with increasing plunge depth. The increase in plunge depth increases the size of the nugget and thus the cooling time, due to the fact that the nugget first cools down at the nugget boundary. This causes that the grains and precipitates have more time to grow in the centre. So an increase in plunge depth decreases the hardness of the SZ. The TMAZ is concentrated below the periphery of the sleeve and consists of elongated grain structures due to moderate deformation and frictional heating as concludes by Shen et al. [19]. It is a transition zone between SZ and HAZ and is seen as a steady slope connecting the low hardness of the HAZ and the higher hardness of the SZ in the microhardness profile. The HAZ undergoes, as the name implies, only a thermal cycle caused by the frictional heating. The effect of this thermal cycle was studied by Santos et al. [20] using electrical conductivity mapping for the EN AW-7075-T6 alloy. They stated that the thermal cycle causes the precipitates in precipitation hardened aluminium alloys to dissolve into the matrix with grain coarsening softening the material [20]. This partially neutralizes the strength increase by the heat treatment during fabrication of the alloy. Softening can also be seen in the TMAZ but Olea [21] attributed this to the higher temperatures rather causing coarsening of the precipitates in precipitation-hardened alloys.

3.1.2 Geometrical patterns and defects The cross-sections of the weld nugget also reveal some geometric patterns associated to the plasticised material flow and common to most welds as examined by Rosendo et al. [5]. These consist of hooking, partial bonding and the bonding ligament. These features can be seen in figure 3.4.

Figure 3.4: Locations of partial bonding, bonding ligament and hooking in a cross section [5]

Hooking is caused by plastic deformation of the lower sheet and the lack of mixing of the sheets at that location. It has an upside down V-shaped appearance. The vertical displacement of the hook can be measured and easily compared with literature with help of the h/t ratio. This is the ratio of the height h, which is the distance between the tip of the hook and the sheets’ intersection, to the thickness of the sheets t. With most materials, the height of the hook decreases and the shape becomes more irregular with the increase of rotation speed at lower welding times [22]. The conclusion made by Rosendo et al. [5] is that the sharpness of the hook can be limited by decreasing

13 the joining time. Arul et al. [23] generalised this further and stated that in the case of an FSSW joint, hooking was more pronounced in the highest levels of the parameters. In contrast Shen et al. [19] found, in case of an EN AW-7075-T6 joint, that the process parameters scarcely affect the hook geometry, mostly due to the difficulty of plasticization. A more recent research, also on EN AW-7075- T6, by Shen et al. [18] revealed a sharper and more pronounced hook with decreasing welding time and plunge depth. This can be seen by looking at the tip of the hook in their recorded images of the hooking phenomenon for different combinations of those parameters (figure 3.5).

Figure 3.5: Hook characteristics at different joining times and plunge depths [18]

Partial bonding is a region where the bonding of the two sheets is not that strong. It resembles a short jagged line underneath the sleeve and together with hooking, play an important role in crack initiation. The bonding ligament can be found centrally in the weld zone, at the bottom of the SZ. It is not a defect but a banded structure where the two sheets have a strong metallurgical bond [5]. In cases where the materials had an alclad layer, a concentrated alclad layer was formed and could clearly be seen at this location. An alclad layer is a thin surface layer of pure aluminium to improve corrosion resistance. Rosendo et al. [5] associated hooking and partial bonding to poor plasticisation of materials and inappropriate process parameters, while Shen et al. [19] attributed these to the FSpW process and stated that joints without these patterns cannot be obtained.

In the weld nugget, imperfections may also be present. Defects like lack of mixing and incomplete refill were found by Rosendo et al. [5] and they attributed these defects to insufficient heat input.

14

Both are found at the path of the sleeve plunging into the workpiece. Shen et al. [19] also noticed voids inside the weld nugget, stating that void formation is mainly due to poor material flow caused by inappropriate welding parameter combinations. Shen et al. [19] also stated that the voids become smaller with increasing welding time and bigger with increasing rotation speed. They pointed out that voids are very unfavourable for the integrity and strength of the weld, so they should be avoided or minimised. Lack of mixing and incomplete refill are visible in the figure below.

Figure 3.6: Close up of two typical weld defects, lack of mixing and incomplete refill [5]

3.2 Mechanical properties

In this section the most important conclusions about the lap shear strength and the reported failure modes are discussed. In the last part, the ageing of the joint for the improvement of its mechanical properties is dealt with.

3.2.1 Lap shear strength The most widely investigated mechanical property of FSpW welds is the lap shear strength. This can be tested by performing a lap shear test following for example EN ISO 14273. Many tests have been performed on different materials, making it possible to compare results with literature. These results need to be interpreted carefully, as purely the lap shear strength value says little. If precise comparisons need to be made, it is critical to keep the thickness of the sheets and the properties of the alloys in mind. One way of comparing joints, is by using the joint efficiency. This ratio is obtained by comparing the tensile strength of the lap shear specimens with the tensile strength of the base material.

There are noticeable changes in strength of the material after the FSpW process. These changes can be attributed to two mechanisms: the before-mentioned grain size reduction and the re-precipitation during cooling of the workpiece post weld in case of precipitation hardened alloys [5]. The overall tensile strength of the weld is mostly dependent on the welding parameters and the presence of imperfections. Zhang et al. [24] investigated the impact of hooking defects on lap shear strength in FSSW welded joints. They reported that tensile shear loads of joints without any hooking defects almost doubled compared to joints with defects. They achieved these hooking defect free welds by friction stir welding (FSW) the edge of the previously made FSSW joints. Direct comparisons were made between the strength of FSSW and refill FSSW in [25]. In this work Uematsu et al. concluded that the refilling process increased the cross sectional area of the nugget, hence increasing the tensile strength. They noticed a 30% improved tensile strength by the refilling process so better results for defect free FSpW welds are also expected.

15

Tier et al. [10] made a statistical analysis to link the lap shear load to the weld parameters in case of 1,5 mm thick EN AW-5042-O sheets. They conducted experiments without dwell time. During preliminary tests, they concluded that there was no reason to treat the plunge time and retraction time as two separate variables and that they should be summed in the total joining time. The analysis resulted in the following equation with R=0,973.

푅푆 2,5 푃퐷 (푃퐷⁄1,45) 푆ℎ퐿 = 퐶 ∗ ( ) + 퐶 ∗ ( ) + 퐶 ∗ (3.2) 1 900 2 1,45 3 (퐽푇⁄1,66)3

With ShL being the lap shear load and the coefficients being: C1=-0,246; C2=5,86; C3=-0,621. The large value for C2 suggests a dominant effect of the plunge depth.

The rotational speed of the tool is an important parameter for the final weld strength. Shen et al. [19] concluded in their investigation regarding refill friction spot welded 2mm thick EN AW-7075-T6 aluminium sheets that the best results were achieved at a lower rotational speed and a shorter joining time. They noticed a maximum when plotting lap shear strength versus the joining time.

The relation between the bonding ligament length (BLL) and the weld strength was investigated by Tier et al. [10] and Suhuddin et al. [8]. They both documented an increase in lap shear strength with the increase of bonding ligament length, revealing a linear relationship as presented in figure 3.7. Suhuddin et al. concluded this by welding AZ32 Mg alloy. Tier et al. conducted the experiments using 1,5mm thick EN AW-5042-O aluminium alloy. They successfully increased the BLL by decreasing the rotation speed, stating that the material flow for welds executed with a higher tool rotational speed is more vertical and thus resulting in a shorter BLL. A bonding ligament line parallel to the surface thus means a higher lap shear strength as opposed to an upwards bending one. Figure 3.8 displays both scenarios, high strength with long BLL and low strength with short BLL.

Figure 3.7: Shear strength as a function of the bonding ligament length [10]

16

Figure 3.8: Indication of the bonding ligament length for a high strength (a) and low strength (b) weld [10]

3.2.2 Failure modes Specimens investigated for lap shear strength are obviously tested until failure. Both Rosendo et al. [5] and Shen et al. [19] documented three basic failure modes schematically represented in figure 3.9. These include: through weld nugget or nugget debonding (a), plug pull-out on upper sheet (b) and plug pull-out on lower sheet (c).

Figure 3.9: Illustrations of the three different failure modes due to shear loads [19]

Partial bonding and hooking are the important factors before the final failure, as reported by Rosendo et al. [5]. These patterns give rise to high stress concentrations and act as crack initiation and propagation sites. This is true for the partial bonding region in particular, where the first tearing occurs due to insufficient bonding. This aids in the formation of an annular crack surrounding the SZ. The direction in which the crack propagates then strongly depends on the integrity of the weld nugget and the properties in the surrounding weld zones. The annular crack is surrounded on one side by the bonding ligament and at the other by the hooking phenomenon. The top sheet can contain the aforementioned defects and the tip of the hook, causing even more stress concentrations. The same can be said for the two extremities of the partial bonding region in the lower sheet. The crack propagates in the weakest part and gives rise to the three different failure modes. This is depicted in the schematic representation of figure 3.10. Also, a weak bonding ligament would cause an interfacial debonding, thus leading to the through weld nugget failure mode. This does not mean that strong welds cannot produce this type of fracture path, as a shear fracture path can also evolve through the nugget.

17

Figure 3.10: Schematic representation of the crack initiation sites and stress concentrations [5]

The more recent study by Shen et al. [18] on the EN AW-7075-T6 aluminium alloys revealed six different fracture modes, including additional fracture modes which are basically combinations of two of the three failure modes mentioned above. Very strong welds even gave rise to the failure of the sheets themselves. An overview of these failure modes with their appearance can be seen in figure 3.11.

Figure 3.11: Different failure modes observed by Shen et al. [18] for EN AW-7075-T6 welds under shear load

18

3.2.3 Artificial and natural ageing As mentioned before, the thermal weld cycle causes a precipitate dissolution in the matrix of the 7xxx series aluminium and precipitate coarsening in the 6xxx series [17] [20]. This causes less desirable mechanical properties in the HAZ. The recovery from this effect has been the focus of some research using artificial post weld heat treatments.

A literature review reveals this research on artificial and natural ageing but mainly limits its focus on the effect on the microhardness in and around the SZ after FSSW and FSW. Only a few studies reported on the effects of post weld heat treatment on the joint properties of the 7xxx series. The work by Gholami et al. [26] on friction stir welded EN AW-7075 aluminium alloy indicates a hardness increase of 30% in the SZ by artificial ageing (at 80°C and 16h) compared to the as-friction-stir-welded material. In their research they investigated the effects of the ageing temperature as well as duration. The results can be seen in figure 3.12. The decrease in hardness when ageing for 16h at 160°C might suggest over-ageing and the existence of an optimum ageing temperature lower than 160°C. On the other hand the microhardness keeps rising with increasing duration.

Figure 3.12: Distribution of microhardness values in the SZ in samples a) aged at different temperatures for 16h and b) aged at 80°C for different durations [26]

19

The work on friction stir welded 7075-T651 sheets by Tufaro et al. [27] investigated the effects of natural ageing on mechanical properties. They reported that both the microhardness and the tensile strength increased with the elapsed time according to a logarithmic relationship. This continued improvement over time was also concluded for the same process and material by Fuller et al. [28].

Studies conducted by Ipekoglu et al. [29] on post weld heat treatments (PWHT) of friction stir welded 7075-T6 revealed a significant recovery in the strength of the joints and thus an increase in joint performance. The post weld heat treatment consisted of solutionising at 485°C for 4 hours followed by quenching and artificial ageing for 6 hours at 140°C. The hardness profiles of both the as-welded samples and post weld heat treated ones are shown in figure 3.13. As mentioned before, the less desirable hardness loss due to the thermal weld cycle is visible. Subjecting the samples to the heat treatment resulted in the recovery of the microhardness to the level of the base material. The effects on the tensile strength are visible in figure 3.14. The restoring effect can be seen by comparing the average tensile properties of the 7075-T6 base material to the as-welded and post weld heat treated samples. The authors also mention that temperature above 140°C might result in a slightly overaged condition and thus in a lesser recovery. As there is still a lack of research about the effects of ageing after friction spot welding (FSpW) in particular, part of this master dissertation will be dedicated to these effects.

Figure 3.13: Microhardness profile of FSW 7075-T6 in as-welded and PWHTed conditions. 1000/150 means FSW at 1000 rpm and 150 mm/s [29]

Figure 3.14: Graph showing average tensile properties of base material, as-welded and PWHTed conditions [29]

20

Chapter 4 Experimental conditions

The previous chapters mainly focused on existing literature related to this master dissertation and their results. The process and its parameters were thoroughly discussed and the basic properties of the studied material were mentioned. The aim of this chapter is to provide the details about the performed experiments, including the used hardware and the procedures followed. This research will aim to report the effects of the three most important parameters on the microstructural and mechanical weld properties. The results of this experimental study are analysed and discussed, separately for each investigated material, in chapter 5.

4.1 Experimental conditions

As already mentioned, the main objective is to study the impact of the parameters on the microstructural and mechanical joint properties. The three most important parameters with respect to these properties are selected and consist of the rotation speed (RS), the dwell time (DT) and the plunge depth (PD). Based on literature, varying each one of these parameters has the most effect on the microstructure and mechanical behaviour. No precise values from literature can be extracted since FSpW is a recent development and research generally focusses on more common alloys. As a result, a wide range of parameter values was chosen in the possible operating range of the FSpW welding machine, to include the optimum welding values. The common parameter range of the test set-up is shown in table 4.1 and is discussed below.

Table 4.1: Common operating range of the FSpW welding machine (t = plate thickness)

Rotation speed (RS) 1000 – 3300 rpm Joining time (JT) 3 – 10 s Plunge depth (PD) t – 2*t

The upper limit for the rotation speed follows from the operating limit of the motor used, which is 3300 rpm. But a lower boundary should be kept in mind as well. This is because the heat input is limited at lower speeds, resulting in too low material temperatures. At these low temperatures the material is not plasticised enough, certainly for the high strength aluminium alloys from the 7xxx series. This leads to too high axial forces for plunging the tool into the material. These high forces demand greater currents from the motors, which can cause interrupted weld cycles, due the motor current protection (safety protection). To prevent this and all the above, a lower speed limit is considered. In literature, rotation speeds for different materials can be found around 2000 rpm.

The lower limit of the joining time follows from the same explanation: a short joining time results in a low heat input and thus in too high forces. The determination of the lower limit followed from advice given by an experienced friction spot welding operator at the Belgian Welding Institute (BWI). There is also an upper limit of the welding time: too long welding times gives rise to excessive heat

21 input with as result an excessive plasticisation of the sheet material under the clamping ring. This is not beneficial since in this case, the high pressure exerted by the clamping ring will deform the sheets. Too long joining times would also make the process less economically feasible. The maximum joining time found in literature, is around 5 s.

The plunge depth is normally varied between the plate thickness and twice the plate thickness for similar material connections. In this master dissertation, these limits are also used. When joining dissimilar materials, a plunge depth smaller than the upper sheet thickness are usually used to prevent mixing of the materials. Setting the plunge depth higher than twice the sheet thickness would obviously result in damage of the tool due to contact with the backing anvil of the machine. Penetration depths around 1,25 times the plate thickness are found a lot in literature.

The plunge time and retraction time are determined by the aforementioned lower limit of the joining time and are kept equal for ease.

The clamping pressure is set lower than the nominal value of 6 bar (force on clamping ring: 16 kN). This had to be done because preliminary tests revealed indentation of the base material at the highest levels of parameters. Without this intervention, indentation under the clamping ring would cause a false determination of the plunge depth, as the machine measures the plunge depth with the top sheet surface as a reference. The surface of the clamping ring needs to be at the same level of the surface of the top sheet. On the other hand, a too low clamping pressure could allow material to get expelled from the weld zone into the area under the clamping ring during the process. Enough force is also necessary to keep the specimen in place as it endures a moment caused by the tool rotation.

A holding time of 5 seconds was also introduced to ensure that the sheets would not buckle during the first seconds of cooldown, so the weld nugget itself can be studied, without the possible adverse effects of buckling.

With this operating range and with literature data in mind, three possible values were assigned to the rotation speed, the dwell time and the plunge depth for the experiments with EN AW-7075-T6 and EN AW-6082-T6. The purpose was to investigate a wide range of parameters and to create a window where a low, medium and high value of these parameters is investigated. The parameter levels for the EN AW-7475-T761 resulted from previous experiments on the used set-up. All this resulted in the parameter matrices shown in tables 4.2, 4.3 and 4.4.

Table 4.2: Experimental parameter levels used for the material EN AW-7075-T6

Rotation speed (RS) 1500 – 2000 – 2500 rpm Dwell time (DT) 0 – 1 – 2 s Plunge depth (PD) 1,8 – 2,2 – 2,7 mm Plunge time (PT) 2 s Retraction time (RT) 2 s Clamping pressure (CP) 5 bar Holding time (HT) 5 s

22

Table 4.3: Experimental parameter levels used for the material EN AW-6082-T6

Rotation speed (RS) 1500 – 2250 – 3000 rpm Dwell time (DT) 0 – 1 – 2 s Plunge depth (PD) 2 – 2,5 – 3 mm Plunge time (PT) 1,5 s Retraction time (RT) 1,5 s Clamping pressure (CP) 5 bar Holding time (HT) 5 s

Table 4.4: Experimental parameter levels used for the material EN AW-7475-T761

Rotation speed (RS) 1500 – 2500 rpm Dwell time (DT) 2,5 – 3,5 – 4,5 s Plunge depth (PD) 1,6 – 2,1 Plunge time (PT) 2 s Retraction time (RT) 1,5 s Clamping pressure (CP) 6 bar Holding time (HT) 0 s

A design of experiments (DOE) approach was opted, based on the work by Schmidt et al. [30]. They proposed a full factorial parameter investigation for experiments studying 3 factors with 3 levels (or less). Preliminary tests had to be performed, before making the final welds, to confirm that the extreme parameter combinations produced sound, but not yet optimised, welds. This is necessary for the design of experiments approach, as unsuccessful joints reveal little information and would result in incomplete data sets. Since the welds of the 7475 series were already produced, two preliminary tests series were performed, one for the 7075 and one for the 6082 material. Unfortunately, the welding machine encountered two failures, resulting in a limited amount of samples for initial testing. One failure happened during the preliminary tests with the 7075 material. In this case, the tool broke, more precisely the sleeve failed by a sudden torsion fracture at the top of the sleeve. This was caused by the rotational force exerted by the motor on the top part of the sleeve, while the bottom part was sticking to the non-rotating clamping ring. The sticking is due to the brazing effect of the aluminium in between the two components. The failed tool can be seen in figure 4.1. Two measures were taken to reduce the stress on the sleeve and to prevent this type of failure in the future. The first was to increase the lower limit of the joining time to 4 seconds when welding the high-strength alloys from the 7xxx series. This value followed from the advice given by the machine manufacturer to set the minimum plunge time at 2 seconds for welding this material. The second measure was to allow the tool to cool down, while it was still in rotation. This prevented the formation of a strong bond between sleeve and clamping ring and was done by programming an additional step (idle state) into the machine.

23

Figure 4.1: New sleeve (left) and fractured sleeve (right)

The second machine failure happened during the preliminary tests with the material EN AW-6082-T6. The zero level of the sleeve varied when the machine settings were kept constant. The cause was a malfunction of the sleeve actuator. These equipment breakdowns initially stopped the production of additional welds for the complete full factorial approach and limited the amount of available samples for investigation. Practically, this only meant that incomplete sets of the 6082 and 7075 weld series were used for metallographic and microhardness testing. The samples available for these tests are graphically presented by red dots in the figures below.

24

Figure 4.2: Graphical representation of available samples in the material EN AW-7075-T6 from series 1

Figure 4.3: Graphical representation of available samples in the material EN AW-6082-T6 from series 1

Figure 4.4: Graphical representation of available samples in the material EN AW-7475-T761 from series 1

25

When the welding machine was repaired, a second series of welds was made with the 6082 alloy in order to study the impact of the parameters on the lap shear strength. Contrary to the previous series, the second series contained all of the possible parameter combinations mentioned before for a full factorial experiment.

4.2 Experimental test procedures

The previous section was devoted to explaining how the parameter windows were determined and which welds resulted from these parameter settings. These welds need to be examined by various testing methods to assess the joint quality. In the following section, the testing methods and procedures followed are explained. All procedures were performed at the Belgian Welding Institute, except for the microhardness maps, which were done in Laboratory Soete of Ghent University.

4.2.1 Welding As already mentioned before, each weld, consisting of two aluminium alloy sheets in the overlap configuration, was made with the RPS 100 FSpW machine from the manufacturer Harms & Wende. The machine can be seen in figure 1.6 and the technical specifications can be found in appendix A.

The RPS 100 welding machine uses a discrete number of process states, programmed by the operator, to perform the FSpW process. These process states are all characterised by a duration, positions for the sleeve and pin and their rotation speed. The positions of the sleeve and pin are coupled by a fixed ratio called the travel factor (1,023). This is necessary because the volume displaced by the plunging sleeve should be the same as the volume of the free space created by the retracting pin and vice versa. The duration given to a process state signifies the amount of time the machine has to reach the specified state positions and rotation speeds. For example: if the tool was not rotating and the next state has a rotation speed of 1000 rpm with a duration of 2 seconds, the tool will increase its rotation speed in the next step until it reaches 1000 rpm after 2 seconds. This method allows the operator to easily set the plunge-, the retraction-, and the dwell time. The change between two state values follows a certain interpolation curve. This curve can be linear, a spline or a sinus. Only the linear interpolation curve was used during the course of this dissertation.

The RPS 100 can use a fan or water cooling to cool down the tool and backing anvil. Water cooling was always switched on during welding and the cooling water temperature was between 15 and 20°C. When switching the machine on, the tool was first brought up to the minimum operating temperature by rotating the pin and sleeve freely for 30 seconds.

For obtaining a flat welded area, the final position of the sleeve and pin have to be in the same plane at the end of the weld cycle. The zeroing of the pin and sleeve was done by making test welds and measuring the height difference between the sheet surface and the welded areas under the sleeve and pin, using a dial gauge. The difference was then subtracted manually to the machine’s zero setting. The maximum allowed difference as specified by the manufacturer is 100 µm. The zero level of the sleeve caused some problems in weld series 1, but those will be discussed in chapter 5.

Prior to this dissertation, the machine was used for different material combinations, including steel, copper and other aluminium alloys. For this reason, the pin and sleeve were cleaned with a 70%

26 sodium hydroxide solution before welding to minimise contamination. After approximately 15 welds, the tool needs to be cleaned mechanically. This can be done by running the cleaning program of the machine. The cleaning program does not allow the weld head to press down and rotates the sleeve and pin at low speeds while moving them up and down in opposite directions for about 15 seconds. This results in the removal of little amounts of stuck aluminium from the clearance between the tool components. When used, it was used two or three times in succession until no more aluminium could be removed.

In order to prevent confusion and facilitate the identification of the numerous welded joints, each completed weld gets a unique identification assigned to. Each identification tag consists of IJ, which refers to the InnoJoin projects which encompassed this master dissertation, followed by the material identifier (75 for the 7075, 74 for the 7475 and 82 for the 6082), followed by R and the test series number and finally the weld number. R1 indicates the welds made during preliminary testing or the aged welds of the 7475 alloy from storage. Only one second series, indicated with R2, was made and that was for the 6082 alloy. This series was used to investigate the influence of the process parameters on the lap shear strength as the first series was used for metallographic examination. The welds from series two were made after the machine was repaired by the manufacturer.

For example, the weld identification tag IJ-75-R1-8 refers to a weld made during the preliminary test series with the 7075 aluminium alloy with parameter combination number eight. This parameter combination together with an overview of all welds, with their identification tag, can be found in appendix C. The numbers in figures 4.2, 4.3 and 4.4 also correspond to the weld condition numbers of the respective materials.

4.2.2 Visual inspection One way of assessing the weld quality is by looking at the weld surface. Surface imperfections are caused by sticking of aluminium to the tool, a bad refill or an incorrect zero setting. A flat surface without imperfections is aimed for, because it improves the corrosion resistance and fatigue strength (less possible crack initiation sites). All weld surfaces were visually inspected and divided into four different classes. These classes are shown in table 4.5. The class of each weld can be found in appendix C. Examples of welds of classes 1 and 3 can be found in figures 4.5 and 4.6 respectively.

Table 4.5: Visual inspection classes

1 Surface without imperfections 2 Surface with minor imperfections and/or small circumferential grooves 3 Surface with large imperfections and/or large circumferential grooves

27

Figure 4.5: Class 1 weld surface (IJ-82-R1-17)

Figure 4.6: Class 3 weld surface (IJ-82-R1-4)

4.2.3 Peel test Some extreme parameter combinations of each material combination were peel tested to see whether these conditions produced sound welds, i.e. welds that cannot be easily separated. This was done manually using a vise-grip and a bench screw. These few results are not discussed in chapter 5, but it can be said that all parameter combinations produced sound welds and all failure modes were plug pull-out on the upper sheet. An example can be seen in figure 4.7.

Figure 4.7: Plug pull-out on upper sheet from peel test on weld IJ-82-R1-13

28

4.2.4 Lap shear strength tests The most used testing method found in literature to assess the joint quality is the lap shear tensile test. This test was performed on the first series of the 7475 welds and on the second series of the 6082 welds, according to EN ISO 14273. The lap shear test specimens consisted of two sheets with dimensions 105 by 45 mm and an overlap of 35mm. The welds were made in the centre of the overlap area with the aid of fabricated positioning plates. The lap shear strength samples are schematically presented in figure 4.8.

Figure 4.8: Schematic representation of the lap shear strength samples

The large free length of each sheet allowed for the secure clamping of the specimens in the tensile test machine (Instron model 8801). Shims, out of the same material, were added to align the samples so that the tensile load was translated into a pure shear load on the weld nugget. The location of these shims is visualised in in the figure below together with the tensile forces.

Figure 4.9: Schematic representation of lap shear strength test

Since this test has large data scatter, for each weld condition, 3 samples were produced. In addition, the order in which the welds for the lap shear tests were made was randomised to increase the

29 reliability and validity of the analysis afterwards. All the results can also be found in appendix C. All samples are tested until destruction and the maximum force is used for further analysis. The displacement speed was fixed and set at 2 mm/min. The failure mode was also documented. The codes used (see appendix C) to indicate which failure mode occurred, can be seen in table 4.6. When the failure mode was a combination of two failure modes from the table below, both modes are specified.

Table 4.6: Failure modes

1 Through weld nugget 2 Plug pull out on upper sheet 3 Plug pull out on bottom sheet

4.2.5 Metallographic examination The preparation and examination of metallographic samples consisted of multiple sequential steps. These steps are discussed below in their correct order. All the following products and machines are from manufacturer Struers, except the optical microscope.

First, the samples are cut into pieces of about 30 by 20 mm containing the weld, without cutting through the weld nugget, as seen in figure 4.10. This results in cut samples that are larger than the weld nuggets themselves in order to include as much of the heat affected zone as possible for further testing.

Figure 4.10: Metallographic samples

These are then put into specimen mounting cups (FixiForm) with a diameter of 40mm and filled up with a mixture of EpoFix Resin and EpoFix Hardener with a 25 to 3 ratio respectively. After briefly putting the samples in a vacuum, to bring entrapped air bubbles to the surface, they are left to harden overnight in a pool of water to extract the produced heat by the exothermic chemical reaction. Epoxy is used because the hardened epoxy is clear, has good adhesion and is resistant to chemicals. The samples are then marked with their identification tags by engraving it on the back of the embedded samples. This ensures that the tags cannot be rubbed off by contact with the etching and polishing chemicals. The next step is to wet grind the samples until the centre of the weld nugget is exposed. This is done on a RotoPol-22 and RotoForce-4 grinding and polishing set-up by starting with coarse 180 grit SiC (silicon carbide) paper and gradually moving on to the finest 4000 grit SiC paper. The final polishing is done in three steps using three different polishing cloths. The first two, MD-Mol and MD-Floc, are both used in combination with a mixture of DiaPro diamond suspension and DP- Blue lubricant. The last cloth MD-Chem is used with the OP-S NonDry colloidal silica suspension for final polishing.

30

These un-etched polished samples are then photographed with a Olympus MX51 optical microscope using its lowest magnification of 12,5 times. This step is to identify the voids and the cracks, as this becomes more difficult after etching. Etching is done by briefly submerging the metallo samples in Keller’s reagent to reveal the grain boundaries and orientations. Care has to be taken not to over-etch the samples, because this would result in a very dark weld nugget cross section. The composition of the Keller’s reagent used can be found in the following table.

Table 4.7: Composition of Keller’s reagent

Component Amount [ml] HF 1 HCL 1,5

HNO3 10

H2O 100

The samples are then photographed a second time but now in the etched condition using the same optical microscope as mentioned before. Pictures are taken from the entire weld area, as well as close-ups of places of interest, geometrical patterns and any anomalies. The microscope’s software is also used to measure some weld characteristics. An example of an imbedded and etched sample can be seen in the following figure.

Figure 4.11: Embedded and etched sample

4.2.6 Microhardness measurements Microhardness measurements of some welds executed with a high or low heat input are performed after the metallographic examination. Two types of hardness measurements were done. The first was a full microhardness map. These maps are made with a Leco AMH43 Automatic Micro/Macro-Indentation Hardness Testing System. This system has the benefit of indenting and measuring an entire specified area autonomously. An example of such an indentation pattern can be seen in figure 4.12. The measurement area is large enough that it includes the base material.

31

Figure 4.12: Example of an indentation pattern

The test performed is the Vickers hardness test with 0,3 kgF load. The amount of individual indentations range from 700 points for the 1,6mm sheet welds to more than 900 points for the thicker 2mm sheet welds. The systems returns the hardness data as scatter points and a programmed MATLAB script is used to visualise the data with 2D contour plots. The second type of hardness measurements performed was a traverse line, as seen extensively in literature. This was done with the Struers Duramin-A300, which requires manual measurements of the indentations. The traverse line was situated in the middle of the top sheets. Indenting was done following Vickers hardness with 0,5 kgF load and 0,5 mm spacing. An example can be seen in figure 4.13. The determination of the hardness of each base metal was done by averaging five measurements over an un-welded single sheet specimen.

Figure 4.13: Traverse hardness measurements on IJ-82-R1-4

32

Chapter 5 Experimental results

In this chapter, the results following from the experiments and procedures explained in the previous chapter are presented and discussed. The results are classified according to the different alloys. The first section deals with the results of the EN AW-7475-T761 alloy, the second with the EN AW-6082- T6 alloy and a final section is devoted to the EN AW-7075-T6 aluminium alloy. Each of these sections is further divided based on the tests performed on the samples. The parameters corresponding with the identification tags are found in appendix C. The final conclusions regarding these results are summarised in chapter 6.

5.1 Weldability of EN AW-7475-T761

This section discusses the experimental results for the 7475-T761 alloy. As mentioned in chapter 4, these welds were made one year before the time of the weld testing. Welding of this alloy went without major issues. There was not much sticking of the alloy to the friction spot welding tool, which results in the conclusion that it is technically possible to weld the 7475 alloy.

5.1.1 Metallographic examination The general structure of the 7475 alloy mostly resembles the ones mentioned in literature. Examples of the weld structure can be seen in figure 5.1. The SZ is mostly characterised by fine equi-axed grains, with a brighter band of coarser grains in the middle (location indicated with a white arrow in figure 5.1). For the welding conditions with lower rotational speed this band is rather thin, whereas a higher rotational speed causes this band to be much broader. An increase of the dwell time results in a narrower coarse central band for both rotational speeds used in the experiments. The effect of the increasing dwell time can be seen in figure 5.1 from left to right and the aforementioned effect of the increasing rotational speed can be seen when comparing the top images versus the lower ones.

IJ-74-R1-3 IJ-74-R1-1 IJ-74-R1-5

IJ-74-R1-10 IJ-74-R1-8 IJ-74-R1-12

Figure 5.1: Effect of dwell time and rotation speed on the central coarse grain band. From left to right: dwell time of 2,5; 3,5 and 4,5 seconds. Top to bottom: 1500 and 2500 rpm

33

There is a gradual change of the microstructure from base metal to the SZ microstructure. The TMAZ can be seen in between the two, starting at the sleeve plunge path and extending towards the base material. The shape of this weld zone resembles a half ellipse, with the flat side against the sleeve plunge path, which varies in size. The TMAZ is characterised by elongated and 90° rotated grain structures due to moderate deformation and heating. It is depicted in figure 5.2. As in literature, the distinction between the HAZ and the base material cannot be seen with optical microscopy.

Figure 5.2: TMAZ of weld condition IJ-74-R1-1

The area just below the pin and sleeve interface (outer diameter of the pin, inner diameter of the sleeve) is characterised by the finest grains. This is due to the intense stirring at that location. These fine grains seem to align due to the material flow to form striations (i.e. a series of long and thin marks). These striations tend to become thinner at higher rotational speeds. These are very narrow in sample IJ-74-R1-8, which is depicted in figure 5.3, or broad, like the striations seen in figure 5.4. When the striations are very narrow, the etching process causes the highly reactive intermetallic layers, which are trapped in between, to become visible as fine black lines. This can be seen in the detail of figure 5.3.

Detail

Detail

Figure 5.3: Fine striations under pin/sleeve interface in IJ-74-R1-8

34

Figure 5.4: Broad striations under sleeve/pin interface in IJ-74-R1-1

The bonding ligament cannot be easily observed. The produced welds exhibit two types of bonding ligaments. The first being the formation of a very fine banded structure, centrally below the SZ, for all welds made with a rotational speed of 1500 rpm (i.e. weld conditions IJ-74-R1-1 up to IJ-74-R1-6). The banded structure consists of alternating finer and coarser grains. This type of bonding ligament is limited in width and can be seen in the top image of figure 5.5. It is a region of good adhesion between the upper and lower sheets, with a strong resistance.

The opposite is observed for the second type of bonding ligament, which is attributed to all welds made with a rotational speed of 2500 rpm. Within these welds, there is a zone where the two sheets are not bonded. At the ends of this un-bonded zone, a normal bonding ligament is observed which continues to the outside of the weld nugget. The debonding never extends underneath the sleeve area, which suggests that bonding under the sleeve area is stronger than bonding centrally under the pin. These un-bonded zones can be seen clearly when examining the un-etched samples, like the one in the bottom image of figure 5.5. No clear correlation can be found between the plunge depth and the extent of the debonding. For the plunge depth of 1,6 mm, the worst debonding is seen for a dwell time of 4,5 seconds (i.e. welding condition IJ-74-R1-6) and the least amount of debonding is noticed for a dwell time of 3,5 seconds (i.e. welding condition IJ-74-R1-2). This is the opposite for samples with a plunge depth of 2,1 mm, where a dwell time of 3,5 seconds (i.e. welding condition IJ- 74-R1-8) is the worst case and a dwell time of 4,5 seconds the better one (i.e. condition IJ-74-R1-12). The shortest dwell time of 2,5 seconds resulted in a mediocre amount of debonding for both plunge depths.

35

IJ-74-R1-9

Detail

IJ-74-R1-6 Detail

Figure 5.5: The two types of bonding ligament

A new kind of geometrical pattern also emerged from the metallographic examination. A dark line is found in the middle of the weld nugget for the welding conditions IJ-74-R1-8 and 12. It connects the outer areas of the nugget surface and has a W-shaped appearance. This line represents an axi- symmetrical plane around the nugget centre, which consists of very fine and compact grains. This W- shaped pattern is the most pronounced in sample IJ-74-R1-8, which is also displayed in figure 5.1 and figure 5.3. In sample IJ-74-R1-12, the bonding ligament line even fused with this fine grain structure at the outer edges. This can be seen in figure 5.6. Looking at the common parameters of both conditions, this pattern thus only appears when using a rotational speed of 2500 rpm together with a plunge depth of 2,1 mm and a sufficiently high dwell time, which causes the necessary material flow.

Figure 5.6: Fusion of BL and W-shaped fine grain structure in IJ-74-R1-12

Hooking was the most pronounced with the lower plunge depth of 1,6 mm. It was the sharpest for welding conditions IJ-74-R1-1, 3, 5 and 6. Weld conditions 2 and 4 showed a broader two pointed M- shaped hook tip. An example of these two hook tips can be seen in figure 5.7 below. No sharp hooking was observed for the samples with the higher plunge depth of 2,1 mm. Hooking was either

36 absent or ,in the cases of IJ-74-R1-10 and 8, very blunt and faint. The height of the hook, from the sheets interface to the tip of the hook (also indicated in figure 5.7), was measured for conditions 1 through 6, as these allowed accurate measurements. This revealed that the hook height increases with increasing rotation speed and dwell time. This could be attributed to the fact that higher levels of these parameters cause more plasticisation of the material, which increases the vertical material flow during the process. When comparing the hook height to the faint ones in the welds made with a plunge depth of 2,1 mm, even a higher hook is present. This might indicate that the hook height also increases with the plunge depth due to the larger vertical displacement and thus a more vertical material flow. In all samples, partial bonding was not visible.

H

IJ-74-R1-3 IJ-74-R1-2

Figure 5.7: Sharp hook (left) and M-shaped hook (right). H = hooking height

A slight form of incomplete refill is seen in figure 5.8. The circumferential groove resulting from the incomplete refill is shallower and less sharp when increasing the heat input. This defect is probably due to an insufficient movement of the pin in the refill stage of the process, as all of the twelve parameter combinations exhibit this kind of incomplete refill and a slightly elevated sheet surface area under the pin is noticed. The viscous flow during this phase did not refill the entire void created by the retracting sleeve. Adjusting the zero level of the pin, should eradicate this.

Other kinds of defects are only found for some welding conditions. Lack of mixing was observed at the sleeve plunge path for conditions IJ-74-R1-1 and 3. Small voids were also present at the edge of the hook for weld condition IJ-74-R1-1. It thus seems that these defects can be attributed to an insufficient heat input and poor material flow. Examples of the defects are shown in the figure below.

IJ-74-R1-8 IJ-74-R1-1 IJ-74-R1-1

Figure 5.8: Incomplete refill (left), lack of mixing (middle) and voids (right)

37

5.1.2 Microhardness Both traverse microhardness lines as microhardness maps were made. The results of the more common traverse microhardness lines are discussed first. One should keep in mind that the 7xxx series has the ability to recover by ageing a great part of the loss of mechanical properties due to the thermal welding cycle. The largest part of this recovery is in the first months, so the following results represent the end state as the microhardness tests were performed more than a year later. Two important phenomena act inside the weld, recrystallisation and precipitate dissolution. The first one causes an increased microhardness. The opposite is true for precipitate dissolution.

Traverse microhardness lines Traverse hardness lines were made for the high and low heat input welding conditions, both with a 1,6 mm plunge depth. These conditions are IJ-74-R1-3 (1500 rpm, 2,5 seconds DT) and IJ-74-R1-6 (2500 rpm, 4,5 seconds DT); their hardness profiles are plotted in figure 5.9. The edge of the SZ is indicated with the black dotted line and the base metal hardness with the red dotted line.

The two profiles resemble the ones found in literature. They have a distinct W-shape, as expected for a heat treatable alloy. Whereas the HAZ could not be spotted with optical microscopy, the start of this zone is visible in the hardness profiles as the downward slope, connecting the base material hardness to the minimum hardness points. The right part of the HAZ of the high heat input sample could not be reached by the indenter, but as the weld zones are axi-symmetrical, one side is enough for interpretation. Some differences are noticed between the two. The first one is found in the extent of the hardness drop seen in between the HAZ and the TMAZ. The low heat input has a hardness drop of 15 HV0,5, while the high heat input sample causes a 37 HV0,5 drop. This means that the higher heat input sample has a 147 % increase of the hardness drop, which shows the sensitivity of the alloy to the heat input. The cause of the increase itself is the more extreme dissolution of the precipitates in the case of a higher heat input. When welding 7475-T761, the parameter combination used should be the one with the lowest heat input possible, while still producing strong joints, to minimise this drop in microhardness. The second major difference is the shape of the SZ profile. The high heat input sample displays a more flat SZ profile with an average of 155 HV0,5, whereas the low heat input sample displays distinct peaks with a slightly higher average of 161 HV0,5. The outer peaks of the SZ in the IJ-74-R1-3 sample can be explained by looking at their location. They are situated under the sleeve and pin interface and as the metallographic examination revealed, this is where the finest grains are located. These finer grains are the cause of the outer hardness peaks. The central peak is caused by the ageing after the solubilisation of the precipitates due to the highest temperature, reached at the nugget centre. This results in precipitation hardening. The flat SZ profile of the IJ-74-R1-6 sample can be attributed to the more intense mixing, which causes a more homogeneous overall grain structure. The excessive heat also causes the recrystallization to extend further than the sleeve plunge path, which causes the SZ profile to appear slightly wider.

38

Figure 5.9: Low heat input (left) and high heat input (right) microhardness profiles

Microhardness maps Microhardness maps were also made for the high and low heat input welding conditions, but as opposed to the traverse lines, both samples had a plunge depth of 2,1 mm. This resulted in hardness maps for conditions IJ-74-R1-9 (1500 rpm, 2,5 seconds DT) and IJ-74-R1-12 (2500 rpm, 4,5 seconds DT). The hardness maps can be seen in figures 5.10 and 5.11. The origin of the maps is located in the middle of the nugget’s surface and the outer edge of the sleeve plunge path is marked with a black dotted line, which contains the SZ. The start of the HAZ and thus the end of the TMAZ can only be seen clearly for the low heat input sample. For this case, these zones are also indicated. The exact location of the boundary between TMAZ and HAZ is not clear for the high heat input condition, so no indication is made.

The following conclusions can be drawn. The weld nugget shows a fairly homogeneous hardness distribution. The absence of well-defined structures may be the consequence of the natural ageing behaviour of the 7475 alloy. These maps were made a year after welding, which could have given the weld enough time to recover and smooth out its hardness distribution. A higher average hardness in the SZ can also be seen for the low heat input sample. This is due to the existence of multiple lower hardness zones in the SZ of the high heat input sample. The highest hardness values are reached along the sleeve plunge path, with this zone extending slightly further for the high heat input sample.

HAZ TMAZ

Figure 5.10: Low heat input microhardness map

39

Figure 5.11: High heat input microhardness map

5.1.3 Lap shear strength As stated before the lap shear strength of all weld conditions were determined, with a full factorial experimental design using three times repetition, in order to study the influence of the process parameters on the mechanical performance of the welds. The 7475 alloy is used mostly in high-end or aerospace applications, so the weld strengths can be assessed based on the traditional spot welding standard for the aerospace industry; ISO 16338:2013. It dictates that the average lap shear strength for 1,6 mm thick aluminium alloy sheets, with an ultimate strength greater or equal than 386 MPa, must be greater than 3,635 kN while individual results must exceed 2,980 kN. The EN AW- 7475-T761 alloy falls under this category due to its ultimate strength of 524 MPa. The individual results from the test ranged from 2 to 6 kN, resulting in averages from 2,5 to 5,4 kN. This means that some welding conditions produce inferior welds while others produce acceptable joints. When using this standard the only acceptable joints resulted from welding conditions IJ-74-R1-7, 11 and 12. The best lap shear strengths were obtained for IJ-74-R1-11 (i.e. RT 1500 rpm, 2,1 mm PD and 4,5 seconds DT). This was expected, because this welding condition produced the most sound metallographic cross-section in terms of homogeneity and absence of weld defects. The worst results, with an average tensile force of 2,5 kN, were obtained for condition IJ-74-R-3. This is due to a too low heat input and the defects associated with it, as this condition has the lowest rotation speed, plunge depth and dwell time of all the conditions.

Main parameter effects The main effect plots can be seen in figure 5.12. The maximum and minimum lap shear strengths are indicated with an upward and downward facing triangle respectively. The average value is indicated with a red dot and the standard deviation is represented by an error bar. The minimum value determined by the standard ISO 16338:2013 is also indicated with the red dotted line.

The large scatter of the test results can be seen immediately and no plot maxima have been reached, so additional tests should be done, outside the current parameter range, to determine the optimal parameter values. An analysis of variance (ANOVA with significance level 0,05) determined that the effect of the rotation speed is statistically insignificant (p-value 0,563), while the effects of the other process parameters are significant. The following trends are noticed. The lap shear strength increased with the plunge depth. This can be attributed to the fact that the increase in plunge depth causes the absence or very faint presence of hooking. As hooking is one of the most important crack initiation sites, a reduction in sharp hook thus results in an increase in lap shear strength. The strength of the joint also increases with increasing dwell time. This can be due to the fact that an increase in dwell time causes a more homogeneous nugget microstructure and the longer welding

40 time allows for more recrystallisation. Although insignificant, a slightly lower strength was observed for the higher rotational speed, which could be the result of the formation of the un-bonded regions, which in term compromises the integrity of the joint.

Figure 5.12: Parameter influences on the lap shear strength

Parameter interactions The benefits of a full factorial experiment is that the interactions between the process parameters can be studied as well. The parameter interaction plots can be seen in figure 5.13. An ANOVA test (significance level 0,05) revealed that there is only a significant interaction between the plunge depth and the dwell time. The increase in lap shear strength with the dwell time is much larger when using a plunge depth of 2,1 mm, compared to the lower plunge depth of 1,6 mm. This interaction works both ways. Other interactions are not discussed as they are not statistically significant.

41

Figure 5.13: Parameter interaction plots

5.1.4 Overview The previous sections confirm that the three main process parameters have a great impact on many different weld properties. It is thus certainly practical and beneficial to summarise these correlations. Table 5.1 below provides the reader with the most significant correlations.

Table 5.1: Overview of significant parameter correlations. 1) only for 1,6 mm PD

Central coarse Striation Central Hooking Hook Defect HAZ hardness LSS band thickness bonding sharpness height1) amount drop width RS ↗ ↗ ↘ ↘ - ↗ ↘ ↗ - DT↗ ↘ - - - ↗ ↘ ↗ ↗ PD↗ - - - ↘ - - - ↗

5.2 Weldability of EN AW-6082-T6

This section presents the results of the experiments performed on the EN AW-6082-T6 aluminium alloy. This alloy is the weakest and the most common of all three alloys investigated. It showed to be the easiest to weld. A new and clean tool was used for these welds because the previous tool broke during the welding experiments for the 7075 series and welding of this alloy commenced thereafter. Two series were welded; the first was used for metallographic examination and the second for lap shear strength testing. The experimental conditions for both series can be found in appendix C. There was little to no material accumulation in the gap between sleeve and pin and very little sticking of the 6082 alloy to the free surface of the tool. This means that more welds could be made without tool

42 cleaning, which is a great benefit when the FSpW process is used to weld the 6082 aluminium alloy in industry.

Welding condition IJ-82-R1-4 was an exception. This welding condition, with its very low rotational speed of 1000 rpm and thus a low heat input, caused an irregular surface underneath the pin. The result can be seen in the previous chapter, figure 4.6. One of the problems experienced during welding of the first series was the unstable zero position of the sleeve, which sometimes caused incorrect penetration and retraction levels, but this could be corrected by changing the zero level manually. The second weld series did not encounter this issue as the machine was repaired. This resulted in all welds from the second series having a class one visual inspection surface. The stable zero setting also allowed the second series samples to be welded in quick succession. This heated up the tool significantly and more sticking of the alloy to the clamping ring was noticed for increased tool temperature.

5.2.1 Metallographic examination The cross sections reveal a weld structure that is similar to the ones found in literature. Examples of the weld structure are shown in figure 5.15. Although individual grains were more difficult to distinguish than in the case of the 7475 welds, the SZ is again characterised by very fine equi-axed grains. They result in a fairly homogeneous nugget microstructure for most welding conditions. The TMAZ has the distinct half ellipse shape, also seen in the 7475 welds, and is depicted in figure 5.14. The lines inside the TMAZ signify the moderately deformed grains present. They are bend downwards and inwards towards the nuggets centre. This reveals that the majority of the grains inside the TMAZ are deformed during the sleeve plunge phase. The grains are pulled down because they still have a strong connection to the yet to be plasticised material under the sleeve. When the sleeve retracts, the material under the sleeve is plasticised, resulting in a less strong connection with the TMAZ grains. This causes the only slight upward bend closest to the SZ. The sheets un-bonded interface bends upwards and fuses around the TMAZ for all conditions with a plunge depth of 3 mm. This is also visible on the left side of figure 5.14. In the samples with a plunge depth of 2 mm, the interface bends downward. The transition between the base material and the TMAZ is less gradual than for the 7475 alloy. It seems that the HAZ zone can be discerned from the base material with optical microscopy, as opposed to literature. The HAZ is defined by a dark band, indicated with a white arrow in figure 5.15. This zone becomes darker and extends more outward for increasing heat input. The effect of increasing the rotation speed, an important factor in the heat input, can be seen in figure 5.15 from left to right.

43

Figure 5.14: TMAZ of IJ-82-R1-14

IJ-82-R1-4 IJ-82-R1-2 IJ-82-R1-3

Figure 5.15: Typical weld cross sections. White arrow indicates dark HAZ band. Left to right: 1000; 2000 and 3000 rpm

The area under the sleeve and pin interface is once again characterised by the finest grains. The striations described for the 7475 samples can also be seen inside the 6082 alloy welds. Weld conditions with a plunge depth of 2 mm and a dwell time of 2 seconds, which are IJ-82-R1-8, 12 and 16 (parameters in appendix C), result in the most visible striations. These become thinner and fade away due to better mixing as the rotational speed increases. This becomes clear in figure 5.16. For all other welding conditions, clear striations are limited in length and confined closer to the surface, which suggests the importance of the plunge depth and the dwell time on the material flow and the alignment of the fine grains.

IJ-82-R1-16 IJ-82-R1-8 IJ-82-R1-12

Figure 5.16: Effect of rotation speed on the fine grain striations. Left to right: 1500; 2000; 3000 rpm

The bonding ligament, which can be seen in detail 1 of figure 5.17, consists of fine grains and forms a zone of good adhesion between the two sheets. It is visible in all welding conditions but the general

44 appearance and behaviour changes with the plunge depth. A plunge depth of 2 mm results in a flat bonding ligament line with partial bonding at the ends. In this case, the width of the bonding ligament and the length of the partial bonding increases with increasing dwell time. These effects can be seen when comparing the top versus the lower image in figure 5.17. The length of the partial bonding, shown in detail 2, also shortens with increasing rotation speed. This translates into a longer bonding ligament. When considering the plunge depth of 3 mm, the bonding ligament is bend upwards and no partial bonding is documented. This behaviour is shown in figure 5.18. In these welds, the width of the bonding ligament does not seem to change noticeably with the dwell time. It can be concluded that the vertical flow at the sleeve plunge path is greater with deeper plunge depths. This causes a better mixing at the expected partial bonding location, this in turn dissolves the partial bonding in the upward motion. A new pattern occurred in welding conditions IJ-82-R1-15 and 17 due to an unknown material flow. This pattern is indicated in detail 1 of figure 5.18. It is a dark thin line, consisting of very fine grains, attached to the lower part of the sleeve plunge path and extending downwards and inwards.

Detail 1

IJ-82-R1-2 Detail 1

Detail 2

IJ-82-R1-8 Detail 2

Figure 5.17: Typical bonding ligament behaviour of welds with a plunge depth of 2 mm

Detail 1 IJ-82-R1-15 Detail 1

Figure 5.18: Typical bonding ligament behaviour of welds with a plunge depth of 3 mm

45

Hooking was the sharpest in welds produced with a plunge depth of 2 mm and became sharper for a decreasing rotational speed. Although the upward bending of the bonding ligament was the highest for a plunge depth of 3 mm, no very sharp hooking was noticed.

The normal FSpW defects, lack of mixing and voids, were also observed in the 6082 welds. Lack of mixing only occurred for IJ-82-R1-4 and 9. These welding conditions have the two lowest rotational speeds (1000 and 1500 rpm), with the deepest plunge depth (3 mm) and without dwell time. This again shows that lack of mixing is a consequence of a too low heat input and the poor plasticisation of the material. Voids were noticed at two locations, at the end of the sleeve plunge path and at the centre of the bonding ligament. The latter mostly occurred at the highest rotational speed of 3000 rpm and these voids become smaller with increasing dwell time. Voids at the end of the sleeve plunge path only occurred for weld conditions IJ-82-R1-4, 6, 8, 9 and 16, all situated in the lower heat input range. There were also some unconventional defects present in sample IJ-82-R1-14 (see figure 5.19). This condition resulted in cracks, originating from the surface just underneath the pin’s centre. This condition consists of the most extreme parameters, with a rotation speed 3000 rpm, a plunge depth of 3 mm and 2 seconds of dwell time. This results in excessive heat, which even causes surface cracks on the free surface of the bottom sheet, also shown in figure 5.19.

Detail 1

Detail 1 Detail 2

Detail 2

Figure 5.19: Surface cracks of IJ-82-R1-14

5.2.2 Microhardness The microhardness properties of the 6082-T6 welds were also investigated using both traverse lines and microhardness mapping methods. The 6082 aluminium alloys does not possess the ability to recover the largest part of the drop in mechanical properties due to the thermal welding cycle. This means that no excessive ageing is required in order to measure the end state. Therefor it is acceptable that these experiments were performed 2 months after welding.

Traverse microhardness lines Traverse hardness lines were also made for low and high heat input welding conditions both with a plunge depth of 2 mm. The low heat input condition was IJ-82-R1-4, with a rotation speed of 1000 rpm and 0 seconds of dwell time, while the high heat input one was IJ-82-R1-12, with a rotation speed of 3000 rpm and 2 seconds of dwell time. Note the very large difference between the rotational speeds. The hardness profiles can be seen in figure 5.20. The outer edge of the sleeve

46 plunge path is indicated with black dotted lines whereas the base metal hardness is indicated with the red dotted line.

Both profiles have the W-shape mentioned in literature, as the 6082 is also a heat treatable alloy. The HAZ, seen as the descending slope in the hardness profile connecting the base metal hardness to the lowest point, is almost entirely included in the range of indentations. This differs from the 7475 alloy, where only a small part of the HAZ could be measured. There are many clear differences between the low and high heat input profiles, due to the fact that the rotational speeds and the heat input lie so far apart. The first difference is that the slope, representing the rise in hardness in the HAZ, is steeper for the low heat input case, meaning that the lower heat input does not impact the material as far as for the high heat input case. The HAZ for IJ- 82-R1-4 stops around 8,5 mm of the joints centre whereas the HAZ for IJ-82-R1-12 stops slightly further than 13,5 mm. The second noticeable difference is the location of the hardness minimum. The minimum is situated at 6 mm in the high heat input profile, much further than the minimum of the low heat input sample, which is situated at the sleeve plunge path. The shape of the profile inside the SZ also significantly changes with the heat input. The increasing slope, which typically represents the TMAZ, can be seen connecting the minimum point to the start of the SZ for welding condition IJ-82-1-R1-12, in the same way as reported in literature. In IJ-82-R1-4 however, as the minimum is located at the start of the SZ, this slope continues through the SZ until the symmetrical slopes meet at the centre, creating a small hardness peak. In contrast to the high heat input condition IJ-82-R1-12, a hardness dimple is seen in the middle of the SZ together with microhardness peaks at either side. These peaks are the result of the finer grains underneath the sleeve and pin interface. It thus seems that a higher heat input causes the TMAZ slopes to move outwards together with the minimum and the edge of the HAZ. This causes a plateau to be formed in between the SZ hardness peaks. The average hardness of the SZ, 78 HV0,5, resulting from the high heat input is also higher than that resulting from the low heat input, 55 HV0,5. The opposite behaviour to that of the 7475 alloy is seen when looking at the extent of the hardness minima. In the 7475 experiments, the minima significantly decreased with increasing heat input, whereas those of the 6082 alloy seem to slightly increase (53 HV0,5 for condition IJ-82-R1-4 versus 63 HV0,5 for condition IJ-82-R1-12). The overall lower microhardness values for the low heat input might be due to the low heat input causing less recrystallisation, resulting in coarser and softer grains.

47

Figure 5.20: Low heat input (left) and high heat input (right) microhardness profiles

Microhardness maps The microhardness maps were made for high and low heat input weld conditions both with a plunge depth of 3 mm. Conditions IJ-82-R1-9 (1500 rpm and 0 seconds dwell time) and IJ-82-R1-14 (3000 rpm and 2 seconds of dwell time) were chosen and their hardness maps are shown in figures 5.21 and 5.22. The origin is situated in the middle of the nugget’s surface and the outer edge of the sleeve plunge path is indicated with a black dotted line, which contains the SZ. It is clear that the following hardness maps contain well-defined structures and zones, caused by the alloy’s inability to naturally age and therefore homogenise the hardness of the cross-section over time. This allows for the clear indication of both TMAZ and HAZ in the figures below. They are indicated on one side, but the zones are symmetrical to the weld centreline as they are bowl shaped. For the 6082 alloy, the following observations can be made based on the different hardness distributions of welding conditions IJ-82- R1-9 and 14.

The first clear difference is that the higher heat input causes the TMAZ to extend further away from the SZ. The width of the TMAZ is approximately 1,5 mm for the low heat input condition versus 2,5 mm for the high heat input condition. This also causes the HAZ to start at a larger distance from the centre. The width of the HAZ seems to increase with the heat input as well. Its width changes from about 4 mm to more than 6 mm. The above observations are in line with the conclusions resulting from the traverse hardness lines. The border between the TMAZ and the HAZ seems to be characterised by a large area of lower hardness in case of weld condition IJ-82-R1-14 whereas this area is rather thin for weld condition IJ-82-R1-9.

The SZ, resulting from the low heat input, has an inhomogeneous appearance. There are distinct zones with a lower hardness and a harder core. This harder centre was also noticed in the traverse hardness lines. It becomes clear, when comparing its microhardness map to its metallographic cross- section shown in figure 5.23, that the areas of lower hardness correspond with the darker areas in the cross-section. On the other hand, the SZ of the high heat input condition has a homogeneous higher hardness distribution around its centre, which has a slightly lower hardness. This softer centre was again visible in the traverse hardness lines. The dark zones seen in the metallographic cross-

48 section of IJ-82-R1-9 that signify the lower hardness zones, are not visible in the cross-section of the higher heat input sample also shown in figure 5.23. This together with the absence of distinct lower hardness zones in the SZ confirms the previously made link. The homogeneity can be attributed to the more intense mixing, resulting from the higher rotational speed and dwell time. This also allows for more extensive recrystallisation which in term causes the higher hardness documented. It thus seems that the area below the sleeve becomes harder and the core softer with increasing heat input.

TMAZ HAZ

Figure 5.21: Low heat input microhardness map

TMAZ HAZ

Figure 5.22: High heat input microhardness map

IJ-82-R1-9 IJ-82-R1-14

Figure 5.23: Low heat input (left) and high heat input (right) metallographic cross-sections

49

5.2.3 Lap shear strength A full factorial experiment, with three times repetition, was performed on the second weld series of the 6082 alloy in order to study the influence of the process parameters on the lap shear strength. Each welding condition was tested three times and the scatter in test results fairly small, with an average standard deviation of 0,22 kN. The ultimate tensile strength of the 6082 alloy is 290 MPa and the used sheets have a thickness of 2 mm. When looking at the ISO 16338 aerospace standard for resistance spot welds, these properties correspond to a minimum average lap shear strength of 4,760 kN and the individual results may not be lower than 3,805 kN. However, the lowest individual result recorded was 5,68 kN and the averages are never below 5,85 kN. This means that according to the ISO 16338 standard, every single parameter combination resulted in a joint with acceptable strength. This already indicates the good weldability of the 6082 alloy when using the friction spot welding process.

Main parameter effects The main effect plots of the process parameters are displayed in figure 5.24. The maximum and minimum lap shear strengths are indicated with an upward and downward facing triangle respectively. The average value is indicated with a red dot and the standard deviation is represented by an error bar.

An ANOVA test (significance level 0,05) was conducted and confirmed that the all the parameter values have a significant effect on the lap shear strength. It is immediately clear from the lap shear strength range that very strong connections were made, much stronger than the 7475 alloy welds. Each parameter influence plot exhibits a maximum, meaning that the optimal welding parameter combination is contained in the suggested parameter window. The highest average lap shear strength (10,52 ± 0,20 kN) is obtained when all these maxima are combined (i.e. RT 2250 rpm, PD 2,5 mm and DT 1 s). This corresponds to weld condition IJ-82-R2-14. The worst lap shear strength (5,58 ± 0,13 kN) was the result of the most extreme parameter combination, IJ-82-R2-27. This might be due to the surface cracks, seen in the previous section, acting as crack initiation sites.

50

Figure 5.24: Parameter influences on the lap shear strength

Parameter interactions The interactions between the different process parameters are plotted in figure 5.25. None of the plots show completely parallel lines, which suggests the presence of interactions. An ANOVA test (significance level 0,05) revealed that all process parameters have a significant effect on each other. The interactions with the largest impact are discussed below, but all of the interactions can be seen in figure 5.25. The largest interaction is between the plunge depth and the rotation speed or dwell time. Increasing the rotation speed or dwell time has a negative effect on the lap shear strength when using a plunge depth of 3 mm. This becomes clear when looking at the second row of plots in figure 5.25. It can be also concluded that increasing the dwell time, causes a maximum in combination with higher rotational speeds (2250 and 3000 rpm), whereas the lap shear strength keeps rising in combination with a rotation speed of 1500 rpm. The last major interaction is between dwell time and rotation speed or plunge depth. Using a dwell time of 2 seconds causes the lap shear strength to decrease with increasing rotation speed or plunge depth. Lower dwell times cause a maximum strength.

51

Figure 5.25: Parameter interaction plots

Failure modes The failure mode of each tested weld was documented and can be found in appendix C. This shows that the most common failure mode was plug pull out on upper sheet, followed by through weld fracture. The through weld fracture path mostly appears for the plunge depth at interface (2 mm), but when the heat input is sufficiently high, this failure modes does not appear again. Plug pull out on lower sheet is only documented in the strongest joints. The nuggets that were pulled out, either from the top or bottom sheet, are slanted. This indicates the strong connection between the nugget and the base material. Two combinations of failure modes are also documented. Through weld and plug pull out on upper sheet are seen together in two welds of condition IJ-82-R2-1. The second combination is plug pull out on upper sheet together with plug pull out on lower sheet, resulting in the complete separation of the weld nugget. This only happened to one weld of condition IJ-82-R2-20. An example of each failure mode can be found in figure 5.26.

52

1 2 3

4 5

Figure 5.26: Failure modes: 1) trough weld 2) plug pull out on upper sheet 3) plug pull out on lower sheet 4) combination of through weld and plug pull out on upper sheet 5) combination of plug pull out on bottom and upper sheet

5.2.4 Overview The many and most significant correlations between the main parameters and the weld properties are again summarised. The result can be seen in table 5.2.

Table 5.2: Overview of significant parameter correlations. 1) only at 2 mm PD

BL Partial HAZ Striation BL Hook Defect BL SZ width bonding HAZ TMAZ hardness LSS thickness bending sharpness amount voids hardness 1) length 1) drop RS↗ ↘ - - ↘ ↘ ↘ ↗ ↗ ↗ ↘ ↗ ↗↘ DT↗ - - ↗ ↗ - ↘ ↘ ↗ ↗ ↘ ↗ ↗↘ PD↗ - ↗ - - ↘ ------↗↘

5.3 Weldability of EN AW-7075-T6

This section deals with the results of the experiments performed on the 7075-T6 aluminium alloy. The alloy is the strongest and also the most difficult of the three alloys to weld. The cleaning program had to be used after every couple of welds due to a lot of material accumulating in the clearance between the sleeve and pin along with extensive sticking to the free surface of the tool. This was more extreme than during welding of the 6082 alloy, which caused the unstable zero level of the sleeve to be a larger issue. A cross-section, resulting from this type of failure can be seen in figure 5.27.

Plasticised material pushed in between the sleeve and clamping ring cooled down in between welds to form a strong bond. The welding machine had to overcome this brazing effect at the start of each weld, resulting in an audible cracking sound. As mentioned in chapter 4 this eventually caused a sudden torsion fracture of the sleeve resulting in the limited amount metallographic samples

53 available. When testing the consistency of the welds made, all the above became even worse when the tool and material gets hotter due to prolonged use. With the current tool, it thus seems that the 7075-T6 alloy is not that suitable for friction spot welding. A new tool, out of Thermodur material, was fabricated and heat treated in an attempt to overcome this excessive sticking of the alloy. The effects were not tested during the course of this master dissertation and therefore additional research should be performed towards possible tool materials or coatings.

Figure 5.27: Cross-section of IJ-75-R1-5 resulting from the unstable sleeve zero level

5.3.1 Metallographic examination The metallographic cross-sections resemble the ones seen from the 7475 aluminium alloy welds and thus the ones found in literature. The SZ is again characterised by fine equi-axed grains and the band of coarser grains down the middle shows some of the same behaviour as well. The coarse band is significantly wider for the higher rotational speed. However it only becomes thinner with the dwell time in combination with a plunge depth of 1,8 mm. An example of the typical microstructure, including the coarse grain band, can also be seen in figure 5.28. The TMAZ has the same appearance as the TMAZ of the 7475 welds shown in figure 5.2. However it appears smaller in size, probably due to the stronger 7075 resisting more to the moderate deformation which defines the TMAZ. Sample IJ-75-R1-7 was an exception. It can be seen in figure 5.28 that the TMAZ consists of two small half ellipse shaped structures instead of one long one. The HAZ is not discernible from the base material with optical microscopy.

54

Detail 1 Detail 2

Detail 1 Detail 2

Figure 5.28: Cross-section of IJ-75-R1-7

The finest grains are again seen under the sleeve and pin interface, together with the same previously mentioned striations, like those observed in figure 5.27. The bonding ligament line was not clearly visible except for IJ-75-R1-7. Figure 5.28 shows the clear presence of the bonding ligament line, consisting of very fine grains with a wider central part. This figure also shows the dark band of fine grains in the middle of the nugget, first spotted in the 7475 series welds. It is only seen for weld condition IJ-75-R1-7. This condition shares the same rotational speed (2500 rpm) and almost the same plunge depth (2,7 vs 2,1 mm). It thus seems that these conditions together with a sufficiently high dwell time cause the necessary material flow to form this pattern. As with the 7475 welds, no partial bonding is documented. A slight upwards bending of the lower sheet is seen, but without sharp hooking. No clear correlation could be found between the extent of the upwards bending and the process parameters. Besides the unstable zero level, other defects are seen as well. Lack of mixing was present in IJ-75-R1- 6 and conditions IJ-75-R1-2, 3 and 6 resulted in voids at the end of the sleeve plunge path. Due to the limited samples, only a few conclusions could be made and no clear correlations could be found between the parameters and the above structures and patterns. Further investigation should be done on full sets of metallographic examples to reveal these correlations.

5.3.2 Microhardness Traverse microhardness lines and microhardness maps were made on low and high heat input samples. This time, both types of tests were performed on the same samples with necessary re- polishing in between tests. The low heat input was represented by weld condition IJ-75-R1-5 (1500 rpm, 0 seconds DT and 2,7 mm PD), while the high heat input sample resulted from condition IJ-75- R1-7 (2500 rpm, 2 seconds of DT and 2,7 mm PD).

Traverse microhardness lines The results of the traverse hardness lines, seen in figure 5.29, are discussed first. The base metal

55 hardness is indicated with a red dotted line and the outer edge of the sleeve plunge path is represented by a black dotted line.

At first glance, it can be seen that the hardness measurements have a larger scatter than the 7475 or 6082 profiles, but the following conclusions can still be made. The W-shape of a heat treatable alloy’s profile can be observed, but to a limited extent. The location of the minimum, the boundary between HAZ and TMAZ, is further away from the centre in the case of the high heat input (6 mm versus 8 mm). This is the result of the extension of the TMAZ with increasing heat input. The hardness values of the minima themselves are only slightly lower in sample IJ-75-R1-7. In both profiles, the HAZ is fairly large, resulting that the base material is not reached in the measurements range. The slope of the HAZ of the low heat input case is steeper than the slowly rising one of the high heat input sample. When comparing the stir zones of the samples, it can be said that the SZ, resulting from the low heat input, is flatter with hardness peaks documented at the sleeve plunge path. The high heat input does not have these hardness peaks, but it has an only slightly lower overall hardness together with a hardness plunge in the middle. This plunge is the result of the very coarse grains seen in the middle of the nugget for sample IJ-75-R1-7, depicted in figure 5.28. As opposed to the 7475 alloy from the same family, where an increased SZ hardness is perceived, welding of the 7075 causes a much lower SZ hardness relatively to the base metal hardness. This might indicate that not recrystallisation, but dissolution of the precipitates is the dominant acting phenomenon.

Figure 5.29: Low heat input (left) and high heat input (right) microhardness profiles

Microhardness maps The microhardness maps of the low and high heat input conditions can be seen in figures 5.30 and 5.31 respectively. The origin is situated in the middle of the nugget’s surface and the SZ is contained within the black dotted lines. The symmetrical TMAZ and HAZ are also indicated on one side. Based on the differences between the two maps, the following conclusions can be made.

It is clear that the hardness distribution is fairly homogeneous inside the weld nugget, just as with the 7475 welds. The small amount of well-defined structures was expected as the two alloys are from the same family, which has the ability to recover through natural ageing. The higher heat input causes the TMAZ and the HAZ to extend further, meaning that more material is influenced by the thermal weld cycle. The boundary between the two exhibits the same behaviour as

56 seen in the 6082 samples. The boundary, which is a zone characterised by the lowest hardness values, widens when increasing the heat input. The SZ resulting from the lower heat input is characterised by a higher average hardness. It seems that the dominant effect is the dissolution of the precipitates. This becomes more extreme with increasing heat input, which causes the overall hardness to be lower. The outline of the bonding ligament zone, indicated with a white dotted line, can be recognised when comparing figure 5.31 to figure 5.28. The bonding ligament resulting from a high heat input parameter combination corresponds with a zone of lower microhardness.

TMAZ HAZ

Figure 5.30: Low heat input microhardness map

TMAZ HAZ

Figure 5.31: High heat input microhardness map

5.3.3 Overview Due to the limited samples in the investigation of the weldability of the 7075 aluminium alloy, only a few significant correlations were found between the main process parameters and the weld properties. However, an overview table was made for the readers’ convenience and can be seen below.

Table 5.3: Overview of significant parameter correlations

SZ TMAZ HAZ hardness RS ↗ ↗ ↗ ↘ DT↗ ↗ ↗ ↘ PD↗ - - -

57

Chapter 6 Conclusion

In the first chapter, an introduction was given to provide the reader with the necessary insights by explaining the process, the equipment and its parameters. Chapter two was devoted to the materials studied in this master dissertation together with their most important properties and emphasises the need for the friction spot welding technique by discussing their poor traditional fusion welding weldability. The third chapter forms the foundation on which this master dissertation was built. It discusses the results obtained by other research groups and more importantly reveals the areas where additional research should be done. After gathering information, an experimental program, aiming to produce high quality connections, was designed and explained in chapter four. The choice of parameters, considered experiments and procedures followed were detailed and motivated together with the experimental scene. The results obtained are analysed in the previous chapter where the metallographic and mechanical properties are discussed. More specifically, the influence of the plunge depth, rotational speed and welding time on the microstructure, shear strength and weld microhardness was investigated. This final chapter summarises the main conclusions.

For the EN AW-7475-T761 aluminium alloy, the following conclusions can be made:

 The strongest welds are made with the following parameters: rotation speed 1500 rpm, plunge depth 2,1 mm and dwell time 4,5 sec. This resulted in the most sound metallographic cross section and a average lap shear strength of 5,38 ± 0,56 kN.  The lap shear strength significantly increases with dwell time and plunge depth and an interaction between the two exists.  A rotation speed of 2500 rpm resulted in an un-bonded zone centrally under the stir zone which resulted in a lower lap shear strength.  Insufficient heat input results in lack of mixing and voids.  Fine grain striations are seen under the sleeve/pin interface characterised by a higher hardness.  A band of coarse grains is seen along the axis of the weld nugget, which narrows with increasing dwell time.  A dark w-shaped band of fine grains forms in the middle of the weld nugget when using: rotation speed 2500 rpm, plunge depth 2,1 mm and a sufficiently high dwell time (3,5 and 4,5 sec).  A higher heat input results in a lower hardness inside the homogeneous weld nugget and a lower hardness drop between TMAZ and HAZ.

The main conclusions resulting from the EN AW-6082-T6 aluminium alloy experiments are:

 The strongest joints are made with the following parameters: rotation speed 2250 rpm; plunge depth 2,5 mm and dwell time 1 sec. This results in an average lap shear strength of 10,52 ± 0,20.

58

 The lap shear strength has a maximum for all process parameters and interactions between all parameters exist.  The alloy has a very good weldability when using the FSpW process.  The most sound metallographic cross-section was made with: rotation speed 1500 rpm, plunge depth 3 mm and dwell time 2 sec.  The HAZ can be discerned from the base material with optical microscopy.  Fine grain striations are seen under the sleeve/pin interface characterised by a higher hardness and dissolve with increasing rotation speed.  The bonding ligament behaviour changes with the plunge depth. The length increases with the rotation speed for a plunge depth of 2 mm. No partial bonding is observed for a plunge depth of 3 mm.  The following parameter combination resulted in surface cracks in both top and bottom sheets: rotation speed 3000 rpm, plunge depth 3 mm and dwell time 2 sec.  Too low heat input results in lack of mixing and voids.  A higher heat input extends the TMAZ and HAZ outwards while increasing the overall SZ hardness.  The lower hardness boundary between TMAZ and HAZ widens with increasing heat input.

The conclusions about the EN AW-7075-T6 aluminium alloy can be summarised as follows:

 More research on tool materials and coatings should be made to eliminate the sticking of the alloy onto the tool.  The weld structure resembles that of the 7475 alloy.  A dark W-shaped band of fine grains forms in the middle of the weld nugget when using: rotation speed 2500 rpm, plunge depth 2,7 mm and dwell time 2 sec.  A higher heat input extends the TMAZ and HAZ outwards while decreasing the overall SZ hardness.  The zone around the BL is characterised by a lower hardness in case of a high heat input.  The lower hardness boundary between TMAZ and HAZ widens with increasing heat input.

59

References

[1] BMW 5 series part, Germany: Riftec gmbh.

[2] C. Schilling and J. dos Santos, “Method and device for linking at least two adjoining work pieces by friction welding”. European patent number Patent EP 1230062 B1, 1999.

[3] C. Schilling, A. Strombeck and J. Santos, “Friction spot welding: new joining process for spot connections,” GKSS Research Center, Geesthacht, 2001.

[4] “Kawasaki Heavy Industries,” [Online]. Available: http://global.kawasaki.com/en/corp/rd/magazine/163/ne163tr03.html. [Accessed 24 november 2015].

[5] T. Rosendo, B. Parra, M. Tier, A. da Silva, J. dos Santos, T. Strohaecker and N. Alcantara, “Mechanical and microstructural investigation of friction spot welded AA6181-T4 aluminum alloy,” Materials & design, vol. 32, no. 3, pp. 1094-1100, 2011.

[6] Y. Zhao, S. Lin, L. Wu and F. Qu, “The influence of pin geometry on bonding and mechanical properties in friction stir welded 2014 Al alloy,” Materials letters, vol. 59, no. 23, pp. 2948-2952, 2005.

[7] Y. Zhao, H. Liu, S. Chen, Z. Lin and J. Hou, “Effects of sleeve plunge depth on microstructures and mechanical properties of friction spot welded alclad 7B04-T74 aluminum alloy,” Materials & design, vol. 62, pp. 40-46, 2014.

[8] U. Suhuddin, L. Campanelli, M. Bissolatti, H. Wang, R. Verastegui and J. dos Santos, “A review on microstructural and mechanical properties of friction spot welds in Al-based similar and dissimilar joints,” in Proceedings of the 1st international joint symposium on joining and welding, Osaka, Japan, 2013.

[9] P. Su, A. Gerlich, T. North and G. Bendzsak, “Energy utilisation and generation during friction stir spot welding,” Science and technology of welding and joining, vol. 11, no. 2, pp. 163-169, 2006.

[10] M. Tier, T. Rosendo, J. dos Santos, N. Huber, J. Mazzaferro, C. Mazzaferro and T. Strohaecker, “The influence of refill FSSW parameters on the microstructure and shear strength of 5042 aluminum welds,” Journal of materials processing technology, vol. 213, no. 6, pp. 997-1005, 2013.

[11] Machine manual: RPS100 K11/K12, Harms & Wende GmbH & Co. KG.

[12] “MatWeb Material Property Data,” [Online]. Available: http://www.matweb.com. [Accessed 12 April 2016].

60

[13] “Certified test report, material certificate 7075,” Kaiser Aluminum, 2014.

[14] “ESAB,” 26 March 2014. [Online]. Available: http://www.esabna.com/us/en/education/blog/how-to-avoid-cracking-in-aluminum- alloys.cfm. [Accessed 8 February 2016].

[15] “Alcotec,” [Online]. Available: http://www.alcotec.com/us/en/education/knowledge/techknowledge/trouble-shooting-for- aluminum-welding.cfm. [Accessed 1 Februari 2016].

[16] G. Matters, The welding of aluminum and its alloys, Woodhead Publishing Limited, 2002.

[17] W. Van Haver, “Cursus IWE: Aluminium en aluminiumlegeringen. Deel 1: metallurgie, lasfouten en lasvoorbereiding,” Belgisch Instituut voor Lastechniek vzw, 2011.

[18] Z. Shen, Y. Chen, J. Hou, X. Yang and A. P. Gerlich, “Influence of processing parameters on microstructure and mechanical performance of refill friction stir spot welded 7075-T6 aluminum alloy,” Science and technology of welding and joining, vol. 20, no. 1, pp. 48-57, 2015.

[19] Z. Shen, X. Yang, Z. Zhang, L. Cui and T. Li, “Microstructure and failure mechanisms of refill friction stir spot welded 7075-T6 aluminum alloy joints,” Materials & design, vol. 44, pp. 476- 486, 2013.

[20] T. Santos, R. Miranda, P. Vilaça, J. Teixeira and J. dos Santos, “Microstructural mapping of friction stir welded AA7075-T6 and AlMgSc alloys using electrical conductivity,” Science and technology of welding and joining, vol. 16, no. 7, pp. 630-635, 2011.

[21] C. Olea, Influence of energy input in friction stir welding on structure evolution and mechanical behaviour of precipitation-hardening in aluminum alloys, Master dissertation: University of Bochum, 2008.

[22] Z. Shen, X. Yang, S. Yang, Z. Zhang and Y. Yin, “Microstructure and mechanical properties of friction spot welded 6061-T4 aluminum alloy,” Materials & design, vol. 54, pp. 766-778, 2014.

[23] S. Arul, S. Miller, G. Kruger, T. Pan, P. Mallick and A. Shih, “Experimental study of joint performance in spot friction welding of 6111-T4 aluminium alloy,” Science and technology of welding and joining, vol. 13, no. 7, pp. 629-637, 2008.

[24] Z. Zhang, J. Li, W. Li, D. Gao, W. Wang and C. Dong, “Improving mechanical properties of pinless friction stir spot welded joints by eliminating hook defect,” Materials & design, vol. 62, pp. 247- 254, 2014.

[25] Y. Uematsu, K. Tokaji, Y. Tozaki, T. Kurita and S. Murata, “Effect of refilling probe hole on tensile failure and fatigue behaviour of friction stir spot welded joints in Al-Mg-Si alloy,” International Journal of Fatigue, vol. 30, pp. 1956-1966, 2008.

61

[26] S. Gholami, E. Emadoddin, M. Tajally and E. Borhani, “Friction stir processing of 7075 Al alloy and subsequent aging treatment,” Transactions of nonferrous metals society of China, vol. 25, no. 9, pp. 2847-2855, 2015.

[27] L. N. Tufaro and H. G. Svoboda, “Effects of FSW procedure on the time-evolution of the properties in AA7075-T651 joints,” Soldagem & Inspecao, vol. 19, no. 4, pp. 333-342, 2014.

[28] C. Fuller, M. Mahoney, M. Calabrese and L. Micona, “Evolution of microstructure and mechanical properties in naturally aged 7050 and 7075 Al friction stir welds,” Materials science and engineering, vol. 527, no. 9, pp. 2233-2240, 2010.

[29] G. Ipekoglu, S. Erim and G. Cam, “Effects of temper condition and post weld heat treatment on the microstructure and mechanical properties of friction stir butt-welded aa7075 Al alloy plates,” International journal of advanced manufacturing technology, vol. 70, no. 1-4, pp. 201- 213, 2014.

[30] S. R. Schmidt and R. G. Launsby, Understanding industrial designed experiments, Colorado Springs, Colorado: Air Academy Press, 1994.

62

Appendix A Technical specifications FSpW machine

63

64

65

66

Appendix B Material certificate EN AW-7075-T6

67

68

Appendix C Welding conditions

Each of the following weld conditions is always used in addition with the constant parameters as specified in chapter 4.

EN AW-7475-T761

Identification RS [rpm] PD [mm] DT [s] Visual class Shear strength [kN] 2,54 3,09 2,64 IJ-74-R1-1 1500 1,6 3,5 2 IJ-74-R1-2 2500 1,6 3,5 1 1,96 2,96 3,17

2,25 1,95 3,38 IJ-74-R1-3 1500 1,6 2,5 2 IJ-74-R1-4 2500 1,6 2,5 1 3,08 2,99 2,56

3,81 3,58 3,15 IJ-74-R1-5 1500 1,6 4,5 1 IJ-74-R1-6 2500 1,6 4,5 2 2,71 3,80 3,89

3,94 3,72 3,93 IJ-74-R1-7 1500 2,1 3,5 1 IJ-74-R1-8 2500 2,1 3,5 1 3,07 3,75 3,81

3,50 3,32 3,78 IJ-74-R1-9 1500 2,1 2,5 1 IJ-74-R1-10 2500 2,1 2,5 2 2,63 4,01 3,07

6,02 4,98 5,14 IJ-74-R1-11 1500 2,1 4,5 1 IJ-74-R1-12 2500 2,1 4,5 2 4,60 5,73 5,21

EN AW-6082-T6

Identification RS [rpm] PD [mm] DT [s] Visual class Use IJ-82-R1-1 1500 2 0 1 Peel test IJ-82-R1-2 2000 2 0 1 Metallo IJ-82-R1-3 3000 2 0 2 Metallo IJ-82-R1-4 1000 2 0 3 Metallo IJ-82-R1-5 2000 3 0 2 Peel test IJ-82-R1-6 2000 3 0 1 Metallo IJ-82-R1-7 2000 3 0 1 Peel test IJ-82-R1-8 2000 2 2 1 Metallo IJ-82-R1-9 1500 3 0 1 Metallo IJ-82-R1-10 1500 3 0 1 Peel test IJ-82-R1-11 1500 2 0 1 Peel test IJ-82-R1-12 3000 2 2 1 Metallo IJ-82-R1-13 3000 2 2 1 Peel test IJ-82-R1-14 3000 3 2 1 Metallo IJ-82-R1-15 1500 3 2 1 Metallo IJ-82-R1-16 1500 2 2 1 Metallo IJ-82-R1-17 3000 3 0 1 Metallo

69

Identification RS [rpm] PD [mm] DT [s] Visual class Shear strength [kN] Failure mode IJ-82-R2-1.1 1500 2 0 1 5,68 1 & 2 IJ-82-R2-1.2 1500 2 0 1 6,13 1 & 2 IJ-82-R2-1.3 1500 2 0 1 6,86 1 IJ-82-R2-2.1 1500 2 1 1 7,87 1 IJ-82-R2-2.2 1500 2 1 1 7,88 1 IJ-82-R2-2.3 1500 2 1 1 8,08 1 IJ-82-R2-3.1 1500 2 2 1 8,63 1 IJ-82-R2-3.2 1500 2 2 1 8,63 1 IJ-82-R2-3.3 1500 2 2 1 8,63 1 IJ-82-R2-4.1 1500 2,5 0 1 8,52 1 IJ-82-R2-4.2 1500 2,5 0 1 8,54 1 IJ-82-R2-4.3 1500 2,5 0 1 8,31 2 IJ-82-R2-5.1 1500 2,5 1 1 8,77 1 IJ-82-R2-5.2 1500 2,5 1 1 9,33 2 IJ-82-R2-5.3 1500 2,5 1 1 9,62 1 IJ-82-R2-6.1 1500 2,5 2 1 10,10 2 IJ-82-R2-6.2 1500 2,5 2 1 10,44 2 IJ-82-R2-6.3 1500 2,5 2 1 10,29 2 IJ-82-R2-7.1 1500 3 0 1 7,97 2 IJ-82-R2-7.2 1500 3 0 1 8,38 2 IJ-82-R2-7.3 1500 3 0 1 8,67 2 IJ-82-R2-8.1 1500 3 1 1 7,80 2 IJ-82-R2-8.2 1500 3 1 1 7,07 2 IJ-82-R2-8.3 1500 3 1 1 7,26 2 IJ-82-R2-9.1 1500 3 2 1 7,21 2 IJ-82-R2-9.2 1500 3 2 1 7,46 2 IJ-82-R2-9.3 1500 3 2 1 7,80 2 IJ-82-R2-10.1 2250 2 0 1 8,72 1 IJ-82-R2-10.2 2250 2 0 1 8,47 1 IJ-82-R2-10.3 2250 2 0 1 8,64 1 IJ-82-R2-11.1 2250 2 1 1 9,52 1 IJ-82-R2-11.2 2250 2 1 1 9,53 1 IJ-82-R2-11.3 2250 2 1 1 9,45 1 IJ-82-R2-12.1 2250 2 2 1 9,35 2 IJ-82-R2-12.2 2250 2 2 1 9,65 2 IJ-82-R2-12.3 2250 2 2 1 9,46 1 IJ-82-R2-13.1 2250 2,5 0 1 9,64 3 IJ-82-R2-13.2 2250 2,5 0 1 9,62 3 IJ-82-R2-13.3 2250 2,5 0 1 9,68 3

70

IJ-82-R2-14.1 2250 2,5 1 1 10,51 3 IJ-82-R2-14.2 2250 2,5 1 1 10,73 2 IJ-82-R2-14.3 2250 2,5 1 1 10,33 3 IJ-82-R2-15.1 2250 2,5 2 1 9,54 2 IJ-82-R2-15.2 2250 2,5 2 1 9,39 2 IJ-82-R2-15.3 2250 2,5 2 1 8,99 2 IJ-82-R2-16.1 2250 3 0 1 7,06 2 IJ-82-R2-16.2 2250 3 0 1 7,06 2 IJ-82-R2-16.3 2250 3 0 1 7,33 2 IJ-82-R2-17.1 2250 3 1 1 6,55 2 IJ-82-R2-17.2 2250 3 1 1 6,63 2 IJ-82-R2-17.3 2250 3 1 1 6,65 2 IJ-82-R2-18.1 2250 3 2 1 6,48 2 IJ-82-R2-18.2 2250 3 2 1 6,21 2 IJ-82-R2-18.3 2250 3 2 1 6,29 2 IJ-82-R2-19.1 3000 2 0 1 8,34 2 IJ-82-R2-19.2 3000 2 0 1 8,37 2 IJ-82-R2-19.3 3000 2 0 1 8,49 1 IJ-82-R2-20.1 3000 2 1 1 9,68 1 IJ-82-R2-20.2 3000 2 1 1 9,36 2 & 3 IJ-82-R2-20.3 3000 2 1 1 9,68 2 IJ-82-R2-21.1 3000 2 2 1 10,36 2 IJ-82-R2-21.2 3000 2 2 1 10,23 2 IJ-82-R2-21.3 3000 2 2 1 10,10 2 IJ-82-R2-22.1 3000 2,5 0 1 9,52 2 IJ-82-R2-22.2 3000 2,5 0 1 9,72 2 IJ-82-R2-22.3 3000 2,5 0 1 9,49 2 IJ-82-R2-23.1 3000 2,5 1 1 9,98 2 IJ-82-R2-23.2 3000 2,5 1 1 9,19 2 IJ-82-R2-23.3 3000 2,5 1 1 9,99 2 IJ-82-R2-24.1 3000 2,5 2 1 8,05 2 IJ-82-R2-24.2 3000 2,5 2 1 8,42 2 IJ-82-R2-24.3 3000 2,5 2 1 8,08 2 IJ-82-R2-25.1 3000 3 0 1 6,01 2 IJ-82-R2-25.2 3000 3 0 1 7,28 2 IJ-82-R2-25.3 3000 3 0 1 7,37 2 IJ-82-R2-26.1 3000 3 1 1 5,89 2 IJ-82-R2-26.2 3000 3 1 1 6,50 2 IJ-82-R2-26.3 3000 3 1 1 6,02 2 IJ-82-R2-27.1 3000 3 2 1 5,71 2 IJ-82-R2-27.2 3000 3 2 1 5,97 2 IJ-82-R2-27.3 3000 3 2 1 5,87 2

71

EN AW-7075-T6

Identification RS [rpm] PD [mm] DT [s] Visual class Use IJ-75-R1-1 1500 1,8 2 1 Peel test IJ-75-R1-2 2500 1,8 2 2 Metallo IJ-75-R1-3 1500 1,8 1 1 Metallo IJ-75-R1-4 1500 1,8 0 2 Peel test IJ-75-R1-5 1500 2,7 0 3 Metallo IJ-75-R1-6 2500 2,7 0 2 Metallo IJ-75-R1-7 2500 2,7 2 2 Metallo IJ-75-R1-8 1500 2,7 2 3 Metallo IJ-75-R1-9 2500 1,8 0 2 Peel test IJ-75-R1-10 2500 1,8 0 2 Peel test IJ-75-R1-11 1500 2,7 0 3 Peel test

72