Extension of Gap Bridgeability and Prevention of Oxide Lines in the Welding Seam through Application of Tools with Multi-Welding Pins

Abstract

Friction stir welding has become important in many areas of production and is increasingly used for joining aluminum components. For long welding seams, conventional tools with only one welding pin reach their technical limitations due to low gap bridgeability. When welding aluminum, the stirred in surface layers, such as oxides, lead to a decrease in static and dynamic strength since linear accumulations are formed in the welding seam. The aim of the present study is to increase the gap bridgeability using tools with various welding pins and to prevent linear accumulation in the welding seam. The results show that a gap bridgeability of up to 2 mm for 4 mm material thickness is possible for the aluminum alloys EN AW 5083 H111 and EN AW 7020 T651. With the help of multi-pin tools, no impact of the gap width on the tensile strength was observed for joint gaps of up to 0.9 mm when using butt joint with a sheet thickness of 4 mm. Furthermore, the use of multi-pin tools showed significant influence on the prevention of linear accumulations in the welding seam. In addition, the oxide layers were finely distributed in welded joints using multi-pin tools.

1. Introduction

Friction stir welding (FSW) is a joining process developed at “The Welding Institute” in the 1990s, mainly used for joining aluminum alloys. During FSW, a rotationally symmetrical, rotating tool immerses in the joint gap until the tool shoulder is located on the material surface and exerts downward pressure. Due to the heat resulting from friction between the tool and the work material, the low melting alloy is plasticized, and the tool can move along the joint. Figure 1 shows a schematic illustration of the process principle [1,2]. Due to rotational and translational movements, the material flow is both in the same direction as the tool feed motion as well as a counter direction (rotational and translational vector is commutated) and a counter direction (rotational and translational vector is opposed) appear. The main process parameters that are adjusted on the machine are the tool rpm, feed rate, work angle, and axial force (in the case of force-controlled welding) [3,4].

Even though FSW offers many advantages over other fusion and solid-state welding methods, it has several process-typical characteristics that lead to deterioration of joint quality. Especially in case of long welding seams, which are used, for example, to produce crane masts and large ship components, there are some requirements in terms of the friction stir welding process and the clamping devices. To meet these requirements, joint gap, tool positioning, and the clamping technique must be carefully selected. This is problematic when producing long welding seams whereby accurate positioning of the components is only feasible with increased preprocessing and inspection efforts. Besides exact positioning and alignment of the edge to the welding seam, a very stiff fixation of the components is necessary for FSW. There is a risk of unsuitable fixation points dependending on the component design and the geometry. The force transmission points, which are not perpendicular to the clamping table, displace the component while clamping, leading to misalignment. However, the restricted gap bridgeability when using FSW with conventional single pin tools requires exact tool positioning and strict adherence to a zero or very small gap. When joining components with high gap tolerances, this could lead to irregularities in the welding seams in the form of worm holes, top layer defects, and lack of fusion due to a non-stationary material flow. Friction stir welding is mainly used for aluminum as work material. In comparison to fusion welding processes, FSW does not require any additional preparation of the edge of the seam immediately before starting the joining process. However, the top layers on aluminum, such as oxides or other coatings, are not removed and are stirred in the welding seam. This can lead to oxid lines in the welding seam (Figure 2) and to a decrease in static and especially dynamic strength. In literature, a decrease of tensile strength of up to ten percent has been reported [6]. Investigations show that oxide lines are a non-negligible problem in friction stir welding in addition to typical irregularities such as degree formation, angular misalignment, edge misalignment, or a void in EN ISO 25239-5.

2. Experimental Procedure

The tests within the scope of this work were performed on a portal friction stir welding machine manufactured by PTG Heavy Industries at the University of Kassel. This is a CNC 5-axis portal system with position, force, and torque control. The possible travels of the machine are 2300/1500/300 mm in X/Y/Z direction. With a welding force of 60 kN in the Z-direction, the machine can achieve welding depths of 0.5 mm to 30 mm in aluminium alloys. In the X and Y direction, the machine can apply forces up to 30 kN. Due to the high machine stability and the 5-axis system, various component geometries up to three-dimensional components can be realised. At spindle speeds of up to 4000 rpm, spindle torques of up to 160 Nm can be achieved at feed speeds of up to 6000/6000/1500 mm/min in the X/Y/Z direction [8].

In the context of this paper, sheets of the two wrought aluminium alloys EN AW 5083 H111 and EN AW 7020 T651 with dimensions of 500 × 150 × 4 mm³ were used for the experiments. The alloys in question were selected because they are frequently used in the automotive industry, rail vehicle construction, and aviation. Particularly with regard to rail vehicle construction, small component tolerances are difficult to maintain and often cause difficulties. These two alloys differ in their composition and the associated hardening mechanisms.

In the aluminium alloy EN AW 5083, the main alloying elements magnesium and manganese are added to pure aluminium according to DIN EN 485-2. Another designation for the material from which the main alloying elements can be derived is AlMg4.5Mn0.7. The exact chemical composition of the material according to DIN EN 573-3 is given in

Table 1. Chemical composition of EN AW 5083 [9].

Element	Si	Fe	Cu	Mn	Mg	Cr	Zn	Ti	Al
Percentage [%]	0.40	0.40	0.10	0.40–1.0	4.00–4.90	0.05–0.25	0.25	0.15	Remaining

.

AlMgMn alloys belong to the group of non-hardenable alloys with medium to high static strength characteristics and rank among the most important structural alloys. They are characterised by good fatigue properties, high cold formability, corrosion resistance, and very good weldability. The strength of the material depends mainly on the magnesium content and the strain hardening burr. The sub alloying element manganese influences the structural properties by a fine precipitation dispersion [10].

The minimum values of the mechanical properties of the material according to DIN EN 485-2 for a sheet thickness of 4 mm can be found in

Table 2. Mechanical properties of EN AW 5083 by DIN EN 485-2 [11].

Properties	EN AW 5083 H111
$\mathrm{Tensile}\mathrm{strength}{R}_m$ [MPa]	275
$\mathrm{Yield}\mathrm{strength}{R}_{p0.2}$ [MPa]	125
$\mathrm{Elongation}{A}_{50}$ [%]	15
Hardness [HBW]	75

.

In the aluminium alloy EN AW 7020 T651, the main alloying elements zinc and manganese are added to the aluminium according to DIN EN 485-2. The designation referring to the alloying elements is AlZn4.5Mg1. The exact chemical composition can be found in

Table 3. Chemical composition of EN AW 7020 [9].

Element	Si	Fe	Cu	Mn	Mg	Cr	Zn	Ti	Al
Percentage [%]	0.35	0.40	0.20	0.05–0.50	1.00–1.40	0.10–0.35	4.00–5.00	-	Remaining

.

AlZnMg alloys belong to the group of hardenable wrought alloys with medium to high static strength properties at good weldability and high welding seam strengths, since hardening of the heat affected zone takes place when stored at room temperature. Zinc as alloying element has a very good solubility in aluminium, which is, however, strongly limited by the addition of magnesium. Thus, the hardening potential of the alloy is significantly increased. The material condition T651 characterises the heat treatment carried out in the manufacturing process [10]. The minimum values of the mechanical properties of the material according to DIN EN 485-2 for a sheet thickness of 4 mm can be found in

Table 4. Mechanical properties of EN AW 7020 T651 by DIN EN 485-2 [11,12].

Properties	EN AW 7020 T651
$\mathrm{Tensile}\mathrm{strength}{R}_m$ [MPa]	350
$\mathrm{Yield}\mathrm{strength}{R}_{p0.2}$ [MPa]	280
$\mathrm{Elongation}{A}_{50}$ [%]	10
Hardness [HBW]	104

.

Further Information on the Methods

In this paper, tensile strength and flexural strength are performed on the Zwick/Roell Z100 universal tensile testing machine (ZwickRoell GmbH & Co. KG, Ulm, Germany) with a positioning accuracy of 2 mm and a load cell up to 30 kN. Measurements were performed at a test speed of 10 mm/min, with tensile force and strain recorded at 10 mm increments. The tensile tests were performed according to DIN EN ISO 4136.

The metallographic examination was performed using Leica Z16 APO A light macroscope (Leica Mikrosysteme Vertrieb GmbH, Wetzlar, Germany). The evaluation is performed with the Leica Application Suite 4.9.0 user software (Leica Application Suite 4.9.0, Leica Mikrosysteme Vertrieb GmbH, Wetzlar, Germany).

3. Results

3.1. Concept of a Multi-Pin Tool

Six different multi-pin tools, which differ in terms of varying pin circle and shoulder diameters were used as the basis for the subsequent tests. In addition, comparative tests were carried out with two single-pin tools to obtain corresponding reference welds or welding results. A concave shoulder shape was used for all tools. As can be seen in Figure 3, there were several welding pins on the shoulder surface, which were arranged eccentrically on a defined pitch circle.

The individual multi-pin tools differed in their dimensions or the various combinations of welding pin diameter, shoulder diameter, and pin circle diameter. The multi-pin tools developed were two-, three-, and four-pin tools, having welding pins with a diameter of 4 mm or 5 mm were on varying pitch circles. The shoulder diameters were 14 mm, 15 mm, and 20 mm.

One way to characterise friction stir welding tools is to show the relationship between the welding pin diameter and the shoulder diameter. It can be said that the diameter of the welding pin should be as small as possible to keep the size of the heat affected zone small, but the pin diameter on the other side must be sufficiently large so that it is stable enough for the process forces that occur. This could be confirmed by preliminary investigations in tool design for both aluminum alloys. Therefore, FSW tools usually have a pin diameter 1/3 of the shoulder diameter [13]. For example, the optimum ratio between pin diameter ($d_{st}$) and shoulder diameter ($d_s$) for 6 mm thick sheets of an aluminium alloy is between 1/2.5 and 1/3 [14]. As the ratio decreases, the welding process becomes more sensitive and the range in which the welding force is allowed to move for a high-quality welding seam decreases [14]. This in turn increases the demands on rigidity and the machine and machine control, as only small deviations of the welding force or axial force $F_A$ can be tolerated. It is also important to note that with increasing sheet thickness $h,$ the heat input into the component decreases while the shoulder diameter $d_s$ remains the same, which means that the welding pin must generate a higher heat input [14].

Since no simple pin diameter can be used to calculate the ratio, the pin diameter $d_{st,res}$ resulting in the process can be used here, for example. This results from the rotational movement of the pin circle diameter $s_{sk}$ plus the simple pin diameter $d_{st}$. The ratio due to the eccentrically arranged welding pins in multi-pin tools, between pin diameter and shoulder diameter in multi-pin tools is far above the usually used ratio of 1/3.

The multi-pin tools are constructed in several parts, which means that the basic tool and the welding pins are not made from one piece. The welding pins are inserted into holes drilled in the basic tool and are secured against slipping in the direction of the rotation axis with an axially seated threaded pin with a hexagon socket. An additional fixing against displacement and torsion is ensured by a threaded pin inserted radially to each welding pin. For this purpose, the cylindrical welding pins are provided with 25 mm long surfaces as support surfaces for the threaded pins, allowing the individual welding pin lengths to be infinitely adjusted and then fixed. The process relevant tool dimensions or the various combinations of shoulder, pin circle, and welding pin dimensions of the individual tools can be found in

Table 5. Process relevant tool dimensions and ratios.

Number of Pins	${\mathsf{d}}_{\mathsf{s}}$	${\mathsf{d}}_{\mathsf{s}\mathsf{k}}$	${\mathsf{d}}_{\mathsf{s}\mathsf{t}}$	${\mathsf{d}}_{\mathsf{s}\mathsf{t},\mathsf{r}\mathsf{e}\mathsf{s}}$	$\frac{{\mathsf{d}}_{\mathsf{s}\mathsf{t},\mathsf{r}\mathsf{e}\mathsf{s}}}{{\mathsf{d}}_{\mathsf{s}}}$
1	15	-	5	5	0.33
1	20	-	5	5	0.25
2	14	5	4	9	0.64
2	14	6	4	10	0.71
2	20	6	5	11	0.55
2	20	8	5	13	0.65
3	20	8	5	13	0.65
4	20	8	5	13	0.65

.

In order to obtain corresponding reference values in comparison with single-pin tools, comparative tests were performed with single-pin tools (15 mm or 20 mm shoulder diameter and 5 mm pin diameter) so that the influence of the multi-pin arrangement on the performance of the weld and the welding results can be evaluated in direct comparison with single-pin tools.

3.2. Influence of a Multi-Pin Tool on the Oxide Line Distribution

The representation of the oxide line distribution or, in general, the verification of the material flow of friction stir welded joints is highly complex, requiring the use of a marker material. The marker material tungsten carbide was evenly applied to the joint edges of the joining partners, which were butt-welded. An example of an aluminum plate with the applied tungsten carbide layer can be seen in Figure 4.

In the following, the designation for the tools is structured as follows: [number of welding pins]-[shoulder diameter $d_s$]-[pin circle diameter $d_{sk}$], which results in the designation 2-20-8 for a two-pin tool with $d_s$ = 20 mm and $d_{sk}$ = 8 mm. These are all tools with 20 mm shoulder diameter and 8 mm pin circle diameter in the case of the marker pen tools.

In order to be able to compare the effect of a weld with a low and a high energy input, two different parameter sets are used for the subsequent welds. A speed to feed ratio of 3 to 1 is used for welds with a high energy input, and a speed to feed ratio of 1 to 1 is used for the welds with a low energy input. The welds were made at a speed 300 rpm and a feed rate f 100 mm/min (hot weld) and 300 mm/min (cold weld). As expected, the marker material was entrained by the rotating tool and distributed in the welding seam. Following the welds, samples were eroded out of the centre and metallographically examined as shown in Figure 5. With the 1-pin (1-20-0) tool, a vertical line is created during the hot (left) as well as the cold (right) weld, which runs through the entire thickness of the aluminium. This line is also called “Lazy S” due to its shape and is considered an irregularity in aluminium welds. When using a multi-pin tool, the “Lazy S” is broken open and stirred. This effect can be attributed to the very high plastification by the examined multi-pin tools. Best results are achieved with the 2-pin (2-20-8) tool. In comparison to a cold weld, the marker material can be broken open and stirred further during a hot weld.

For better clarification, a CT image (Figure 6) of the cold welded sample of the WZ 2-20-8 was taken. The micro computer tomograph Zeiss XRADIA Versa 520 was used, and the image shows, matching the microscopic images, the finely distributed tungsten carbide within the joining zone.

Based on the results, it may be deducted that the oxide lines can be avoided using multi-pin tools. The cold welds show the direct difference between the use of a multi-pin and single-pin tool. Oxides, lacquers etc., that are stirred in the welding seam are finely distributed using a multi-pin tool and the mechanical-technological properties are not impaired by their presence.

Exchange of Material between the Welding Pins

In multi-pin tools, material is not only transported around the pins but also between the welding pins during plasticising. To support this thesis, random tests were carried out to prove the material flow between the welding pins. Figure 7 shows on the left side the mentioned two-pin tool. First a weld was performed in aluminium (EN AW 7020 T651) and then in copper (ETP).

The material adhesions on the tool show that the material has apparently been completely replaced. For a more detailed analysis, the tool was first used for welding in aluminium (EN AW 5083 H111) and prepared centrally with tungsten carbide according to Figure 8. The preparation serves to prove the material transport through the welding pins.

After a 200 mm long weld with prepared tool in EN AW 5083 H111, samples were taken at two points of the welding seam and examined metallographically. Figure 9 shows the microsections taken after 10 mm (left) and 50 mm (right) from the start of the weld.

At the tapping point at 10 mm after the start of welding, large accumulations of tungsten carbide can also be seen at tunnel defects, caused by insufficient aluminium exchange through the welding pins, which have already become considerably smaller at the tapping point after 50 mm from the start of welding. Accordingly, a direct exchange of material between the welding pins takes place along the welding seam. With increasing welding distance, the tungsten carbide is completely exchanged. Figure 10 shows a microsection taken after 200 mm of the weld.

There are apparently no more tungsten carbide particles, so it has been proven that a complete material exchange between the welding pins takes place in a multi-pin tool. Figure 11 can also be used to confirm the thesis presented at the beginning. The remaining material on the multi-pin tool can be seen.

The cross section shows no residues of the tungsten carbide, which was introduced centered at the beginning of the test. This shows that the material was completely replaced within the 200 mm weld.

3.3. Parameter Study for the Use of Multi-Pin Tools

Since there was no parameter study on the multi-pin tools yet, a welding parameter window for the tools had to be worked out first. The first welds were initially evaluated by means of a visual inspection in accordance with DIN EN ISO 17637. Subsequently, the welding process window was worked out by means of a detailed parameter study. For this purpose, investigations were carried out with a tool speed between 100 and 750 rpm and a welding feed rate of 100 to 750 mm/min. The parameter examinations showed that it is necessary to reduce the speed $n$ and the feed $f$ of the multi-pin tools compared to the single-pin tools in order to achieve satisfactory welding results. Furthermore, the tendency can be seen that both parameters have to be further reduced with increasing number of welding pins. The ratio of speed to feed remains identical, but both absolute values are further reduced. The parameters leading to flawless welds for the individual tools and alloys can be found in

Table 6. Speed n and feed rate f for individual tools as a result of parameter determination.

Tool Name	EN AW 5083 H111		EN AW 7020 T651
Tool	Rotation Speed [rpm]	$\mathbf{Feed}\mathit{f}$ [mm/min]	Rotation Speed [rpm]	$\mathbf{Feed}\mathit{f}$ [mm/min]
1-15-0	600	200	600	200
1-20-0	600	200	600	200
2-14-5	300	100	900	300
2-15-6	300	100	600	200
2-20-6	750	250	600	200
2-20-8	450	150	300	100
3-20-8	300	100	300	100
4-20-8	300	100	300	100

.

In the following, as in the previous chapter, the designation for the tools (Table 6) is structured as follows: [number of welding pins]-[shoulder diameter $d_s$]-[pin circle diameter $d_{sk}$], resulting in the designation 2-20-8 for a two-pin tool with $d_s$ = 20 mm and $d_{sk}$ = 8 mm.

3.4. Increased Gap Bridgeability through the Use of Multi-Pin Tools

In this chapter, the results of the welding tests on gap bridgeability are presented and evaluated. Here, first the tensile strengths of the welded joints are discussed, used in the paper as a fundamental criterion for assessing the mechanical weld strengths. Then the quality of the welded joints is assessed on the basis of the cross sections produced. Furthermore, after an initial evaluation of the most promising welds, an overview of the results of the hardness and bending tests is presented. Within the scope of the evaluation, important characteristic values of the welded joints produced are also compared with the values required by DIN EN 485-2 for the base material of the respective alloy in order to be able to make well-founded statements on the mechanical characteristic values achieved. For the seam evaluation on the basis of the cross sections, DIN EN ISO 25239-5 is used to check the admissibility of the geometric properties of the joints welded with the multi-pin tools.

Due to the different test results caused by the respective material properties of the aluminium alloys used, the welding tests of the two materials are first evaluated individually before a comparison is then made between the test results on both materials and the most important findings are highlighted.

3.4.1. Evaluation of the Tests with the Material EN AW 5083 H111

Tensile Tests

Primarily, the tensile strengths of the tests welded with different joint gap widths should give an overview of the welding seam qualities. For this purpose, the mean values of the tensile strengths of the three tensile test specimens are to be taken from a weld of Figure 12. The evaluation focuses on the results of the multi-pin tools. The results of the tensile strength $R_m$ of the base material, which is represented by the green line in the diagram, serve as reference values.

When looking at the tensile strength, a tendency can be seen that the tensile strength decreases with increasing joint gap. However, the multi-pin tools 2-20-8, 3-20-8 and 4-20-8, whose welds each have almost constant tensile strengths, are an exception. For welds with these tools, despite increasing joint gap widths up to 0.9 mm or 22.5% of the sheet thickness, no influence of the joint gap width on the tensile strength $R_m$ was observed. The tool 2-20-8, whose welds show a high and constant level of tensile strength $R_m$ of about 300 MPa, is particularly noteworthy. This value corresponds to a welding seam strength of about 92% of the base material evaluated as reference value and is not reached by any of the other tools. However, tools 3-20-8 and 4-20-8 with a tensile strength $R_m$ of approximately 280–290 MPa and about 86–89% of the base material strength also show good welding seam strengths at the set joint gap widths of up to 0.9 mm. The minimum values for the tensile strength $R_{m }$ of 275 MPa required by DIN EN 485-2 are achieved for all welded joints welded with tools 2-20-8, 3-20-8 and 4-20-8. This indicates a high-quality welding seam. The lowest tensile strength level, which is strongly dependent on the joint gap width, is shown by the two-pin tool 2-14-5.

By means of the error indicators shown in Figure 12, the minimum and maximum values of the achieved tensile strength $R_m$ of the tensile specimens 1–3 of a single welding seam with a certain underlying joint gap width in each case can be identified. The difference between the respective minimum and maximum tensile strength values indicates the consistency of the strength in the course of a welding seam. The fluctuation within a welding seam is relatively high, e.g., in the case of tool 2-20-6, especially for the joint gap widths of 0.6 mm and 0.9 mm. The maximum stresses of the tensile specimens 1–3 within the welding seam, which was welded with a joint gap of 0.9 mm, vary from 200 MPa up to 272 MPa. The fluctuating tensile strengths indicate that stable and constant welding seam properties over the entire length of the welding seam cannot be achieved with the tool. The multi-pin tools 2-20-8, 3-20-8, and 4-20-8, on the other hand, show relatively constant mean values of the tensile strengths over the different gap widths up to 0.9 mm. Furthermore, the fluctuations within the individual welding seams are also very small. This indicates that the weld has stable and uniform properties over the entire length of the seam and that solid welded joints can be produced.

When looking at the fractured surfaces, it can be seen that the tensile specimens break exclusively in the area of the welding seam and mostly in the area of the stirring zone in the middle of the seam. On closer inspection, it is noticeable that the tensile specimens from the welding tests carried out with tool 2-20-8, as shown in Figure 13, are more constricted than the other tensile specimens.

The tensile test, which is based on the welding with tool 2-20-8, suggests a more ductile behaviour in the area of the welding seam. This is confirmed by looking at the elongation at break $A_{80}$, Figure 14. The elongation $A$ indicates the achieved extension of the sample from the initial measuring length [15]. The additional index 80 at elongation at break $A_{80}$ identifies the specimen dimensions and indicates the initial measuring length, to which the achieved elongation at break is referred, in mm. The tensile test specimens of the welds of the two-pin tool 2-20-8 achieve the highest elongation at break values of about 11%. These are comparable with those of the evaluated base material, whose reference value is shown as the green line in Figure 14. Moreover, the welding seams welded with the multi-pin tools 3-20-8 and 4-20-8 still show good elongation at break of more than 8%. In comparison, tool 2-20-6 with a smaller pin circle diameter does not reach even half of these values.

It has been shown that the multi-pin tools 2-20-8, 3-20-8 and 4-20-8 show significantly higher and more constant elongations at break compared to the single-pin tools. However, the minimum elongation at break of 15% required by DIN EN 485-2 for the nominal thickness of the material of 4 mm was not achieved in any of the welds. The tensile test specimens of the base material also only show a lower elongation at break $A_{80}$ of 11.3%.

Cross Sections

Based on the cross sections produced for each welding seam, internal welding seam irregularities could be identified, and the dimensions of the respective seam and the welding seam zones could be measured. Figure 15 shows the cross section of a weld using tool 1-15-0. In the cross sections, the welding seam zones can be clearly recognised and are additionally labelled in the Figure.

The welded joint shown in Figure 15 was welded with a joint gap of 0.6 mm or 15% of the sheet thickness. The base material shows a dark discoloration on the horizontal plane. This discoloration occurs partly in rolled materials, such as the EN AW 5083 H111 shown. It should be explicitly noted that this is not a material defect. On closer inspection of the micro section, it is noticeable that even at this joint gap width, a cavity with a diameter of about 0.3 mm can be seen in the welded joint. This probably represents a tube pore, which, according to the DIN EN ISO 25239-5 standard, would only be classified as impermissible from a diameter of about 0.8 mm and, therefore, does not represent a welding seam defect. Parallel to this, the diagram in Figure 12 shows that the tensile strength $R_m$ decreases slightly from this joint gap width for the welds of tool 1-15-0 as the irregularity occurs. In comparison to this, no irregularities in the welding seam area were detected up to a joining gap of 0.9 mm, for example, when welding with the two-pin tool 2-20-8, with which the highest and most constant tensile strengths were achieved in the welding tests. The cross section of a welding seam with this tool and a joining gap of 0.9 mm is shown in Figure 16. Compared to welding with the single-pin tool, a layered structure of the stirring zone is clearly visible here. In principle, the individual layers can be described as concentric ovals around the centre of the weld. The layered structure is mainly visible in the lower welding seam area.

Upon evaluation of the welding seams on the basis of the cross sections, it can be ascertained that, in contrast to single pin tools, a welding seam without irregularities can be produced with the two-pin tools, even with a joint gap of 0.9 mm. Similar to the two-pin tool 2-20-8, the three-pin tool 3-20-8 can also produce flawless welding seams up to a joint gap width of 0.9 mm, however, the tensile strength of the joint is lower than that of the two-pin tool. Nevertheless, the fluctuations of the tensile strength values are smaller when using a three-pin tool. With the four-pin tool 4-20-8, small irregularities can occur in the lower right-hand area of the welding seam, comparable to the single-pin tool 1-15-0 (see Figure 15), but according to DIN EN ISO 25239-5 these are not to be classified as impermissible.

Hardness Measurement

In order to highlight a possible influence of the multi-pin arrangement on the hardness profile across the width of the welding seam, hardness measurements according to Vickers were carried out and the hardness HV is measured. For this purpose, test points were arranged on a line and placed at a distance of 2 mm over the width of the cross sections. The distance between the two hardness measurement lines is also 2 mm apart, enabling the measurements to be carried out in both the upper and lower welding seam areas. The results of the hardness measurements are shown in Figure 17. The centre of the welding seam is shown in the diagrams as a dashed line.

By means of the hardness profiles, it can be seen that the multi-pin arrangement has no significant influence on the hardness HV within the welding seam, since the hardness in the welding seam area is at a similar level for all the welds shown. Furthermore, no clear difference between the hardness values in the upper and lower welding seam area can be seen. However, it can be clearly seen from the profiles that the hardness in the right and left area of the cross section of the welds welded with the single-pin tool is lower than in the centre of the welding seam. In contrast, the hardness increases in these areas of the welds with the multi-pin tools, especially in the upper area of the welding seam. This tendency is also slightly present in the lower welding seam area with the multi-pin tools.

The different hardness profiles for the single-pin and multi-pin tools can be attributed to the different dimensions of the tools. In the case of the multi-pin tools, the resulting pin diameter is approximately 2.6 times that of the single-pin tool. This results in different dimensions of the different welding seam zones, which is also confirmed by the hardness measurements. Accordingly, the multi-pin tools form a considerably wider area of the stirring zone.

Bending Test

This section deals with the results of the comparative three-point bending test on welds welded with a joint gap of 0.9 mm. The welds with smaller joint gap widths are not considered due to the fact that the gap bridgeability is the main focus of the tests. The bending test is used to highlight the welding seam properties under an applied bending stress by means of a test stamp. For this purpose, the bending samples are placed on two support rollers. The maximum bending force $F_{max},$ which can be absorbed by the welding seam under bending stress and the distance travelled by the test stamp used at this force can be stated as characteristic values of this test, which are included in the following evaluation. As the force opposing the test stamp decreases again after the maximum force $F_{max}$ has been reached, it can be assumed that the bending sample begins to crack at the maximum force reached. The distance covered by the test stamp up to this force can, thus, be used as a characteristic value for the deformation capacity of the bending sample under bending stress.

Figure 18 shows that weldment with tool 2-20-8 has the highest maximum bending force $F_{max}$. The welding seams produced with this tool have also shown the best properties in the other tests so far in the evaluation of the test welds. A similar maximum bending force and the corresponding deformation is achieved by welding tool 1-15-0, the lowest values were achieved by tool 1-20-0.

Although the welds with the multi-pin tools 3-20-8 and 4-20-8 have lower maximum forces, these two bending samples show the highest deformations until the maximum force $F_{max}$ is reached. This suggests that the welding seam can be further deformed without the occurrence of a defect but can still absorb lower forces. The lowest values in relation to the maximum force $F_{max}$ and the deformation at this bending force are shown by welding with tool 1-20-0, for which the tensile strength $R_m$ and the elongation at break $A_{80}$ were already low.

3.4.2. Evaluation of the Tests with the Material EN AW 7020 T651

The quasi-static tensile tests were carried out several days after welding. The samples were milled for this purpose and not removed from storage.

Tensile Test

When looking at the tensile tests of the welding tests with the alloy EN AW 7020 T651, it can be seen in Figure 19 that the welds performed with the single-pin tool 1-15-0 have the highest tensile strength values. In addition, the variations between the different joint gap widths and within a welding seam are lowest in the test welds with this tool. This is the tool with the smallest shoulder diameter $d_s$ compared to the other tools.

All welds with a shoulder diameter $d_s$ of 20 mm show similar tensile strength values, which are in the range of the minimum value of 350 MPa for the base material according to DIN EN 485-2. By far the lowest tensile strength values are achieved by the welds of the two-pin tool 2-14-5, which is also the only tool where the tensile strength is directly dependent on the joint gap width. The tool will not be discussed in detail below with regard to the alloy EN AW 7020 T651, as the tool dimensions do not permit satisfactory welded joints.

When looking at the fractures from the tensile test, it can be seen that the tensile specimens of alloy EN AW 7020 T651 break mainly in the base material outside the welding seam. The fractures occur about 10 mm beside the welding seam. An exception to this is the welds welded with tool 1-20-0, in which almost all fractures are within the welding seam. Figure 20 shows an example of two fractures that were welded with tools 2-20-8 and 4-20-8, where the tensile specimen is broken outside the welding seam. On the basis of the seam surfaces, it can be seen that despite the same speed and feed parameters, the surface of the welding seam that was welded with the four-pin tool is much smoother and finer scaled than the surface of the weld that was welded with the three-pin tool. Thus, a correlation between the force and torsional moment courses and the formation of the seam surface can be established. The comparison shows that the welding seams, in which stronger fluctuations in the course of the force and torsional moment courses occur during the welding process, also have a stronger or more irregularly scaled seam surface.

When looking at the diagram of the achieved elongation at break $A_{80}$ in Figure 21, no significant difference can be found between the welds with the different tools. However, the welds with tool 1-15-0 show the smallest variations in elongation at break within a welding seam and between the welding seams with the different joint gap widths. The minimum values of elongation at break A₈₀ for the base material according to DIN EN 485-2 are not reached with any of the FSW tools.

Cross Sections

The individual welding seam zones can also be easily identified in the cross sections of the weld samples made of the alloy EN AW 7020 T 651. In line with the knowledge gained from the tensile tests, in which the tensile specimens of the different gap widths have similar strengths, no impermissible defects according to DIN EN ISO 24239-5 can be detected in the welded joints up to a joint gap width of 0.9 mm when welding with the single-pin tool 1-15-0. An exemplary cross section can be seen in Figure 22. When looking at other cross sections, however, small cavities are sometimes visible, comparable to the welds that were welded with tool 1-15-0 for material EN AW 5083. However, these are permissible according to DIN 485-2.

In comparison to the results of the welds on the material EN AW 5083, flawless welding seams are also possible with the tools 2-20-8, 3-20-8, and 4-20-8 according to DIN EN ISO 25239-5. On the basis of the cross sections produced, the scaling on the seam surface in the form of elevations and depressions can also be clearly seen. The smaller the elevations, the finer the seam scaling on the surface. When looking at the cross sections of the welds welded with the multi-pin tools and shown in Figure 23, it is noticeable that the seam surfaces differ in the flatness of the surface and the scales formed. The thickness of the scaling decreases significantly from the welds with the two-pin tool to the four-pin tool. This results in a very smooth welding seam surface with fine shingling in the four-pin tool.

Another difference between welding with the two-, three-, and four-pin tools can be seen in relation to the formed structure in the area of the stirring zone. When welding with the 2-20-8 tool, there is a narrow homogeneous area with a fine-grained structure in the upper area. Below this, the layered onion skin-like formation of the structure in a concentric oval shape is clearly visible, which extends over a large part of the welding seam surface. When welding tool 3-20-8, the structure of the welding seam is similar, but it can be seen that the layered areas increasingly run into each other. It is also noticeable that the individual layers are more finely distributed compared to the two-pin tool. Compared to the two-pin and three-pin tool, the weld with the 4-20-8 tool shows a larger homogeneous area with a fine-grained structure in the upper part of the stirring zone. A small area with a finer layered structure can only be seen in the lower area of the stirring zone. The cross sections, thus, suggest a finely distributed and better mixed structure when welding with the four-pin tool.

Hardness Measurement

The results of the hardness measurements for the welds on the alloy EN AW 7020 T651 are shown in Figure 24.

From the hardness profiles it can be seen that the hardness HV increases slightly in the area of the welding seam centre. The strongest increase can be seen when welding with tool 2-20-8. The hardness in the area of the nugget is in the range between 115 HV and 120 HV for all tools. The hardness courses are similar when welding with different tools and no hardness course stands out significantly.

Bending Test

Based on the characteristic values from the comparative three-point bending test (Figure 25), it can be seen that the welds of tools 1-15-0 and 3-20-8 achieve the highest maximum forces at bending stress. The welds of tools 1-20-0 and 4-20-8 reach slightly lower maximum forces, with the greatest deformation, which must be produced until the maximum force $F_{max }$ is reached, occurring in the bending test welded with tool 1-20-0. The lowest maximum force $F_{max}$ and the lowest deformation up to this force is shown by welding with tool 2-20-8.

3.5. Increase of the Joint Gap

Due to the fact that both the single-pin tools and the multi-pin tools can be used to produce flawless welds with good mechanical properties with a sheet thickness of 4 mm up to a joint gap width of 0.9 mm according to DIN EN ISO 25239-5, no clear statement can be made on the influence of the multi-pin tools on the gap bridgeability on the basis of these tests. Therefore, tests with larger joint gap widths of up to 3 mm were carried out on material EN AW 5083. The achieved tensile strengths from these additional tests are shown in Figure 26 and are compared with the values of the previous welds with a joint gap of 0.9 mm.

In order to get an impression of the tools shown in the illustration Figure 26 with regard to the gap bridgeability at higher gap widths, the tests were carried out with 2 mm and 3 mm wide joining gaps. This corresponds to a joint gap width of 50% and 75% of the sheet thickness. While the tensile strengths within a welding seam are still relatively constant up to a joint gap width of 0.9 mm, especially in the case of multi-pin tools, stronger fluctuations can be observed for the 2 mm and 3 mm wide joint gap. These are clearly shown in the form of the error indicators in the diagram. Nevertheless, in the cross sections to the welds with a 2 mm joint gap with the tools 2-20-8 and 4-20-8 no irregularities in the welding seam were detected. The welding of tool 2-20-8 also achieves the tensile strength $R_m$ for the base material of 275 MPa required by DIN EN 485-2. In contrast, tools 1-20-0 and 3-20-8 produce impermissible cavities in the form of tube pores in the 2 mm wide joint gap. Although the weld of tool 2-20-8 still has relatively good tensile strength at a joint gap width of 3 mm, a closed welding seam cannot be produced here. In principle, however, a decrease of the tensile strength $R_m$ can be seen with increasing joint gap widths over 0.9 mm.

The achieved elongations at break of welds with a larger joint gap are shown in Figure 27. It can be clearly seen that the welds with no impermissible defects up to a joint gap width of 2 mm also achieve the best elongation values. These are the welds with tools 2-20-8 and 4-20-8.

On the basis of these tests, it is shown that with the multi-pin tools 2-20-8 and 4-20-8, in contrast to the single-pin tools, flawless welding seams with good welding seam properties are possible with joint gap widths of up to 50% of the sheet thickness. For larger joint gap widths of 3 mm, for example, the gap is so large that the material that is included in the process by the larger resulting welding pin diameter in the multi-pin tools is not sufficient to produce a flawless welding seam.

4. Discussion

4.1. Microstructure

The concentrically running flow lines in the area of the stirring zone show zones of different plastic deformations that occur during the feed motion of the rotating welding tool [16]. This applies to the two aluminum alloys used. In the welds with the four-pin tool 4-20-8, a smaller area of this onion skin-like structure can be detected, while the area of the homogeneous structure increases in the upper part of the welding seam. The microstructure, thus, suggests that the variations in the burr of plastic deformation decrease in the four-pin tool compared to the two- and three-pin tools. As the number of welding pins is the only change in comparison to the other tools, it can be assumed that the welding pin arrangement is responsible for this. One explanation could be that there are more welding pins for material transport per tool revolution and, thus, the material to be transported behind the tool is better distributed.

4.2. Welding Seam Properties

The fact that the fractures in the tensile tests of the welds of the alloy EN AW 7020 T651 mainly occur outside the welding seam in the base material and that the tensile strength $R_m$ is lower than that of the unmachined base material suggests that softening of the material occurs in the area of the heat-affected zone. The alloy is a hardenable aluminium alloy, which achieves its base material strength through heat treatment. Thus, the strength properties can be influenced by heat input in the friction stir welding process. In the case of alloys that are subjected to the T6 heat treatment with solution annealing and artificial ageing, the friction stir welding process can lead to an overaging of the microstructure in the area of the heat-affected zone and, among other things, the hardness in this area can decrease compared to the unaffected base material. [14]. It can be confirmed that the hardness HV decreases slightly in the heat-affected areas to the right and left of the welding seam. However, due to the wide welding seam formation caused by the tool, the cross sections are not sufficiently wide to produce a complete hardness profile up to the base material adjacent to the heat-affected zone. Due to overaging, the structure coarsens and the strength in this area decreases [10]. As a consequence of the reduction in strength, the heat-affected zone represents the weakest area of the welded joint due to the low strength properties and hardness values, which is why the fractures in the tensile specimens occur at this point. This also explains the lower tensile strength $R_m$ of the welded joints compared to the base material.

The fact that all multi-pin tools with a shoulder diameter of 20 mm have comparable tensile strengths for the material EN AW 5083 H111 can be explained by the fact that the tensile specimens are mainly broken in the base material. The number and spacing of the welding pins, thus, seems to have a minor influence on the welding seam strength. On the basis of the tensile strength $R_m$, a greater influence of the tool shoulder on the welding seam strength can be seen, as the welds with the single-pin tool 1-15-0 show slightly higher tensile strengths compared to the tools with a shoulder diameter of 20 mm. It should be noted that the joint gap in the welds of alloy EN AW 5083 H111 was only 0.3–0.9. The tests with high gap bridgeability were carried out only with EN AW 7020 T651. Despite different feed and speed parameters, the temperatures in the welding seam or in the welding pin are comparable during the process with the single-pin tool 1-15-0 and the multi-pin tools. Possibly the higher strength values of the single-pin tool result from the higher feed rate and the resulting shorter dwell time of the material at high temperatures. Thus, it is possible that this would have an influence on the strength in the heat-affected zone and, as a consequence, the strength drop in the heat-affected zone would be lower.

5. Conclusions

In this paper, it was shown that by using multi-pin tools gaps of up to 75% of the thickness of the joining part can be bridged and linear accumulations in the welding seam can be avoided.

Within the scope of fundamental feasibility studies, it was proven that largely closed, flawless welding seam surfaces could be realised with all the six multi-pin tools that were tested. The investigations also showed that the ratio between pin circle diameter and shoulder diameter is the decisive factor. The number of pins causes a change with regard to the general production of a flawless welding seam but also the straightness of the mixing of the welding material.

Hot welding with a speed to feed ratio of 3:1 can increase the ability to bridge small joint gaps in single-pin tools. A disadvantage is the tendency to form linear accumulations in the welding seam during these “hot” welds. By applying a marker layer of tungsten carbide to the joint joints, the formation of linear accumulations or the plastification of the respective tools could be simulated. The investigations have shown that the plastification of multi-pin tools is significantly higher than that of single-pin tools. Therefore, the use of multi-pin tools can reduce or avoid linear accumulations. The mechanical-technological properties of sheet metal with oxide layer and without oxide layer were examined in keywords. An influence of the oxide layer could not be determined.

The investigation of the gap bridgeability was a central point in this research project. For this purpose, butt joint weldings were performed under the gradations of 0.3 mm/2.0 mm and 3.0 mm, at a sheet thickness of 4 mm. With the multi-pin tools, good to very good results were achieved throughout. This can be attributed to the high plasticization with the addition of sufficient material in the welding zone. The two-pin tool with 8 mm pin circle diameter proved to be outstanding. Despite an increasing joint gap, no decrease in tensile strength was observed. The elongation at break is also far above that of the other tools.