Computed Tomography and Scanning Electron Microscopy Analysis of a Friction Stir Welded Al-Cu Joint

The study aimed touse3D computed tomography (CT) to analyse a joint between two dissimilar materials produced by friction stir welding (FSW). As the materials joined, i


Introduction
Joining dissimilar materials is critical in numerous applications, especially when there is a need for lower energy consumption, higher efficiency, and lighter weight [1].Al-Cu joints are of particular importance in the electronics, transportation, and thermal power industries [2][3][4].In traditional welding, large amounts of energy are required to generate heat so that the materials to be joined can be melted and fused.Friction stir welding is a solid-state joining technology that needs a lower heat input [1].FSW is applied to butt weld metal sheets and plates.A special tool with a profiled pin, also called a probe, is required.It rotates while traversing forward along the contact line.This rotary motion generates friction heat, which is responsible first for heavy plastic deformation and then blending of the materials joined.The welding involves forging of the material at the rear of the tool [5].The friction stir welding process is illustrated in Figure 1.
The weld is formed as a result of mechanical mixing of the materials in contact.To prevent discontinuities, such as entrapped oxide defects, it is recommended that the workpiece surfaces should be cleaned prior to FSW [7].The resistance of the weld surface to oxidation can also be improved by properly designing the tool shoulder [8].In friction stir welding, the mixing of the plastically deformed metals is due not only to an increase in their temperature but also the formation of a stress field.The diffusion processes taking place in FSW lead to the friction blending of plastically deformed and displaced particles.The resulting weld has high strength properties [5], which are largely dependent on the microstructures of the base metals.Fig. 1.A schematic diagram of the FSW process [6].
One of the key benefits of FSW is the fact that it is a solidstate joining process.No solidification shrinkage or hot cracking is observed.There is also no undesirable segregation of alloying elements or intermetallic phases as is the case with classical welding.The joint is assumed to have a dense non-porous structure.Another advantage of FSW, which makes it superior to traditional welding processes, is low energy consumption [9][10][11][12].Although the economic and environmental factors are not of the utmost importance in the welding sector today, they may become relevant in the future.
Because of the occurrence of friction heat and plastic deformation induced by the rotation of the FSW tool, the following zones can be distinguished in a friction stir welded joint between two dissimilar metals: base metal 1 (BM1), base metal 2 (BM2), the heat affected zone (HAZ), the thermo-mechanically affected zone (TMAZ), and the stir zone (SZ), also known as the nugget zone (NZ) [2].There are differences in material composition, microstructure, and properties between the zones, and all this will affect the weld quality, strength and efficiency [13][14][15].
Extensive research has been done to understand the FSW process, especially to determine the relationship between different welding parameters and the microstructure and mechanical properties of FSW joints [1,[16][17][18][19][20][21][22][23][24][25][26].However, there has been hardly any qualitative analysis conducted along the whole length of an FSW joint.Generally, such joints are characterized on the basis of data obtained for several thin cross-sectional samples.Weld testing and analysis can be done using 3D computed tomography.This technique requires multiple measurements of X-ray attenuations, which are taken from different angles.The data are then processed to determine the quality and correctness (strength) of the weld.Radiography is suitable to assess not only the quality of welds but also castings and forgings [27].Unlike surface and near-surface inspection methods (e.g., visual, liquid penetration, magnetic particle inspection) [28], X-ray testing provides information about the whole object.This volumetric analysis method is often used to test unconventionally machined products [29][30][31][32][33].
Multiple cross-sectional images produced by computed tomography are combined into one 3D model so that the whole object of interest or a selected "virtual slice" can be visualized [34][35][36].Computed tomography uses radiation detectors to determine the object's capacity to absorb radiation.This requires measuring the intensity of X-rays that have passed through the object [37]; obviously, it is lower than the initial one.It can be expressed as a function of radiation energy, material type and object thickness.The relationship can thus be written as: (1) where: Ithe intensity of X-rays which have passed through a body, I0the initial intensity of X-rays, µthe linear radiation absorption coefficient, characteristic of the material and X-ray wavelength, xthe workpiece thickness.
X-ray radiation is created when accelerated electrons are suddenly stopped by matter.X rays are produced by two mechanisms: braking radiation and characteristic radiation [38].Braking radiation is a result of the interaction of free electrons in the electric fields of the atomic nuclei and the electrons at the surface of matter decelerating the moving electrons.Accelerated electrons have enough energy to overcome Coulomb repulsion of electrons from the external orbits and get closer to the positively charged atomic nuclei, where they are scattered.Electrons scattered in this way lose their energy non-elastically while changing the travel directions and speeds [12].The difference in the electron energy before and after deceleration is released in the form of photons of electromagnetic radiationbremsstrahlung radiation [38].A continuous spectrum is thus produced; all possible energy losses can be observed.
Characteristic X-rays are induced by a pass of an orbital electron from a higher energy level to a lower energy level, where an electron was earlier removed as a result of electron bombardment.Surplus electron energy is radiated in the form of a radiation quant characterized by a specific wavelength (K, L, M) [38].The length of the emitted wave depends on the type of matter of the braking surface.The plots of the characteristic spectrum and that of the continuous bremsstrahlung spectrum define the relationship between the X-ray intensity and the wavelength.The source of radiation in CT is an X-ray tube.The operating parameters, such as power and voltage, are selected according to the requirements.Unlike medical CT, industrial CT can be employed to inspect materials with higher densities.Industrial testing of large metal (e.g., steel or aluminum) elements produced, for instance, by casting, metal forming or welding, requires highenergy radiation.Metals are tested with high-voltage X-ray tubes because of their high absorption capacity [27].In industry, this method is being used on an increasing scale.CT is a common technique in defectoscopy (e.g., to check the internal structure of engineering materials), micromechanics (e.g., to inspect micromechanisms or precision components), geology (e.g., to determine the porosity of reservoir rocks, or to study fossils and minerals), archaeology (e.g., to visualize archaeological objects), and biology (e.g., to determine the structure of soft and hard biological tissues).
The aim of this study was to check if computed tomography would be suitable for analyzing the quality of a friction stir welded joint between two dissimilar materials with different properties (e.g., different densities).Volumetric analysis of such FSW joints, including the analysis of the formation of inter metallic compounds, is extremely interesting.The CT results were verified by scanning electron and optical microscopy, which provided information about the microstructure and chemical composition of the weld.

Materials and methods
The study focused on a friction stir welded joint between A1 aluminium and M1E copper.Both materials were in the form of 25 mm thick plates.As dissimilar materials, they have different chemical compositions and different properties.Aluminum, for example, is lighter and cheaper, while copper is a better heat and electricity conductor.The density of Al is 2.7g/cm3 whereas that of copper is 9.8g/cm3.Acorrding to Axon et al. [39] the lattice constant, for solid solution of Al is 0.4050nm.For a solid solution of Cu, the parameter is lower (a=0.363nm), as determined by Pearson [40].
The joint was produced on a DMG DMU50 milling machine.The surfaces to be in contact were ground along the whole length before welding.The plates were clamped firmly in a vice and the friction stir welding was performed using a Ø32 mm HSS tool with a tapered pin at a speed of 900 rev/min, a feed of 30 mm/min, and a tool tilt angle of 0°.Then, the workpiece was mounted in an upside-down position so that the process could be repeated for the root side.The same parameters were used.When the joint was completed, a 55x55 mm sample was cut out for analysis.
The CT scanning and analysis were carried out at the Radiography and Computed Tomography Laboratory of the Kielce University of Technology, Poland, applying the world's first computed tomography system that combines three radiation sources, i.e., two micro-and one minifocus X-ray sources (225 kV, 450 kV, and 450 kV, respectively).The 450kV/450W microfocus X-ray source was selected for the tests.The examinations were conducted using a 440 kV and 232 µA (102.8Wats of power) 450 kV X-ray tube with a 4 mm thick copper filter.A VAREX XRD 1611 detector was employed.The resolution of the detector is 4048 x 4048 pixels with a pixel size of 0.2 mm.
The scanning data were then processed and visualized using VG Studio 3.5.2.software.The images of the FSW Al-Cu joint were segmented using gray-scale thresholding [27].The results were compared to the Al-Cu phase diagram.
The joint was then prepared for metallographic micro structural analysis by sectioning, mounting in cold setting resin (VariDur 10), and polishing first with a STRUERS automatic polisher and then by applying a 0.05 μm Al2O3 (microdiamond) suspension and colloidal silica.The preparation of the specimens for the OM and SEM examinations did not include etching.A JEOL JSM 7100F (field emission) scanning electron microscope was employed to study the weld structure, whereas an Oxford Instruments X-MAX EDS spectrometer was applied to determine its chemical composition.

Results and discussion
The heat generated during FSW is a result of the friction between the rotating tool and the materials joined, adhesion of the tool to these materials, and their plastic deformation around the tool.Figure 3 shows the macro-photograph of the FSW Al-Cu connection.The SEM analysis revealed local thermal changes and the accompanying diffusion of elements formed joints as a result of mixing materials.It was found that the heat produced by the rotating tool can lead to the transformation of the two metals into intermetallic compounds.Since the lattice constants, also called lattice parameters, of aluminum are much higher, the diffusion of copper in aluminum is easier than the opposite.Ionizing radiation analysis helped determine the amount and size of inclusions on both sides.
The Al-Cu phase diagram (Figure 5) can be used to predict the phases formed provided that the electron-matter interaction volume is smaller than the volume of the phase observed.The multiphase microstructure and the changes in the material flow at the Al-Cu interface were determined from the magnified images of the metallographic specimens -Figure 6.The compositions of the characteristic regions were determined by SEM-EDS.The results obtained for different points (Table 1) were compared with those of the phases in the Al-Cu diagram.The SEM analysis of the Al-Cu junction area allows us to distinguish three different-looking areas, i.e., the first with an average content of about 60.4% Al and 39.0% Cu (points 1.1 and 1.2), the second on the aluminium side with an average content of about 54.6% Al and 45.0% Cu (points 2.1 and 2.2), and a third area with an average content of approximately 31.1% Al and 68.6% Cu (points 3.1 and 3.2).Silicon should be treated as an impurity.The workpieces of aluminium and copper probably contained trace amounts of silicon from the metallographic process (contamination with colloidal silica during polishing).
Figure 6 shows three different chemical composition observed at the interface between pure aluminum and pure copper.From the microphotographs, it is evident that no welding flaws (pores.voids.cracking.etc.) are present.The area adjacent to pure aluminum is 0,25÷0,8 µm in thickness; the average chemical composition at points 1.1 and 1.2 is 60.4% Al, 39.0% Cu and 0.6% Si, which suggests the presence of Al4Cu9 (δ).On the Cu side, there is an area with a thickness of about 0,5÷1 μm, where the average chemical composition is: 31.2%Al, 68.6% Cu, and 0.2% Si (points 3.1 and 3.2), which most probably is Al2Cu (θ).
From the Al-Cu phase diagram (Figure 5), it is apparent that locally the temperature may exceed 950 K.This temperature is close to the eutectoid temperature, at which Al4Cu9 (δ) forms.
The intensity of friction heat generated during FSW is dependent on the tool type and the material thickness.If necessary, the weld microstructure can be modified.It is possible, for example, to reduce the formation of certain intermetallic phases by selecting the right process parameters.
Computer-based reconstruction of the area of interest required using VG Studio software, which helped determine the size, amount and distance from the joint line of copper inclusions in aluminum and of aluminum inclusions in copper.Graphic visualization provided information about the material in the stir zone, including the distribution of elements mixed due to friction heat and plastic deformation.
The heat affected zone (HAZ) is also visible on the image (Figure 7).It has a thickness of approx.2.5 mm on the Cu-side and approx.4 mm on the Al-side, where the material is denser and has greater thermal conductivity.The type of materials used to create the joint using the FSW method and their properties -especially the melting point will be essential for the resulting joint.It may happen that the process parameters will be selected in such a way that the lower melting material will go into the liquid state, and the higher melting material will be solid.Figure 7 illustrates the plastically deformed plates joined together by FSW.The type of materials used (differences in the properties of copper and aluminium) affects the tendency to create material discontinuities.The main cause remains the high shrinkage during the solidification of aluminium.The amount of Cu particles on the Al side is lower than the amount of Al particles on the Cu side.On the Al side, there are only single inclusions of copper while on the Cu side, the area adjacent to Al has numerous agglomerates of Al.The occurrence of voids and pores in the weld (Figure 8) is directly dependent on the surface quality of the elements joined.If oxides are present at or near the workpiece surfaces, they should be removed because they may contaminate the weld and accordingly reduce its strength.

Conclusions
From this study, it appears that Friction Stir Welding (FSW) can be an effective method of joining dissimilar materials and that computed tomography (CT) is well-suited to assess the quality of such joints.

Fig. 3 .
Fig. 3. Macrophotography of the FSW Al-Cu joint.The investigations aimed to visualize the volumetric structure of the joint by means of computed tomography -Figure 4to assess the mixing of the metals.The results were then compared with those obtained by SEM and OM.

Fig. 4 .
Fig. 4.A CT model depicting the Al-Cu joint / 3D visualization of the Al-Cu joint produced by FSW

Fig. 6 .
Fig. 6.A SEM image of the Al-Cu joint

Fig. 7 .
Fig. 7.The heat affected zone analyzed using VG Studio software

Table 1 .
Results of the EDS analysis (in wt%) at the points marked in