Strengthening of AA5754 aluminum alloy by DRECE process followed by annealing response investigation

materials Article Strengthening of AA5754 Aluminum Alloy by DRECE Process Followed by Annealing Response Investigation Przemysław Snopi´nski 1,* , Tomasz Ta´nski 1, Klaudiusz Gołombek 1, Stanislav Rusz 2, Ondˇrej Hilser 2, Tibor Doniˇc 3, Paweł M. Nuckowski 1and Marcin Benedyk 4 1Division of Material Processing Technology, Management and Computer Techniques in Materials Science, Institute of Engineering Materials and Biomaterials, Silesian University of Technology, 44-100 Gliwice, Poland; [email protected] (T.T.); [email protected] (K.G.); [email protected] (P.M.N.) 2Department of Mechanical Technology, Faculty of Mechanical Engineering, VŠB-Technical University of Ostrava, 17. listopadu 15, 708 33 Ostrava, Czech Republic; stanislav[email protected] (S.R.); [email protected] (O.H.) 3Research Centre of University of Zilina, University of Zilina, 010 26 Zilina, Slovak Republic; [email protected] 4Paks’D Sp z.o.o. Strzelecka 74, 43-100 Tychy, Poland; [email protected] *Correspondence: przemyslaw[email protected] Received: 3 December 2019; Accepted: 6 January 2020; Published: 10 January 2020   Abstract: In this study, a dual rolls equal channel extrusion (DRECE) process has been applied for improving the mechanical properties of the 5754 alloy. Supplementary experiments involving metallography, electron backscattered diffraction (EBSD), and XRD tests were carried out to evaluate the effect of the DRECE process. XRD analysis showed that the maximum dislocation density was achieved after six DRECE passes, which were accompanied by the formation that is typical for low-strain structures. The increasing dislocation density, as well as grain refinement throughout DRECE deformation, resulted in an increase in the mechanical properties. Annealing of the as-deformed sample resulted in grain growth and strength reduction. Keywords: DRECE; XRD; EBSD; structure 1. Introduction Severe plastic deformation (SPD) processes are regarded as appropriate for the manufacturing of bulk samples having a nano or ultrafine-grained microstructure via imposing significant plastic strain (which is difficult to obtain using the conventional deformation technique). Several different SPD processing methods are now available i.e., HPT (high-pressure torsion), TE (twist extrusion), or ARB (accumulative roll bonding) [ 1 – 3 ]. However, in recent years, the ECAP (equal channel angular pressing) method attracted the greatest interest. In this procedure, the workpiece, in the form of a rod or bar having a circular cross-section, is pressed through a die that consists of two channels that meet at a predetermined angle: usually 90 degrees [ 4 ]. A significant limitation of conventional ECAP is the size of the work samples and process discontinuity. The work sample must be repeatedly reinserted in the die channel for several consecutive passes to impose high strain. This type of procedure from a practical point of view has a disadvantage of high labor input requirement during the deformation process, and it may be used easily only for fundamental laboratory investigations. The ECAP process is discontinuous and slow, which makes it impractical in industrial applications. Due to that, critical attention has been focused on the possibility of alternative processing procedures development that is continuous Materials 2020,13, 301; doi:10.3390/ma13020301 www.mdpi.com/journal/materials Materials 2020,13, 301 2 of 14 and precludes the need for removing and reinserting the work samples after every deformation cycle. Several attempts have been made to enhance the effectiveness of the SPD processes. For instance, continuous cyclic bending [ 5 ], repetitive corrugation and straightening (RCS) [ 6 ], accumulative roll bonding [ 7 ], ECAP conform [ 8 ], or dual rolls equal channel extrusion (DRECE) [ 9 ] have been used successfully to produce fine-grained microstructures on a large scale. The DRECE method is a modification/combination of two commonly known SPD techniques: the DCAP method (dissimilar channel angular pressing) and ECAP-CONFORM (continuous equal-channel angular pressing), and it has been developed for forming metal sheets with maximum dimensions of 1000 × 60 × 2 mm. In this technique, the sample is inserted into the working space and then extruded through the roller without changing their dimension. The critical factor that influences the mechanical properties of the material after the DRECE process is a proper selection of the output geometry in the deformation zone–deformation angle (123 ◦ , 118 ◦ or 108 ◦ ) [ 10 ]. When comparing the DRECE method to the other commonly known SPD techniques, we must remember that the strain accumulation after a single pass is not as high as i.e., in the ECAP process. Similar to the ECAP process, the repetitive deformations increase the accumulated strain, which affects the microstructure and thus changes the mechanical properties. The main advantage of the DRECE method is the length of the work specimen and the form of the batch material sheet. Such form of the batch material is typical i.e., for automotive industry applications. The 5754 aluminum alloy used in this study is typically supplied to automotive manufacturers, where it is used for car body parts and door structural components. This is because of its unique properties, including moderate strength, high corrosion resistance, weldability, and easy forming characteristics [ 11 ]. These properties can be increased by severe plastic deformation processes by employing the grain refinement process [ 12 , 13 ]. However, conventional SPD processes provide only laboratory-scale samples, which cannot be applied industrially. Therefore, in this investigation, we focused on the DRECE method and its possible application to enhance the mechanical properties of the 5754 aluminum–magnesium alloy. 2. Materials and Methods The investigated material was an aluminum–magnesium alloy 5754 that had the chemical composition given in Table 1. Table 1. Chemical composition of the 5754 aluminum alloy. Al Mg Si Mn Fe Cr Cu Zn Ti Bal. 2.6–3.6 ≤0.4 ≤0.5 ≤0.4 ≤0.3 ≤0.1 ≤0.2 ≤0.15 The DRECE process with the principle is shown in Figure 1was conducted at VŠB-Technical University in Ostrava in the Department of Mechanical Engineering. In this experiment, we set the deformation αangle to be 108◦. Materials 2020,13, 301 3 of 14 Figure 1. Principle of the dual rolls equal channel extrusion (DRECE) process. The 5754 alloy samples covered with a Gleit grease were subjected up to six DRECE passes without changing the sample orientation between subsequent passes. The pressure of the feed roller was set to 150 bar. The sample was processed with a constant deformation speed of 40 mm/min. To investigate the thermal stability of the obtained microstructure, the sample was subjected to isothermal annealing at 150 ◦C, 180 ◦C, 200 ◦C, 250 ◦C, 300 ◦C, and 350 ◦C respectively for 30 min after six DRECE passes. The microstructures were examined using a light microscope Axio Observer Z1. To reveal the grain size and morphology, the samples were mechanically ground and polished and then electrolytically etched using Barker’s reagent. To study the microstructural evolution in a greater detail, the 5754 aluminum alloy samples were twin-jet electropolished in an electrolyte containing 20% nitric acid and 80% methanol at a temperature of − 30 ◦ C and 20 V for 15 s. Then, the grain structure was recorded by orientation imaging microscopy (OIM) using the electron backscattered diffraction (EBSD) technique integrated with a Zeiss Supra 35 Scanning Electron Microscope controlled and analyzed using OIM software (EDAX, Inc., Mahwah, NJ, USA). An area 60 × 60 µ m was scanned on the normal direction plane, and the step size was taken to be 0.2 µ m in all scans. A neighbor orientation correlation data cleaning process (level 4), followed by a grain confidence index (CI) standarization cleanup with a grain tolerance angle threshold set to 2 ◦ , was used. After the data cleaning, a coincidence index filter of 0.05 was used to remove the points that were not indexed correctly. X-ray diffraction analysis was carried out on the electropolished sections of the samples by employing a PANalytical X’Pert Pro diffraction system (Malvern Panalytical Ltd., Royston, UK) equipped with a CuK α radiation source, 30 kV, and 30 mA with a scan rate of 0.005 ◦ /s. The average crystallite size was calculated from the four Bragg reflection peaks of fcc-Al: (111), (200), (220), and (311) using the Scherrer and Wilson equation [14]. Dv=0.9λ BcosθB (1) where D v is the effective crystallite size, λ is the X-ray wavelength, θ is the Bragg angle, and Bis the line broadening. The microstrain was calculated using the Williamson Hall equation [15]. Bcosθ=0.9λ Dv+[4εsinθ](2) where ε is the root mean square of microstrain. According to the Rietveld method, the dislocation density can be estimated from the following equations: ρ=(ρd·ρs)1/2(3) Materials 2020,13, 301 4 of 14 ρd=3/D2 v(4) ρs=ε2/b2(5) where ρd and ρs are the dislocation density due to domains and the dislocation density due to the microstructure, respectively. Finally, the microhardness (HV) of the specimens were measured along the thickness by a microhardness tester Future-Tech FM-ARS under a load of 300 g for 15 s. Tensile tests were carried out using Zwick Z100 equipment (ZwickRoell GmbH & Co.KG, Ulm, Germany) with a strain rate of 6.7 ×10−4s−1. 3. Results and Discussion 3.1. Initial Microstructure Figure 2is an optical micrograph of 5754 alloy in an unprocessed state. This figure shows a typical wrought microstructure built of equiaxed grains. According to Cabibbo [ 16 ], the 5754 alloy matrix consist of a Mg in Al solid solution with fine dispersoids of Mg2Si and Al6Mn phases. Figure 2. Microstructure of the 5754 aluminum alloy in an initial state. 3.2. Effect of DRECE on Dislocation Density Figure 3shows the X-ray diffraction profiles of 5754 alloy samples in a different condition. As shown, the main reflections for the fcc Al (111), (200), (220), (311), and (400) were detected. In addition, two supplementary peaks at 2 θ =40.21 ◦ and 58.11 were identified to be characteristic of the Mg 2 Si phase. The diffraction peaks of the as-deformed samples are broadened, and their intensities change throughout the DRECE process. According to Ungar [ 17 ], this can be due to an increased amount of dislocations accumulated during severe plastic deformation and the small size of the diffracting grains Additionally, the peak shift expresses that the DRECE process causes a substantial amount of distortion in the lattice structure of Al5754 alloy. This can be related to long range stresses generated by severe plastic deformation. The summarized results of the structural parameters obtained from XRD analysis are presented in Table 2. Throughout DRECE processing, the average domain size parameter decreased from ~35 nm in an initial state to ~29 nm after six DRECE passes. In addition, after DRECE processing, the dislocation density increased from 3.46 ×1014 m−2in the unprocessed state to 6.47 ×1014 m−2after the first pass. Then, the dislocation density showed a steady decrease with an increasing number of passes, indicating the annihilation of dislocations despite increased strain accumulation [ 18 ], reaching a saturation level Materials 2020,13, 301 5 of 14 of 7.55 × 10 14 m −2 after six DRECE passes. This result is in accordance with the results of other works for similar Al–Mg alloys. Dinkerl [ 19 ] has documented that the dislocation density of AlMg2 alloy was calculated by XRD to be ~5 × 10 14 m −2 ~3 × 10 14 m −2 and 12 ECAP passes, respectively. According to Liu [ 20 ], the dislocation density of an AA5182 alloy after five turns of the HPT process increased to 12.8 ×1014 m−2. Figure 3. X-Ray diffraction profiles of 5754 aluminum alloy in different conditions. Table 2. Structural parameters of the 5754 aluminum alloy determined by X-ray diffraction peak analysis. D v =domain size; ρd =dislocation density due to lattice microstrain; ρs =dislocation density due to size effect; ρ=total dislocation density. Sample Dv(nm) ρ(m−2)ρd(m−2)ρ(m−2) Initial state 35 2.46·× 1015 4.86·× 1013 3.46 ×1014 1×32 2.98·× 1015 1.40·× 1014 6.47 ×1014 2×34 2.63·× 1015 1.35 ×1014 5.96 ×1014 3×31 3.13 ×1015 1.04·× 1014 5.73 ×1014 4×30 3.15 ×1015 1.67·× 1014 7.25 ×1014 5×32 2.84 ×1015 1.50·× 1014 6.54 ×1014 6×29 3.36 ×1015 1.69·× 1014 7.55 ×1014 3.3. Evolution of Microstructure Figure 4a–f shows an optical micrograph (longitudinal plane) of 5754 aluminum alloy subjected up to six DRECE passes. As compared with the initial state of the microstructure (Figure 2), the grain size is not reduced significantly. Areas of different crystallographic orientations inside individual grains are visible. For the micrographs shown, it appears that the increase in the number of DRECE passes is not accompanied by noticeable changes in the microstructure. Throughout subsequent passes, the grains remain equiaxed and almost constant in size. Materials 2020,13, 301 6 of 14 Figure 4. Examples of grain structures of 5754 aluminum alloy subjected to ( a ) one DRECE pass, ( b ) two DRECE passes, ( c ) three DRECE passes, ( d ) four DRECE passes, ( e ) five DRECE passes, ( f ) six DRECE passes. Figure 5shows the EBSD colored inverse pole figure (IPF) maps of the 5754 aluminum alloy samples in an initial state and subjected to one, four, or six DRECE passes, respectively. In these microstructures, the grain color/shade corresponds to the individual grain orientation denoted by the unit triangle legend in Figure 5b. The red lines indicate the locations of high-angle grain boundaries (θ>15◦) , while green lines indicate the locations of low-angle grain boundaries (3 ◦ > θ >15 ◦ ). The IPF image of the initial state sample (Figure 5a) demonstrates that the microstructure is composed of equiaxed grains. The measured average intercept length prior deformation is ~7.916 µ m. Figure 5b shows the microstructure after the first DRECE pass. According to the unit triangle legend used, several orientations can be observed even in the grain interiors. This indicates the variation in crystal orientation in the grain interiors at the early stages of deformation. The average intercept length remains almost unchanged: ~7.862 µ m. Figure 5c shows the effect of four DRECE passes. In this condition, the grain interiors are covered with many elongated bands of cells, having a low-angle grain boundary misorientation. These deformation bands are almost parallel to the TD direction. The measured average intercept length decreased to ~6.872 µ m. Figure 5d shows the microstructure of the sample subjected to six DRECE passes. In this state, the microstructure consists of grains covered with deformation bands having a low angle misorientation. This type of microstructure is typical for aluminum alloys subjected to low strains [21,22]. The measured average intercept length after six DRECE passes slightly increases to ~6.958 µm. Detailed extracted values of the fraction of the grain boundaries are presented in Figure 6and Table 3. In the initial state, the fraction of HAGBs is very high ~88.4% and slightly decreases after the first DRECE pass to ~86.6%. Along with the increase in the number of deformation passes, the fraction of HAGBs decreases. After four DRECE passes, approximately 58.6% of the grains have a grain boundary misorientation of more than 15 ◦ , indicating HAG (high-angled grains). Finally, the opposite trend is visible. After six DRECE passes, the fraction of HAGBs (high-angle grain boundaries) slightly increases to 59.6%. Materials 2020,13, 301 7 of 14 Figure 5. Electron backscattered diffraction (EBSD) inverse pole figure (IPF) maps of 5754 aluminum alloy (a) initial state, (b) one DRECE pass, (c) four DRECE passes, (d) six DRECE passes. Table 3. Summarized results of the EBSD analysis. Grain Size (Intercept Lengths) [µm] Fraction of Low Angle Grain Boundaries, % Fraction of High Angle Grain Boundaries, % Average Misorientation Angle, θAV [◦] Initial state 7.916 11.6 88.4 36.98 1×7.862 13.4 86.6 37.76 4×6.872 41.4 58.6 25.50 6×6.958 40.4 59.6 27.27 Materials 2020,13, 301 8 of 14 Figure 6. Misorientation distribution of the 5754 alloy samples ( a ) initial state, ( b ) one DRECE pass, (c) four DRECE passes, (d) six DRECE passes. 3.4. Mechanical Properties after DRECE Process Figure 7presents the indentation Vickers microhardness for an initial state and DRECE processed 5754 alloy samples taken across the thickness of the samples. Additionally, a summary of mechanical properties as a function of DRECE passes is given in Table 4. The observed high hardness gradient between the surface and center of the initial state sample is due to the inhomogeneity of deformation during rolling. The measured mean microhardness value is ~79 Hv. The hardness of the sheet increases throughout DRECE deformation. After the first DRECE pass, this value raises to about ~92 Hv. In addition, in this state, the hardness gradient between the surface and center increases—the hardness values lie in the range of 82–111 Hv. The observed hardness increase is accompanied by a dislocation density growth, as presented in Table 2. Along with increasing strain accumulation, the hardness values slightly increase, reaching a maximal value of ~99 Hv after six passes. It is worth mentioning that the hardness gradient between the surface and center of the samples decreases throughout DRECE deformation—the hardness values lie in the range of 92–108 Hv, indicating that the non-homogeneous hardness distribution (non-uniform deformation) in the as-rolled initial sample remains in the DRECE processed samples. This may be attributed to the nature of the rolling process, in which the imposed strain is higher near the surface, while a lower amount of deformation is applied at the center. Materials 2020,13, 301 9 of 14 Figure 7. Hardness distribution along the sheet thickness of the 5754 aluminum alloy samples. Table 4. Summary of mechanical properties of 5754 alloy. Condition Average Vickers Microhardness Yield Strength (MPa) Tensile Strength (MPa) Elongation, % Initial state 79.6 112.3 216 30.1 1×91.7 184.7 235.4 13.2 2×93.6 188.1 239.5 10.6 3×93.7 189.4 241.2 10.0 4×96.3 193.3 246.1 9.1 5×97.5 194.1 247.0 9.0 6×99.5 198.2 251.8 8.4 The mechanical properties of 5754 alloy after the DRECE process were also tested in a tensile test at room temperature, as shown in Table 4. It is clear that both yield and tensile strength increase throughout DRECE deformation. For the initial state, the mean yield strength value is 112.3, and this value increases rapidly to 184.7 MPa after the first pass. Then, it rises slightly to 198.2 MPa after six passes. This means that the work-hardening rate reduces with the strain increase. However, the observed strength improvement, which was lower than that with the ECAP method [ 23 , 24 ], was achieved at the expense of ductility, which decreased from 30.1% to about 13.2% after the first DRECE pass; then with an increase in strain accumulation, it decreased to 8.4% after the six DRECE passes. Such changes in strength and ductility are typical for severely deformed metals, as proved in numerous research papers [4,16,25,26]. 3.5. Annealing Response of the As-Deformed 5745 Alloy After the DRECE deformation, a thermodynamically unstable microstructure containing lattice defects and subgrain boundaries was produced. This distorted microstructure increased the mechanical properties. To allow further work operation such as shaping, stamping, or forming, the as-deformed alloy must be softened by annealing. The microstructural changes that occur upon annealing are commonly described with reference to recovery processes, the nucleation of new grains, and their