Investigation of single point incremental forming parameters and forming limit curves prediction for heterostructured aluminum sheets

Journal Pre-proof Investigation of single point incremental forming parameters and forming limit curves prediction for heterostructured aluminum sheets Danielle Cristina Camilo Magalhães, Luciana Montanari, Sergio Alberto Elizalde Huitron, Jose Maria Cabrera Marrero, José Benaque Rubert, Sergio Henrique Evangelhista, Andrea Madeira Kliauga PII: S1044-5803(25)00282-7 DOI: https://doi.org/10.1016/j.matchar.2025.114993 Reference: MTL 114993 To appear in: Materials Characterization Received date: 3 December 2024 Revised date: 28 February 2025 Accepted date: 28 March 2025 Please cite this article as: D.C.C. Magalhães, L. Montanari, S.A.E. Huitron, et al., Investigation of single point incremental forming parameters and forming limit curves prediction for heterostructured aluminum sheets, Materials Characterization (2024), https://doi.org/10.1016/j.matchar.2025.114993 This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2025 Published by Elsevier Inc. Journal Pre-proof Journal Pre-proof Investigation of single point incremental forming parameters and forming limit curves prediction for heterostructured aluminum sheets Danielle Cristina Camilo Magalhães1*, Luciana Montanari2, Sergio Alberto Elizalde Huitron3, Jose Maria Cabrera Marrero3,4, José Benaque Rubert5, Sergio Henrique Evangelhista5, Andrea Madeira Kliauga1 1Department of Materials Engineering, Federal University of São Carlos, Rod. Washington Luís, km 235, 13565-905, São Carlos, SP – Brazil. 2Department of Mechanical Engineering, São Carlos School of Engineering - State University of São Paulo, Av. Trabalhador São-Carlense, 400, 13566-590, São Carlos, SP – Brazil. 3Departamento de Ciencia e Ingeniería de Materiales, EEBE - Universitat Politècnica de Catalunya, c/Eduard Maristany 10-14, 08019, Barcelona – Spain. 4Fundacio Centre CIM, c/ Llorens I Artigas, 12, 08028, Barcelona-Spain. 5Department of Mechanical Engineering, Federal University of São Carlos, Rod. Washington Luís, km 235, 13565-905, São Carlos, SP – Brazil. *Corresponding author: [email protected] ABSTRACT This study explores the formability and fracture behavior of Al-based heterostructured materials (HM) fabricated using Accumulative Roll-Bonding (ARB) and assessed through Single Point Incremental Forming (SPIF). The HM sheets, consisting of alternating AA1050 and AA7050 layers, were processed at preheating temperatures of 450 °C and 500 °C. Detailed characterization using SEM, EBSD, and TEM revealed variations in grain size, crystallographic texture, and precipitate distributions between processing temperatures, influencing the strength, hardness and ductility of HM sheets. Tensile tests showed that sheets processed at 500 °C exhibited higher elongation and improved ductility compared to those processed at 450 °C, attributed to changes in size and distribution of precipitates. Forming Limit Curves (FLC) and SPIF experiments demonstrated superior formability at 500 °C, with the HM sheets achieving higher critical wall angles and fracture strains compared to AA7050 layers alone. The results Journal Pre-proof Journal Pre-proof highlighted the influence of ARB to induce microstructural features, including residual stresses and shear strain localization, on the forming behavior of HM sheets. The study underscores the potential of optimizing processing conditions and material compositions to enhance the mechanical performance and manufacturability of HM aluminum sheets. Keywords: single point incremental forming; accumulative roll-bonding; heterostructured sheets; forming limit curves; texture; aluminum alloys. 1. INTRODUCTION Heterostructured materials (HM) represent a novel class of materials characterized by the presence of heterogeneous zones with significantly different mechanical or physical properties [1]. These materials were developed to address challenges associated with the very low ductility and toughness found in homogeneous nanostructured or ultrafine-grained materials, while still maintaining high strength levels [2,3]. One remarkable feature of HM is that their integrated properties often surpass predictions made by the rule-of-mixtures, which considers that the strength is given by the volumetric fraction of the heterogeneous zones. In this context, several types of HM have been developed, including multilayered structures comprising dissimilar materials, such as Al alloys [4-6]. For instance, Al heterostructured sheets have attracted significant attention in recent years due to their combination of strength and ductility, making them candidates for the automotive and aerospace industries [7, 8]. Recent studies have significantly enhanced the understanding of the mechanical response of HM, demonstrating their potential to overcome the traditional strength-ductility trade-off and improve the overall mechanical properties of multilayered sheets. However, these investigations have predominantly concentrated on tensile properties, such as yield strength, ultimate tensile strength, and elongation. This narrow focus has left a critical gap in the comprehension of HM formability, including their behavior under complex loading conditions, forming limits, and fracture mechanisms, which is a major technological aspect for manufacturing components. Journal Pre-proof Journal Pre-proof Although numerous studies have been performed on single-layer and bimetallic sheets, research on the formability of multilayered sheet metals remains scarce. The earliest investigations into Forming Limit Curves (FLC) for bimetallic sheets were pioneered by Semiatin and Piehler [9], examining the formability of stainless steel/aluminum sheets. Their work revealed that the initiation of necking and fracture is highly dependent on the specific arrangement of the layers within the sheet. Building on this, Mori and Kurimoto [10] further explored the formability behavior of Al1100/SUS340 bilayered sheets. They found that placing aluminum as the outer layer significantly enhances formability compared to alternative configurations. Similarly, Tseng et al. [11] studied the limiting strains of Al/Cu bimetallic sheets during deep drawing processes, examining the effect of thickness ratios between the layers. Their findings highlighted the detrimental impact of residual stresses at the interface of the layers on overall formability. Other works based on Al/Mg and Al/Cu were also performed in this topic [12, 13]. These studies emphasize the complexity of the bimetallic sheets behavior and the need to consider factors such as material compatibility, layer thickness, spatial distribution and residual stresses when predicting formability in such systems. Among various methods for assessing sheet formability, the Single Point Incremental Forming (SPIF) technique emerges as a promising approach for manufacturing complex geometries for several industries [14-16]. Particularly, the SPIF process has demonstrated a substantial enhancement in the attainable stretch and formability of sheets compared to conventional forming or stamping processes when formability/fracture curves are analyzed [17]. Consequently, it becomes feasible to predict FLC using data acquired through SPIF experimentation. In this context, SPIF can be applied to investigate the formability of several alloys, and particularly Al alloys. The results have demonstrated that the formability of Al alloys under different processing and testing conditions is higher using SPIF than the hemispherical dome test [18-22]. Regarding HM, only a few studies have used SPIF to analyze deformation and fracture behavior. Some studies predicted the necking and fracture strain during SPIF using bior tri-metallic sheets produced by explosive welding [23], or by Accumulative Roll-Bonding (ARB) [24, 26]. In general, deformation and fracture behaviors are strongly affected by the layers' Journal Pre-proof Journal Pre-proof arrangement. In addition, the forming angle may also be affected by the layer in contact with the tool. Kumar et al. [27] investigated SPIF for Al-Cu bimetallic sheets, emphasizing the role of process parameters in achieving high formability and surface quality without delamination. Similarly, Li et al. [28] explored the formability and fracture behavior of multilayered sheets Al/steel and Al/Ti sheets, highlighting the importance of layer thickness and toolpath strategies in optimizing forming limits and reducing fracture. These studies underscore SPIF's capability to form complex geometries in multilayered systems while maintaining structural integrity, though further research is needed to refine process parameters for diverse material combinations. Despite very little research about FLC and/or SPIF in HM, the development of formability curves enables the prediction of critical process limits and facilitates the design of efficient manufacturing processes tailored to specific applications of this new class of materials. The current literature highlights the importance of material properties and geometric factors, such as layer thickness, in determining the formability and performance of HM sheets. However, a significant gap remains in the literature, as the aforementioned studies primarily focus on layers with thicknesses in the millimeter range, leaving the behavior of multilayer systems with micrometer-scale layers largely unexplored. Understanding the formability, fracture mechanisms, and overall performance of such thin multilayer configurations is critical for industrial applications. Further research is needed to bridge this gap and provide insights into the unique challenges and opportunities presented by micrometer-scale multilayered HM sheets. In this study, we investigated the formability and fracture behavior of HM Al sheets composed of alternating AA1050 and AA7050 layers with average thickness below 25 μm, fabricated using ARB and assessed through SPIF. HM sheets were processed by ARB at preheating temperatures of 450 °C and 500 °C, and their macroand microstructural features, including layers’ thickness, mean grain size, crystallographic texture, and precipitate distributions, were characterized. In addition, SPIF experiments were performed to assess formability, measuring critical wall angles and fracture strains. Thus, FLC were predicted using SPIF data. These results were correlated with microstructure features and mechanical properties to provide deeper insights into the Journal Pre-proof Journal Pre-proof deformation mechanisms, formability and fracture behavior of HM sheets. The aim of this study lies in its potential to advance the adoption of multilayered aluminum sheets in diverse industrial sectors, offering lightweight, high-strength solutions for complex component fabrication. By elucidating the intricate interplay between material properties, process parameters, and forming behavior, this research contributes to the optimization of incremental forming processes. 2. EXPERIMENTAL PROCEDURE In this investigation, HM sheets were fabricated using ARB processing, which provides a good bonding strength between layers. Commercial sheets of AA1050 (0.001 wt.% Zn, 0.003 wt.%Mg, 0.013 wt.% Cu, 0.07 wt.% Si, 0.12 wt.% Fe, and Al balance) and AA7050 (6.45 wt.% Zn, 2.38 wt.% Mg, 2.48 wt.% Cu, 0.08 wt.% Si, 0.09 wt.% Fe, 0.09 wt.% others, and Al balance) were employed in this investigation. The plates were cut down into specimens with a width of 50.0 mm and a length of 80.0 mm, and then the initial sheets were produced for ARB.For this work, AA1050 and AA7050 sheets were cleaned and stacked alternately, and then processed up to six ARB cycles with preheating for 5 min at 450 °C and 500 °C, with a final pass with 50% thickness reduction, which resulted in 1 mm sheets. The characterization of the layered macrostructure of the HM AA1050/AA7050 was performed by means of scanning electron microscopy (SEM) in a FEI-Inspect S50 microscope. All analyses were taken on the RD-ND plane (rolling direction – normal direction). To measure the individual thickness variation in each layer, at least six straight lines were marked on the photomicrographs at the same magnification, covering approximately ten layers in the same frame. The thickness of individual layers and the mean standard deviation values were estimated for both studied conditions. In addition, electron backscatter diffraction (EBSD) analysis was performed in these samples to measure the average grain size and grain boundary character as well as to determine the mesotexture into the layers. For this purpose, a Tescan MIRA3 SEM microscope was used, with operating voltage of 25 kV and step size of 0.10 μm on the RD-ND plane. Prior to EBSD, the samples were mechanically polished using diamond suspension, followed by electrochemical polishing at -30 °C using Journal Pre-proof Journal Pre-proof a solution with 14% of distilled water, 6% of perchloric acid and 80% of ethanol. Transmission Electron Microscopy (TEM) samples were prepared from the multilayered composite sheets using a twinjet electro polisher with a solution of 60 mL of perchloric acid, 140 mL of distilled water and 800 mL of ethanol, with an operating voltage of 35 V. Observation was performed in an FEI TECNAI G²F20 TEM microscope operated at 200 kV. The crystallographic texture was also measured by X-ray diffraction at the LNNano Synchrotron National Laboratory in Campinas/Brazil on a PhilipsX’Pert MPD diffractometer using Co-Kα radiation with the Schulz method to obtain incomplete pole figures (PF), background and defocusing curves for further correction. The orientation distribution functions (ODFs) were calculated from the pole figures for the {111}, {200} and {220} measured at the sub-surface of the sheets after removing ¼ of the thickness by grinding and electrochemical polishing. To complete the initial characterization of the layers, Vickers microhardness (HV) and tensile tests were performed. For Vickers microhardness, a FutureTech hardness tester was used with a load of 100 gf applied on the polished samples’ surface (RD-ND plane) during 15 s. More than ten indentations were measured in each layer to establish the average hardness. Prior to formability tests, the sheets were tested in uniaxial tensile tests. Subsize tensile specimens were machined with dimensions of 3.0 × 2.0 × 7.0 mm in the gauge length, following the RD. They were tested in an Instron 5500 machine, at room temperature and under an initial strain rate of 1.0 × 10-3 s-1. At least three specimens were tested for each condition and the presented results are average values. After tensile tests, the fractured surfaces were analyzed by SEM (FEI-Inspect S50 microscope). To estimate the formability of the sheets, initially Nakazima tests were carried out using five different specimen geometries to produce different plane stress conditions. Three specimens of each geometry were used to estimate the formability curves. These tests used a hemispherical punch with a 50 mm diameter, installed in a Zwick BUP200 machine, with a punch speed of 1 mm/s until specimen fracture. To determine the minor and major strain values, an in situ digital image correlation (DIC) system was used during Nakazima tests, in which specimens’ surfaces were previously painted with a white coat with black Journal Pre-proof Journal Pre-proof speckle. Thus, the limit strains were estimated by the position-dependent method using the DIC system. For these tests, lubricant oil and Teflon layers were used to minimize friction. After tests, the fractured surfaces were analyzed by SEM (FEI-Inspect S50 microscope). The SPIF experiments utilized blanks measuring 55 mm × 55 mm × 1 mm, which were prepared by electrochemically etching grids of 3 mm diameter circles. This preparation facilitated the measurement of in-plane strains from deformed ellipses. The experimental setup was housed within the CNC machining center (ROMI D600). The initial forming diameter was 42 mm, and the tool had a tip diameter of 4 mm. The tool path strategy involved initiating with Ψi and incrementally increasing the drawing angles by ΔΨ = 5° until reaching either fracture or the tool tip diameter limit for the benchmark conical shape. Different initial Ψi angles were tested varying from 20° to 55°. Figure 1. Incremental sheet forming. (a) Schematic representation of the tooling set-up. (b) Drawing angle (Ψ) and depth (h). The forming tool had a hemispherical tip with 4 mm diameter and was made from VC131 tool steel, hardened and tempered to 60 HRC. The tests were performed with helical tool paths with a feed rate of 0.25 mm/rev (downward feed) equal to a linear speed of 125 mm/min. The rotation of the forming tool was 1750 rpm. 3. RESULTS 3.1 Initial characterization of the HM AA1050/AA7050 sheets Fig. 2 shows the SEM images for HM AA1050/AA7050 sheets produced at 450 °C and 500 °C, and their respective average thickness and Vickers microhardness. In Figs. 2(a) and 2(b), it can be observed that the layers remained approximately straight and continuous, with slight waviness. The observed waviness is due to differences in flow stress between the materials, leading to deformation heterogeneities and the formation of shear bands that cut through Journal Pre-proof Journal Pre-proof several layers. Another relevant point that can be observed in these micrographs is the excellent adhesion between the layers. In particular, the last layer generated in the final cycle of ARB is well bonded due to the additional pass with 50% thickness reduction. The strain in the diffusion process, known as deformation-induced interdiffusion, enhances element diffusion into the interface zone through mechanisms such as vacancy generation, pipe diffusion along dislocations, and stress-induced atomic displacements. In the ARB process, compression and shear deformation promote interdiffusion, contributing to interface formation and bonding strength, where mechanical bonding occurs rapidly during ARB, while diffusion bonding during annealing plays a critical role in forming AA1050/AA7050 interfaces [29]. Regarding the layer thickness (Fig. 2(c)), it is observed that there is a dispersion of values, with layers varying in thickness. For the case of sheets processed at 450 °C, the average thickness of AA1050 layers is about 23 ± 7 μm, and for AA7050 layers is 17 ± 6 μm. For samples produced with preheating at 500 °C, the average thickness of AA1050 and AA7050 layers was 24 ± 7 μm and 18 ± 8 μm, respectively. Figure 2. Macrostructures obtained for HM AA1050/AA7050 sheets produced by ARB at: (a) 450 °C e (b) 500 °C. (c) Thickness variation of the internal layers. (d) Average Vickers hardness measurements in individual layers. Regarding the hardness evolution of individual layers, Fig. 2(d) shows that the AA1050 Al layers exhibit similar hardness levels at both pre-heating temperatures, around 50 ± 5 HV0.1. Initially, the hardness level of AA1050 Al sheets was 20 ± 1 HV0.1. This suggests that the increase in hardness is due to either residual strain from the recovered substructure and/or solute strengthening from diffusion (originating from AA7050 Al layers) during processing. Conversely, the AA7050 layers displayed a hardness level lower than their initial condition (150 ± 7 HV0.1) at both processing temperatures. The primary factor contributing to the hardness of this alloy is the precipitates’ size and distribution. Sheets processed at 450 °C reached a lower hardness value, approximately 121 ± 4 HV0.1. When processed at 500 °C, the highest hardness value of about 137 ± 5 Journal Pre-proof Journal Pre-proof sheets, with their dominant shear components, exhibited higher resistance to deformation, while the AA1050 layers displayed greater formability, even with a more pronounced rolling texture. This interaction leads to anisotropic behavior, highlighting the need for further exploration of stress distribution and structural integrity after forming. Interestingly, HM sheets achieved higher depths for all drawing angles than AA7050 alloy alone. In addition, as FLC results indicated, the sheets processed at preheating temperature of 500 °C showed better formability than that at 450 °C. Fig. 11 presents photographs of the specimens after SPIF experiments with 𝜂i = 25o, showing that HM sheets with an intermediate behavior between AA1050 and AA7050 separately. The fracture initiated at the outer surface and then propagated towards the inner surface. Because of the initial geometry of the experiment, the fracture of the AA1050 sheet was only observed at initial wall angles higher than 35°. The fracture was ductile with the presence of necking and reduction of the wall thickness. The critical wall angle achieved values between 82° and 86°. The fracture in the AA7050 sheet was characterized by the absence of necking, and a much smaller wall angle of 32° to 36°. The fracture of the HM structures was characterized by delamination and the critical wall angles increased to 39° - 44° and 39° - 50° for the process temperatures of 450 °C and 500 °C, respectively. Table 3. Experimental results for tests produced by SPIF. Initial wall angle (Ψi) and final depth and final wall angle (Ψf). No fracture ◯; fracture ⚫. Material Ψi (º) Depth (mm) Ψf (º) Fracture AA1050 20 10.25 40.00 ◯ 25 13.50 51.50 ◯ 30 19.50 58.50 ◯ 35 27.25 67.00 ⚫ 40 25.00 85.00 ⚫ Journal Pre-proof Journal Pre-proof Figure 11. Representative specimens’ photographs after SPIF experiments initiated with 𝜂i = 25°: (a) AA1050 sheet - no rupture; (b) AA7050 sheet - rupture at 𝜂 = 36o; (c) HM sheet produced at 450 °C - rupture at 𝜂 = 44°; (d) HM sheet produced at 500 °C - rupture at 𝜂 = 46°. In SPIF the wall thickness diminishes with the increment of the depth and wall angle. According to Houssain et al. [33] the wall thickness t can be estimated by the geometrical conditions of the experiment, given by Eq (1): 𝑡=𝑡0𝑡𝑡𝑡 𝑡𝑡 (1) Where t0 is the initial wall thickness. For each material, there is a thinning limit, and the estimated strain in the wall direction is shown in Fig. 12. The critical strain for the AA1050 alloy was -0.65 and for the AA7050 -0.17. The HM materials with a volumetric fraction of 0.50 of each alloy yielded -0.23 and -0.30 for HM 450 °C and HM 500 °C, respectively. Figure 12. Rupture strains through the wall direction (εwall) as a function of the initial wall angle (ψi). 45 22.75 83.50 ⚫ 50 20.25 82.50 ⚫ 55 14.75 86.00 ⚫ AA7050 20 6.50 32.50 ⚫ 25 5.75 36.00 ⚫ 30 5.75 36.00 ⚫ HM sheet - 450 °C 20 10.00 39.50 ⚫ 25 11.25 44.00 ⚫ 30 6.50 42.50 ⚫ HM sheet - 500 °C 20 10.00 39.50 ⚫ 25 10.75 46.00 ⚫ 30 10.00 49.00 ⚫ Journal Pre-proof Journal Pre-proof The results from the SPIF experiments on HM AA1050/AA7050 sheets provided valuable data for predicting FLCs. These curves highlight the material's capacity to undergo plastic deformation under incremental forming. Comparisons were made with FLCs obtained from conventional Nakazima tests, as shown in Fig. 13 and with data published in the literature for the AA1050 [34] and AA7050 alloys [35]. Figure 13. Comparison of FLC data from SPIF experiments and FLC results for: (a) AA1050, (b) AA7050, (c) 450 °C and (d) 500 °C. For reference, see Fig. 9: dashed lines refer to FLC curves of multilayered materials processed at 450 °C and 500 °C. 4. DISCUSSION FLC curves predict the occurrence of necking in conventional sheet forming and are used as a practical limit for ductile fracture [36]. However, the strain to fracture presents different failure mechanisms. The FLC cannot be used as a limiting criterion for SPIF, and the Fracture Forming Limit line (FFL) is generally utilized in this case [37]. The Hill criterion for necking failure [38] predicts that the point of failure depends on the hardening exponent n of the Hollomon’s hardening law (Eq. 2) that correlates effective stress (𝜎𝑡𝑡𝑡), effective strain (𝜀𝑡𝑡𝑡) and a strength coefficient (K): 𝜎𝑡𝑡𝑡 =𝑡𝑡𝑡𝑡𝑡 𝑡 (2) Localized necking happens at a condition ε1 = n for plane strain tensile straining, as shown in Table 2, and diffuse necking occurs at ε1 = ε2 = n for equibiaxial tensile straining. Because the onset of necking and void nucleation occurs differently depending on the stress state, two sets of equations are normally applied to predict the FLC curve: one for the left side (ε2 < 0), and another for the right side (ε2 > 0). Paul [36] proposed the following Eqs. (3) and (4): For 𝛽=𝑡2 𝑡1<0: 𝜀1=𝑡𝑡𝑡0−𝑡2 (3) Journal Pre-proof Journal Pre-proof For 𝛽=𝑡2 𝑡1>0: 𝜀1=(1+𝑡𝑡𝑡0)(1+𝑡2)𝑡−1 (4) FLC0 and p are fitting parameters, obtained by linear regression using UTS, n, 𝑡1 from the tensile test, r, and sheet thickness t, given by Eq. (5) and (6): 𝑡𝑡𝑡0= 7.702𝑡𝑡𝑡(−0.0122 𝑡𝑡𝑡)−0.1124 𝑡−0.690𝑡𝑡𝑡(−12.4187𝑡1)+ +0.01149𝑡 +0.0823 𝑡+0.3011 (5) 𝑡=1.0834 𝑡𝑡𝑡(−1.4114 𝑡𝑡𝑡0)−0.361 (6) Fig. 14 compares the predicted FLC using Equations (3-6) and data from Tables 1 and 2. The experimental data is shifted to the right, which is a characteristic of pre-strained material. It was shown by the residual strain analysis in Fig. 3 that the AA1050 layers have higher density of geometrically necessary dislocations, and thus accumulate higher misorientation within the grains than in the AA7050 layers. Due to the difference in YS values of the two layers, the different strain levels lead to the buildup of shear components at the layer boundaries and to higher strain levels in the soft layer [31]. Kümmel et al. [39] have also measured different residual stress levels in the aluminum MH produced by ARB: compressive residual stresses form in the soft layers and tensile residual stresses in the hard layers. The residual stresses increase with more pronounced differences in YS of the constituent materials. At the processing temperatures, a higher hardening stress difference is achieved at 450 °C in comparison with 500 °C. Therefore, higher residual stress gradients are expected in the HM AA1050/AA7050 sheet processes at 450 °C. Figure 14. Comparison between predicted FLC curves using equations suggested by Paul [35] and the experimental data. The same results of the FLC may be presented in the form of effective strain versus stress triaxiality η, defined as the ratio between the mean principal Journal Pre-proof Journal Pre-proof stress and the von Mises effective stress (𝜂=𝑡𝑡 𝑡𝑡𝑡𝑡). This can be calculated from the main strain values by Eq. (7) [40]: 𝜂=√3 3 1+𝑡 √1+𝑡+𝑡2 (7) Where 𝜂 is the ratio of the minor to major in-plane strains (𝜂2/𝜂1). The effective strain according to the Hill criterion [38] is calculated by Eq. (8): 𝜀𝑡=1+𝑡𝑡 √1+2𝑡𝑡√𝑡1 2+𝑡2 2+2𝑡𝑡 (1+𝑡𝑡)𝑡1𝑡2 (8) This limit for the ductile fracture is shown on Fig. 15. The fracture mode in tensile plane stress experiments leads to ductile damage by void nucleation– void growth–void coalescence (in moderate to high triaxialities). The ARB bonding of the pair AA1050/AA7050 is constituted of a ductile, low strength material and a high strength with low ductility material. The fracture of the HM sheet in tension and in plane stress is characterized by ductile fracture in the AA1050 layer and fragile fracture in the AA7050 layer. The principal idea is that the ductile phase should improve the formability of the HM, but crack initiation in the AA7050 layers should also be avoided. It can be observed in Figs. 8, 10 and 13 that the area of ductile fracture increases at a higher annealing temperature. It is shown in Fig. 4 that at both temperatures the T-phase (Al3ZnMg) and Sphase (Al2CuMg) coalesce with the increment of ARB cycles and the volumetric fraction of precipitates is higher in the materials produced at 450 °C. The T-phase is coarser, and the S-phase is partially dissolved in the HM500 material. In addition, because the solvus temperature of the ηor η′-phase (MgZn2) lays at 475 °C, there will be dissolution and reprecipitation at 500 °C and coalescence of this phase at 450 °C. Thus, the improved formability of HM sheets processed at 500 °C can be attributed to the refined precipitate distribution and larger grain size when compared to the 450 °C process, as shown in Figures 2, 3, and 4. Finer precipitates act as effective barriers to dislocation motion, enhancing strength, while the larger grain size improves ductility. These microstructural changes Journal Pre-proof Journal Pre-proof result in a unique balance of strength and ductility, which is critical for the formability of HM sheets under SPIF. Figure 15. Comparison of the ductile fracture limit in the triaxiality vs. necking strain diagram for the materials of this study. The main advantages of the SPIF process are the simplicity of the process configuration and the extension of the strain to fracture observed in ductile materials. However, the stress-strain state is very complex, and the localization and fracture happen at higher equivalent strains due to the nature of the deformation. Eykens et al. [41] and Malhotra et al. [42] observed that the suppression of the necking mechanism is due to bending plus shear parallel to the surface caused by the tool translation. According to them, this results in a fracture due to an out-of-plane shear. After evaluation of measured and simulated strains, they conclude that a pure shear mechanism is unlikely because of the bending stiffness of the sheet. The local bending of the sheet around the tool causes a greater plastic strain at the outer side of the sheet while the inner part is subjected to shear and hydrostatic pressure (compression in the contact point and the meridional and circumferential directions). There is also a through thickness shear gradient component. Therefore, the observed strains at the outer surface are no longer the principal strain directions and it is not possible to determine a single stress state across the sheet thickness. For example, Mirnia and Shamsari [43] calculated a stress triaxiality of -0.6 at the inner surface and +0.6 at the outer surface in the region of contact with the SPIF tool and a strong variation of intensity as the tool progresses. Li et al. [44] concluded that there is a combination of shearing, bending and stretching. In cone-forming, the major deformation, according to the authors, is the stretching perpendicular to the forming direction. The major shear strain lies in the forming direction and is maximum at the middle of the sheet. They also found that the in-plane shear strain is not negligible. Journal Pre-proof Journal Pre-proof In addition, the contribution of bending to fracture in the SPIF process is characterized by a critical wall angle (𝜂). For the AA1050 alloy the reported critical angle for conical frustum and conical shapes lay between 6585° [45, 46], and the present results at approximately 85o are in accordance with the literature. Hussain et al. [33] have also noted that the point of wall angle in the failure depends on the geometry of the shape formed and may exceed the estimate obtained by constant wall angle parts by ∼4°. The estimated strain in the direction of the wall was -0.65. The AA7050 alloy on the contrary shows very low plasticity, and the fracture occurs without necking with an estimated strain in the wall direction of -0.17, and at low critical wall angles. The HM sheet tends to be the same behavior, but the strain limit was extended twofold at 500 °C. Martins [47] examines the fracture modus and loci under plane stress conditions. For the AA1050 alloy, the SPIF process is characterized by a mode I fracture mode with suppression of necking and the effective strain is limited by a constant though-thickness effective strain (as seen in Fig. 11) and by fracture toughness. The fracture forming limit (FFL) is a straight line falling from left to right in the principal strain space. For the AA1050 alloy this limit yield is given by Eq. (9) [48]: 𝜀𝑡1 =1.36 −0.7𝑡𝑡2 (9) The fracture is related to void growth, which is associated with the fracture of the intermetallic particles of diameter d embedded in the aluminum matrix and the stress triaxiality. In the case of plane stress and a normal plastic anisotropy 𝑡𝑡 (Eq. 10) [49]: 𝑡𝑡𝑡𝑡𝑡 𝑡=𝑡𝑡 (1 𝑡)=∫𝜎𝑡 𝑡𝑡𝑓𝑡 𝜀𝑒𝑓𝑓 0𝑡𝜀=1+𝑟𝑚 3(𝜀1𝑡+𝜀2𝑡) (10) Madeira et al. [48] have measured strain necessary to propagate a crack in tension (mode I) for samples of the AA1050 alloy and compared the results Journal Pre-proof Journal Pre-proof with the SPIF experiments with different part/toll diameters. When this ratio was larger than 40, no necking was observed and the SPIF limit coincided with the mode I fracture limit. But, for values of part/toll diameters smaller than 25 (10 in the present work), failure was preceded by necking. In this case, the onset of failure is delayed by dynamic bending under tension. The fracture in bending is then controlled by the ratio between the sheet thickness and the radius of the forming tool. For a part/tool diameter ratio of 10, they obtained a critical effective strain to fracture of 0.77. Fig. 16 shows the iso-strain locus for failure in plane stress (FLC), for mode I fracture (FFL), and the locus of effective strain 𝜂ef = 0.8 in comparison with the major and minor strain values obtained in the present results. In the present results, a fracture locus for a mixed fracture mode was obtained, but it is noticed that for smaller initial 𝜂 angle (35-40°) the strain to fracture was reduced, and the strain slope 𝜂 shifts to the compression-tension quadrant, because in these conditions the bending component is increased with respect to the stretching component. Figure 16. Fracture loci obtained in the SPIF experiments for the AA1050 alloy in this work compared with the FLC and FFL limits obtained by Madeira et al. [47]. Because of the lower plasticity, the SPIF experiments for the AA7050 alloy, and the HM materials were performed with a low initial wall angle. Under this condition the bending component is higher than the stretching component. The strain path and the fracture locus occurred in the compression - tension quadrant. The fracture of the HM sheets took place at the outer surface and with sights of delamination as shown in Fig. 11. Muhammad et al. [49, 50] analyzed the bendability of an age hardened AA6061 aluminum alloy and observed that the initiation and propagation of cracks were associated with the formation of shear bands and through-thickness grooving along the outer tensile edge at the tensile side of the sheet. Shear bands lead to dislocation accumulation at grain boundaries, and in the presence of secondary phase particles, decohesion of the boundaries took place. Shear bands formed at the surface may propagate inwards through multiple grains, spanning across the entire width of the specimen. Eventually, this intense Journal Pre-proof Journal Pre-proof shearing promotes the formation of surface cracks which propagate inwards in a transgranular manner through the sheared regions. This implies that at the outer surface plastic strain localizes in the directions of maximum shear stress and inplane shear acts as a mechanism of failure. Cladding with low alloyed aluminum led to an increase of bendability because the crack nucleation mechanism was shifted to the softer material. Thinner cladding layers lead to failure at the cladding interface. This fracture mode is similar to the ones observed in the AA7050 and HM materials. Silva et al. [51] discussed the locus of fracture in the 𝜂1-𝜂2 strain diagram in shear (mode II). The defect spacing 1/d is related to the shear strain 𝜂 by Eq. (11): 𝑡𝑡 (1 𝑡)=√1+𝑡2 (11) And the onset of damage (𝑡𝑡𝑡𝑡𝑡 𝑡𝑡 ) will take place at the maximum shear strain is given by Eq. (12): 𝑡𝑡𝑡𝑡𝑡 𝑡𝑡 =𝑡𝑡 (1 𝑡)=∫𝜏 𝜎𝑒𝑓𝑓 𝜀𝑒𝑓𝑓 0𝑡𝜀= 1(1+𝑡𝑡) 2(1+2𝑡𝑡)(𝑡1𝑡−𝜀2𝑡) (12) The critical strain values for damage in plane shear (SFFL) are in straight lines rising from left to right with slope +1 in the compression tension quadrant. Fig. 17 shows the onset of fracture obtained for the AA7050 and the HM sheets. The damage for the AA7050 alloy and the HM alloys is similar (the shear strains 𝜂 yield between 0.13 and 0.17), which indicates that the fracture limits are controlled by the AA7050 layers. Additionally, the delamination observed in SPIF experiments was more pronounced in samples processed at 450 °C, where interfacial bonding was weaker, as shown in Fig. 11. This phenomenon negatively impacts the overall performance of the material by reducing load transfer between layers and promoting crack propagation. To mitigate delamination, future work will focus on optimizing ARB processing conditions and exploring the use of intermediate layers to enhance interfacial bonding. For these materials a similar strategy as in for bending applications [50] using a thicker layer of the AA1050 alloy at the sheet surface either by cladding or ARB may improve the output in the SPIF process. Due to the higher shear components in this Journal Pre-proof Journal Pre-proof process, the adhesion between layers is a critical parameter, much more than in conventional forming, as shown by the FLC curves (see Fig. 16). Figure 17. Fracture loci obtained in the SPIF experiments for the AA7050, HM450 and HM500 materials. A deeper understanding of the delamination process necessitates further investigation. One factor that enhances adhesion during ARB at elevated temperatures (450 and 500 °C) is the reduction in the flow stress difference between the layers [31]. At room temperature, however, the disparity in the hardening behavior of the AA1050 and AA7050 layers becomes more pronounced, leading to higher interfacial shear strains that may contribute to failure. Future research will focus on employing advanced characterization techniques, numerical simulations, and strategies to mitigate delamination, thereby further optimizing the performance of HM sheets in the SPIF process. 5. CONCLUSIONS In this work, the potential of adopting multilayered aluminum sheets in forming processes was analyzed. A layered structure consisting of equal parts of AA1050 and AA7050 alloys, with a mean thickness below 25 μm, was produced by ARB at process temperatures of 450 °C and 500 °C. The study highlights the critical role of residual stresses, particularly those arising from the yield strength differences between the layers, in influencing formability and fracture behavior. The key findings of this work are summarized as follows: i) The AA1050 alloy exhibits ductile fracture, while the AA7050 alloy shows brittle fracture at room temperature, with the HM sheets displaying intermediate behavior and improved ductility and formability when produced by ARB with a preheating temperature of 500 °C. In HM, the AA1050 layers are characterized by a strong rolling texture (e.g., Brass and S components), exhibited higher ductility and formability, contributing to the overall stretchability of the HM sheets. In contrast, the AA7050 layers, with a higher fraction of shear texture Journal Pre-proof Journal Pre-proof Declaration of interests ☒ The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. ☐ The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Journal Pre-proof Journal Pre-proof HIGHLIGHTS  SPIF formability of heterostructured aluminum sheets was systematically explored.  ARB at 450 °C and 500 °C reveals key microstructural influences on formability.  Enhanced ductility and critical wall angles were achieved after ARB at 500 °C.  Strategies to facilitate the SPIF of heterogeneous structures are suggested. 52.