Keywords

Introduction

In the pursuit of developing a substitute for Duralumin [1], ALCOA invented aluminum alloy AA2024-T3 in 1931 with improved yield strength and fatigue performance compared to Duralumin and their first generation aluminum alloy such as AA2017-T4 [2]. Because of its high specific strength, AA2024 became one of the most useful aerospace alloys applied to aircraft structures such as fuselage skin, wing tension member, rib, shear web, etc. [3]. AA2024 is now well developed and optimized in terms of heat treatment (temper heat treatment), chemistry, production method of extruded profiles and rolled sheets/plates, etc., to the extent that the mechanical properties have reached plateau as documented in standard handbooks [3, 4]. In addition, AA2024 is also rated as one of the most difficult-to-extrude aluminum alloys in the extrudability scale however, slightly less difficult than any precipitation hardened AA7XXX alloy and some work hardened AA5XXX alloys [5]. The maximum extrusion speed of AA2024 was reported as 3.5 m/min for a conventional extrusion technique [6, 7]. The incipient melting of the secondary hardening phases (S-Al2CuMg) of AA2024 resulted in a surface tearing or cracking at an elevated temperature (well below the melting temperature of bulk material i.e., at 502 ℃) due to high strain that limit the extrusion speed of this alloy [5, 8, 9].

The present study aims at increasing the extrusion speed of AA2024 using Shear Assisted Processing and Extrusions (ShAPE) [10] such that higher production rate of this alloy might be a cost-effective and competitive solution beyond just the aerospace applications, for example, commercial vehicles, high-speed passenger trains, etc. ShAPE technology has been successfully demonstrated on two types of Al alloys at the extremes of the extrudability scale (very easy vs. very difficult to extrude), that is, AA6063 [11] and AA7075 [12]. In this study, AA2024 is extruded at an extrusion speed of 7.4 m/min. It is also well established that post-extrusion cold working and natural/artificial aging of AA2024 improves the extrusion parts by enhancing dispersion of strengthening precipitates S/S′ phase (Al2CuMg) [3, 5, 13, 14]. Different heat treatment sequences are also applied in this study to enhance the mechanical properties of ShAPE-extruded AA2024 tubes. The material properties are compared with standard test results for relevant heat treatments/tempers. This study also reports on preliminary microstructural characterizations of the ShAPE-extruded AA2024 tubes.

Experimental Procedure

Materials

The cold finish round bar of AA2024-T351 aluminum alloys were procured commercially and employed in the ShAPE process as the starting billet material. The as-received AA2024-T351 was tested in-house per ASTM E8 [15] on an MTS 810 222 kN test frame to confirm the ultimate tensile strength, yield strength, and elongation as 471 ± 1 MPa, 363 ± 1 MPa, and 25%, respectively. The chemical compositions of AA2024 are also listed in Table 1. The AA2024 cylindrical billets were machined to have feedstock dimensions as 31.8 mm outer diameter (OD), 10.1 mm inner diameter (ID), and a length of 102 mm.

Table 1 Chemical compositions of AA2024
Full size table

ShAPE Process

The AA2024 tubes were extruded using one-of-its-kind ShAPE machine designed at the Pacific Northwest National Laboratory and built by BOND Technologies. The AA2024 billets were placed into the billet container (ID = 31.85 mm) and mandrel (diameter = 10 mm) assembly shown in Fig. 1. The rotating die head contains spiral groove to enhance plastic deformation and flow of materials as the die plunges into the AA2024 billet. The ID of the die is 12 mm and its clearance from the mandrel dictates the thickness of the ShAPE tube, 1 mm thick in this study. The die, container, and mandrel are made of H13 steel that were heat treated to austenitized temperature and quenched to achieve hardness >50 HRC. A K-type thermocouple was attached at about 0.5 mm from the die face at a radius of 15.9  mm to sense die temperature during ShAPE processing. It should be emphasized that this temperature measurement may not be interpreted as the exact temperature of the material being extruded. The maximum hydraulic ram speed of 360 mm/min was applied during ShAPE which resulted in an extrusion speed of 7.4 mm/min at the die exit for the studied extrusion ratio of 20.6:1. It has been reported in friction stir welding (FSW) literature [16, 17] that the maximum strength of the recrystallized zone may occur when processed temperature is ~480 ℃. Several ShAPE trials were run at different rotational speeds (not reported here). It was found that 75 RPM provided the desired processing temperature (482 ℃) for this study.

Fig. 1
figure 1

ShAPE process and components for extruding AA2024 tubes

Full size image

Post-extrusion Heat Treatment

Post-extrusion heat treatments were performed immediately after the extrusion to obtain T3510 and T8510 temper conditions [4, 13]. Table 2 summaries the heat treatment sequences for representative samples in this study. It should be noted here that the only difference between T3510 and T8510 tempers are applying natural aging and artificial aging respectively, after the stretching with 3.5% total strain (cold working). The heat treatment was performed in a Neytech Vulcan 3–550 multi-state programmable furnace with an ambient environment.

Table 2 Post-extrusion heat treatment of ShAPE-ed tubes [4, 13]
Full size table

Mechanical Testing

Tensile testing of the ShAPE+T3510 and ShAPE+T8510 tubes was performed using an MTS312.21 servo hydraulic test frame with a 25 kN load cell at 5.1 mm/min cross head speed. At least four specimens of each category of the heat-treated samples were cut into 127 mm long sections and tested per ASTM B557-15 [18]. The gage length of tensile test specimen was 50 mm and an Epsilon non-contact optical extensometer was employed for the strain measurement.

Microstructural Characterization

The microstructures of the ShAPE+T3510 and ShAPE+T8510 tubes were investigated along the transverse and longitudinal sections. The samples were mounted in epoxy and sequentially ground and polished using standard SiC grit papers with the final polishing sequence with colloidal silica (0.05 μm) using vibratory polisher. The polished specimens were analyzed under the scanning electron microscopy (SEM) for microstructural investigation. A JEOL 7001F FESEM with dual Bruker X-Flash 6|60 EDS detectors, Bruker e-Flash EBSD detector, and Esprit 2.1 software were used for chemical and microstructural evaluation of the samples. The weighted average grain size of as-received AA2024 and ShAPE-extruded AA2024 was measured per ASTM E 1382–97 [19].

Results and Discussion

Tensile Strength of AA2024 Tubes

Figure 2 compares the ultimate tensile strength (UTS), 0.2% yield strength (YS) and % elongation at break of the AA2024 tubes produced using ShAPE followed by T3510 (white bars) and T8510 (gray bars) tempering. The relevant standard of ASM-typical, ASTM-minimum, and maximum reported industry values for the thin walled (<6 mm) tubes [3, 20, 21] are shown for comparison. The UTS of ShAPE + T3510 and ShAPE+T8510 are 467 ± 15 MPa and 523 ± 14 MPa (Fig. 2a). The average UTS values are increased by ~18% as compared with the ASM-typical and ASTM-minimum reported values and 0–8% higher than maximum reported industrial values for the corresponding temper conditions. Also, the 0.2% YS of ShAPE + T3510 and ShAPE + T8510 tempers are 380 ± 14 MPa and 496 ± 14 MPa, which far exceed ASM-typical and ASTM-minimum values (increased by ~30%, in Fig. 2b). Also these YS are about 17% and 11% higher than the maximum reported industrial values for corresponding T3510 and T8510 tempers. The elongation at break ShAPE + T3510 and ShAPE + T8510 tempers are ~100% higher than ASM-typical/ASTM-minimum values and on the par of industry values. An ongoing investigation will attempt to elaborately explain the underlying reason for this property improvement. However, following microstructural investigation will provide a glimpse on the relationship among the mechanical properties, grain structures, and distribution of strengthening precipitated of the ShAPE AA2024 tubes with respect to the base materials structures.

Fig. 2
figure 2

Tensile properties of different tempered AA2024 tube and compared with relevant standard value [3, 20, 21]: a ultimate tensile strength, b 0.2% yield strength, and c % elongation

Full size image

Grain Structures of Tubes from ShAPE + Post-Extrusion Heat Treatment

The pattern quality map of grain structures of as-received AA2024 bar and ShAPE+T3510 and ShAPE+T8510 AA2024 tubes in the longitudinal direction are presented in Fig. 3. The pancake-shaped elongated grain structure in Fig. 3a is commonly observed in the conventionally extruded AA2024 round bars. The minimum and maximum grain sizes in as-received AA2024 was measured from Fig. 3a as ~60 and ~190 μm, respectively, in the extrusion direction shown in Fig. 3d. The grains are significantly refined after the ShAPE plus post-extrusion heat treatment process (Fig. 3b–c). However, the trend of grain size refinement in the radial direction of ShAPE-extruded tubes possess different spreading between naturally aged (T3510) and artificially aged (T8510) AA2024 tubes. Table 3 summarizes the grain size variance through the wall thickness with 0 μm being the tube ID and the tube OD at about 1000 μm. Except the very few layers of grain near the ID of the tube, the grain growth occurred to some extent in artificially aged sample as presented the average grain size in Table 3.

Fig. 3
figure 3

Pattern quality map of grain structures of a as-received AA2024 extruded bar, b ShAPE + T3510, c ShAPE + T8510 tubes in longitudinal direction, and d extrusion direction

Full size image
Table 3 Grain size measured along the radial direction of AA2024 tubes
Full size table

Furthermore, Fig. 4 presents the pattern quality map of grain structures of as-received AA2024 bar and ShAPE+T3510 and ShAPE+T8510 AA2024 tubes in the transverse direction. The conventionally extruded AA2024 bar exhibited grain size within the range of 40–100 μm in the transverse direction (see Fig. 4a). The substantial grain refinement was observed after the ShAPE of AA2024 tubes with average grain sizes of 19.91 ± 9.16 μm and 26.57 ± 19.47 μm for respective tempering of T3510 and T8510 conditions. While the difference in grain size due to temper difference is statistically insignificant in the transverse direction, a phenomenon of grain growth as peripheral coarse grain (PCG) [22] along the edge of the tube OD was evident with artificial aging (T8510). In essence, the grain size refinement observed in both Figs. 3 and 4 is one of the contribution factors for improving the ductility (% elongation at break) of ShAPE-ed tubes with different post-extrusion heat treatments.

Fig. 4
figure 4

Pattern quality map of grain structures of a as-received AA2024 extruded bar, b ShAPE + T3510 and c ShAPE + T8510 tubes in transverse direction

Full size image

Strengthening Precipitation of ShAPE + Post-Extrusion Heat Treatment Tubes

Backscattered electron (BSE) SEM images of two tempered AA2024 ShAPE extruded tubes were evaluated to compare the secondary phase in Fig. 5a–b. The distribution of bright spots in low-magnification SEM micrographs of Fig. 5 a and b possess insignificant difference between naturally aged (T3510) and artificially aged (T8510) tubes. This phenomenon might attribute to the breakdown and uniform redistribution of secondary phases after the ShAPE process with the occurrence of trivial grain growth during artificial aging compared to natural aging. High-resolution SEM images with EDS mapping further reveal the precipitates distribution in which complex secondary phases of AlCuFeMeSi and AlCuFeMn were observed in addition to the presence of the principal hardening phase in the AA2024 system which are S (Al2CuMg) and Ω (Al2Cu) phases. Although, the cluster of S phase is commonly observed in both naturally and artificially aged samples, the dispersion of this phase was noticeable in AA2024-T8510 tube (Fig. 5b) compared to AA2024-T3510 tube (Fig. 5a) while comparing the Cu and Mg distributions. In addition, the T (Al20Cu2Mn3)-phase could not be precisely detected in the current study because of resolution limitations, however, might be estimated from the randomly distributed Mn element in both temper conditions (see Mn element mapping in Fig. 5a–b). Studies have shown that, these nano-sized T-phases formed during SHT and remain stable and assist nucleation of Ω-phase within it during the cold working and artificial aging [23,24,25,26]. It has also been reported that, both T- and Ω-phases activate the pinning and accumulation of dislocation during large plastic deformation and are responsible for improving the UTS and YS of AA2024-T8510 after artificial aging [16, 23,24,25,26]. This might also be true for ShAPE + T8510 condition in this study.

Fig. 5
figure 5

SEM BSE images with EDS mapping illustrating the S phase in longitudinal sections of AA2024 ShAPE tubes after different post-extrusion heat treatment with a T3510 and b T8510 temper

Full size image

Conclusions

The ShAPE has been demonstrated to produce 1 mm thick-walled AA2024 tubes at an extrusion speed of 7.4 m/min. This extrusion speed is twice that of otherwise commercially achievable speed of conventional extrusion processes. Moreover, the UTS, YS and % elongation of ShAPE-ed AA2024 are about 18, 30 and 100% higher than the ASM typical and ASTM minimum values for respective T3510 and T8510 conditions. In addition, the YS of ShAPE-extruded AA2024 are about 17% and 11% higher than the industry reported maximum values for the corresponding T3510 and T8510 temper. In essence, the ShAPE has enabled the refinement and re-distribution of second phases and strengthening precipitates at a faster extrusion speed with improved mechanical properties for precipitation-hardened AA2024 in this study.