2019 Volume 60 Issue 11 Pages 2353-2360
The paper reports the research results on precipitation hardening potentials of Al–Mn–Mg alloys containing a small amount of Cu, during aging at 150°C. The age-hardening response and phase transformation are significantly affected by the Si content of the order of 0.1%. The hardening response increases during aging with increasing Si content. The hardening potential decreases significantly by tensile deformation prior to heat treatment. The pre-strain causes an inhomogeneous nucleation of precipitation on dislocations and cell walls and restricts trans-granular precipitation during aging. In 10% pre-strained alloys, short-term aging leads to dislocation recovery, resulting in an increase in the n-value. Further decrease in dislocation density and increase in the n-value occur due to a small degree of precipitation hardening in a 0.15%-Si-added alloy, whereas dislocation recovery is suppressed by competitive precipitation in a 0.36%-Si-added alloy during prolonged aging.
This Paper was Originally Published in Japanese in J. Japan Inst. Light Metals 68 (2018) 473–479.
The Al–Mn–Mg alloy used for beverage cans belongs to a non-heat treatment type alloy, which can exhibit age hardening due to the formation of precipitation during heat treatment, when a small amount of Cu is present. Several combined studies of alloy composition and aging conditions for forming precipitates have been reported, for example: (i) aging at 150 to 200°C in an Al–1.2Mg–1.1Mn–0.22Cu–0.40Fe–0.15Si (mass%, hereafter) alloy,1) (ii) aging at around 225°C in an Al–1.4Mg–1.0Mn–0.30Cu–0.31Fe–0.17Si alloy,2) and (iii) aging at 160°C in Al–1.3Mg–1.1Mn–0.23Cu–0.40Fe–0.25Si and Al–1.1Mg–1.0Mn–0.1Cu–0.4Fe–0.2Si alloys.3,4) It is believed that the main element Mg and trace amounts of Cu and Si form precipitates which contribute to age hardening. However, the alloy composition and aging conditions investigated in these conventional studies are quite limited. Furthermore, with regard to identification of the precipitate phase, it is limited to estimation of the S′ or S phase from the rod-like characteristics in transmission electron microscope (TEM) images. Industrially, the additional amount of trace elements is a factor that largely fluctuates depending on the blending ratio of scrap. Age hardening behavior with changes in Si content and aging time have hardly been reported for Al–Mn–Mg alloys.
Along with age hardening, tensile deformation behavior such as elongation and n-value (work hardening coefficient) also changes. For example, in Al–Mg–Si alloys and Al–Mg alloys used as automotive body panels, the relationships between the elongation, n-value, and microstructure have been investigated.5) In addition, when used in metal cans for beverages, the material is generally in as-rolled condition and is work-hardened by cold rolling. Therefore, it is necessary to clarify the influence of dislocations introduced in cold working on the age hardening behavior. However, as mentioned above, few studies have reported on the precipitation behavior in Al–Mn–Mg alloys. Regarding the tensile deformation behavior, changes in tensile strength, proof stress, and elongation during aging have been reported, but the influence of pre-strain has not been reported.
In the present study, the effects of Si content and pre-strain on age hardening behavior during isothermal aging are systematically investigated in 0.2%-Cu-added Al–Mn–Mg alloys.
Ingots (their chemical composition is shown in Table 1 were prepared, subjected to soaking at 590°C for 4 h, hot rolled immediately after being removed from the furnace, and cold rolled to a plate thickness of 0.3 mm. Following solution treatment at 500°C for 600 s, isothermal aging at 150°C for up to 2419.2 ks was carried out. With regard to the aging temperature, Ref. 1)) used an alloy with chemical composition similar to that used in the present work. Based on that reference, the temperature of 150°C was selected for observing the hardness change from the start of age hardening to over-aging. In order to investigate the influence of pre-strain, samples with 3% and 10% tensile strain were manufactured using an autograph machine AG-50 kN (Shimadzu Co., Ltd). This was done before isothermal aging at 150°C for the 0.36Si alloy. When the material is used as a can for beverage, the strain introduced by cold rolling is generally higher. However, a relatively low rate of 10% or less was selected in this study in order to evaluate the dispersion state of precipitates after tensile deformation and aging and to clarify the influence of differences in the deformation rate on dislocation structure and density.
For the evaluation of age hardening behavior, samples with different aging times were prepared, and the same autograph machine was used to conduct tensile tests at a crosshead speed of 5 mm/min. From each sample, two JIS-5 specimens were collected, and 0.2% proof stress, tensile strength, and elongation were calculated from the nominal stress versus nominal strain curve obtained from the tensile test. The tensile direction was parallel to the rolling direction of the sheets.
TEM measurements were carried out using Hitachi H-800 and JEOL JEM-2010F microscopes operated at 200 kV. Dislocation density in the samples was measured using a Rigaku X-ray diffractometer (SmartLab) equipped with Cu Kα radiation and operated at 45 kV and 200 mA. Diffraction peaks were fitted with the Voigt function, and the half width was determined. Dislocation density was then calculated using the Williamson-Hall method.6)
The changes in 0.2% proof stress and tensile strength during isothermal aging at 150°C are shown in Fig. 1. The proof stress before aging (A.Q.) of the 0.36Si alloy is higher than that of the other two alloys. It is thought that the additional amounts of Mg and Cu, which have high solid solution strengthening ability, are slightly large, in addition to the large amount of Si. In the 0.03Si alloy, which has the least amount of Si, the proof stress hardly changes until about 10 ks and then increases gradually. The peak value appears after aging for about 1200 ks. In the 0.36Si alloy, which has the largest amount of Si, the proof stress significantly increases from approximately 10 to 100 ks, and its final value is 176 MPa. The larger the amount of Si added, the larger is the increment of proof stress during aging. Tensile strength shows almost the same change as proof stress, but in the 0.36Si alloy, the tensile strength gradually increases while the proof stress hardly changes at short holding times.
Change of (a) 0.2% proof stress and (b) tensile strength during aging at 150°C in the alloys with different Si content.
As an example of the nominal stress versus nominal strain curve, Fig. 2 shows the curves for three types of alloys before aging (A.Q.) and after aging for 345.6 and 2419.2 ks. The local elongation is small under any conditions, approximately in the range of 1–3%, and the difference between alloys and conditions is small. The change in total elongation during aging is shown in Fig. 3. In the 0.03Si alloy, which shows relatively small age hardening response, elongation is almost constant at about 20% and decreases slightly at a holding time of about 1200 ks or more. Elongation in the 0.15Si alloy is also almost constant at short holding times and decreases gradually as the aging time increases. In the 0.36Si alloy, short-time aging increases the elongation. At a holding time of 28.8 ks or more, where the proof stress increases significantly, the elongation decreases as the aging time increases.
Stress-strain curves in the alloys with different Si content in the as-quenched condition and after aging at 150°C for 345.6 ks and 2419.2 ks.
Change of total elongation during aging at 150°C in the alloys with different Si content.
Figure 4 shows the change in the n-value with an increase in strain during tensile deformation in the samples before aging. After eliminating the effect of serration on the true stress versus true strain curve, we calculated the n-value for each strain by the two-point method.7) The n-value increases in the early stage of tensile deformation in all three alloys. The n-value shows a maximum at a true strain of about 0.03 and then decreases. Uchida et al.5) investigated the change in n-value and dislocation structure in tensile deformation in Al–Mg–Si and Al–Mg alloys. It was revealed that the rate of change in the n-value in the true strain range (0.10–0.15), which decreases after reaching its maximum value, shows relatively good agreement with elongation. It was reported that in this true strain range, dislocation cell structures form and dynamic recovery occurs. Following that report, in this study, we calculate the n-value in the true strain range of 0.04–0.06, which is estimated to have a large effect on elongation due to dynamic recovery. Figure 5 shows the change in n-values during aging for three types of alloys. The n-value gradually decreases as the holding time increases (when it exceeds about 100 ks for the 0.03Si alloy and about 50 ks for the 0.15Si alloy). In the 0.36Si alloy, the n-value increases slightly after short-term aging and then decreases significantly as the aging time increases. Therefore, the decrease in total elongation during aging roughly corresponds to the decrease in the n-value, and the decrease in the n-value is possibly due to the promotion of dynamic recovery. It is also found that when the increase in proof stress exceeds about 40 MPa, both elongation and n-value decrease clearly during aging. The causes of the increase in elongation and n-value in the 0.36Si alloy will be discussed later.
n-value variation during tensile deformation in the alloys with different Si content.
Change of n-value during aging at 150°C in the alloys with different Si content.
TEM bright-field images and typical nanobeam electron diffraction (NBD) patterns of the precipitates after aging for 345.6 ks are shown in Fig. 6. The electron beam is close to the [001] zone axis. In the 0.03Si alloy, granular precipitates with a diameter of several nanometers are slightly dispersed. From the position of its diffraction spots, the precipitate is identified as an S′ phase.8) The precipitates observed in the 0.15Si and 0.36Si alloys extend in the ⟨010⟩ and ⟨100⟩ directions, and the number density of the precipitates increases as the Si content increases. Plate-like and rod-like precipitates are mixed within the extended precipitates. From the positions of the diffraction spots on the electron diffraction pattern shown in Fig. 6(b) and (c), the β′′ phase9) is identified, and a Q′ phase10) is also identified in the 0.36Si alloy. The conventional studies on the precipitation behavior in Al–Mn–Mg alloys containing about 0.15% Cu have proposed the formation of the S′ or S phase on the basis of rod-like precipitates seen in TEM images.1,2) In general, β′′ and Q′ phases are reported to be formed in Al–Mg–Si–Cu alloys.11) In summary, the present work has revealed for the first time that in Al–Mn–Mg alloys, the main precipitation phase changes depending on the difference in Si content by about 0.1%. Even if the Si content is only about 0.15%, β′′ and Q′ phases can form, and the formation of S′ phase is suppressed.
TEM bright-field images and typical NBD patterns from the precipitates after aging at 150°C for 345.6 ks in the alloys with different Si content.
In the 0.03Si alloy in which fine precipitates are slightly dispersed, the change in elongation and n-value is relatively small. It is believed that both the resistance to dislocation motion and the promotion of inhomogeneous dislocation multiplication by the precipitates are low. In the 0.36Si alloy, both elongation and n-value increase with short-term aging (Figs. 3 and 5). The increase in n-value, that is, the increase in the amount of work hardening during tensile deformation, corresponds to an increase in tensile strength. The reason why the n-value increases during short-term aging only in the 0.36Si alloy is believed to be that the solute Si content is larger than that in other alloys, and therefore, the amount of work hardening due to tensile deformation increases. It is suggested that dislocations are more uniformly distributed, and deformation is not localized, resulting in an increase in elongation.
3.2 Effect of pre-strain on aging behaviorFigure 7 shows the effect of pre-strain on the 0.2% proof stress change during isothermal aging at 150°C in the 0.15Si and 0.36Si alloys. 3% and 10% pre-strain increases the proof stress before aging by approximately 65 and 120 MPa, respectively. In both the 0.15Si and 0.36Si alloys, the pre-strained material softens in the early stage of aging. In the pre-strained 0.15Si alloy, the proof stress gradually increases with aging time. The increment is small, and the proof stress after aging does not exceed the value before aging in the 10% pre-strained material. Aging for 28.8 ks or more gives a higher proof stress than before aging in the pre-strained 0.36Si alloy. The difference between the maximum value and the minimum value of the proof stress, that is, the amount of age hardening, is 76, 41, and 38 MPa for the non-strained, 3% pre-strained, and 10% pre-strained 0.36Si alloys, respectively. Pre-straining slightly shortens the aging time to reach the peak proof stress and significantly reduces the amount of age hardening.
Change of 0.2% proof stress during aging at 150°C in the pre-strained (a) 0.15Si and (b) 0.36Si alloys.
Changes in total elongation and n-value during aging in the pre-strained 0.15Si and 0.36Si alloys are shown in Fig. 8. In both alloys, as the aging time increases, the total elongation decreases gradually in the 3% pre-strained materials, and it hardly changes, or increases slightly, with long-term aging in the 10% pre-strained materials. The n-value decreases as the amount of pre-work increases before aging, and increases with short-term aging. The smaller the amount of Si and the larger the amount of pre-strain, the longer the aging time before the n-value starts to decrease. In light of the change in proof stress shown in Fig. 7, the n-value hardly decreases as the amount of age hardening decreases. The n-value after aging does not fall below the value before aging in all pre-strained materials.
Change of (a), (c) total elongation and (b), (d) n-value during aging at 150°C in the pre-strained 0.15Si and 0.36Si alloys.
TEM bright-field images after aging for 28.8 ks in the 0.36Si alloys are shown in Fig. 9. The non-strained material exhibits age hardening of 30 MPa during aging for 28.8 ks. The amount of hardening in both the 3% and 10% pre-strained materials is approximately 20 MPa. A large number of fine precipitates are observed, and plate-like and rod-like forms are also recognized in the non-strained material. On the other hand, the pre-strained ones have extremely few fine precipitates. Low-magnification TEM bright-field images after aging for 28.8 ks in the pre-strained 0.36Si alloys are shown in Fig. 10(a) and (b). The dislocation density is higher in the 10% pre-strained material than that in the 3% pre-strained material. In addition, as shown by the circles in the figure, intermetallic compounds having a diameter of more than 100 nm can be observed. This is presumed to be the residue consisting of Al–(Fe, Mn)–Si compounds after the solution treatment. Dislocation density around the compounds is high. It is believed that the compounds affect the inhomogeneous distribution of dislocations. Figure 10(c) and (d) shows the dark-field image (the same field of view as in Fig. 10(b)) and the electron diffraction pattern in the 10% pre-strained material, respectively. The compounds described above and the fine precipitates less than 50 nm are displayed brightly. Many fine precipitates are observed in the region where the dislocation density is high.
TEM bright-field images after aging at 150°C for 28.8 ks in the non-strained and pre-strained 0.36Si alloys.
(a), (b) Low-magnification TEM bright-field images after aging at 150°C for 28.8 ks in the pre-strained 0.36Si alloys. Intermetallic compounds are indicated by circles. (c) Dark field image using the encircled reflection in the diffraction pattern (d) in the 10% pre-strained 0.36Si alloy.
Figure 11 shows TEM dark-field images of the pre-strained materials after aging for 345.6 ks. The aging time makes the proof stress almost reach the peak value. The larger the pre-strain, the larger is the amount of precipitates. In the 10% pre-strained material, the precipitates are distributed along the dislocation cell walls formed by the accumulation of dislocations.
TEM dark-field images after aging at 150°C for 345.6 ks in the pre-strained 0.36Si alloys.
With regard to age hardening of cold-worked materials, it has been reported that the peak hardness after aging increases as the deformation rate increases. But the amount of age hardening decreases in various pre-strained alloys under various aging conditions such as: (i) Al–Mg–Si alloys (cold rolling with 1–60% reduction, followed by aging at 200°C,12) and cold rolling with 30% reduction, followed by aging at 25°C–170°C13)), (ii) Al–Cu–Mg–Ag alloys (cold rolling with 3–60% reduction, followed by aging at 190°C14)), (iii) Al–Li–Cu–Mg alloys (cold rolling with 10% reduction, followed by aging at 100°C and 190°C13)), and (iv) Al–Ag alloys (cold rolling with 70% reduction, followed by aging at 100°C15)), etc. Pre-straining allows many solute atoms to adhere to dislocations, reducing the number of solute atoms that contribute to precipitation during aging. In the pre-strained materials of the Al–Mn–Mg alloy investigated in this study, it is believed that the solute atoms are spent for work hardening, resulting in a decrease in the amount of age hardening. In addition, the dislocation density is high around the compounds with a diameter of more than 100 nm (as shown in Fig. 10), which promotes inhomogeneous distribution of the precipitates. It is suggested that the heterogeneity also affects the decrease in the amount of age hardening.
In the pre-strained material, precipitates along the dislocation cell walls were observed after aging, where the proof stress almost reached the peak value (Fig. 11). It is presumed that some of the solute atoms adhere to the dislocation aggregates during aging, and cell formation proceeds, resulting in formation of precipitates on the cell walls.
Focusing on the behavior at the early stage of aging, hardening is not seen at the initial stage of aging at 150°C, even for the 0.36Si alloy (with the highest Si content). Ding et al.4) investigated the effect of solution treatment temperature before aging on the age hardening behavior at 160°C in an Al–1.2Mg–1.0Mn–0.2Cu–0.4Fe–0.2Si alloy. It is reported that the lower the solution temperature, the smaller the hardening amount at the early stage of aging, and hardening is hardly seen when the solution treatment temperature is 510°C or lower. The solution treatment temperature in this study is 500°C, and intermetallic compounds with a diameter of more than 100 nm, which are presumed to remain after solution treatment, have been identified, as shown in Fig. 10. It is assumed that the lower the solution treatment temperature, the more easily these Al–(Fe, Mn)–Si compounds remain. It is believed that the compound acts as a sink of vacancies during quenching and aging, resulting in delayed precipitation overall. Dislocations are expected to promote atomic diffusion and precipitation in pre-strained materials, but the age hardening start time is only slightly shortened compared to that of non-strained materials. Figure 12 shows TEM bright-field images in the 10% pre-strained 0.36Si alloys before and after aging for 1.2 ks. Dislocation cells are formed by 10% pre-straining, and regions with low dislocation density are clearly observed before aging. The development of such a dislocation cell structure is presumed to be due to the reaction between dislocations caused by the cross slip of dislocations and the rearrangement of dislocations due to dynamic recovery. On the other hand, the dislocation density in the dislocation cell walls is lower, and the dislocation distribution is somewhat uniform in the material aged for 1.2 ks. The change in dislocation density measured by X-ray diffraction during aging in the 10% pre-strained 0.15Si and 0.36Si alloys is shown in Fig. 13 as quantitative evaluation results. The longer the aging time, the lower the dislocation density in the 0.15Si alloy. On the other hand, the dislocation density significantly decreases during aging for 1.2 ks and hardly changes with longer aging in the 0.36Si alloy. In other words, the dislocation density decreases during aging for about 1.2 ks in both the 10% pre-strained 0.15Si and 0.36Si alloys. It is thought that the decrease in the dislocation density during short-term aging causes a delay in dislocation cell formation and dynamic recovery during tensile deformation, resulting in an increase in the n-value. In the 0.15Si alloy, even if the aging time is long, the amount of age hardening is small, and the recovery of dislocation progresses further. On the other hand, in the 0.36Si alloy with a relatively large amount of precipitation during aging, recovery of dislocations and decrease in dislocation density is suppressed when the aging time exceeds 1.2 ks, contributing to the increase in proof stress. It is believed that the n-value decreases with hardening as the aging time increases.
TEM bright-field images in the 10% pre-strained 0.36Si alloys (a) without aging and (b) with aging at 150°C for 1.2 ks.
Change of the dislocation density during aging at 150°C in the 0.15Si and 0.36Si alloys.
Systematic investigation has been carried out of the effects of Si content and pre-straining on age hardening behavior during isothermal aging in Al–Mn–Mg alloys containing a small amount of Cu. The conclusions are:
Insightful discussion with Dr. Tatsuo Sato, Professor Emeritus of Tokyo Institute of Technology is greatly appreciated. The authors would also like to thank Dr. Taku Moronaga, former Kobelco Research Institute Co., Ltd., for TEM observation and identification of precipitated phases.
