A. Evolution of as-cast microstructure with various Mn contents

Figure 1 shows the as-cast microstructure of experimental 3004 alloys with various Mn contents. It can be seen that when the Mn is lower than 1.5% [Alloy M15 in Fig. 1(c)], the as-cast microstructure is mainly composed of Al₆(MnFe) intermetallics with minor Mg₂Si, as shown in Figs. 1(a)–1(c). However, the primary Al₆Mn/Al₆(MnFe) and high volumes of fine Al₆(MnFe) intermetallics were formed in Alloy M20 with 2.0% Mn, ^{Reference Bahadur16,Reference Kang, Li, Su, Liu, Guo and Fu17} as shown in Figs. 1(d) and 1(e). Beside the majority of Al₆(MnFe) intermetallics, stable but low volumes of α-Al(MnFe)Si intermetallics (∼0.4%) is also observed in all experimental alloys, but they are not specifically shown in Fig. 1 due to their similarity with Al₆(MnFe) intermetallics. The difference between these two phases can be found in our previous paper. ^{Reference Liu and Chen2}

FIG. 1.

As-cast microstructure of experimental alloys: (a) Alloy M10, (b) Alloy M12, (c) Alloy M15, and (d)–(e) Alloy M20.

As shown in Fig. 1, it is found that the volume fraction of Al₆Mn/Al₆(MnFe) intermetallics gradually increases from Alloy M10 to Alloy M15 but sharply rises in Alloy M20, which is 1.8% in Alloy M10 to 2.1% in Alloy M12 and 2.9% in Alloy M15 but rapidly rises to 6.1% in Alloy M20 from the image-analysis results. This is further confirmed by the X-ray diffraction shown in Fig. 2. The principal peaks of Al₆Mn/Al₆(MnFe) are indicated in the red square according to the literature. ^{Reference Muggerud, Li and Holmestad10,Reference Liu, Huang, Sun, Zhang, Huang, Wang and Liu18} It can be seen that there is a moderate increase from Alloy M10 to Alloy M15 but a remarkable increase in Alloy M20, confirming the increasing volume fraction of intermetallics with Mn addition, especially with higher Mn (2.0%).

FIG. 2.

XRD results of experimental alloys in as-cast condition.

Figure 3 shows the simulated equilibrium-phase diagram of experimental alloys as well as the mole fraction of Al₆(MnFe) intermetallics with various Mn contents from Thermo-Calc calculations. As shown in Fig. 3(a), the eutectic point is at 1.43% Mn, and it most likely shifts to higher Mn content due to the nonequilibrium condition with high cooling rates in the permanent mold casting used in the present work. Therefore, it is reasonable to classify Alloys M10, M12, and M15 as hypoeutectic alloys, in which Al dendrites first solidify before the eutectic intermetallics. Therefore, the Al₆(MnFe) intermetallics are always formed after Al dendrites and are present in the interdendrites area. However, when the Mn content increases to 2% in Alloy M20, which is higher than the eutectic point even though under nonequilibrium solidification, it becomes a hypereutectic alloy. Therefore, the first solidified phase changes to primary Al₆Mn/Al₆(MnFe) in Alloy M20, as shown in Fig. 3(a) and appears as the blocky form of intermetallics in Fig. 1(d). This is also confirmed by Fig. 3(b), which shows that the formation temperature of Al₆Mn/Al₆(MnFe) in Alloy M20 is higher than that of Al (∼650 °C). Moreover, the mole fraction of Al₆(MnFe) intermetallics in Fig. 3(b) also confirms the observation in Fig. 1 that the volume fraction of intermetallics initially gradually increases from Alloy M10 to Alloy M15 but rapidly rises in Alloy M20. One small peak on the evolution of Al₆(MnFe) intermetallics can be observed in Fig. 3(b), which can be attributed to the formation of α-Al(MnFe)Si intermetallics during solidification [Fig. 3(a)]. However, the difference between peak and plateau in four experimental alloys is similar, meaning that a similar and small volume of α-Al(MnFe)Si intermetallics is formed, which is consistent with our observation.

FIG. 3.

The phase diagram (a) and mole fraction of Al₆(MnFe) intermetallics (b) of experimental alloys.

The evolution of as-cast microstructure is greatly related to the various solidification sequences with increasing Mn contents. As shown in Fig. 3(a), the Al₆(MnFe) intermetallics are solidified after the Al dendrite network in Alloys M10, M12, and M15. Therefore, they only occur in the interdendrite areas. As an example, the distribution of intermetallics in Alloy M12 is shown in Fig. 4(a). It can be clearly seen that all of the intermetallics distribute in the interdendrite areas. In addition, it can be found that the solid solution level of Mn in the Al dendrites is high and Mn-rich areas are large (Keller etching effect ^{Reference Mohammadtaheri13} ), providing the necessary condition for the precipitation of dispersoids during heat treatment. On the other hand, if the Mn content passes over the eutectic point, such as 2% in Alloy M20, primary Al₆Mn/Al₆(MnFe) intermetallics will first solidify, which consumes most of the Mn in the liquid. After that, most of the Mn left in the unsolidified liquid is still participated to the formation of fine Al₆(MnFe) intermetallics in the aluminum matrix. Therefore, the solid solution level of Mn in the aluminum matrix is generally low. As shown in Fig. 4(b), only a small percentage of Mn-rich areas in the matrix can be found in Alloy M20, leading to an unfavorable condition for the precipitation of dispersoid during subsequent heat treatment.

FIG. 4.

Distribution of intermetallics in experimental alloys (Keller etched): (a) Alloy M12 and (b) Alloy M20.

B. Evolution of dispersoids with addition of Mn

In the present work, the peak precipitation treatment of 375 °C/48 h was applied to experimental alloys according to the previous works, ^{Reference Liu and Chen2,Reference Liu, Ma and Chen6} and the microstructures etched with 0.5% HF solution after heat treatment are shown in Fig. 5. The distribution tendency of the dispersoid zone and DFZ is clearly illustrated in Fig. 5. There is a gradual decrease of DFZ from Alloy M10 to Alloy M15 but a sharp increase in Alloy M20 with increasing Mn contents. As shown in Figs. 5(a)–5(c), DFZ is only present in the interdendrite area surrounding Al₆(MnFe) intermetallics, and its volume fraction is gradually decreased from 28% in Alloy M10 to 20% in Alloy M15, as shown in Table II. However, in Alloy M20 [Fig. 5(d)], DFZ seems to be present over a large area of the aluminum matrix, and its volume fraction is much higher than that of the other three alloys in Figs. 5(a)–5(c), which reaches as high as 45% (Table II).

FIG. 5.

Microstructure of experimental alloys treated after 375 °C/48 h (0.5% HF etched): (a) Alloy M10, (b) Alloy M12, (c) Alloy M15, and (d) Alloy M20.

TABLE II.

Characters of dispersoids and DFZ in experimental alloys.

* Note: Standard deviation is shown in bracket.

The evolution of DFZ can be attributed to the different solid solution levels of Mn in the matrix with Mn additions. When the Mn content is lower than the eutectic point (Alloys M10, M12, and M15 in the present work), only the formation of Al₆(MnFe) intermetallics in the interdendrite areas consumes Mn, leading to the presence of interdendrite DFZ. As shown in Fig. 1, a moderate increase in the volume fraction of Al₆(MnFe) among Alloys M10, M12, and M15 is observed, which is from 1.8% to 2.9%, meaning a similar consumption of Mn. On the other hand, the Mn content in the alloys increases from Alloy M10 to Alloy M15, resulting in an increased Mn solid solution. Therefore, the distribution of Mn-rich areas in the matrix increases with increasing Mn contents, resulting in the gradually decreasing volume of DFZ from Alloy M10 (28%) to Alloy M15 (19%). However, in Alloy M20 with Mn higher than the eutectic point, a large amount of primary Al₆Mn/Al₆(MnFe) and fine Al₆(MnFe) form during solidification, which consumes much more Mn solute atoms in the matrix compared to Alloys M10, M12, and M15, leading to a lower Mn solid solution level in the matrix. Therefore, a much higher volume fraction of DFZ (41%) is present and DFZs appear almost everywhere in Alloy M20 [Fig. 5(d)].

The distribution of dispersoids in the dispersoid zone was studied in detail using TEM due to their submicro size, and results are shown in Fig. 6. It is to mention that all the dispersoids in the present work are the α-Al(MnFe)Si dispersoids based on the TEM-EDS results and the previous works. ^{Reference Liu and Chen2,Reference Li and Arnberg9} It is shown that the size of dispersoids slightly decreases from Alloy M10 to Alloy M15, but it becomes bigger in Alloy M20. Meanwhile, the volume fraction of dispersoids initially increases from Alloy M10 [Fig. 6(a)] to Alloy M12 [Fig. 6(b)] and further to Alloy M15 [Fig. 6(c)]. However, much less dispersoids can be observed in Alloy M20 shown in Fig. 6(d). As shown in Table II, the size of dispersoids is found to be 71 nm in Alloy M10, 64 nm in Alloy M12, and 62 nm in Alloy M15 but 77 nm in Alloy M20. On the other hand, the volume fraction of dispersoids increases from 2.68% in Alloy M10 to 3.16% in Alloy M15, but it drops to 1.67% in Alloy M20.

FIG. 6.

Distribution of dispersoids in experimental alloys after 375 °C/48 h: (a) Alloy M10, (b) Alloy M12, (c) Alloy M15, and (d) Alloy M20.

Due to the fact that rapid decomposition of supersaturated solid solutions of Mn in the matrix during heat treatment provides the essential condition for the precipitation of dispersoids, the evolution of dispersoids is greatly related to the gradient difference in Mn concentration between as-cast and after-treated conditions. According to the literature, ^{Reference Li and Arnberg4,Reference Muggerud, Mørtsell, Li and Holmestad5,Reference Zhao, Holmedal and Li19} the concentration of Mn solutes in the solid solution can be estimated by the evolution of EC with the following Eq. (2):

(2) $$1{\rm{/EC}} = 0.0267 + 0.032{\rm{F}}{{\rm{e}}_{{\rm{ss}}}}\% + 0.033{\rm{M}}{{\rm{n}}_{{\rm{ss}}}}\% + 0.0068{\rm{S}}{{\rm{i}}_{{\rm{ss}}}}\% + 0.003{\rm{M}}{{\rm{g}}_{{\rm{ss}}}}\% + 0.0021{\rm{particle}}\% \quad ,$$

where Fe_ss%, Mn_ss%, Si_ss%, Mg_ss%, and particle% are all the solid solution levels of various alloying elements (wt%) while particle% is the volume fraction of particles (vol%). As can be seen from Eq. (2), Si_ss, Mg_ss, and particles have less effect on EC than Mn_ss and Fe_ss. In addition, almost all of the Fe is precipitated in the form of intermetallics during solidification. Therefore, the change of EC during the heat treatment mainly depends on the concentration of Mn in the solid solution. Table III lists the measured EC, volume fraction of all intermetallics including dominant Al₆Mn/Al₆(MnFe) as well as minor Mg₂Si and α-Al(MnFeSi) intermetallics, and the calculated Mn concentrations in the solid solution at both as-cast and after heat-treatment conditions.

TABLE III.

Calculated Mn concentrations in the solid solution.

Note: “AC” and “P” stands for “as-cast” and “after-treated” conditions, respectively. Standard deviation is shown in bracket.

As shown in Table III, the concentration of Mn in the as-cast condition increases with Mn addition until to 1.5%. However, with further addition of Mn to 2% (Alloy M20), the formation of a large amount of primary and fine Al₆(MnFe) intermetallics [Fig. 1(d)] consumes most of the Mn in the matrix, leading to the lowest Mn concentration in the matrix, which is 0.82% compared to 1.06% in Alloy M15. The Mn concentration after heat treatment (375 °C/48 h) and the difference of Mn concentration (ΔC) between as-cast and after-treated conditions are also calculated in Table III. It is found that ΔC increases from Alloy M10 (0.65%) to M12 (0.76%) and further to M15 (0.89%). Because most Mn solute atoms are precipitated in dispersoids, the volume fraction of dispersoids increases with rising Mn contents in Alloys M10, M12, and M15. However, when the Mn is higher (2% in Alloy M20), ΔC even becomes smaller than that of Alloy M10, resulting in the lowest volume fraction of dispersoids in the experimental alloys.

C. Elevated-temperature properties

To evaluate the influence of Mn contents on the mechanical properties, the evolution of microhardness at RT was measured in different Mn-containing alloys [Fig. 7(a)]. It can be found that the microhardness initially increases from Alloy M10 (1.0% Mn) to M12 (1.2% Mn) and further to M15 (1.5% Mn), indicating the positive effect of Mn addition on improvement of mechanical properties at RT. However, the microhardness decreases with further addition of Mn to 2% (Alloy M20) when compared to Alloy M15.

FIG. 7.

Evolution of microhardness at RT (a) and YS at 300 °C (b) with Mn contents.

Meanwhile, the YS at 300 °C was also measured and results are shown in Fig. 7(b). Similar to the tendency of microhardness at RT, the YS at 300 °C increases with increasing Mn contents up to 1.5%, but it decreases when 2.0% Mn was added. For instance, the YS at 300 °C first increases from 78 MPa in Alloy M10 to 80 MPa in Alloy M12 and 84 MPa in Alloy M15 with increasing Mn contents from 1% to 1.5%, but it drops to 75 MPa in Alloy M20 with further increasing the Mn content to 2%.

It is reported ^{Reference Liu and Chen2,Reference Es-Said, Zeihen, Ruprich, Quattrocchi, Thomas, Shin, O’Brien, Johansen, Tijoe and Ruhl20} that the microhardness at RT and the YS at 300 °C in 3004 alloys with 1.0–1.2% Mn were 50 HV and 55 MPa, respectively, after the industrial homogenization treatment (600 °C/24 h), in which there was almost no dispersoids precipitated. Although the properties at both RT and 300 °C in Alloy M20 after 375 °C/48 h are the worst among four alloys, they are still higher than that after the homogenization-treated 3004 alloys, which is still 59 HV on the microhardness at RT and is 75 MPa on the YS at 300 °C, indicating the important effect of dispersoid strengthening on mechanical properties both at RT and elevated temperatures. Furthermore, it is noted that the improvement on mechanical properties at 300 °C is larger than at RT, confirming the remarkable enhancement on elevated-temperature properties from dispersoids. For instance, the microhardness at RT is improved from 50 HV to 67 HV (34% increase), and the YS at 300 °C is increased from 55 to 84 MPa (54% increase) compared to the homogenization condition without the presence of dispersoids to the optimized condition in the present work (Alloy M15 after 375 °C/48 h).

For applications at elevated temperatures, the creep resistance is one of the most important considerations in alloy selections. ^{Reference Arzt21,Reference Knipling, Dunand and Seidman22} Therefore, the creep properties at 300 °C of experimental alloys after 375 °C/48 h were assessed. The typical creep curves of four alloys are shown in Fig. 8, and the creep properties are summarized in Table IV.

FIG. 8.

Typical creep curves at 300 °C of experimental alloys after 375 °C/48 h.

TABLE IV.

Creep properties of experimental alloys.

As shown in Fig. 8, it is clear to observe that the total creep strain decreases from Alloy M20 to M10 and M12 and finally to M15. As listed in Table IV, Alloy M20 possesses the highest values of the total creep strain and minimum creep rate. The total creep strain is 0.022 in Alloy M20 but decreases to the lowest value of 0.005 in Alloy M15. On the other hand, the calculated minimum creep rate is 5.3 × 10⁻⁸ s⁻¹ in Alloy M20, and it decreases to 3.5 × 10⁻⁸ s⁻¹ in Alloy M10, 3.1 × 10⁻⁸ s⁻¹ in Alloy M12, and further to the lowest value of 5.5 × 10⁻⁹ s⁻¹ in Alloy M15, showing the highest creep resistance in Alloy M15, which is in the agreement with the mechanical properties shown in Fig. 7. One thing to be mentioned is that though the creep resistance of Alloy M20 is the worst among four experimental alloys, it is still higher than that in the homogenization-treated 3004 alloy (600 °C/24 h), which was 1.4 × 10⁻⁶ s⁻¹ from the literature. ^{Reference Liu and Chen2}

For the evolution of mechanical properties, two important mechanisms should be considered in the present work, which are solid solution hardening and dispersoid strengthening. ^{Reference Liu and Chen2,Reference Liu and Chen3} As shown in Table III, the concentration of Mn in the matrix after 375 °C/48 h in four experimental alloys is more or less similar, indicating the similar contribution from Mn solid solution strengthening. Therefore, it is reasonable to believe that the changes of elevated-temperature properties (YS and creep resistance) principally resulted from the difference of dispersoids (volume fraction and size) with Mn contents. Due to the partial coherence of dispersoids with the Al matrix, ^{Reference Li, Muggerud, Olsen and Furu1} they can make remarkable contributions on the improvement of alloy properties. According to the Orowan strengthening mechanism on strength and creep resistance, ^{Reference Li and Arnberg4,Reference Dieter23,Reference Karnesky, Meng and Dunand24} the elevated-temperature properties increase with the increasing volume fraction of dispersoids and the decreasing size of dispersoids (Table II), explaining the evolution of strength and creep resistance with various Mn contents in the present work. In brief, as shown in Figs. 5–8, proper control of the Mn content (∼1.5% in the present work) can provide the best condition for dispersoid precipitation and hence the highest properties at both RT and elevated temperatures, making a further progress in exploring the elevated-temperature applications using Al–Mn–Mg 3004 alloys.