Journal of Materials Science & Technology 39 (2020) 135–143

Contents lists available at ScienceDirect

Journal of Materials Science & Technology

journal homepage: www.jmst.org

Research Article

Improvement of elevated-temperature strength and recrystallization

resistance via Mn-containing dispersoid strengthening in Al-Mg-Si
6082 alloys
Chen Li, Kun Liu ∗ , X.-Grant Chen
Department of Applied Science, University of Québec at Chicoutimi, Chicoutimi, QC G7H 2B1, Canada

a r t i c l e i n f o a b s t r a c t

Article history: The precipitation behavior of Mn-containing dispersoids in Al-Mg-Si 6082 alloys with different Mn con-
Received 26 June 2019 tents (0, 0.5 and 1.0 wt%) during various heat treatments (300–500 ◦ C) was investigated. The effects
Received in revised form 5 August 2019 of dispersoids on elevated-temperature strength and recrystallization resistance during hot-rolling and
Accepted 12 August 2019
post-rolling annealing were evaluated. The results showed that the dispersoids in the Mn-containing
Available online 26 September 2019
alloys (0.5 and 1.0%) began to precipitate at 350 ◦ C and reached the optimum conditions after 2–4 h
at 400 ◦ C. However, the dispersoids coarsened with increasing holding time at temperatures above
Keywords:
450 ◦ C. After the peak precipitation treatment at 400 ◦ C for 2 h, the yield strength at 300 ◦ C increased
Al-Mg-Si 6082 alloy
Mn addition from 28 MPa (base alloy free of Mn) to 55 MPa (alloy with 0.5% Mn) and 70 MPa (alloy with 1% Mn), respec-
Dispersoid precipitation tively, demonstrating a significant dispersoid strengthening effect at elevated temperature. In addition,
Elevated-temperature strength the dispersoids were thermally stable at 300 ◦ C for up to 1000 h holding owing to its relative high precip-
Recrystallization resistance itation temperature (350–400 ◦ C), leading to the superior constant mechanical performance at elevated
temperature during the long service life. During hot rolling and post-rolling annealing, the presence of
a large amount of dispersoids results in the higher Zener drag PZ compared with base alloy and then
significantly improved the recrystallization resistance. The alloy containing 0.5% Mn exhibited the high-
est recrystallization resistance among three experimental alloys studied during the post-rolling process,
likely resulted from the lower coarsening rate of dispersoids and the lower dispersoids free zone.
© 2020 Published by Elsevier Ltd on behalf of The editorial office of Journal of Materials Science &
Technology.

1. Introduction cerning the improvement of the elevated-temperature properties

and recrystallization resistance of 6xxx alloys are of great technical
Al-Mg-Si 6xxx aluminum alloys are one of the most popular alu- interest.
minum wrought alloys with a wide range of applications in the Introducing the thermally stable dispersoids via appropriate
aircraft, automotive and construction industries due to their high alloying elements and heat treatments is reported to be one of the
strength to weight ratio, excellent corrosion resistance and low most efficient and applicable approaches to enhance the elevated-
cost [1–3]. Recently, there have been increasing demands for alu- temperature properties of aluminum alloys. It has been reported
minum alloys to be applied at elevated temperatures (250–350 ◦ C) that the yield strength (YS) at 300 ◦ C of Al-Mn-Mg 3004 alloys could
in the automotive industries. However, the mechanical properties be improved by 42% in comparison to that of the as-homogenized
of 6xxx alloys deteriorate rapidly due to the coarsening and dis- sample due to the formation of dispersoids and their excellent
solution of traditional hardening precipitates of Mg2 Si at 250–350 thermal stability [5]. On the other hand, the dispersoids are well
◦ C [4]. In general, the 6xxx alloys are subjected to the rolling and recognized on retarding recovery and inhibiting recrystallization
extrusion processes to achieve a desirable shape. The occurrence during hot deformation and post-deformation heat treatment
of recrystallization during hot deformation inevitably weakens the [3,6–9]. In the AA6013 alloy [10], it was found that the recrys-
mechanical strength of the final products. Therefore, studies con- tallization resistance during hot rolling and post-deformation
heat treatment increased due to the increasing Zener-drag force
induced by Mn-containing dispersoids. Generally, the efficiency
of these dispersoids is highly dependent on their size and dis-
∗ Corresponding author.
tributions. The dispersoids with fine size and dense distribution
E-mail address: [email protected] (K. Liu).

https://doi.org/10.1016/j.jmst.2019.08.027
1005-0302/© 2020 Published by Elsevier Ltd on behalf of The editorial office of Journal of Materials Science & Technology.
136 C. Li et al. / Journal of Materials Science & Technology 39 (2020) 135–143

Table 1 samples were preheated to 400 ◦ C for 30 min and then hot-rolled
Chemical composition (wt%) of the experimental alloys.
at 400 ◦ C for several passes until reaching a final thickness of 3 mm
Alloy Fe Mg Si Mn Al from the original 30 mm. The reduction in thickness was controlled
A 0.26 0.99 0.66 0.00 Bal. at 3 mm for each pass. After a few passes of rolling, the samples were
B 0.27 1.07 0.67 0.52 Bal. returned to the furnace for reheating at 400 ◦ C for 30 min to keep
C 0.23 1.12 0.69 1.06 Bal. the rolling at a similar temperature. Then, the as-rolled samples
were annealed at 500 ◦ C for up to 8 h followed by water quenching. WHy？ 500C
To characterize the microstructure after rolling and annealing, for 8h?
anealing?
are considered to contribute to superior strength at elevated tem- all samples were sectioned and polished parallel to the rolling
perature and a high inhibition effect on recrystallization [11]. The direction. The grain structure was analyzed using the electron
formation of thermally stable dispersoids can be promoted by alloy- backscattered diffraction (EBSD) technique with scanning elec-
ing elements, such as Mn, Fe and Mo [12–14], and proper heat tron microscopy (SEM, JOEL JSM-6480LV). Three different areas of
treatment parameters [3]. However, limited research has been 0.24 mm2 for each condition were scanned with a step size of 2 ␮m.
performed in 6xxx alloys about the evolution of dispersoids at The evolution of the dispersoids formed during various heat treat-
different heat treatments and their influence on the elevated- ments was characterized using optical microscopy (OM), SEM and
temperature properties. In addition, although several works have transmission electron microscopy (TEM, JEOL JEM-2100). To clearly
addressed the effect of secondary particles on recrystallization dur- reveal the dispersoids, the polished samples were etched with 0.5%
ing annealing [15–18], the impact of thermally stable dispersoids HF for 30 s for the OM and SEM observations.
on the recrystallization resistance during a thermomechanical pro-
cess and post-deformation heat treatment in 6xxx alloys is less
clear. 3. Results and discussion
In the present work, Al-Mg-Si 6082 alloys with different Mn con-
tents (0 to 1 wt. %) were subjected to various heat treatments from 3.1. Microstructure in the as-cast and heat-treated conditions
300 to 500 ◦ C for different soaking times. The investigation aimed
to discover the optimum heat treatment process under which the The microstructures of the experimental alloys in the as-cast and
best dispersoid formation conditions with a large density and fine heat-treated conditions were fully studied in the present work. As
size can be achieved to maximize the elevated-temperature prop- an example, the microstructures of Alloys A and C are shown in
erties. After selecting the optimum heat treatment, all samples Fig. 1 to illustrate their evolution with the Mn addition. Fig. 1(a-b)
were subjected to the hot rolling at 400 ◦ C followed by the post- show that the as-cast microstructures of Alloys A and C are similar,
rolling annealing at 500 ◦ C. The recrystallization behavior after consisting of ␣-Al dendrites, dark primary Mg2 Si and gray Fe-rich
hot rolling and during post-rolling annealing was further stud- intermetallics [19]. Based on the results of the image analysis,
ied. the volume fractions of primary Mg2 Si and Fe-rich intermetallics
in Alloys A and C were 0.35 vol.% and 0.47 vol.% for Mg2 Si and
2. Experimental 1.01 vol.% and 1.23 vol.% for the Fe-rich intermetallics, respectively.
The Fe-rich intermetallics in Alloy C contained a small amount of
Three Al-Mg-Si 6082 alloys with different Mn contents were Mn to replace Fe, and the increase in the volume fraction of Fe-
prepared in the present study. The chemical compositions of the rich intermetallics is not significant compared to that in Alloy A.
experimental alloys are listed in Table 1. Approximately 3 kg of Therefore, it suggests that most of Mn in Alloy C remained in the
materials for each alloy was melted in an electrical resistance supersaturated solid solution of the aluminum matrix during solid-
furnace and then cast in a permanent metal mold to obtain the ification. This was further confirmed by the higher microhardness
rectangular ingots (30 mm × 40 mm × 80 mm). and lower EC of Alloy C in the as-cast condition, which was 69.1 HV
Cast ingots were subjected to various heat treatments rang- and 28.9%IACS compared with 62.5 HV and 37.7%IACS for Alloy
ing from 300 to 500 ◦ C for up to 24 h with a heating rate of 100 A.
◦ C/h followed by water quenching. To evaluate the evolution of Fig. 1(c-g) shows the microstructures of Alloys A and C after heat
the dispersoids during the heat treatments and their influence on treatment at 450 ◦ C/24 h. For both alloys, some of the primary Mg2 Si
the mechanical properties at room temperature (RT) and elevated is dissolved during heat treatment, while little change in the Fe-rich
temperature, electrical conductivity (EC), Vickers microhardness intermetallics is observed. However, a significant difference can be
and compression yield strength measurements were conducted. observed in the aluminum matrix between the two alloys. As shown
The EC was measured using a Sigmascope SMP10 electrical con- in Fig. 1(c-d) for Alloy A, no dispersoids can form without Mn addi-
ductivity unit at RT, and the average value of 3 measurements tion and only sparely coarsen Mg2 Si, which is likely formed during
was recorded for each condition. Microhardness tests were per- the heating and holding process, is distributed in the matrix. How-
formed on an NG-1000 CCD microhardness test machine with a ever, a large amount of dispersoids formed in Alloy C (Fig. 1(e-j)).
load of 10 g and a dwell time of 20 s on polished samples. The The dispersoid zone (indicated by blue circles) and dispersoid-free
indentations were applied inside the dendrite cells to evaluate the zone (DFZ, indicated by red circles) can be observed in the etched
strengthening effect of dispersoids in the aluminum matrix, and the optical image (Fig. 1(e)). The details of the dispersoids are shown
average value was obtained from 15 measurements for each sam- in the SEM (Fig. 1(f)) and TEM (Fig. 1(g)) images. The dispersoids in
ple. The elevated-temperature YS was obtained from compression Fig. 1(g) are confirmed as Mn-containing ␣-Al(MnFe)Si dispersoids
tests performed on a Gleeble 3800 thermomechanical testing unit by the TEM-EDS results (results not shown here) and the data from
at 300 ◦ C with a total strain of 0.2 and a fixed strain rate at 10−3 the literature [5,12]. It is noteworthy that the characterization of
s-1 . A cylindrical specimen with dimensions of 15 mm in length the dispersoids by SEM and TEM is comparable. For instance, the
and 10 mm in diameter was used. Prior to the compression tests, average size of the dispersoids after 450 ◦ C/24 h is 112 nm from the
the specimens were annealed at 300 ◦ C for 100 h to stabilize the SEM images while they are 96 nm from the TEM images, and the
microstructure. number density of dispersoids in both cases is similar. Since lim-
For hot rolling tests, ingot samples were cut into a size of ited observations were performed with TEM, the evolution of the
30 mm × 40 mm × 40 mm and subjected to an optimum heat treat- dispersoids for two Mn-containing alloys during heat treatment is
ment (400 ◦ C/2 h) for the dispersoid precipitation. Before rolling, all based on the SEM characterization in the following section.
 C. Li et al. / Journal of Materials Science & Technology 39 (2020) 135–143 137

该作者想获得买高温合金，所以HOM温度选择很低。 ALFeMnSi不能粗化，附带的B-a也没有发生
Fig. 1. Microstructures of Alloys A and C in as-cast condition (a–b) and after heat treatment at 450 ◦ C/24 h (c–g).

3.2. Evolution of dispersoids during heat treatment with various Although the values of ␳ are different, a similar tendency of ␳
Mn contents was found between Alloy B and Alloy A (Fig. 2(a)) and between
Alloy C and Alloy A (Fig. 2(b)). When heat treated at 300 ◦ C, ␳
As shown in Fig. 1, a remarkable difference in the microstructure was high but stable over the soaking time, indicating that most of
between Alloys A and C is observed after 450 ◦ C/24 h because of the the Mn atoms are still remaining in the aluminum matrix in Alloys
precipitation of a high number of dispersoids in the Mn-containing B and C. At 350 ◦ C, ␳ decreased sharply with increasing soaking
Alloy C due to the decomposition of supersaturated Mn in the alu- time, indicating a continuous decrease in the Mn solute concentra-
minum matrix during heat treatment [5]. It is reported that the tion in the matrix and the formation of dispersoids. Variations of
supersaturated solid solution level of Mn in the matrix can be indi- ␳ at 400 ◦ C and 450 ◦ C were similar, which exhibited an initial
cated by the change in the EC according to the following equations fast drop followed by a plateau after 4–8 h holding, indicating the
[5,12]: full decomposition of the Mn supersaturated solid solution. The ␳
at 500 ◦ C showed the lowest value among the given heat treatment
1/ECAlloyA = 0.0267 + 0.032Fess % + 0.0068Siss % + 0.003Mgss % (1)
conditions and paralleled with the plateau achieved at 400 ◦ C and
450 ◦ C, suggesting that decomposition of the Mn solutes in the solid
solution was complete even before 1 h at 500 ◦ C.
1/ECAlloyB/C = 0.0267 + 0.032Fess % + 0.033Mnss % The corresponding difference in Vickers microhardness (HV)
between Alloy B/C and Alloy A in the temperature range of 350–500
+0.0068Siss % + 0.003Mgss % (2) ◦ C is shown in Fig. 3. The hardness curves in Fig. 3 can be regrouped

where Fess %, Mnss %, Siss % and Mgss % are the weight percentages of into two types: Type I is in the temperature range of 350–400 ◦ C,
the solute elements Fe, Mn, Si and Mg, respectively. Siss and Mgss where HV continues to increase until reaching a plateau; Type II
have a much weaker effect on the EC than Mnss and Fess , and most is in the temperature range of 450–500 ◦ C, where the HV begins
of Fe was in the form of intermetallics during solidification, such to decrease after reaching the peak value. As shown in Fig. 3(a and
as Al6 (FeMn) and ␣-Al(MnFe)Si [2]. Therefore, the change in the EC b), HV at 350 ◦ C in Alloys B and C increases gradually up to 24 h,
difference between Alloy B/C and Alloy A can be used to estimate indicating the start of the dispersoid precipitation at that temper-
the evolution of the solid solution concentration of Mn in Alloy B/C ature (see also Fig. S1 in the supplementary materials), while HV
during heat treatment by Eq. (3): reaches the highest value at 400 ◦ C after 2 h due to the almost
complete precipitation of dispersoids, which is 20 HV for HVB-A
␳ = 1/ECAlloyB/C -1/ECAlloyA = 0.033Mnss % (3) and 25 HV for HVC-A. Since the change in the hardness is mainly
Fig. 2 shows the differences in 1/EC (␳) between Alloy B/C related to the formation of dispersoids, the evolution of dispersoids
and Alloy A during the heat treatments at various temperatures. during the heat treatment at 400 ◦ C (Type I) and 500 ◦ C (Type II)
138 C. Li et al. / Journal of Materials Science & Technology 39 (2020) 135–143

6082合金 300C以下， Mnss不动

350C开始慢慢 析出

400C，3-4h达到平衡

Fig. 2. Differences in 1/EC during the heat treatments at various temperatures: (a) between Alloy B and Alloy A and (b) between Alloy C and Alloy A.

δHV 400最快最明显，温度高则dispersoids OVer aged

500C 固溶强化

不含Mn

400以下的 Mg2Si粗化，HV降低

Fig. 3. Differences in Vickers hardness (HV) during heat treatments between Alloy B and Alloy A (a), and between Alloy C and Alloy A (b);and (c) HV values of Alloy A as a
function of soaking time.

was selected for the detailed discussion in the following section. As were already uniformly distributed in the matrix after 1 h holding
a reference, the microhardness of Alloy A as a function of soaking and remained stable after 24 h of soaking (Fig. 4(d and g)). The
time is shown in Fig. 3(c) to show the hardness value of the experi- image analysis results (Fig. 5) show that the number density of
mental alloys during various heat treatments. The microhardness of the dispersoids initially increases in the first 2 h and then remains
Alloy A decreases with soaking time when the temperature is lower steady, while the size of the dispersoids increases in the first 4 h and
than 450 ◦ C due to the precipitation of coarsen Mg2 Si precipitates is almost constant up to 24 h with an average diameter of 74 nm in
out of the matrix (Fig. 1(d)), while it increases with the soaking time Alloy B and 85 nm in Alloy C. The above results indicate that the
when treated at 500 ◦ C due to the dissolution of Mg2 Si precipitates dispersoids started to precipitate shortly after reaching 400 ◦ C and
back into the matrix [20]. reached the optimum and stable condition after 2–4 h. In fact, as
Fig. 4 shows the evolution of dispersoids during the heat treat- shown in Figs. 2 and 3, the dispersoids could precipitate at low
ment at 400 ◦ C. In Alloy A, no dispersoids could be observed after temperatures, such as starting from 350 ◦ C. By comparing Alloy B
24 h but only a few coarsened Mg2 Si precipitates (Fig. 4(a)). How- and Alloy C, a denser distribution of dispersoids is observed in Alloy
ever, in Alloy B (Fig. 4(b-d)) and Alloy C (Fig. 4(e-g)), fine dispersoids C with a high Mn content. The number density of dispersoids after
 C. Li et al. / Journal of Materials Science & Technology 39 (2020) 135–143 139

Fig. 4. Microstructure and evolution of dispersoids when treated at 400 ◦ C: (a) Alloy A; (b–d) Alloy B and (e–g) Alloy C.

Fig. 5. Evolution of the number density (a) and equivalent diameter (b) of dispersoids when treated at 400 ◦ C.

400 ◦ C/24 h increases from ∼9 ␮m−2 in Alloy B to 13 ␮m−2 in Alloy in Alloy C because the number density of the dispersoids decreases
C, while the equivalent diameter increases moderately from 74 nm more sharply and the size increases faster in Alloy C than in Alloy
in Alloy B to 85 nm in Alloy C. B (Figs. 6 and 7). Upon further increasing the time to 24 h, a large
Figs. 6 and 7 present the evolution of dispersoids in Alloys B difference can be observed in terms of the number density and size
and C at 500 ◦ C. Compared to the condition treated at 400 ◦ C of the dispersoids, confirming the higher coarsening rate of disper-
(Figs. 4 and 5), it is apparent that the size of the dispersoids at 500 ◦ C soids in Alloy C. In addition, the evolution of dispersoids at 450 ◦ C
was much larger. In addition, the number density of the dispersoids is shown in Fig. S2 as the supplementary materials. In connection
decreases and the size increases with the soaking time, indicating with the changes in the EC (Fig. 2) and HV (Fig. 3), it is evident that
that the coarsening of the dispersoids dominated over time at 500 the coarsening of the dispersoids occurred starting from 450 ◦ C.
◦ C, leading to the lowest HV at 500 ◦ C for both Alloy B and Alloy The influence of the heat treatment temperature and Mn con-
C (Fig. 3). As shown in Figs. 6 and 7 at 500 ◦ C/1 h, the initial size tent on the dispersoids can be attributed to the rapid increase in the
of the dispersoids in Alloys B and C is not largely different, but the diffusion rate of Mn with increasing temperature and Mn concen-
number density in Alloy C is higher. However, after 2 h, the disper- tration. At relatively low temperatures (350–400 ◦ C), the diffusion
soids became coarsening, and the dispersoid coarsening was faster of Mn solutes is limited to a short distance, and thus, the growth
140 C. Li et al. / Journal of Materials Science & Technology 39 (2020) 135–143

Fig. 6. Evolution of dispersoids after treated at 500 ◦ C: (a–c) Alloy B and (d–f) Alloy C.

Fig. 7. Evolution of the number density (a) and equivalent diameter (b) of dispersoids when treated at 500 ◦ C.

of the dispersoids is restricted [21]. However, at high temperatures

(450–500 ◦ C), long-distance diffusion of Mn solutes is possible, and
the growth and coarsening of dispersoids is promoted [21,22]. On
the other hand, the addition of higher Mn in Alloy C increases the
concentration gradient at the interface of the dispersoids, and its
influence on the diffusion rate is mitigated by a smaller diffusion
coefficient at low temperatures but amplified by a larger diffusion
coefficient at high temperatures, leading to the faster coarsening of
dispersoids in Alloy C than Alloy B [21,22].

3.3. Elevated-temperature strength and thermal stability of

dispersoids

As shown in Figs. 4–7, different precipitation behaviors of the

dispersoids in Alloys B and C can be observed, including the changes
in the number density and size with the heat treatment tempera-
ture and time. Therefore, a few typical heat treatment conditions,
such as the initial precipitation stage (350 ◦ C/24 h), the peak pre-
Fig. 8. Yield strength at 300 ◦ C of experimental alloys at typical heat treatment
cipitation condition (400 ◦ C/2 h), the stable precipitation stage (450
conditions.
◦ C/2 h) and the coarsening stage (450 ◦ C/16 h), were selected to

investigate the effect of dispersoids on the YS at elevated temper-

ature (300 ◦ C). To stabilize the microstructure, all samples were 23 and 30 MPa depending on the heat treatment conditions. A sig-
held at 300 ◦ C for 100 h prior to mechanical testing. The results are nificant improvement in YS in the Mn-dispersoid-containing Alloys
shown in Fig. 8. The YS of the base alloys (Alloy A) varied between B and C is observed compared with the base alloy under all given
 C. Li et al. / Journal of Materials Science & Technology 39 (2020) 135–143 141

heat treatment conditions. For Alloy B (0.5% Mn), the highest value and between the reheating processes. Alloy A had only some coarse
of YS is 55 MPa at 400 ◦ C/2 h and 450 ◦ C/2 h. For Alloy C (1%Mn), Mg2 Si particles but was free of dispersoids (Fig. 4(a)) and therefore,
the highest YS is 70 MPa under 350 ◦ C/24 h and 400 ◦ C/2 h condi- it exhibited almost no retarding effect on recrystallization during
tions compared to 30 and 28 MPa for Alloy A under such conditions, hot rolling [29]. On the other hand, the grain structure in Alloys
which represents a more than 130% improvement in the YS at 300 B and C is a typical deformed and recovery structure consisting
◦ C, confirming the essential contribution of the precipitation of of deformed bands and highly elongated grains paralleling to the
dispersoids on the elevated-temperature strength [14]. With the rolling direction with a very small amount of fine recrystallized
increasing heat treatment temperature, a gradual decrease in YS is grains. As described in Section 3.2, a large amount of fine disper-
observed for Alloy C at 450 ◦ C/2 h and 450 ◦ C/16 h, whereas the YS soids precipitated after the heat treatment at 400 ◦ C/2 h prior to
of Alloy B decreases slightly at 450 ◦ C/16 h due to the coarsening rolling (Fig. 4(c and f)). These dispersoids are thermally stable at 400
of dispersoids at high temperature [5,23]. On the other hand, the ◦ C (Fig. 4(d and g)) and play a significant role in retarding recrystal-

difference in YS between Alloy B and Alloy C also changes with the lization during hot rolling, resulting in the elongated and deformed
precipitation treatment conditions. At the peak condition of 400 grain structure remaining in Alloys B and C [3,30,31].
◦ C/2 h, the YS difference between Alloy B and Alloy C is 15 MPa due After hot rolling, the samples were annealed at 500 ◦ C for
to the higher number density of dispersoids but with a similar size up to 8 h, and Fig. 11 shows the evolution of the microstructure
in Alloy C (Figs. 4(c,f) and 5). However, the difference in YS between of the experimental alloys during the post-rolling annealing. For
Alloy B and Alloy C decreases to 4 MPa at 450 ◦ C/16 h due to the Alloy A, a recrystallized microstructure is almost achieved after
higher coarsening rate of dispersoids in Alloy C (Fig. 3 and Fig. 2(S)). 1 h (Fig. 11(a)), although a few subgrains remain in some gains.
For the precipitate-strengthened aluminum alloys, the room- The completion of recrystallization occurs after 2 h (Fig. 11(d)).
temperature strength can be well explained by the classical Orowan With increasing annealing time, the grains slightly grow, and the
bypass mechanism for the large precipitate size. Approaches have equiaxed grains remain almost unchanged up to 8 h (Fig. 11(g)).
been made to estimate the strengthening effect of dispersoids at On the other hand, no obvious recrystallization can be observed in
elevated temperatures using Orowan mechanism [24–26], but it Alloy B even up to 8 h. As shown in Fig. 11(b and e), the highly elon-
is still too far to provide satisfactorily quantitative results. The gated grains are the main characteristic of the microstructure after
strengthening mechanisms at elevated temperature are more com- 1–2 h annealing. With increasing time, the number of subgrains
plex than the room temperature because there is sufficient thermal decreases, and the recovery is in progress. As shown in Fig. 11(h)
energy to allow more activities happen, such as the higher vacancy after 500 ◦ C/8 h, the microstructure of Alloy B still exhibits a main
and solute diffusion, higher grain boundary movement and the recovery grain structure with a small amount of fine and isolated
activation of multi-slip systems for dislocations, etc. Therefore, recrystallized grains on the grain boundaries (indicated by the red
only the experimental evidences of the improvement on elevated- circle), indicating the excellent recrystallization resistance of Alloy
temperature strength from dispersoids is presented here. To better B. For Alloy C, it shows a partially recrystallized microstructure after
understand the strengthening mechanisms at high temperature 1–2 h (Fig. 11(c and f)), presenting as some fine recrystallized grains
and to quantify the elevated-temperature strength induced by ther- that nucleated and grew heterogeneously on the grain boundaries
mally stable dipsersoids is subject of our ongoing investigation.” between elongated gains (see the white circle in Fig. 11(c and f)).
After the peak precipitation condition (400 ◦ C/2 h) to obtain the However, the number of newly recrystallized grains increases only
maximum strength (Fig. 8), a long-term thermal exposure at 300 ◦ C slightly with time; therefore, the elongated recovery grains are still
for up to 1000 h was performed for Alloy C to assess the ther- dominant in the microstructure up to 8 h (Fig. 11(i)), indicating the
mal stability of the dispersoids. Fig. 9 displays the evolution of good recrystallization resistance of Alloy C, which is much better
the microhardness and dispersoids during long-term thermal hold- than that of Alloy A but less than that of Alloy B.
ing. The microhardness remains fairly constant between 68–71 HV Generally, the recrystallization process is initiated from the
(Fig. 9(a)), and the number density and size of dispersoids hardly formation of strain-free subgrains followed by the subsequent
change during the 1000 h thermal holding (Fig. 9(b-d)). This result growth into recrystallized grains at the expense of the surround-
confirms that the ␣-Al(MnFe)Si dispersoids in 6082 alloys have ing deformed matrix. The results in Figs. 10 and 11 reveal that
excellent thermal stability at 300 ◦ C, which can be attributed to the formation of dispersoids prior to rolling is of particular impor-
1) the relative high precipitation temperature (around 350–400 ◦ C, tance to control the recrystallization behavior during hot rolling
mentioned in the section 3.1), and 2) ␣-Al(MnFe)Si dispersoids and post-rolling annealing. The effectiveness of dispersoids in pre-
precipitated from the supersaturated solid solution are the sta- venting recrystallization can be commonly expressed by using the
ble phase and partially coherent with aluminum matrix [27]. The Zener drag PZ that provides a back-driving pressure on the moving
results are consistent with those obtained. The results are consis- grain boundary, which can be calculated in the following equation
tent with those obtained in dispersoid-containing 3xxx [5,28]. [31,32]:

3GB Fv
3.4. Recrystallization behaviors during hot rolling and Pz = (4)
D
post-rolling annealing
where ␥GB is the specific grain boundary energy, FV and D are the
Prior to rolling, ingot samples were subjected to a heat treatment volume fraction and average diameter of dispersoids, respectively.
at 400 ◦ C/2 h to reach the peak condition for dispersoid precipi- As indicated in Eq. (4), a high amount of dispersoids (i.e., a high
tation (Fig. 4). The rolling was performed at 400 ◦ C, and Fig. 10 FV /D value) is essential to achieve a high PZ on grain boundary
shows the as-rolled microstructure of the experimental alloys. It migration for retarding the growth of recrystallized grains. As men-
can be found that the grain structure in Alloy A is almost equiaxed, tioned above, Alloy A had only a few coarse Mg2 Si particles in the
and some grains are slightly elongated in the rolling direction aluminum matrix and was free of dispersoids; therefore, the Pz is
(Fig. 10(a)), indicating partial or full recrystallization during the very small, resulting in a poor recrystallization resistance. After the
rolling process. Dislocation bands and subgrain boundaries can still heat treatment at 400◦ C/2 h prior to rolling, a large number of dis-
be observed in the more and less equiaxed grains, which were likely persoids were present in Alloys B and C, and then PZ increases
formed by the last rolling passes. According to the microstructure significantly, leading to much higher recrystallization resistance
in Fig. 10(a), it is reasonable to make the assumption that at least than Alloy A. However, Alloy B had a lower number density of dis-
the partial recrystallization occurred during each pass of the rolling persoids but their size was smaller compared to Alloy C (Fig. 5). The
142 C. Li et al. / Journal of Materials Science & Technology 39 (2020) 135–143

Fig. 9. Evolution of the microhardness (a) and SEM images showing no remarkable changes of dispersoids during long-term thermal exposure (b–d).

Fig. 10. As-rolled microstructure of experimental alloys.

Fig. 11. Evolution of microstructure during the post-rolling annealing of experimental alloys.
 C. Li et al. / Journal of Materials Science & Technology 39 (2020) 135–143 143

opposite effect of the number density and size of the dispersoids in benefitting from the lower coarsening rate of dispersoids and
Alloys B and C might lead to a similar Pz and thus result in a similar smaller fraction of dispersoids free zones.
recrystallization resistance after hot rolling (Fig. 10(b and c)).
The general tendency of the recrystallization behaviors of Acknowledgements
the three alloys during post-rolling annealing is similar to that
during hot rolling; e.g., the recrystallization resistance of both Mn- The authors acknowledge the financial support of the Natural
dispersoid-containing alloys (B and C) is much better than that of Sciences and Engineering Research Council of Canada (NSERC) and
the Alloy A free of dispersoids. However, the recrystallization resis- Rio Tinto Aluminum through the NSERC Industry Research Chair in
tance of Alloy C is less than that of Alloy B (Fig. 11), presenting as the Metallurgy of Aluminum Transformation at the University of
a higher amount of fine recrystallized grains during post-rolling Quebec at Chicoutimi.
annealing. This difference might be attributed to the higher coars-
ening rate of dispersoids in Alloy C at a high holding temperature Appendix A. Supplementary data
(500 ◦ C) than that in Alloy B (Fig. 7). Meanwhile, a higher vol-
ume fraction of intermetallic particles and the larger dispersoid Supplementary material related to this article can be found, in
free zone C is observed in Alloy due to its high amount of Mn con- the online version, at doi:https://doi.org/10.1016/j.jmst.2019.08.
tent in our previous study [33], which would likely contribute to 027.
more available nucleation sites for recrystallization, leading to a
relatively lower recrystallization resistance compared to Alloy B. References
However, the exact reasoning is not clear at the moment and needs
to be further studied. [1] J. Dutkiewicz, L. Litynska, Mater. Sci. Eng., A 324 (2002) 239–243.
[2] L.P. Troeger, E.A. Starke, Mater. Sci. Eng., A 277 (2000) 102–113.
[3] R. Hu, T. Ogura, H. Tezuka, T. Sato, Q. Liu, J. Mater. Sci. Technol. 26 (2010)
4. Conclusions 237–243.
[4] M. Usta, M.E. Glicksman, R.N. Wright, Metall. Mater. Trans. A 35 (2004)
435–438.
(1) With Mn additions (0.5%–1.0%) to Al-Mg-Si 6082 alloys, ␣-
[5] K. Liu, X.G. Chen, Mater. Des. 84 (2015) 340–350.
Al(MnFe)Si dispersoids began to precipitate at 350 ◦ C and [6] D. Tsivoulas, P.B. Prangnell, Acta Mater. 77 (2014) 1–16.
reached the optimum conditions at 400 ◦ C for 2–4 hours with a [7] S. Tangen, K. Sjølstad, T. Furu, E. Nes, Metall. Mater. Trans. A 41 (2010)
high number density (10–15 ␮m−2 ) and fine size (70–80 nm). 2970–2983.
[8] P. Mukhopadhyay, M. Loeck, G. Gottstein, Acta Mater. 55 (2007) 551–564.
However, the dispersoids became coarse with increasing hold- [9] F. Hichem, G. Rebai, Appl. Phys. A 119 (2015) 285–289.
ing time at temperatures above 450 ◦ C. [10] R.A. Jeniski, B. Thanaboonsombut, T.H. Sanders, Metall. Mater. Trans. A 27
(2) After the peak precipitation treatment at 400 ◦ C for 2 h, the (1996) 19–27.
[11] E. Nes, N. Ryum, O. Hunderi, Acta Metall. 33 (1985) 11–22.
precipitation of a large amount of dispersoids greatly enhanced [12] K. Liu, X.-G. Chen, J. Mater. Res. 32 (2017) 2585–2593.
the elevated-temperature strength. The yield strength at 300 ◦ C [13] K. Liu, X.G. Chen, Metall. Mater. Trans. B 47B (2015) 3291–3300.
increased from 28 MPa (the base alloy free of Mn) to 55 MPa (the [14] K. Liu, H. Ma, X.G. Chen, J. Alloys. Compd. 694 (2017) 354–365.
[15] M. Somerday, F.J. Humphreys, Mater. Sci. Technol. 19 (2003) 20–29.
alloy containing 0.5% Mn) and 70 MPa (the alloy containing 1% [16] I. Nikulin, A. Kipelova, S. Malopheyev, R. Kaibyshev, Acta Mater. 60 (2012)
Mn), respectively. 487–497.
(3) During a long-term thermal exposure at 300 ◦ C, ␣-Al(MnFe)Si [17] B. Forbord, L. Auran, W. Lefebvre, H. Hallem, K. Marthinsen, Mater. Sci. Eng., A
424 (2006) 174–180.
dispersoids in 6082 alloys are thermally stable up to 1000 h [18] A.R. Eivani, H. Ahmed, J. Zhou, J. Duszczyk, Metall. Mater. Trans. A 40 (2009)
owing to its relative high precipitation temperature, resulting 2435–2446.
in the superior constant mechanical performance during the [19] G. Mrówka-Nowotnik, J. Sieniawski, J. Mater. Process. Technol. 162-163
(2005) 367–372.
long service life at elevated temperature.
[20] A.R. Prabhukhot, IJAMSE 3 (2015) 287–294.
(4) During hot rolling and post-rolling annealing, the presence of [21] A. Rodin, N. Goreslavets, Defect Diff. Forum 383 (2018) 31–35.
a large amount of dispersoids significantly increase the Zener [22] N.N. Goreslavets, A.O. Rodin, Phys. Met. Metallogr. 118 (2017)
drag PZ and therefore improved the recrystallization resistance. 1120–1126.
[23] K. Liu, X.G. Chen, Mater. Sci. Eng. A 697 (2017) 141–148.
In the as-rolled condition, a high volume of recrystallized grains [24] G.S. Wang, K. Liu, S.L. Wang, Materials 11 (2018) 1158.
was observed in the base alloy, but it was dominated by highly [25] Z. Li, Z. Zhang, X.G. Chen, Mater. Sci. Eng., A 729 (2018) 196–207.
elongated grains in the Mn-containing alloys. During post- [26] Z. Li, Z. Zhang, X.G. Chen, Mater. Sci. Eng., A 708 (2017) 383–394.
[27] Y.J. Li, A.M.F. Muggerud, A. Olsen, T. Furu, Acta Mater. 60 (2012)
rolling annealing at 500 ◦ C, a full recrystallized structure was 1004–1014.
reached only after 1 h in the base alloy, whereas the main recov- [28] K. Liu, H. Ma, X.G. Chen, Metall. Mater. Trans. B 49 (2018) 1588–1596.
ery grain structure remained up to 8 h in the Mn-containing [29] S. De La Chapelle, P. Duval, Texture Microstruct. 35 (2002) 55–70.
[30] C. Shi, X.G. Chen, Mater. Sci. Eng., A 613 (2014) 91–102.
alloys. [31] J. Lai, C. Shi, X.G. Chen, Mater. Charact. 96 (2014) 126–134.
(5) Under the same post-rolling annealing conditions, the alloy [32] Z. Guo, G. Zhao, X.G. Chen, Mater. Charact. 102 (2015) 122–130.
containing 0.5% Mn exhibited the highest recrystallization [33] C. Li, K. Liu, N. Parson, X.-G. Chen, Proceeding of 16th International Aluminum
Alloys Conference (ICAA16), Mcgill University, Montreal, Canada, June 17-21,
resistance among all three experimental alloys studied, likely 2018.