Improved tensile properties of a new aluminum alloy for high pressure die casting

Author’s Accepted Manuscript Improved tensile properties of a new aluminum alloy for high pressure die casting Peng Zhang, Zhenming Li, Baoliang Liu, Wenjiang Ding, Liming Peng www.elsevier.com/locate/msea

PII: DOI: Reference:

S0921-5093(15)30590-6 http://dx.doi.org/10.1016/j.msea.2015.10.127 MSA32978

To appear in: Materials Science & Engineering A Received date: 18 July 2015 Revised date: 30 October 2015 Accepted date: 31 October 2015 Cite this article as: Peng Zhang, Zhenming Li, Baoliang Liu, Wenjiang Ding and Liming Peng, Improved tensile properties of a new aluminum alloy for high pressure die casting, Materials Science & Engineering A, http://dx.doi.org/10.1016/j.msea.2015.10.127 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. 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.

Improved tensile properties of a new aluminum for high pressure die casting Peng Zhang1, Zhenming Li1, Baoliang Liu2, Wenjiang Ding1, Liming Peng1 1

National Engineering Research Center of Light Alloys Net Forming and State Key Laboratory of Metal

Matrix Composite, Shanghai Jiao Tong University, Shanghai, 200030, P. R. China 2

National Engineering Research Center of Light Alloys Net Forming Feng yang branch center, Chuzhou,

233100, P. R. China Abstract This paper investigates the effects of strain rate and test temperature on the tensile properties and deformation behavior of a recently developed high-ductility cast aluminum alloy Al-5Mg-0.6Mn. The as-cast alloy tested at room temperature and the lowest strain rate of ~1.67 × 10-4 s-1 shows the highest yield strength of ~212 MPa, ultimate tensile strength of ~357 MPa and elongation (~17.6%). Increasing strain rate reduces the ultimate tensile strength and ductility of the as-cast alloy. With the increasing of test temperature, the as-cast alloy shows significantly decreases in tensile strengths and improvements in elongation. The tensile failure of the alloy is mainly originated from the cracking and debonding eutectic particles. The Portevin-Le Chatelier effect occurs in the alloy tested at RT. Strain rate in current study ranges does not significantly affect the work-hardening behavior of the alloy. Increasing test temperature apparently reduces the strain-hardening exponent and coefficient. For the alloy tested at RT, all tensile failures occur prior to global instability, indicating the existence of localized damage. In contrast, for the alloy tested at HT, the global instability occurs at strains below the logarithmic fracture strains, suggesting that there is still a postnecking damage. Keywords: Al-5Mg-0.6Mn aluminum alloy; high pressure die casting; tensile properties; strain rate; high temperature

1. Introduction Die cast aluminum alloy are increasingly used in automotive, aerospace and other transportation industries for light weighting and better performance



[1-4]

. Among various kinds of casting process,

Corresponding author. Tel.: +86 021 54742618; fax: +86 021 34202794. E-mail address: [email protected] (Z.M. Li) -1-

high pressure die casting (HPDC) is more suitable for mass production due to its higher productive efficiency, capacity of producing intricate shapes and thin-walled components, good dimensional accuracy and surface finish, and good mechanical properties

[5,6]

. Because of high filling speed and

fast cooling rate, however, the gases have no enough time to escape from the cavity. As a result, these gases in cavity are inevitably involved in the metal liquid, leading to casting defects such as pores and oxide inclusions forming in the components

[7-10]

. During monotonic or cyclic loading, these

defects serve as stress concentration sites for tensile or fatigue crack initiation, resulting in lower mechanical properties of the castings

[11-13]

. At the high temperature environment, the gases in the

pores will dilate, which induces the surfaces of the castings to bubble. This result apparently influences the appearance quality of the products and deteriorates their mechanical properties. Therefore, aluminum alloy die castings containing pores generally cannot undertake heat treatment, and also do not work under high temperature. Besides reasonable component structure and perfect design mold, the alloy selected for high pressure die casting is also one of the most important elements to obtain excellent die cast productions. The increasing use of cast aluminum components has demanded that these alloys possess high tensile strength, elongation as well as fatigue property. Current commercial aluminum alloys (such as Al-Si-Mg, Al-Si-Cu-Mg and Al-Si-Cu) without sufficient ductility (δ< 5%) for structural applications cannot meet the requirement of modern automobile and aerospace industries [6]

. Heat treatment has been recognized as one of the most important factors in influencing tensile

properties of cast aluminum alloys. These alloys tend to experience heat treatment in order to obtain high tensile strength and ductility [14-18]. For the large die cast components with thin wall, however, it is necessary to carry out the rectification works after heat treatment including solutionized and aged conditions because these treatment technologies often cause the changes in size for these parts. Therefore, it would have more interesting to develop new die cast aluminum alloys which do without heat treatment so as to widen practical applications of cast aluminum alloys. The basic requirements for die casting aluminum alloys can be summarized as follows

[19-21]

: (1) Excellent fluidity is very

important to fill with the cavity and obtain good surface of the castings; (2) Better mechanical properties at room and high temperatures can meet the needs to produce larger and complicated castings with thin-wall; (3) Narrow range of crystallization temperature can reduce the shrinkage and improve the quality of the products; (4) Low tendency of the interaction between the alloy and the -2-

mold cavity will decrease the probability of the die-sticking; (5) Good mechanical properties and corrosion resistance are able to satisfy the demands of the structural applications. Al-Mg aluminum alloy exhibits good corrosion and high mechanical properties and is very attractive for sub-frame and door-frame of the automotive

[22-25]

. Generally, the fluidity and oxidation behavior of the

conventional Al-Mg aluminum alloys is certainly not as good as those of the Al-Si based aluminum alloys. During die casting process, because of the lower casting forming ability and oxidation resistance in the Al-Mg based alloys compared with those of the Al-Si alloys, it has long been challenge to obtain the die castings of the Al-Mg based alloys. In addition, there are few researches and reports focusing on the developments of the new Al-Mg based alloys for high pressure die casting. As a new developed die-casting alloy, Al-5Mg-0.6Mn (all composition in wt.% except otherwise stated) is very attractive to automotive powertrain and structural applications due to its high strength and ductility as good as resistance. More importantly, the tensile performance of this alloy without experiencing heat treatment also can meet the requirements for the components. The composition was optimized for the alloy to provide hardness of ~90 (HB, Brinell hardness), yield strength of ~210 MPa, ultimate tensile strength of ~340 MPa, and elongation of ~15% (at room temperature). In order to improve the formability, mechanical strength and ductility of Al-5Mg alloys, chemical compositions of the alloy applied in the present study were designed as follow: (1) It has reported that in cast aluminum alloys, Fe content is above 0.8%, the molten metal has little or no tendency to dissolve and solder die steel

[26]

. On the other hand, Fe tends to form some brittle

intermetallic compounds which decrease mechanical properties of the castings. In this present study, Fe content is reduced to as low as 0.14% to avoid its negative effect. The main effect of manganese (Mn) (0.6%) added into Al-5Mg alloy is to replace the iron (Fe) and prevent the die-sticking; (2) Besides the silicon (Si) content (< 0.3%), some other impurity elements (i.e. Cu and Zn) are strictly controlled in low level (< 0.04% and 0.08%); (3) RE (rare earth) ( 0.1%) are added in the alloy in order to reduce the oxide inclusions and purify the microstructure of the alloy [27-29]. This new alloy also exhibits the suitable fluidity performance and oxidation resistance during die casting process. The increasing use of cast aluminum components under high stress and high temperature (HT), has drawn considerable interest in their tensile properties and deformation behavior which are generally associated with tensile rate and test temperature [30-37]. In this work, the effects of strain rate and test temperature on the tensile properties, deformation behavior and fracture mechanism of the -3-

as-cast Al-5Mg-0.6Mn aluminum alloy have been studied so as to widen practical applications of this new die casting alloy.

2. Experimental procedure 2.1. Material and sample preparation The alloy with a nominal composition of Al-5Mg-0.6Mn was prepared from high purity Al and Mg, Al-10Mn, Al-10Ti, Al-8Si and Al-10RE master alloys in an electrical resistance furnace, and then cast into a metal mold at pouring temperature of 690~700°C and mold temperature of 180~200°C. Before pouring, the melt was held in the furnace at 690 ± 3°C for 30 min to ensure homogeneity and dissolution of the present intermetallic. Die casting parameters including filling speed (mm/s) and filling time (s) at different stages are shown in Fig. 1. Standard tensile samples according to ASTM B557-06 (gauge diameter of 6.4 mm and gauge length of 64 mm), as shown in Fig. 2, were obtained by using a LK400S 4000KN HPDC machine. The actual composition was measured by an Optima 7300DV inductively coupled plasma optical emission spectroscopy (ICP-OES) (as shown in Table 1). 2.2. Tensile testing Tensile testing was performed on a Zwick/Roell-20kN tensile machine with an attached high temperature furnace controlled within ± 3°C. Stress-strain curves were obtained by attaching a knife-edge extensometer (50 μm) to the gauge section of the specimens. The ultimate tensile strength (UTS), 0.2% proof stress (YS), and elongation (A) of specimens were determined from the test data. In this work, all tensile samples were kept at room temperature for at least two months before testing their tensile properties. Five different strain rates of 1.67 × 10-4 (designated as B1), 3.33 × 10-4 (B2), 6.67 × 10-4 (B3), 1.33 × 10-3 (B4), and 2.67 × 10-3 s-1 (B5) were used to study the effects of the different rate on the tensile properties, deformation and fracture behaviors of this alloy. All the tests were performed at room temperature (RT~20°C) and at least four tests were performed for each strain rate. In addition, five test temperatures including 20°C (designated as C1), 120°C (C2), 150°C (C3), 185°C (C4) and 220°C (C5) were applied to investigate the effect of test temperature on the tensile behaviors of the alloy. Tensile samples were first heated to the desired temperature and held at the test temperature for 15 min, and then tested at a strain rate of 6.67 × 10-4 s-1. In addition, at least -4-

four tensile tests were finished for each test temperature.

2.3. Fractographic Analysis & Microstructural The identification of the various phases was carried out by using a D/MAX-ШA X-ray diffractometer (XRD). The fracture surfaces of all samples were further examined using a JEOL JSM-6460 filed emission scanning electron microscope (SEM) with an attached energy dispersive spectroscope (EDS) (Japan Electron Optics Ltd., Tokyo, Japan). X-ray energy disperse spectrum (EDS) was used in determining the chemical elements and constituents of the phases. In order to investigate the microstructures of the alloy, metallographic samples were taken from the grip areas or fracture areas of the tensile samples using a water-cooled low-speed diamond wheel. The cold mounted samples were carefully ground, polished and etched in a solution composed of 0.5 ml HF and 100 ml H2O. The microstructure observation of the alloy samples was also conducted by a SEM.

3. Results and discussion 3.1. Microstructures The typical microstructure of the as-cast Al-5Mg-0.6Mn alloy, as shown in Fig. 3a, mainly consists of soft -aluminum matrix and hard eutectic phase compounds (such as Al12Mg17, Fig. 3c). A very small amount of porosities can also be seen in the Al-5Mg-0.6Mn alloy samples (marked by the blue arrows in Fig. 3a). The detailed observation in Fig. 3b reveals that the as-cast Al-5Mg-0.6Mn alloy specimens also contain a few massive Alx(Fe, Mn)ySiz intermetallic phases [26,38-40]

(marked by the red arrow), which is confirmed by the EDS analysis (Fig. 3d) and the results

of XRD analysis (as shown in Fig. 4). In addition, no RE-rich phases were seen in the samples due to lower RE content added in this alloy (actual content < 0.05%). Furthermore, low-magnification microstructure observation reveals that nearly all die casting tensile samples contain defect bands (as shown in Fig. 5a), which has been reported in some literatures for HPDC aluminum and magnesium alloys [41-48]. It can be seen that the average width of these defect bands located close to the specimen center is about 200 μm. The average distance between the defect bands and the specimen surface is ~2000 μm. The detailed observations in Fig. 5b clearly indicate that these defect bands mainly comprise inclusions (as shown in Figs. 5c and 5d) and porosities. These defect bands are not the commonly-known segregation bands reported in some aluminum and magnesium die-casting alloys -5-

[41-48]

. For an Al-Si aluminum alloy

[41]

, it has been reported that these segregation bands contain a

higher liquid fraction and higher solute content than their surroundings. Dahle et al.

[42]

pointed out

that for high pressure die castings of magnesium alloys, the banded defects have been shown to be caused by shear deformation of the semi-solid mush entering the die cavity. Gourly et al.

[43]

also

suggested that the formation of these defect bands in alloys is attributed to localized deformation within partially solid material during HPDC. For the Al-5Mg-0.6Mn aluminum alloy, the defect bands containing the impurities are not compatible with the mechanism of segregation bands mentioned in some Al-Si aluminum alloys

[41]

. In this work, to discern the formation reason of these

defect bands, the simulation analysis of mold filling process using a Flow 3D analysis software were performed and the results were shown in Fig. 6. The result in Fig. 5a clearly indicates that the location and symmetry of these defect bands in the Al-5Mg-0.6Mn alloy are fully compatible with the commonly-known segregation bands reported in some aluminum and magnesium alloys

[41-48]

.

This result is mainly attributed to the symmetry of the thermal field acting during solidification, which can be appreciated by simply looking at the results of the thermal field analysis (as shown in Figs. 6a-d). The changing trends of the volume fraction of the entrained air in the Al-5Mg-0.6Mn alloy castings are shown in Figs. 6e-h. During the filling process of the casting, the eddy occurs at the bottom of the bar and the liquid tends to fill the cavity in the form of spiral rise under high pressure and high speed. These results lead to air entrainment and porosities which occur close to the center of the specimen bar during the filling process. In addition, compared with the Al-Si based alloys, the Al-5Mg-0.6Mn alloy also exhibits the lower oxidation resistance due to higher magnesium content in this alloy. In the rising process of the fluid, the molten metal located at the exterior of the casting exhibits the higher flow speed along the mold under the external pressure. In contrast, the flow speed of the molten metal in the middle position of the casting is lower. The different flow speed and solidification rate between the inner and outer of the casting lead to the formation of these defect bands. Any innovative process, which could reduce the level of casting defects, can surely improve the mechanical properties of the castings. Therefore, it would have been more interesting to discuss deeply the mechanism of air entrainment and seek the ways to reduce or eliminate defect bands in the further.

3.2. Effect of strain rate on tensile properties -6-

Fig. 7 and Table 2 show the tensile properties of the as-cast Al-5Mg-0.6Mn alloy tested at different strain rates (test temperature of ~20°C). For the alloy tested at the various strain rates studied in this work, the yield strength (YS) does not show considerable difference and the average YS is about 210 ± 3 MPa. The ultimate tensile strength (UTS) of the alloy slightly decreases with the increasing strain rate. The specimens tested at the strain rate of ~1.67 × 10-4 s-1 (B1) exhibits the highest ultimate tensile strength of ~ 357 MPa, an approximate 30 MPa increase compared with that of the specimens tested at the strain rate of ~2.67 × 10-3 s-1 (B5). In addition, increasing strain rate significantly reduces the ductility of the alloy (from 17.6% to 10.2%). The specimens tested at the lowest strain rate (B1) exhibit the highest ductility, improved by about 72% in elongation, compared with the specimens tested at the strain rate of ~2.67 × 10-3 s-1.

3.3. Effect of test temperature on tensile properties Fig. 8 and Table 3 show the tensile properties of the as-cast Al-5Mg-0.6Mn alloy tested at different test temperatures (strain rate of ~6.67 × 10-4 s-1). The data of the specimens tested at room temperature (C1) are also included for comparison. It is seen that both YS and UTS of the alloy decrease with the increase of test temperature. The specimens tested at the highest temperature (C5~220°C) exhibit the lowest tensile strengths, reduced by 47 MPa in YS and 149 MPa in UTS, respectively, compared with the specimens tested at RT (YS ≈ 210 MPa, UTS ≈ 340 MPa). In contrast, the ductility of the alloy significantly increases with increasing test temperature, as shown in Fig. 8. The specimens tested at the highest temperature of 220°C exhibit the highest ductility (elongation of ~23.9%), which is increased by ~100% compared with the specimens tested at RT (A~12.1%). The differences between tensile properties of the alloy at RT and HT can be greatly attributed to different matrix strengths and different phase particle strengths under different temperature environments. It is commonly accepted that slip bands are easily activated at HT in comparison with RT. Non-basal slip systems can also be activated at elevated temperatures. Moreover, the phase particles tend to act as barriers for dislocation motion and constrain the deformation of the matrix alloy during tensile test

[49]

. However, the Al12Mg17 and Al(Fe, Mn)Si

phase particles are easy to soften at HT (> 130°C), which deteriorates the tensile properties of Al-5Mg-0.6Mn alloy [50,51]. Therefore, the samples tested at RT show significantly improvements in tensile strengths and decreases in ductility compared with the samples tested at HT. -7-

3.4. Fractographic observations Fig. 9 shows the typical fractographs of the as-cast Al-5Mg-0.6Mn alloy tested at RT (20°C), 120°C, 150°C and 220°C. The results clearly indicate that there are lots of dimples on the fracture surface of the samples tested at both RT and HT (as shown in Fig. 9a, c, e and g), corresponding to the high ductility of these samples. The small or large dimples is on the order of the size of phase particles (Fig. 9b, d, f and h), indicating that the particles cracking and debonding are the main damage mechanism prior to the final fracture for the as-cast alloy tested at both RT and HT.

3.5. Strain-hardening behavior of the as-cast alloy tested at different strain rates The effect of strain rate on the flow behavior of the as-cast Al-5Mg-0.6Mn alloy is shown in Fig. 10a. For a better view, curves B1 (strain rate of ~1.67 × 10-4) and B2 (~3.33 × 10-4) have been shifted up 20 MPa and 10 MPa than the raw curves, respectively. In contrast, curves B4 (~1.33 × 10-3) and B5 (~2.67 × 10-3) have been shifted down 10 MPa and 20 MPa than the raw curves, respectively. It was obviously seen that the true stress-logarithmic strain curves for these samples tested at RT and different strain rates exhibited the Portevin-Le Chatelier effect [57-65] (as shown in Fig. 10a), which is considered as a consequence of dynamic strain ageing (DSA) with a certain range of temperature and strain rate

[66-75]

. The normal PLC effect is mainly controlled by pinning of solute atoms and its

diffusion process. It is believed that the PLC effect in the as-cast Al-5Mg-0.6Mn alloy is probably attributed to the interactions between the magnesium atoms in solution and the glide dislocations. According to the load serrations or band propagation characteristics, the PCL effect in this work is confirmed as Type A behavior [76]. Generally, a systematic statistical analysis of the characteristics of the PLC effect in aluminum alloys mainly includes the amplitude of stress drop and the reloading time

[64]

. In the present study, stress drop amplitude, stress increase amplitude, stress decrease time,

reloading time (as shown in Fig. 10b) were analyzed in order to reveal the strain rate dependence of PLC effect. σ1, σ2, σ3, and σ2 +σ3 in Fig. 10b are the stress drop amplitude, stress increase amplitude at rapid increase stage, stress increase amplitude at stable increase stage and total stress increase amplitude, respectively. t1, t2, t3 and t2 + t3 are the stress decrease time, rapid reloading time, stable reloading time, total reloading time, respectively (Fig. 10b). Fig. 11 shows σ1, σ2, σ3, σ2 +σ3 and t1, t2, t3, t2 + t3 as a function of logarithmic strain for the as-cast Al-5Mg-0.6Mn alloys. The results indicate -8-

that σ1 (stress drop amplitude), σ2 (stress increase amplitude at rapid increase stage), σ2 +σ3 (total stress increase amplitude) and t3 (stable reloading time), t2 + t3 (total reloading time) increase with the increase of logarithmic strain. On the other hand, it is seen that σ3 (stress increase amplitude at stable increase stage), t1 (stress drop time) and t2 (rapid reloading time) remain almost with increasing the logarithmic strain. As shown in Figs. 11a and 11d, it can be seen that the total stress increase amplitudes (σ2 +σ3) are apparently higher than the stress drop amplitudes. Similar results are also obtained for the time, in which the total stress increase times (t2 + t3) are also longer than the stress drop times (t1) (as shown in Figs. 11e and 11g). In addition, the results in Figs. 11e-g clearly indicate that t1, t2, t3 and t2 + t3 decrease with increasing the strain rate. For all studied alloy samples except for the samples tested at the lowest strain rate of ~1.67 × 10-4 s-1, however, there is nearly no significantly difference of stress increase amplitude and stress drop amplitude (Figs. 11a-d). Fig. 10c shows the true stress-logarithmic plastic strain curves of the as-cast Al-5Mg-0.6Mn alloy tested at different strain rates and room temperature (~20°C). The effect of strain rate on the strain-hardening rate dσ/dε at a low strain of 0.0015 is shown in Fig. 12a. The results indicate that increasing strain rate does not significantly influence the strain-hardening rate dσ/dε (~4 GPa) at low strains. The Hollomon equation (σ = kεn) was assumed to represent the stress-strain curves of aluminum alloys in plastic deformation stage. The strain-hardening exponent (n) and the strain-hardening

coefficient

(k)

were

determined

from

a

linear

fit

to

the

log(true

stress)-log(logarithmic plastic strain) data over a strain range from about 0.01 up to instability, as shown in Fig. 12b. Figs. 12c-d show the influence of strain rate on the n value and the k value at the plastic strain from 0.01 to instability, respectively. It can be seen that there is nearly no apparently difference of the strain-hardening exponent or the strain-hardening coefficient (n~ 0.16, k~ 480 MPa) among the samples tested at different strain rates. It is, therefore, believed that strain rate in current study ranges does not significantly affect the work hardening behavior of the as-cast Al-5Mg-0.6Mn alloy.

3.6. Strain-hardening behavior of the as-cast tested at different temperatures The effect of test temperature on the flow behavior of the as-cast Al-5Mg-0.6Mn alloy is shown in Fig. 13a. The curve of the samples tested at RT is also included for comparison. The result shows that the flow stress decreases and the total strain increases with increasing test temperature. It is also -9-

clearly seen that the true stress curves of the alloy tested at high temperature all show softening after the flow stress reaches a maximum with increasing the strain. The logarithmic strain corresponding to the maximum true stress decreases with increasing test temperature, as shown in Fig. 13b. The ratios of logarithmic strain at the maximum flow stress (εm) to fracture strain (εf) are summarized in Fig. 13c. The results indicate that the εm /εf ratio for the samples tested at different temperatures decreases with the increasing test temperature. According to the curves in Fig. 13a, it can also be seen that there are two regimes consisting of linear hardening and parabolic hardening. The transition between the two regimes likely controlled by the relatively strength of the soft matrix and the harder eutectic particles

[76-81]

. Logarithmic strain at the transition site between two regions (εt) is also

shown in Fig. 13b. It can be seen that for the samples tested at different temperatures, the increase of test temperature improves the strain corresponding to the transition site. For the as-cast Al-5Mg-0.6Mn alloy, the ratio of εt /εf shows a similar trend as that of logarithmic strain at the transition site relative to various test temperatures, as shown in Fig. 13c. The samples tested at the highest temperature of 220°C exhibit the highest εt /εf ratio (~0.124), which increases by 400% compared with the samples tested at RT. Moreover, it is also seen that for the as-cast Al-5Mg-0.6Mn alloy, the PLC behavior disappears at the high temperature environment (Fig. 13a). PLC effects were indeed eliminated because it is hard to form the effective solution atmospheres locking mobile dislocations. The true stress- logarithmic plastic strain curves of as-cast alloy tested at different test temperatures and the same strain rate (~1.67 × 10-4 s-1) are shown in Fig. 13d. The effect of test temperature on the strain-hardening rate dσ/dε at a low strain of 0.0015 is shown in Fig. 14a. The as-cast samples tested at different test temperatures show the similar strain-hardening rate of ~4 GPa, which is same to that of the samples tested at RT. The n and k values were also determined form a linear fit to the log(true stress)-log(logarithmic plastic strain) data over a strain range from about 0.01 up to instability, as shown in Fig. 14b. Fig. 14c illustrates the influence of test temperature on the n values. The results clearly show that increasing test temperature in the as-cast Al-5Mg-0.6Mn alloy significantly reduces the strain-hardening exponent n. For example, the strain-hardening exponent (n) of the as-cast alloy decreases from 0.157 to 0.043 when the test temperature increases from 20°C to 220°C. The strain-hardening coefficient (k) also decreases with increasing test temperature, as shown in Fig. 14d. The specimens tested at RT exhibits the highest k value of ~465 MPa, improved by ~243 - 10 -

MPa, compared with the samples tested at 220°C. Furthermore, it can also be seen that there are good correlations between strain-hardening exponent or strain-hardening coefficient and test temperature. The best fit line for the present data indicates the relationship of the n and k values via test temperature (T(K)) as: n= -0.008T + 0.23 (R2= 0.991)

[1]

k = -1.673T + 1035 (R2= 0.962)

[2]

The data of the samples tested at RT (20°C) is a little below the fitting line, as shown in Figs. 14c-d. Fig. 15 shows the influence of test temperature on the n and k values at different plastic strain ranges from 0.01 to 0.015, from 0.02 to 0.025, from 0.04 to 0.045, from 0.08 to 0.085, and from 0.12 to 0.125, respectively. For the as-cast alloy, the increase of test temperature significantly reduces the n and k values in the present plastic strain ranges including both lower strains (0.02~0.015) and higher strains (0.12~0.125). Moreover, it can also be seen that at the same test temperature, the n and k values of the alloy decrease with the increase of plastic strain, especially a significant decrease of them at the higher strains (≥ 0.04). It is to be noted that the work hardening rates of the as-cast alloy tested at RT are considerably higher than that obtained at HT. Accumulation of dislocations is considered as the possible mechanisms responsible for the observed variations in work-hardening behaviors. Increasing the temperature accelerates the annihilation processes resulting in lower work hardening rates. With the increase of test temperature, the reduction in the uniform deformation corresponds to the decrease of the n and k values.

3.7. Tensile instability The tensile instability behaviors of the as-cast Al-5Mg-0.6Mn alloy tested at different strain rates and different test temperatures are shown in Fig. 16. The results indicate that for all samples tested at RT except the samples tested at B1, the strain-hardening curves do not intersect with the tensile flow curves prior to failure. The ratios of tensile instability (εi = n) to fracture strain (εf) for the samples tested at different rates are shown in Fig. 17a. It can be seen that the εi /εf values for all samples are close to or larger than 1. This behavior indicates that all tensile samples tested at different rates and RT occur failure prior to the occurrence of global instability. Premature damage and high damage rate (particle cracking) occurring in the samples tested at RT are mainly attributed to localized deformation of the inhomogeneous microstructure. In addition, it can also be seen that - 11 -

the εi /εf ratio of these samples increases with the increase of strain rate. For the samples tested at the lowest strain rate (~1.67 × 10-4 s-1), the ratio of εi /εf is almost equal to 1. In contrast, for all samples tested at HT, the strain-hardening curves do in fact intersect the flow curves prior to failure, as shown in Fig. 16b. The results in Fig. 17b also indicate that for the samples tested at HT, the global instability occurs at strains apparent below the logarithmic fracture strains (εi > n). The damage accumulation rate in the samples tested at HT is lower than that in the samples tested at RT. These results indicate that there is still a postnecking damage

[81]

. Furthermore, it is also seen that

the εi /εf ratio of the samples increases with the increase of test temperature (Fig. 17b). The high ductility in the samples tested at HT is attributed to the low strain hardening rate from the soft matrix and eutectic phases. Like A356/A357 aluminum castings [81], the fracture strain of the as-cast Al-5Mg-0.6Mn alloy also corresponds to the critical amount of damage by particle cracking locally or globally. In the samples tested at RT, the critical amount of damage for failure is easily reached prior to global necking. This result is attributed to the large and hard particles in the as-cast alloy. On the other hand, it is difficult to obtain the critical amount of damage before the global necking takes place in the samples tested at HT, which is attributed to the soft matrix and particles with much lower damage rate. Fig. 18 shows the macroscopic fracture morphologies of the specimens tested at different temperatures. It can also be clearly seen that necking occurs in the samples tested at HT of 220°C, corresponding to better ductility with lots of large and deep dimples. On the other hand, the failure features of the samples tested at RT are macroscopic shear fracture with a large number of shallower shear dimples and some big and deep dimples on the microscale.

5. Conclusion (1) The as-cast Al-5Mg-0.6Mn alloy samples tested at RT and the lowest strain rate of ~1.67 × 10-4 s-1 exhibits the highest tensile properties (yield strength of ~212 MPa, ultimate tensile strength of ~357 MPa and elongation of ~17.6%), improved by ~10% in YS, ~15% in UTS and ~250% in elongation, respectively, compared with the as-cast Al-Mg based alloy. (2) The ultimate tensile strength and the ductility of the as-cast Al-5Mg-0.6Mn alloy decrease with the increase of strain rate. Increasing test temperature reduces the tensile strengths and improves the

- 12 -

ductility of the alloy. For the as-cast Al-5Mg-0.6Mn alloy, phase particles cracking and debonding are the main damage mechanism prior to the final fracture at both room and high temperatures. (3) The Portevin-Le Chatelier effect is observed in the as-cast Al-5Mg-0.6Mn alloy tested at room temperature. This behavior is probably attributed to the interactions between the magnesium atoms in solution and the glide dislocations. With temperature increasing above 120°C, PLC effects disappear because it is hard to form effective solution atmosphere locking mobile dislocations. (4) Strain rate in current study ranges does not significantly affect the work-hardening behavior of the as-cast Al-5Mg-0.6Mn alloy. The strain-hardening exponent and strain-hardening coefficient of the alloy decrease with increasing test temperature. (5) Both global and local instability determine the final failure of the as-cast Al-5Mg-0.6Mn alloy. The ratio of the tensile instability (εi = n) to the fracture strain (εf) is larger than 1 for the as-cast Al-5Mg-0.6Mn alloy tested at RT, indicating that all samples fail prior to global instability. In contrast, for the alloy tested at HT, the global instability occurs at strains apparent below the logarithmic fracture strains. The damage accumulation rate in the samples tested at HT is lower than that in the samples tested at RT. These results indicate that for the as-cast Al-5Mg-0.6Mn alloy tested at HT, there is still a postnecking damage. (6) If we can solve the realistic problems (i.e. HPDC castability: fluidity, oxidation, die soldering and cost issues), it would have been more interesting to use the Al-5Mg-0.6Mn alloy for industrially produced HPDC components.

ACKNOWLEDGMENTS This work was supported by the Project Funded by China Postdoctoral Science Foundation (2015M571562). The authors are grateful to Prof. Qigui Wang (General Motors Company) and Prof. Alan.A Luo (Ohio State University) for their helpful discussions.

REFERENCES [1] A.I. Taub, P.E. Krajewski, A.A. Luo, J.N. Owens, JOM 59(2) (2007) 48-57. [2] S.X. Ji, D. Watson, Z.Y. Fan, M. White, Mater. Sci. Eng. A 556(2012) 824-833. - 13 -

[3] Z.Q. Hu, L. Wan, S.L. Lü, P. Zhu, S.S. Wu, Mater. Design. 55(2014) 353-360. [4] S. Farahany, A. Ourdjini, T.A.A. Bakar, M.H. Idris, Thermochim. Acta 575(2014) 179-187. [5] V.S. Zolotorevsky, N.A. Belov, M.V. Glazoff, Casting Aluminum Alloys, Elsevier, Oxford, 2007. [6] S.X. Ji, F. Yan, Z.Y. Fan, Mater. Sci. Eng. A 626(2015) 165-174. [7] S.X. Ji, Z.Y. Fan, Metall. Mater. Trans. A 33(11)(2002) 3511-3520. [8] V.D. Tsoukalas, Mater. Design. 29(10) (2008) 2027-2033. [9] S.G. Lee, A.M. Gokhale, Scripta Mater. 55(4)(2006) 387-390. [10] H.D. Zhao, F. Wang, Y.Y. Li, W. Xia, J. Mater. Process Tech. 209(9)(2009) 4537-4542. [11] M. Avalle, G. Belingardia, M.P. Cavatora, R. Doglione, Int. J. Fatigue 24(1)(2002) 1-9. [12] H. Mayer, M. Papakyriacou, B. Zettl, S.E. Stanzl-Tschegg, Int. J. Fatigue 25(2003) 245-256. [13] Z. Ma, A.M. Samuel, F.H. Samuel, H.W. Doty, S. Valtierra, Mater. Sci. Eng. A 490(1-2)(2009) 36-51. [14] R. Sharma, Anesh, D.K. Dwivedi, Mater. Sci. Eng. A 408(1-2)(2005) 274-280. [15] B. Lin, W.W. Zhang, Z.H. Lou, D.T. Zhang, Y.Y. Li, Mater. Design. 59(2014) 10-18. [16] I.A. Luna, H.M. Molinar, M.J.C. Román, J.C.E. Bocardo, M.H. Trejo, Mater. Sci. Eng. A 561(20)(2013) 1-6. [17] Y. Han, A.M. Samuel, H.W. Doty, S. Valtierra, F.H. Samuel, Mater. Design. 58(2014) 426-438. [18] Z.X. Wang, H. Li, F.F. Miao, B.J. Fang, R.G. Song, Z.Q. Zheng, Mater. Sci. Eng. A 607(2014) 313-317. [19] S.F. Radtke, D.C. Herrschaft, Inter. J. Mater. Eng. Appl. 1(4)(1979) 209-226. [20] A.A. Luo, J. Magnes. Alloys 1(1)(2013) 2-22. [21] F. Casarotto, A.J. Franke, R. Franke, in: J. Rowe (ed.), Advanced Materials in Automotive Engineering, Woodhead Publishing limited, Cambridge, 2012, pp.109-149. [22] L. Kramer, M. Phillippi, W.T. Tack, C. Wong, J. Mater. Eng. Perform. 21(2012) 1025-1029. [23] R.H. Jones, D.R. Baer, M.J. Danielson, J.S. Vetrano, Metall. Mater. Trans. A 32(2001) 1699-1711. [24] J.L. Searles, P.I. Gouma, R.G. Buchheit, Metall. Mater. Trans. A 32(2001) 2859-2867. [25] C.Y. Meng, D. Zhang, C. Hua, L.Z. Zhuang, J.S. Zhang, J. Alloys Comp. 617(2014) 925-932. [26] G. Timelli, B. Franco, Mater. Sci. Eng. A 528(2010) 273-282. [27] X.W. Hu, F.G. Jiang, F.R. Ai, H. Yan, J. Alloys Comp. 538(2012) 21-27. [28] Q.L. Li, T.D. Xia, Y.F. Lan, W.J. Zhao, L. Fan, P.F. Li, J. Alloys Comp. 562(2013) 25-32. [29] W.H. Yuan, Z.Y. Liang, C.Y. Zhang, L.J. Wei, Mater. Design. 34(2012) 788-792. [30] Y.L. Duan, G.F. Xu, D. Xiao, L.Q. Zhou, Y. Deng, Z.M. Yin, Mater. Sci. Eng. A 624(2015) 124-131. - 14 -

[31] M.M. Sharma, J.D. Tomedi, T.J. Weigley, Mater. Sci. Eng. A 619(2014) 35-46. [32] A.C. Magee, L. Ladani, Mater. Sci. Eng. A 582(2013) 276-283. [33] H. Toda, Z.A.B. Shamsudin, K. Shimizu, K. Uesugi, A. Takeuchi, Y. Suzuki, M. Nakazawa, Y. Aoki, M. Kobayashi, Acta Mater. 61(2013) 2403-2413. [34] S. Banerjee, P.S. Robi, A. Srinivasan, L.P. Kumar, Mater. Sci. Eng. A 527(2010) 2498-2503. [35] G. Han, W.Z. Zhang, G.H. Zhang, Z.J. Feng, Y.J. Wang, Mater. Sci. Eng. A 633(2015) 161-168. [36] Y.Y. Li, W.H. Wang, Y.F. Hsu, S. Trong, Mater. Sci. Eng. A 497(2008) 10-17. [37] Q.G. Wang, Metall. Mater. Trans. A 35A(2004) 2707-2718. [38] S. Murali, K.S. Raman, K.S.S. Murthy, Mater. Sci. Eng. A 190A(1995) 165-172. [39] G. Davignon, A. Serneels, B. Verlinden, L. Delaey, Metall. Mater. Trans. A 27A(1996) 3357-3361. [40] X. Fang, G. Shao, Y.Q. Liu, Z. Fan, Mater. Sci. Eng. A 445-446(2007) 65-72. [41] H.I. Laukli, C.M. Gourlay, A.K. Dahle, O. Lohne, Mater. Sci. Eng. A 413-414(2005) 92-97. [42] A.K. Dahle, S. Sannes, D.H.S. John, H. Westengen, J. Light Met. 1(2001) 99-103. [43] C.M. Gourlay, H.I. Laukli, A.K. Dahle, Metall. Mater. Trans. A 38(2007) 1833-1844. [44] S. Otarawanna, C.M. Gourlay, H.I. Laukli, A.K. Dahle, Metall. Mater. Trans. A 40(2009) 1645-1659. [45] E. Fiorese, F. Bonollo, G. Timelli, L. Arnberg, E. Gariboldi, Int. J. Metalcast 9(1) (2015) 55-66. [46] W.G. Walkington, Die casting defects – causes and solutions, North American Die Casting Association, 1997. [47] E. Gariboldi, F. Bonollo, M. Rosso, La Metallurgia Italiana 99(2007) 39-46. [48] Aluminium and aluminium alloys-Classification of defects and imperfections in high Pressure, Low Pressure and Gravity Die Cast Products, CEN/TR 16749, 2014. [49] Q.G. Wang, C.H. Cáceres, J.R. Griffiths, Metall. Mater. Trans. A 34A(2003) 2901-2912. [50] B. Kondori, R. Mahmudi, Metall. Mater. Trans. A 40A(2009) 2007-2015. [51] F. Kabirian, R. Mahmudi, Metall. Mater. Trans. A 40A(2009) 116-127. [52] C.D. Lee, Mater. Sci. Eng. A 464(1-2) (2007) 249-254. [53] C.D. Lee, K.S. Shin, Mater. Sci. Eng. A 599(2014) 223-232. [54] H.R. Ammar, A.M. Samuel, F.H. Samuel, Inter. J. Fatigue 30(6)(2008) 1024-1035. [55] Y.X. Gao, J.Z. Yi, P.D. Lee, T.C. Lindler, Fatigue Fract. Eng. Mater. Struct. 27(7) (2004) 559-570. [56] Q.G. Wang, D. Apelian, D.A. Lados, J. Light Met. 1(2001) 73-84. [57] M.A. Le Chatelier, Revue de Metallurgie 6 (1909) 914. - 15 -

[58] A. Portevin, F. Le Chatelier, Comptes Rendus Acad. Sci. Paris 176(1923) 507-509. [59] Q.G. Wang, M. Praud, A. Needleman, K.S. Kim, J.R. Griffiths, C.J. Davidson, C.H. Cáceres, A.A. Benzerga, Acta Mater. 58(2010) 3006-3013. [60] M. Niinomi, T. Kobayashi, K. Ikeda, J. Mater. Sci. 5(9)(1986) 847-848. [61] H.S. Rosenbaum, D. Turnbull, Acta Metall. Mater. 6(1958) 653-659. [62] J. Coër, P.Y. Manach, H. Laurent, M.C. Oliveira, L.F. Menezes, Mech. Res. Commun. 48(2013) 1-7. [63] S. Zhang, P.G. Mccormick, Y. Estrin, Acta Mater. 49(2001) 1087-1094. [64] H.F. Jiang, Q.C. Zhang, X.D. Chen, Z.J. Chen, Z.Y. Jiang, X.P. Wu, J.H. Fan, Acta Mater. 55(2007) 2219-2228. [65] T.A. Lebedkina, M.A. Lebyodkin, Acta Mater. 56(2008) 5567-5574. [66] A.H. Cottrell, Dislocations and plastic flow in crystals, Oxford University Press, London, 1953, pp. 134. [67] Y. Estrin, L.P. Kubin, J. Mech. Behav. Mater. 2(1989) 255. [68] C. Fressengeas, A.J. Beaudoin, M. Lebyodkin, L.P. Kubin, Y. Estrin, Mater. Sci. Eng. A 400-401(2005) 226-300. [69] L.P. Kubin, Y. Estrin, J. Phys. Ⅲ France 1(1991) 929-943. [70] A. Nortmann, C.H. Schwink, Acta Mater. 45(1997) 2043-2051. [71] A. Nortmann, C.H. Schwink, Acta Mater. 45(5)(1997) 2051-2058. [72] R.A. Mulford, U.F. Kocks, Acta Metall. Mater. 27(7)(1979) 1125-1134. [73] P.G. McCormick, Acta metall. Mater. 20(1972) 351-354. [74] L.P. Kubin, Y. Estrin, Acta Metall. Mater. 33(1985) 397-407. [75] A. van den Beukel, Physica status. solidi. (a) 30(1975) 197-206. [76] P. Rodriguez, S. Venkadesans, Solid State Phenom. 42-43(1995) 257-266. [77] C.H. Cáceres, J.R. Griffiths, P. Reiner, Acta Mater. 44(1996) 15-23. [78] L.M. Brown, W.M. Stobbs, Phil. Mag. 23(1971) 1185-1199. [79] L.M. Brown, W.M. Stobbs, Phil. Mag. 23(1971) 1201-1233. [80] S.F. Corbin, D.S. Wilkinson, Acta Metall. Mater. 42(4)(1994) 1311-1318. [81] Q.G. Wang, Metall. Mater. Trans. A 35A(2004) 2707-2718.

- 16 -

Tables Table 1 Chemical composition of the Al-5Mg-0.6Mn aluminum alloy (wt.%) Table 2 Tensile properties of as-cast Al-5Mg-0.6Mn alloy tested at different strain rates (Test temperature of 20°C) Table 3 Tensile properties of as-cast Al-5Mg-0.6Mn alloy tested at different temperatures (Strain rate of 6.67 × 10-4 s-1) Table 1 Chemical composition of the Al-5Mg-0.6Mn aluminum alloy (wt.%) Element Mg Mn Ti RE Si Fe Zn Cu Al Al-5Mg-0.6Mn 4.39 0.57 0.15 0.05     Bal.  Table 2 Tensile properties of the as-cast Al-5Mg-0.6Mn alloy tested at different strain rates (Test temperature of 20°C) Strain rate (s-1) Symbol Yield strength (MPa) Ultimate strength (MPa) Elongation (%) 1.67 × 10-4 B1 212 357 17.6 -4 3.33 × 10 B2 213 356 17.5 -4 6.67 × 10 B3 210 340 12.1 -3 1.33 × 10 B4 207 333 10.8 -3 2.67 × 10 B5 207 325 10.2 Table 3 Tensile properties of the as-cast Al-5Mg-0.6Mn alloy tested at different temperatures (Strain rate of 6.67 × 10-4 s-1) Test temperature (°C) 20 120 150 185 220

Symbol C1 C2 C3 C4 C5

Yield strength (MPa) 210 202 188 165 163 - 17 -

Ultimate strength (MPa) Elongation (%) 340 281 250 205 191

12.1 19.0 20.3 21.9 23.9

Figures Fig. 1. Die Casting parameters including filling speed (mm/s) and filling time (s) at different stages. Fig. 2. Diagram of die casting for the standard tensile testing samples of cast aluminum alloy according to the specification defined in ASTM B557-06. Fig. 3. (a) Secondary SEM image of the microstructure from the grip section of the as-cast Al-5Mg-0.6Mn alloy sample. A very small amount of porosities are also seen in the samples (marked by the blue arrows). The location of image (b) is shown by the square in image (a). (c) Energy dispersive spectroscope (EDS) showing the Al-Mg phases (marked by the white arrow in figure b). (d) Energy dispersive spectroscope (EDS) of the phase in location A (marked by the red arrow). Fig. 4. XRD curve of the Al-5Mg-0.6Mn alloy. Fig. 5. (a) Microstructure observation of the as-cast Al-5Mg-0.6Mn alloy showing defect bands. The location of image (b) is shown by the square in image a. The inclusions and porosities were marked by red and white arrows, respectively. (c) High magnification image of the inclusions in image b. (d) Energy dispersive spectroscope spectrum of the inclusions in figure b (location A). Fig. 6. Numerical flow simulation images about (a, b, c, d) thermal field and (e, f, g, h) volume fraction of entrained air based on Flow-3D analysis software showing defect bands formation process. Fig. 7. Effect of strain rate on tensile properties of the as-cast Al-5Mg-0.6Mn alloy. Fig. 8. Effect of test temperature on tensile properties of the as-cast Al-5Mg-0.6Mn alloy. Fig. 9. SEM fractographs showing the effect of test temperature on the fracture in the as-cast Al-5Mg-0.6Mn alloy: (a, b) 20°C; (c, d) 120°C; (e, f) 150°C; and (g, h) 220°C. Fig. 10. (a) Effect of strain rate on true stress-logarithmic strain curves of the as-cast Al-5Mg-0.6Mn alloy (B1-1.67 × 10-4, B2-3.33 × 10-4, B3-6.67 × 10-4, B4-1.33 × 10-3 and B5-2.67 × 10-3). (b) The magnified curve of location A in figure 9a. σ1 is the stress drop amplitude, σ2 is the stress increase amplitude at rapid increase stage, σ3 is the stress increase amplitude at stable increase stage; t1 is the stress decrease time, t2 is the rapid reloading time, t2 is the stable reloading time. (c) True - 18 -

stress-logarithmic plastic strain curves showing the effect of strain rate on flow behavior of the as-cast alloy. For a better view, Curves B1 and B2 have been shifted up 20 MPa and 10 MPa than the raw curves, respectively; Curves B4 and B5 have been shifted down 10 MPa and 20 MPa than the raw curves, respectively. Fig. 11. (a) σ1, (b) σ2, (c) σ3, (d) σ2 +σ3 and (e) t1, (f) t2, (h) t3, (g) t2 + t3 in figure 9b, as a function of logarithmic strain for the as-cast Al-5Mg-0.6Mn alloys. Fig. 12. (a) Strain-hardening rate measured at a plastic strain of 0.0015, as function of strain rate for the as-cast Al-5Mg-0.6Mn alloy. (b) Determination of n and K values for the alloys tested at different strain rates by linear fit to the log true stress-log logarithmic plastic strain curves. The logarithmic plastic strain ranges from about 0.01 up to instability. For a better view, Curves B1 and B2 have been shifted up 0.1 and 0.05 than the raw curves, respectively; Curves B4 and B5 have been shifted down 0.05 and 0.1 than the raw curves, respectively. (c) Strain-hardening exponent n and (d) strain-hardening coefficients k measured at a logarithmic plastic strain range from about 0.01 up to instability, as function of strain rate for the as-cast alloys. Fig. 13. (a) Effect of test temperature on true stress-logarithmic strain curves of the as-cast Al-5Mg-0.6Mn alloy (C2-120°C, C3-150°C, C4-185°C and C5-220°C). (b) Logarithmic strain corresponding to the transition site between the two regions and the maximum true stress. (c) The ratio of strain at the transition site (εt) to fracture strain (εf) and the ratio of strain at maximum flow stress (εm) to fracture strain (εf). (d) True stress-logarithmic plastic strain curves showing the effect of test temperature on flow behavior of the as-cast alloy. The curve of the samples tested at room temperature (C1-20°C) is included for comparison. Fig. 14. (a) Strain-hardening rate measured at a plastic strain of 0.0015, as function of test temperatures for the as-cast Al-5Mg-0.6Mn alloy. (b) Determination of n and K values for the alloys tested at different test temperatures by linear fit to the log true stress-log logarithmic plastic strain curves. The logarithmic plastic strain ranges from about 0.01 up to instability. (c) Strain-hardening exponent n and (d) strain-hardening coefficients k measured at a logarithmic plastic strain range from about 0.01 up to instability, as function of test temperatures for the alloys. Fig. 15. (a) Strain-hardening exponent n and (b) strength coefficient k (at different logarithmic plastic strain ranges) of the as-cast Al-5Mg-0.6Mn alloy, as a function of different test temperature. Fig. 16. Tensile instability plots for the as-cast Al-5Mg-0.6Mn alloy tested at different strain rate and test temperature: (a) B1-1.67 × 10-4, B2-3.33 × 10-4, B3-6.67 × 10-4, B4-1.33 × 10-3 and B5-2.67 × 10-3; (b) C1-20°C, C2-125°C, C3-150°C, C3-185°C and C4-220°C. Fig. 17. The ratio of tensile instability strain εi to fracture strain εf of the as-cast Al-5Mg-0.6Mn alloy tested at different strain rates (a) and test temperatures (b). εi /εf > 1 indicates that local fracture governs tensile instability, while εi /εf < 1 indicates more uniform damage accumulation where tensile instability occurs at the onset macroscopic necking.

- 19 -

Fig. 18. Macroscpic tensile fracture features of the as-cast Al-5Mg-0.6Mn alloy tested at different temperatures (C1-20°C and C4-220°C).

4500 0.005s

4000

Filling speed (mm/s)

3500 3000 2500

Casting parameters Stage 1 Stage 2 Stage 3

2000 1500 1000

0.015s

500 0 -500 1.10

1.12s

1.11

1.12

1.13

1.14

1.15

1.16

Filling time (s)

Fig. 1. Die casting parameters including filling speed (mm/s) and filling time (s) at different stages.

- 20 -

Fig. 2. Diagram of die casting for the standard tensile testing samples of cast aluminum alloy according to the specification defined in ASTM B557-06.

Fig. 3. (a) Secondary SEM image of the microstructure from the grip section of the as-cast Al-5Mg-0.6Mn alloy sample. A very small amount of porosities are also seen in the samples (marked by the blue arrows). The location of image (b) is shown by the square in image (a). (c) Energy dispersive spectroscope (EDS) showing the Al-Mg phases (marked by the white arrow in figure b). (d) Energy dispersive spectroscope (EDS) of the phase in location A (marked by the red arrow).

- 21 -

8000 7000

  

Intensity(CPS)

6000

 



 Al(Fe,Mn)Si







5000



 Al12Mg17  Al3Ti  Al6Mn Al18Mg3Ti2

4000 3000 2000 1000 20

30

40

50

60

70

80

90

100

110

2-Theta/(degree)

Fig. 4. XRD curve of the Al-5Mg-0.6Mn alloy.

Fig. 5. (a) Microstructure observation of the as-cast Al-5Mg-0.6Mn alloy showing defect bands. The location of image (b) is shown by the square in image a. The inclusions and porosities were marked by red and white arrows, respectively. (c) High magnification image of the inclusions in image b. (d) Energy dispersive spectroscope spectrum of the inclusions in figure b (location A).

- 22 -

- 23 -

500

25

400

20

300

15

200

10

100

YS UTS Elongation

o

Test temperature of 20 C

0 0.0

-4

5.0x10

-3

1.0x10

-3

1.5x10

-3

2.0x10 -1

-3

2.5x10

Elongation (%)

Strength (MPa)

Fig. 6. Numerical flow simulation images of (a, b, c, d) thermal field and (e, f, g, h) volume fraction of entrained air based on Flow-3D analysis software showing defect bands formation process.

5

0 -3 3.0x10

Strain rate (s )

Fig. 7. Effect of strain rate on tensile properties of the as-cast Al-5Mg-0.6Mn alloy.

- 24 -

30

500

25

400

20

300

15

200

10

100 Strain rate of 6.67 10

YS UTS Elongation

-4

0

Elongation (%)

Strength (MPa)

600

5

0 0

30

60

90

120

150

180

210

240

Test temperature (Degree celsius)

Fig. 8. Effect of test temperature on tensile properties of the as-cast Al-5Mg-0.6Mn alloy.

- 25 -

Fig. 9. SEM fractographs showing the effect of test temperature on the fracture in the as-cast Al-5Mg-0.6Mn alloy: (a, b) 20°C; (c, d) 120°C; (e, f) 150°C; and (g, h) 220°C. a

b 340

400 B1 B2

350

300

335

B3

3

True stress (MPa)

True stress (MPa)

A B4 B5

250

200

-4

-1

-4

-1

-4

-1

-3

-1

-3

-1

1.67  10 s 3.33  10 s 6.67  10 s

150

1.33  10 s

o

Test temperature of 20 C

2.67  10 s

330

t3 1 2

325

320

t1 t2 315

100 0.00

0.03

0.06

0.09

0.12

0.15

0.18

0.094

0.21

Logarithmic strain

0.096

0.098

0.100

Logarithmic strain

- 26 -

0.102

0.104

c

390 B1 B2

360 B3

True stress (MPa)

330 B4 B5

300 270 240

-4

-1

-4

-1

-4

-1

-3

-1

-3

-1

1.67  10 s 3.33  10 s

210

6.67  10 s

180

1.33  10 s

o

Test temperature of 20 C

150 0.00

0.03

0.06

0.09

2.67  10 s

0.12

0.15

0.18

0.21

Logarithmic plastic strain

Fig. 10. (a) Effect of strain rate on true stress-logarithmic strain curves of the as-cast Al-5Mg-0.6Mn alloy (B1-1.67 × 10-4, B2-3.33 × 10-4, B3-6.67 × 10-4, B4-1.33 × 10-3 and B5-2.67 × 10-3). (b) The magnified curve of location A in figure 9a. σ1 is the stress drop amplitude, σ2 is the stress increase amplitude at rapid increase stage, σ3 is the stress increase amplitude at stable increase stage; t1 is the stress decrease time, t2 is the rapid reloading time, t2 is the stable reloading time. (c) True stress-logarithmic plastic strain curves showing the effect of strain rate on flow behavior of the as-cast alloy. For a better view, Curves B1 and B2 have been shifted up 20 MPa and 10 MPa than the raw curves, respectively; Curves B4 and B5 have been shifted down 10 MPa and 20 MPa than the raw curves, respectively. a

28

b 20

o

Test temperature of 20 C

o

Test temperature of 20 C

Stress increase aplitude 2 (MPa)

Stress drop aplitude

1 (MPa)

24 20 16 12 8 4 0

-1

-4

-1

3.33  10 s

-4

-4

6.67  10 s

-8 -12 -0.03

-4

1.67  10 s

-3

-1

-3

-1

1.33  10 s 2.67  10 s

0.00

0.03

0.06

0.09

0.12

0.15

-1

0.18

15

10

5

0

-4

-1

-4

-1

-4

-1

-3

-1

-3

-1

1.67  10 s 3.33  10 s 6.67  10 s

-5

1.33  10 s 2.67  10 s

-10 -0.02 0.00

0.21

Logarithmic strain

0.02

0.04

0.06

0.08

0.10

Logarithmic strain

- 27 -

0.12

0.14

0.16

30

c

d 35

o

Test temperature of 20 C

2 + 3 (MPa)

Stress increase aplitude 3 (MPa)

25 20

Stress increase aplitude

15 10 5 -4

-1

-4

-1

-4

-1

-3

-1

-3

-1

1.67  10 s

0

3.33  10 s 6.67  10 s

-5

1.33  10 s 2.67  10 s

-10 -0.02 0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

30 25 20 15 10 5

-5

-1

2.67  10 s

0.02

0.04

0.06

0.08

0.10

10

0.12

0.14

0.16

o

Test temperature of 20 C

0

10

-4

-1

-4

-1

-4

-1

1.67  10 s -1

10

3.33  10 s 6.67  10 s -3

10

0

10

-1

-4

-1

3.33  10 s -4

6.67  10 s

-1

-3

-4

1.67  10 s

-1

10

1.33  10 s

-3

-1

-1

1.33  10 s

-1

2.67  10 s

-3

2.67  10 s

-2

-2

10

0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20

-1

0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20

Logarithmic strain

Logarithmic strain

3

3

h 10

o

Test temperature of 20 C

Total reloading time t2 + t3 (s)

Stable reloading time t3 (s)

-1

-3

1.33  10 s

1

Rapid reloading time t2 (s)

Stress decrease time t1 (s)

1

10

2

10

1

10

-4

-1

-4

-1

-4

-1

1.67  10 s 0

10

3.33  10 s 6.67  10 s -3

1.33  10 s -3

o

Test temperature of 20 C

2

10

1

10

2.67  10 s

-4

-1

-4

-1

1.67  10 s 0

10

3.33  10 s -4

6.67  10 s -3

-1

1.33  10 s -3

-1

-1

-1

2.67  10 s

-1

-1

-1

10

-1

-3

6.67  10 s

2

f

o

Test temperature of 20 C

g 10

-1

-4

Logarithmic strain

2

10

10

-1

-4

3.33  10 s

-10 -0.02 0.00

0.16

-4

1.67  10 s

0

Logarithmic strain

e

o

Test temperature of 20 C

10

0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20

0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20

Logarithmic strain

Logarithmic strain

Fig. 11. (a) σ1, (b) σ2, (c) σ3, (d) σ2 +σ3 and (e) t1, (f) t2, (h) t3, (g) t2 + t3 in figure 9b, as a function of logarithmic strain for the as-cast Al-5Mg-0.6Mn alloys. - 28 -

a

12

b

8

2 0

0.163

= 490

0.161

= 490

2.5

2

R = 0.9936 0.155

= 483 2

R = 0.9933

2.4

0.158

= 478 2

R = 0.9932

2.3

o

-1

-4

-1

-4

-1

-3

Test temperature of 20 C

-1

-3

0.0

-4

5.0x10

-3

1.0x10

-3

1.5x10

-3

2.0x10

-3

-1

2.5x10

6.67  10 s

2.2

applied strain = 0.0015

1.33  10 s

2.1 -2.0

-3

3.0x10

-1

2.67  10 s

-1.8

0.28

d

o

Test temperature of 20 C

-1.4

-1.2

-1.0

-0.8

-0.6

750

Strain hardening coefficients (MPa)

700

0.20

0.16

0.12

0.08

-1.6

Log (logarithmic plastic strain)

0.24

Strain hardening exponet (n)

-4

1.67  10 s 3.33  10 s

Strain rate (s )

c

0.157

= 465

R = 0.9933

2.6

4

-4

2

R = 0.9937

2

6

-2

o

Test temperature of 20 C

2.7

Log (true stress)

Strain hardening rate (GPa)

10

2.8

Logarithmic plastic strain range from 0.01 up to instability

o

Test temperature of 20 C

650 600 550 500 450 400 350 300 Logarithmic plastic strain range from 0.01 up to instability

250 200

0.04 0.0

-4

5.0x10

-3

1.0x10

-3

1.5x10

-3

2.0x10

-3

2.5x10

0.0

-3

3.0x10

-1

-4

5.0x10

-3

1.0x10

-3

1.5x10

-3

2.0x10

-3

2.5x10

-3

3.0x10

-1

Strain rate (s )

Strain rate (s )

Fig. 12. (a) Strain-hardening rate measured at a plastic strain of 0.0015, as function of strain rate for the as-cast Al-5Mg-0.6Mn alloy. (b) Determination of n and K values for the alloys tested at different strain rates by linear fit to the log true stress-log logarithmic plastic strain curves. The logarithmic plastic strain ranges from about 0.01 up to instability. For a better view, Curves B1 and B2 have been shifted up 0.1 and 0.05 than the raw curves, respectively; Curves B4 and B5 have been shifted down 0.05 and 0.1 than the raw curves, respectively. (c) Strain-hardening exponent n and (d) strain-hardening coefficients k measured at a logarithmic plastic strain range from about 0.01 up to instability, as function of strain rate for the as-cast alloys.

- 29 -

400 b1

350

C1

True stress (MPa)

300

b2

b5

a2 a3 a4 a5

150

b2

C4 C5

100

0.12

b4 b5

0.09 0.06 0.03

Strain rate of 6.67 10

a3

120

150

a5

a4

a1

-0.03

-4

0 0.00

0.03

0.06

0.09

0.12

0.15

0.18

0.21

0.24

Strain rate of 6.67 10

-4

-0.06

0.27

0

30

Logarithmic strain

60

90

1.5

300

True stress (MPa)

0.6 0.3 0.0 -0.3

60

90

120

150

180

210

Test temperature (Degree celsius)

C3

200

C4

150

C5

100

0 0.00

240

C2

250

50

-4

-0.6 30

240

20C 120C 150C 185C 220C

C1

350

0.9

0

210

d 400

t / f m / f

Strain rate of 6.67 10

180

Test temperature (Degree celsius)

1.2

The ratio

a2

0.00

50

c

b3

b1

b4

a1

Corresponding to transition between the two regions Corresponding to maximum true stress

0.15

C3

200

0.21 0.18

C2

b3

250

b

20C 120C 150C 185C 220C

Logarithmic strain

a

Strain rate of 6.67 10

0.04

0.08

-4

0.12

0.16

0.20

0.24

Logarithmic plastic strain

Fig. 13. (a) Effect of test temperature on true stress-logarithmic strain curves of the as-cast Al-5Mg-0.6Mn alloy (C2-120°C, C3-150°C, C4-185°C and C5-220°C). (b) Logarithmic strain corresponding to the transition site between the two regions and the maximum true stress. (c) The ratio of strain at the transition site (εt) to fracture strain (εf) and the ratio of strain at maximum flow stress (εm) to fracture strain (εf). (d) True stress-logarithmic plastic strain curves showing the effect of test temperature on flow behavior of the as-cast alloy. The curve of the samples tested at room temperature (C1-20°C) is included for comparison.

- 30 -

a

12 -4

Strain rate of 1.67  10 s

b

-1

2

R = 0.9937

0.157

= 465

10

Strain hardening rate (GPa)

2.6

2.5

instability

8

2

R = 0.9981

Log (true stress)

0.120

6 4 2

2.4

0.099

60

90

120

150

180

210

240

2

2.2

2.0 -2.0

270

= 252

R = 0.9981

2.1

30

0.066

2.3

-2

0

2

R = 0.9992

= 322

0

-4 -30

= 379

-1.8

Test temperature (Degree celsius)

c

-4

d

-1

Strain hardening coefficients (MPa)

Strain hardening exponet (n)

0.24 0.20 0.16 0.12 0.08 0.04 0.00 -0.04 -30

Logarithmic plastic strain range from 0.01 up to instability

0

30

60

90

120

150

180

210

= 222 2

R = 0.9971

-1.6

-1.4

-1.2

-1.0

-0.8

-0.6

Log (Logarithmic plastic strain)

0.28 Strain rate of 1.67  10 s

0.043

20C 120C 150C 185C 220C

240

800 -4

600 500 400 300 200 100 0 -100 -30

270

-1

Strain rate of 1.67  10 s

700

Logarithmic plastic strain range from 0.01 up to instability

0

30

60

90

120

150

180

210

240

270

Test temperature (Degree celsius)

Test temperature (Degree celsius)

Fig. 14. (a) Strain-hardening rate measured at a plastic strain of 0.0015, as function of test temperatures for the as-cast Al-5Mg-0.6Mn alloy. (b) Determination of n and K values for the alloys tested at different test temperatures by linear fit to the log true stress-log logarithmic plastic strain curves. The logarithmic plastic strain ranges from about 0.01 up to instability. (c) Strain-hardening exponent n and (d) strain-hardening coefficients k measured at a logarithmic plastic strain range from about 0.01 up to instability, as function of test temperatures for the alloys.

- 31 -

0.3 -4

-1

Strain rate of 1.67  10 s Logarithmic plastic strain range from 0.01 up to instability

0.2

0.1

0.0

-0.1

-0.2

0.01-0.015 0.02-0.025 0.04-0.045 0.08-0.085 0.12-0.125

-0.3 -50

0

50

100

150

200

250

b

700

Strain hardening coefficients (MPa)

Strain hardening exponet (n)

a

600

-4

500 400 300 200 100 0 -50

300

Test temperature (Degree celsius)

-1

Strain rate of 1.67  10 s Logarithmic plastic strain range from 0.01 up to instability

0.01-0.015 0.02-0.025 0.04-0.045 0.08-0.085 0.12-0.125

0

50

100

150

200

Test temperature (Degree celsius)

250

300

Fig. 15. (a) Strain-hardening exponent n and (b) strength coefficient k (at different logarithmic plastic strain ranges) of the as-cast Al-5Mg-0.6Mn alloy, as a function of different test temperature. 4500

True stress of hardening rate (MPa)

4000

b

True stress (B1) Hardening rate (B1): n= 0.152, K= 485 True stress (B2) Hardening rate (B2): n= 0.163, K= 490 True stress (B3) Hardening rate (B3): n= 0.161, K= 490 True stress (B4) Hardening rate (B4): n= 0.155, K= 483 True stress (B5) Hardening rate (B5): n= 0.158, K= 478

dd

3500 3000 2500 2000

True stress of hardening rate (MPa)

a

1500 1000

k

500

n

0 -500 -0.05

4500 4000

True stress (C1) Hardening rate (C1): n= 0.157, K= 465 True stress (C2) Hardening rate (C2): n= 0.120, K= 379 True stress (C3) Hardening rate (C3): n= 0.099, K= 322 True stress (C4) Hardening rate (C4): n= 0.066, K= 252 True stress (C5) Hardening rate (C5): n= 0.043, K= 222

dd

3500 3000 2500 2000 1500 1000

kn

500 0

0.00

0.05

0.10

0.15

0.20

-500 -0.05

0.25

0.00

0.05

0.10

0.15

0.20

0.25

Logarithmic plastic strain

Logarithmic plastic strain

Fig. 16. Tensile instability plots for the as-cast Al-5Mg-0.6Mn alloy tested at different strain rate and test temperature: (a) B1-1.67 × 10-4, B2-3.33 × 10-4, B3-6.67 × 10-4, B4-1.33 × 10-3 and B5-2.67 × 10-3; (b) C1-20°C, C2-125°C, C3-150°C, C3-185°C and C4-220°C.

- 32 -

3.0

as-cast

2.5

b

o

Test temperature of 20 C

-4

Strain rate of 1.67  10 s

2.5

2.0

2.0

1.5

if

if

a

1.5

1.0

1.0

0.5

0.5

0.0

0.0

0.0

-4

5.0x10

-3

1.0x10

-3

1.5x10

-3

2.0x10

-3

2.5x10

-0.5 -30

-3

3.0x10

-1

Strain rate (s )

-1

0

30

60

90

120

150

180

210

240

270

Test temperature (Degree celsius)

Fig. 17. The ratio of tensile instability strain εi to fracture strain εf of the as-cast Al-5Mg-0.6Mn alloy tested at different strain rates (a) and test temperatures (b). εi /εf > 1 indicates that local fracture governs tensile instability, while εi /εf < 1 indicates more uniform damage accumulation where tensile instability occurs at the onset macroscopic necking.

Fig. 18. Macroscpic tensile fracture features of the as-cast Al-5Mg-0.6Mn alloy tested at different temperatures (C1-20°C and C4-220°C).

- 33 -