Professional Documents
Culture Documents
J Msea 2016 07 009
J Msea 2016 07 009
www.elsevier.com/locate/msea
PII: S0921-5093(16)30762-6
DOI: http://dx.doi.org/10.1016/j.msea.2016.07.009
Reference: MSA33840
To appear in: Materials Science & Engineering A
Received date: 7 January 2016
Revised date: 26 April 2016
Accepted date: 3 July 2016
Cite this article as: X. Li, S.M. Xiong and Z. Guo, Failure behavior of high
pressure die casting AZ91D magnesium alloy, Materials Science & Engineering
A, http://dx.doi.org/10.1016/j.msea.2016.07.009
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.
Failure behavior of high pressure die casting AZ91D magnesium alloy
a
School of Materials Science and Engineering, Tsinghua University, Beijing 100084, China
b
Key Laboratory for Advanced Materials Processing Technology, Ministry of Education,
*
Corresponding author: zhipeng_guo@mail.tsinghua.edu.cn
Abstract
The failure behavior of high pressure die casting AZ91D magnesium alloy during both tensile
and fatigue tests was studied in situ by using scanning electron microscope. Attention was
focused on the role of microstructure played in crack initiation and propagation. Results
showed that the defects in castings, including gas pore, shrinkage pore and defect band, were
the crack initiation sources. In tensile test, the crack propagated in a combination of
intergranular and transgranular modes, and the specimen fractured by connecting defects at
the section with minimum effective force bearing area. In fatigue test, the crack propagated in
a transgranular mode at specific crystalline planes. When the crack was in contact with the
β-phase, the crack would pass through, and fracture the network β-phase, whereas bypass the
island β-phase by detaching it from the surrounding α-Mg grains. Besides, defects in front of
the crack would act as the secondary crack initiation sources, from which new cracks would
initiate and propagate. With the propagation of the fatigue crack, the actual maximum cyclic
1
stress would increase to the fracture stress of the left cross section and lead to the final
Keywords: High pressure die casting, AZ91D magnesium alloy, Failure behavior, Crack
1 Introduction
Magnesium alloys are very attractive for industrial applications because of their light
weighting, energy saving, and environment-friendliness [1, 2]. As one of the most popular
casting methods for processing magnesium alloys, the high pressure die casting (HPDC)
process has various advantages including faster prototyping, better casting dimensional
Because porosity is one of the main defects in die castings, studies have been conducted to
reveal the correlation between porosity and mechanical properties. Weiler et al [3]
investigated the relationship between internal porosity and fracture strength of die-cast
AM60B magnesium alloy by using the X-ray tomography, and concluded that the local area
fraction of porosity was the primary factor in determining the tensile properties of specimens.
The variability in tensile ductility of HPDC AM50 and AE44 magnesium alloy was studied
by Lee et al [4, 5], and the results showed that the ductility was correlated to the area fraction
of porosity measured in the fracture surface, not the average volume fraction of porosity in
deformation of HPDC AM50 alloy to study the changes of micro-voids and β phase during
tensile deformation, and suggested that the fracture tended to occur at larger micro-voids or in
2
the cluster micro-voids area. On the other hand, Weiler and Wood [8] used an analytical
tensile failure model to predict the tensile ductility of die cast magnesium alloy. By
employing the size and location of the porosity in fracture surfaces of die-cast AZ91 and
AE44 specimens, the tensile elongation was determined with an average error of 5.3%. The
effect of porosity in HPDC magnesium alloy on the ductility was studied by Sun et al [9] via
that for the regions with lower pore size and lower volume fraction, the ductility generally
decreased as the pore size and pore volume fraction increase, whereas, for the regions with
larger pore size and larger pore volume fraction, other factors such as the mean distance
Though porosity was proved to be the key factor in determining the mechanical properties in
many studies, there were still studies showed that the mechanical properties were influenced
by the whole microstructure not only the porosity. Leo Prakash et al [10] studied the effect of
position and section thickness on the tensile properties in HPDC magnesium alloy, and
suggested that the tensile properties were influenced by the average size, area fraction and
clustering tendency of pores and β phase particles as well as average grain size. Mechanism
for fatigue crack growth in HPDC magnesium alloys was studied by El Kadiri et al [11], and
results showed that these fatigue micromechanics were manifested by the concomitant effects
of casting pores, interdendritic Al-rich solid solution layer, β-phase particles, Mn-rich
inclusions, rare earth-rich intermetallics, dendrite cell size, and surface segregation
phenomena. A skin layer in super vacuum die casting which contained 6%-12% more eutectic
phase and was ~14% harder than those of the HPDC material was considered to be
3
responsible for the improved bending fatigue and corrosion properties of a Mg-Al-Mn
alloy[12]. To improve the mechanical properties of die castings, some new developed HPDC
suggested that the improved mechanical properties could be attributed to the fine and uniform
It is clear that the mechanical properties of die castings are influenced by the microstructure,
including porosity, β phase, grain size et al, and current attention has been focused on the
correlation between the primary factor (e.g. porosity) and mechanical properties, and the
methods to improve the mechanical properties. However, very limited studies have been
performed to study the changes of the microstructure during failure process in HPDC
magnesium alloy, and the related roles of typical structures played in determining the
In this study, the changes of different microstructure features during the crack initiation and
propagation in HPDC AZ91D magnesium alloy were studied under unidirectional and cyclic
loads, based on which roles of these structures played in determination of the mechanical
2 Experimental
During the experiment, a specific casting (see Fig. 1a) was produced by a TOYO BD–350V5
cold chamber die casting machine with AZ91D magnesium alloy. Processing parameters
adopted during HPDC are listed in Table 1. Samples for further analysis were extracted at
location A, as shown in Fig. 1a. To observe the crack initiation and propagation during the
4
tensile and fatigue tests, in situ observation was conducted in the scanning electron
microscope (SEM) chamber on the tensile specimens (Fig. 1b) and axial fatigue specimens
(Fig. 1c). There were many stages with different stress levels and numbers of cycling, at
which the tensile test and fatigue test was paused for the SEM observation of the changes of
different structures. Both the fatigue and tensile tests were conducted under ambient
temperature (20 oC). Tensile tests were carried out with displacement control at a deformation
speed of 0.1 mm/min. The axial fatigue tests were performed with a strain ratio of R = 0.1, a
maximum cyclic stress of Smax = 170MPa, and a constant frequency of f = 10 Hz. Three
tensile specimens (Fig. 1b) and three fatigue specimens (Fig. 1c) were tested. Properties
including ultimate tensile strength (UTS), yield strength (YS), elongation (EI), fatigue life (N)
and fatigue crack growth rate (FCGR) were evaluated. The FCGR was calculated by the
L2 L1
FCGR (1)
N 2 N1
Where L1 and L2 are the crack length at the stages of the crack initiation and the end of crack
propagation, respectively. Accordingly, the N1 and N2 are the numbers of cycling during the
For metallography observation of the microstructure, the specimens were sectioned, mounted,
and polished directly from the in situ specimen, then etched with a diluted acetic acid solution
of 50 ml distilled water, 150 ml anhydrous ethyl alcohol and 1 ml glacial acetic acid to reveal
the microstructure. The microstructure of the in situ test specimens was observed with a
ZEISS scope A1 optical microscope (OM) and a Hitachi S-4500 SEM. The fracture surfaces
of the in situ test specimens were examined using a Hitachi S-4500 SEM to identify the crack
5
initiation sites and fractography. To observe the size and morphology of the different
microstructure features, the diameter (d) and roundness (R) were analyzed by employing a
software named MIAPS (Micro-image Analysis&Process system), which were defined by the
d 4 A (2)
R P 2 4 A (3)
Where A and P are the actual area and perimeter of the target, respectively.
3.1 Microstructure
Fig. 2 shows the typical microstructure of the HPDC AZ91D magnesium alloy, which was
characterized by externally solidified crystals (ESCs), α-Mg grains, porosity (gas pore and
shrinkage pore), defect band and β-phase (island and network). The quantitative
characterization of these microstructure features is listed in Table 2. The HPDC process was
composed of pouring, slow shot stage, fast shot stage, and intensification pressurization. Due
to the fast heat transfer [22], the solidification started immediately when the melt was poured
into the shot sleeve. Crystals nucleated, grew and subsequently the so-called ESCs were
created during the following slow shot stage. In the fast shot stage, the melt with ESCs was
injected into the die cavity. Some ESCs were broken or remelted, and the rest ESCs
continued to grow with the solidification of melt in the die cavity, forming large (~16.30 μm)
and dendritic (R=~5.43) ESCs (Figs. 2a and 2c). Due to much higher cooling rate in die
cavity, large number of α-Mg grains nucleated in die cavity, and had very little time to grow
6
before the end of solidification, leading to smaller (~5.32 μm) and rounder (R=~3.05) α-Mg
grains (Fig. 2e). Similar to [23], there were two types of β-phase, including island β-phase
and network β-phase (Fig. 2e). The island β-phase was small in size (~1.96 μm) and round in
morphology (R=~3.57), while the network β-phase was large in size (~4.7 μm) and complex
in morphology (R=~10.21).
Because of the fast filling speed and high cooling rate in HPDC process, defects including
gas pores (gas entrapment) and shrinkage pores (solidification contraction) were inevitable
(Figs. 2a, 2b and 2c). On the other hand, the air in gas pores was an efficient heat-insulating
medium, the local solidification rate could be lowered, subsequently the shrinkage formed
and usually connected to the gas pore [24]. Thus, there were always shrinkage pores around
the gas pores (Fig. 2b). Similar to the study conducted by Gourlay et al [25], the defect band
of positive porosity commonly following the surface contour of castings could be observed in
Figs. 2a and 2d. By comparing the gas pores, shrinkage pores and defect band in Fig. 2 and
Table 2, it can be concluded that the shrinkage pores were the most irregular (R=~3.59), and
had the largest size (28.64 μm). As for the gas pores and porosity in defect band, the average
diameter of the gas pores was the largest (13.01 μm), and the average roundness of the
aggregated porosity in defect band (Fig. 2d) was the smallest (R=~1.80).
Results of the tensile tests are listed in Table 3, and the stress-strain curves of the three tensile
tests are shown in Fig. 3. Fig. 4 shows the changes of typical gas pore, shrinkage pore, defect
band and β-phase in specimens during tensile deformation. It can be observed that the crack
7
initiated at the gas pore, shrinkage pore, and defect band, and propagated along the grain
boundaries. As shown in Figs. 4c, 4g and 4k, the fracture morphology of the gas pore,
shrinkage pore and defect band was rough. A further observation of the zoom-in area at these
crack initiation sites showed that the local fracture was torn because of the complex
In a finite plate with a center crack, the stress intensity factor K I depends on the plate width
2W, crack length 2a, and applied tensile stress , and can be calculated by the empirical
KI Y a (4)
whereby
a a a
Y 1 0.256( ) 1.152( ) 2 12.2( )3 (5)
W W W
Assuming that the gas pores, shrinkage pores and defect band are the crack sources during
tensile deformation, the diameters of different porosities are the crack length, and the
specimen is simplified as a finite plate with some cracks. The stress intensity factors of the
different porosities were calculated and listed in Table 4. It can be observed that the largest
and average K I of different porosities were very close, thus, the gas pores, shrinkage pores
and defect band all would act as the crack initiation sources to initiate during the tensile
deformation.
Because the gas pore, shrinkage pore and defect band all had complex morphology, there
were many tip regions on the porosities, which were mostly not in a plane. Under the tensile
stress, the tip regions of these defects would cause large stress concentration. If the local
stress exceeded the strength limit of the material around, crack would initiate and propagate
8
in different directions at different planes (Figs. 4b, 4f, 4j and 4c, 4g, 4k), leading to the rough
fracture morphology[27]. On the other hand, the β-phase hardly changed during the tensile
transgranular modes along the direction roughly perpendicular to the tensile stress. Thus, the
crack would propagate roughly in a plane (Fig. 4o), leading to much flatter fracture
morphology. A further observation of the zoom-in area at the β-phase showed that there were
dimples on the fracture surface. The elastic modulus of the β-phase and ESCs or α-Mg grains
was much different, and the different strain levels would be induced by the same stress at
grain boundaries[7], leading to the formation of the plastic void between β-phase and ESCs
(or α-Mg grains). The plastic void would get expanded when in contact with the crack,
The crack propagation in specimen during tensile deformation is shown in Fig. 5. In Fig. 5a,
it can be observed that there was some porosity on the crack before fracture, and the crack
propagated roughly along the direction perpendicular to the tensile direction. As shown in Fig.
5b, further investigation revealed that the crack propagated by connecting the porosity in
front. At locations without porosity, the crack propagated in the transgranular mode, leading
to a flat fracture morphology. However, when the crack encountered with the shrinkage pore,
the crack would propagate along the grain boundaries in a intergranular mode, leading to
According to the study of the tensile test, it can be concluded that the specimen would
fracture by connecting the porosity at the section with minimum effective force bearing area
during the tensile deformation. The defects in die castings, including gas pore, shrinkage pore,
9
and defect band would act as the crack initiation sources to initiate. During crack propagation,
the intergranular mode and transgranular mode would occur at the locations with and without
porosity, respectively.
Fig. 6 shows the typical crack initiation sites of specimens during the fatigue test. Similar to
the tensile test, the crack initiated at the gas pore, shrinkage pore and defect band. In fatigue
test, the stress applied to the specimen was not uniformly distributed in the cross section, and
there would be the stress concentration phenomenon around the defect tip. Besides the
maximum cyclic stress Smax (170MPa) was larger than the yield stress y (~ 160MPa [21]),
during continuous cycling, plastic strain would accumulate at the defect tip, leading to the
permanent damage with the persistent slip bands in α-Mg grains, and consequently the crack
would initiate.
The crack propagation during fatigue test is shown in Fig. 7. It can be observed from Fig. 7a
that the crack propagated roughly along the direction perpendicular to the applied stress. A
further observation of the zoom-in area of the crack showed that the crack propagated in a
transgranular mode, i.e. through the specific crystalline plane, as shown in Figs. 7b and 7c.
According to [7, 28], the main deformation mechanisms in hexagonal Mg were deformation
twinning and dislocation slip. At room temperature, there were only the primary basal slip of
{1000} , deformation twinning of {1012} and double twinning of {1011} {1012} in tension
fatigue test. Prismatic slip was activated, and substantial cyclic hardening was also observed
above the fatigue limit, which was considered to be caused by cross-slipping of prismatic
10
dislocations as well as interactions between prismatic dislocations and {1012} twins. As the
cyclic hardening progressed, double twinning of the {1011} {1012} type occurred.
Localized deformation in the double twins and the associated strain incompatibility would
lead to the crack initiation and propagation. In this fatigue test, the actual stress at the crack
tip was larger than the yield stress of ESCs and α-Mg grains due to the stress concentration.
Under the cyclic stress, the strains induced by the actual stress would accumulate at the crack
tip, which would cause the formation of slip band and twinning at the specific crystalline
Fig. 7d shows the changes of β-phase during crack propagation. During crack propagation,
the crack would pass through and fracture the network β-phase, whereas it would bypass the
island β-phase by detaching it from the surrounding α-Mg grains. This was because the
β-phase was hard and brittle, and difficult to be plastically deformed [29]. When the crack
encountered the island β-phase, because the island β-phase was small and round in
morphology, the strains induced by the stress would accumulate at the boudaries between
island β-phase and ESCs or α-Mg grains [7], not in the island β-phase, leading to the
detachment of island β-phase from ESCs or α-Mg grains. On the other hand, when the crack
was in contact with the network β-phase, the crack tip was encircled by network β-phase, the
strains induced by the stress would accumulate in the network β-phase, then the network
β-phase was plastically deformed, leading to brittle fracture of network β-phase at the crack
tip. Afterwards, when the propagation of crack was hindered by the large network β-phase,
the defects in front of the crack would act as the secondary crack initiation source, from
which new cracks would initiate. Detail of this behavior is shown in Figs 7e and 7f. With
11
further cycling, the propagation of the secondary crack would play a dominate role, and the
actual maximum cyclic stress would increase due to the propagation of crack. When the
actual maximum cyclic stress reached the fracture stress of the specimen, the specimen would
fracture instantaneously.
Fig. 8 shows typical fracture morphology of a fatigue specimen. The fatigue fracture
comprised three typical zones, including the crack initiation site (Fig. 8b, I-zone), crack
propagation zone (Fig. 8c, II-zone) and instantaneous fracture zone (Fig. 8d, III-zone).
Results of the fatigue tests are listed in Table 5, which showed that the crack normally
initiated at the gas pore or shrinkage pore, which was similar to [30]. The fatigue life (N) and
fatigue crack growth rate (FCGR) were shown to be inversely correlated to the area fraction
of the I-zone and II-zone (SI+II/ SI+II+III). Fig. 8c shows the striation existed in the crack
propagation zone. In the fatigue test, the crack propagation would be hindered if the local
stress was lower than certain magnitudes (yeild strength of the microstructure at crack tip),
e.g. during the relaxation stage of the fatigue cycle. Whereas, the plastic deformation would
occur at the crack tip and the crack would continue to grow if the local cyclic stress was
maximized. The cyclic hindering and growth of the crack subsequently disturbed the fracture
surface, and led to the formation of striation in the crack propagation zone. In the
instantaneous fracture zone as shown in Fig. 8d, the fracture morphology comprised porosity
and dimples, which was similar to that of the tensile specimen. When the local stress
exceeded certain magnitude (strength limit of the rest cross section), the fracture would occur,
12
3.4 Comparison of the tensile test and fatigue test
Fig. 9 shows the comparison of the crack initiations in tensile test and fatigue test. In tensile
test, the crack initiated at the grain boundaries around porosity in an intergranular mode,
whereas in fatigue test, the crack initiated at the specific crystalline plane around porosity in a
transgranular mode. The primary difference between the tensile test and fatigue test was the
loading mode. During tensile deformation, the stress gradually increased until the final
fracture. The porosity with complex morphology, possibly located at the ESC boundaries,
would cause the stress concentration. With the tensile stress increased, the concentrated stress
would reach or exceed the bonding force between ESCs and other phases, leading to crack
initiation at the grain boundaries around the porosity. However, during the fatigue test, the
maximum stress was lower than the bonding force between ESCs and other phases, and the
crack would not initiate at the grain boundaries. On the other hand, the stress around or/and at
the tip of the defects would be much larger than the yield stress due to the stress
concentration, leading to the plastic deformation of ESCs or α-Mg grains. Under cyclic
plastic deformation, persistent slip bands and twins would form at specific crystalline planes
of both ESCs and α-Mg grains, and subsequently lead to crack initiation.
The fracture morphology of the tensile specimen and fatigue specimen is shown in Fig. 10. It
can be observed that the fracture morphology of the tensile specimen was rough. While in the
fatigue specimen, the fracture morphology in the crack propagation zone was very flat,
whereas it was rough in the instantaneous fracture zone. the difference of the fracture
morphology can be attributed to the different crack propagation modes during the tensile test
13
and fatigue test. In tensile test, the crack would propagate in a combination of intergranular
and transgranular modes, and the tensile specimen would fracture at the cross section with
minimum force bearing area [27]. Because of the non-uniform distribution of the defects in
specimen, the rough fracture morphology was created. In fatigue test, the crack propagated in
a transgranular mode along the direction roughly perpendicular to the stress in the crack
propagation zone, leading to the flat fracture morphology. While in the instantaneous fracture
zone, similar to the tensile test, the crack propagated in a combination of intergranular and
transgranular modes by connecting the defects in the rest cross section with minimum
4 Conclusions
In this paper, the failure behavior during the tensile test and fatigue test of the HPDC AZ91D
magnesium alloy was investigated. According to the related results, the following conclusions
can be drawn:
(1) During the tensile test and fatigue test, the defects in specimen, including gas pore,
shrinkage pore and defect band, were shown to be the crack initiation sources.
(2) In tensile test, the specimen fractured by connecting the defects at the section with
intergranular and transgranular modes. At locations where defects were present, the crack
would propagate along the grain boundaries of ESCs, whereas the crack would propagate
along the direction roughly perpendicular to the tensile stress if no defects were present.
(3) In fatigue test, the crack propagated in a transgranular mode at the specific crystalline
14
planes in crack propagation zone. In early stage, when the crack was in contact with the
β-phase, the crack would pass through, and fracture the network β-phase, whereas bypass the
island β-phase by detaching it from the surrounding α-Mg grains. In the later stage, when the
propagation of the main crack was hindered by large network β-phase, the defects in front of
the main crack would act as the secondary crack initiation sources to initiate. With the
propagation of the fatigue crack, the actual maximum cyclic stress would increase to the
fracture stress of the left cross section, and lead to the final fracture of the specimen.
Acknowledgements
The authors would like to thank the National Natural Science Foundation of China (No.
20121087918), and the National Science and Technology Major Project of the Ministry of
References
[1] H. Friedrich, S.J. Schumann. Research for a new age of magnesium in the automotive
between internal porosity and fracture strength of die-cast magnesium AM60B alloy.
15
[4] S. G. Lee, G. R. Patel, A. M. Gokhale, A. Sreeranganathan, M. F. Horstemeyer. Variability
in the tensile ductility of high-pressure die-cast AM50 Mg-alloy. Scr. Mater. 2005; 53:
851-856.
[6] J. Song, S. M. Xiong, M. Li, J. Allison. The correlation between microstructure and
mechanical properties of high-pressure die-cast AM50 alloy. J. Alloys Compd. 2009; 477:
863-869.
[8] J. P. Weiler, J. T. Wood. Modeling the tensile failure of cast magnesium alloys. J. Alloys
[9] X. Sun, K. S. Choi, D. S. Li. Predicting the influence of pore characteristics on ductility
of thin-walled high pressure die casting magnesium. Mater. Sci. Eng. A. 2013; 572:
45-55.
[10] D. G. Leo Prakash, D. Regener. Micro-macro interactions and effect of section thickness
39: 190-205.
16
[12] W. Wei, A. A. Luo, Z. Tongguang, J. Yan, C. Yang-Tse, I. Hoffmann. Improved bending
fatigue and corrosion properties of a Mg-Al-Mn alloy by super vacuum die casting. Scr.
[13] S. Ji, Z. Zhen, Z. Fan. Effects of rheo-die casting process on the microstructure and
die-cast alloy by intensive shearing. Mater. Sci. Eng. A. 2010; 527: 2929-2934.
properties and microstructure of AZ91D alloy motorcycle wheels formed by die casting
behavior of a super-vacuum die cast magnesium alloy. Mater. Sci. Eng. A. 2012; 546:
72-81.
engineered in UTS and elongation of aluminum alloy high pressure die castings through
the alteration of runner geometry and plunger velocity. Mater. Sci. Eng. A. 2013; 559:
276-286.
[19] S. Ji, W. Yang, B. Jiang, J. B. Patel, Z. Fan. Weibull statistical analysis of the effect of
melt conditioning on the mechanical properties of AM60 alloy. Mater. Sci. Eng. A. 2013;
17
566: 119-125.
[20] W. Li, H. Zuqi, W. Shusen, L. Xueqiang. Mechanical properties and fatigue behavior of
vacuum-assist die cast AlMgSiMn alloy. Mater. Sci. Eng., A. 2013; 576: 252-258.
[22] R. Helenius, O. Lohne, L. Arnberg, H.I. Laukli. The heat transfer during filling of a
high-pressure die-casting shot sleeve. Mater. Sci. Eng. A. 2005; 413-414: 52-55.
[23] M. Wu, S. Xiong. Microstructure characteristics of the eutectics of die cast AM60B
[25] C. M. Gourlay, H. I. Laukli, A. K. Dahle. Defect band characteristics in Mg-Al and Al-Si
[26] W. Brown,J. Srawley, Plane strain crack toughness testing of high strength metallic
[27] X. Li, S.M. Xiong, Z. Guo. Correlation between Porosity and Fracture Mechanism in
High Pressure Die Casting of AM60B Alloy. J. Mater. Sci. Technol. 2016; 32:54-61.
dislocation slip in the fatigue failure mechanism of AZ31 Mg alloys. Scr. Mater. 2010; 63:
747-750.
modeling of fatigue crack growth mechanisms in a die-cast AM50 magnesium alloy. Acta
18
Mater. 2006; 54: 5061-5076.
Behavior of Die-Cast Mg Alloys AZ91 and AM60. Metall. Mater. Trans. A. 2012; 43A:
2260-2274.
Figure 1 Configuration of (a) the specific casting including three tensile test bars (diameter at
the center was 6.4mm) and one plate sample (thickness was 2.5mm), (b) tensile specimen, (c)
microstructure in OM and (b) gas pore, (c) shrinkage pore and ESCs, (d) defect band, (e)
Figure 4 Morphology of the crack and fracture during tensile deformation. From top to
bottom, the four rows were related to gas pore, shrinkage pore, defect band and β-phase,
respectively. The first and second column corresponded to an applied stress level of 0 MPa
and 180 MPa, respectively. The third and fourth column shows the overall and zoom–in
Figure 5 Propagation of crack during tensile deformation, showing (a) before fracture, (b)
zoom-in area of (a), (c) after fracture and (d) zoom-in area of (c).
Figure 6 Crack initiation sites in fatigue tests. (a) and (d) gas pore with shrinkage pore, (b)
Figure 7 Propagation of crack during fatigue test in specimen, showing (a) overall, (b)
19
zoom-in area of (a), (c) zoom-in area of the crack front, (d) β-phase in the crack, (e)
secondary cracks in the crack front and (f) the next step of (e).
Figure 8 Fracture morphology of a fatigue specimen, (a) overall, (b) crack initiation site, (c)
crack propagation zone with striation and (d) instantaneous fracture zone with porosity and
dimple.
Figure 9 Comparison of the crack initiation in (a) and (b) tensile test and (c) and (d) fatigue
test.
Figure 10 Comparison of the fracture morphology in (a) and (b) tensile specimen and (c) and
Intensification
Melting Initial mold Slow shot Fast shot
casting pressure
temperature (C) temperature (C) speed (m/s) speed (m/s)
(MPa)
Figures 1, 2, 4-10 were created in Microsoft office Visio 2007, Figure 3 was plotted in OriginPro9.0.
21
(a) (b)
A
(c)
100mm
Figure 1
40μm 40μm
(d) (e)
22
225
Sample-1
200 Sample-2
175
Sample-3
150
Stress (MPa)
125
100
75
50
25
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
Strain (%)
Figure 3
Figure 4
23
(a) (b)
100μm 25μm
(c) (d)
50μm 25μm
Figure 5
Figure 6
24
(a) (b)
100μm 50μm
(c) (d)
20μm 25μm
(e) (f)
40μm 40μm
Figure 7
25
(a) (b)
500μm 100μm
(c) (d)
50μm 50μm
Figure 8
(a) (b)
50μm 100μm
(c) (d)
25μm 25μm
Figure 9
26
(a) (b)
500μm 500μm
(c) (d)
500μm 500μm
Figure 10
27