Materials Science and Engineering A 456 (2007) 4351

Effect of composition on the microstructure and

mechanical properties of MgZnAl alloys
Jing Zhang a,b, , Z.X. Guo b , Fusheng Pan a , Zhongsheng Li a , Xiaodong Luo a
a

College of Materials Science and Engineering, Chongqing University, Chongqing 400044, PR China
Department of Materials, Queen Mary, University of London, Mile End Road, London E1 4NS, UK

Received 25 June 2006; received in revised form 16 November 2006; accepted 17 November 2006

Abstract
Magnesium is receiving great attention for transport applications, particularly its cast alloys. This investigation focuses on the as-cast microstructure and mechanical properties of permanent-mould cast MgZnAl alloys with typical compositions within the high zinc castable domain. Three
types of alloys were identified and characterized by Mg32 (Al, Zn)49 , also known as the phase; MgZn phase, also known as the phase; and
a ternary icosahedral quasi-crystalline phase, denoted as the Q phase, respectively. A schematic phase diagram is proposed to show the change
of microstructral constituents with element content and the Zn/Al ratio. The diagram reveals that the microstructral constituent is dominated by
both the content of Zn or Al and the Zn/Al mass ratio; alloys with a high Zn/Al ratio and a low Al content fall into the -type; alloys with an
intermediate Zn/Al ratio and an intermediate Al content favour the -type; and those with a low Zn/Al ratio and a high Al are dominated by the
icosahedral quasi-crystalline phase. No Mg17 Al12 () phase was found in those ZA series alloys. The solidification process and its effects on the
phase constituents were discussed. Preliminary mechanical property testing showed that all the experimental alloys possess comparable ultimate
strength and yield strength with the AZ91 alloy at ambient temperature, but show far superior creep resistance at elevated temperatures. Moreover,
while ambient-temperature properties solely depend on the total element contents, the - and the Q-type alloys show greater potential than the
-type alloys on the improvement of elevated temperature properties.
2006 Elsevier B.V. All rights reserved.
Keywords: Mg alloys; As-cast microstructure; Casting; Intermetallic compounds; Creep; Mechanical properties

1. Introduction
Magnesium alloys are attracting increasing attention for
transport applications in the automotive and aerospace industry
for weight reduction and higher fuel efficiency. The most popularly used magnesium alloys are those based essentially on the
MgAl system, such as AZ91D, AM60B and AM50A. Although
offering a good combination of mechanical properties, corrosion
resistance and castability at ambient temperature, these alloys
are all prone to excessive creep deformation when exposed to
even low levels of load at temperatures above 100 C. The poor
elevated-temperature properties of low-cost magnesium alloys
have now become a critical issue for wide-spread applications
of magnesium alloys.

Corresponding author at: College of Materials Science and Engineering,

Chongqing University, Chongqing 400044, PR China. Tel.: +86 23 65111167;
fax: +86 23 65127306.
E-mail address: jingzhang@cqu.edu.cn (J. Zhang).

0921-5093/$ see front matter 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.msea.2006.11.089

In the past few years, a substantial effort has been made to

elucidate the creep behavior of AZ and AM alloys. The prevailing interpretation attributes the poor creep resistance to the
discontinuous precipitation of the -Mg17 Al12 phase at grain
boundaries [1,2]. The general approaches used hitherto have
been to modify the existing MgAl alloys by additions of minor
elements [35] through either formation of new stable precipitates or reduction of the discontinuous precipitation of the
Mg17 Al12 phase in order to suppress grain boundary sliding at
elevated temperature. However, only limited improvements of
creep resistance have been achieved so far. A alternative solution is to design new alloys that do not contain any Mg17 Al12
particles in the microstructure.
The ternary addition of zinc [614] to binary MgAl alloys
has shown complete suppress of the formation of the Mg17 Al12
phase and the resultant MgZnAl alloys (ZA) have yielded
an improved creep resistance, surpassing the AZ and the AM
alloys. The phase constituents in the microstructure of the ZA
alloy, however, is not clear. It is always difficult to predict
microstructure constituents of a certain composition, since the

44

J. Zhang et al. / Materials Science and Engineering A 456 (2007) 4351

Fig. 2. Specimens used for tensile and creep tests (unit: mm).

Fig. 1. Castability of the MgZnAl alloy system [16] and compositions of the
experimental alloys.

Mg-rich MgZnAl ternary phase diagrams are very complex

and are not well established [15]. In addition, phases in alloys
solidified under practical casting conditions are often at variance with those in the equilibrium phase diagram and between
those under different casting conditions as well. Although there
are several investigations on the phases present in high-zinc
alloys, no consensus has been reached. Zhang et al. [12,13]
examined permanent mould cast ZA(10,14)2, ZA(10,14)4 and
ZA(10,14)6 alloys by X-ray diffraction and EDX analysis and
reported that intermetallic phases contained in these alloys
depend solely on the aluminum content: alloys containing 2%
Al consisted of (Mg32 (Al, Zn)49 ) and (MgZn); those of 4%
Al only contained the phase; whereas those of 6% Al was
dominated by the (Mg5 Zn2 Al2 ) with a small amount of .
Bourgeois et al. [6,7] argued that the composition of these casting alloys is in the (-Mg + ) two phase field of the MgZnAl
phase diagram, and there is no evidence of the phase from
transmission electron microscopy of the as-cast microstructure of the Mg8%Zn(46)%Al alloys; Instead, most primary
intermetallic particles were found to be a metastable icosahedral phase, which gradually transformed to equilibrium
phase during prolonged homogenization treatment at 325 C.
Whereas Anyanwu et al. [14] claimed that both and are
observed by X-ray diffraction of the as-cast microstructure
of all the examined MgZnAl alloys with a broad composition range of 614%Zn and 28%Al, and with a total
element content of 13%, a Zn/Al ratio of 2:1, and 0.1% or
0.6% Ca addition. Moreover, little work has been carried out
on the relationship between composition, microstructure, and
mechanical properties of the MgZnAl alloys. Attempts to
improve creep resistance through microstructure modification
by identifying more promising chemical constitution are thus
restricted.
A series of ZA alloys with typical compositions within
the castable domain of high-zinc concentration (Fig. 1) [16]
were developed in this study. The aim of the work is to
identify the general rule of the phase variation with composition of high-zinc magnesium alloys and the relationships
between composition, microstructure and mechanical properties, for improved understanding of the fundamental mechanisms necessary to develop new creep resistant magnesium
alloys.

2. Materials and experimental procedure

Within the wide castable domain of high-zinc concentration,
experimental alloys were selected (see Fig. 1) based on the following rules: the content of zinc is 1% higher than that of the
boundary between the hot cracking area and the castable area in
the high-zinc side; and the content of alloy elements was kept
as low as possible in order to maintain the advantage of weight
reduction. The ZA82 alloy and the ZA74 alloy were thus determined. Four other compositions, ZA102, ZA122, ZA104, and
ZA75 were also selected, either with a fixed Al or Zn content, or a
fixed Zn/Al ratio, to compare the tendency of phase change with
the element content and the ratio. The nominal compositions of
the studied alloys are given in wt.% throughout the text.
Commercial high-purity Mg (>99.9%), Zn (>99.9%) and Al
(>99.8%) were used to prepare the ZA series alloys. An electric
resistance furnace and a steel crucible were used for melting and
alloying operations, during which continuous protection with a
gas mixture of SO2 and CO2 was provided by an on-line protective gas generation apparatus. The melt was homogenized by
mechanical stirring. After complete mixing, the melt was kept at
700 C for 5 min before being poured into a permanent mould at
300 C. Chemical compositions of ingots were inspected by ICP.
Bar samples as shown in Fig. 2 were machined from the castings
for tensile and creep tests. The AZ91 alloy ingots were also cast
and machined into the same dimensions and tested under the
same conditions as the above samples for comparison.
The microstructure of specimens was examined by optical
microscopy and scanning electron microscope (SEM), after the
samples were etched in a 8% nitric acid solution in distilled
water. Detailed structural characterization was conducted on a
Tecnai 20 transmission electron microscope (TEM) operating
at 200 kV. X-ray diffraction (XRD) characterization was conducted using a D5000 Siemens diffractometer equipped with
Cu K radiation to identify the phases present in the experimental alloys. Differential scanning calorimetry (DSC) was carried
out using a NETZSCH STA 449C system. Samples weighted
around 30 mg were heated in a flowing argon atmosphere from
30 to 700 C, and then kept at 700 C for 5 min before being cold
down to 100 C. Both heating and cooling curves were recorded
at a controlled speed of 15 K/min.
Tensile properties were determined from a complete stressstrain curve. 0.2% yield strength (YS), ultimate tensile strength
(UTS) and elongation to failure (elongation) were obtained
based on the average of three tests. Constant-load tensile creep
tests were performed at 150 C and 60 MPa for creep extension

J. Zhang et al. / Materials Science and Engineering A 456 (2007) 4351

45

Fig. 3. Optical microstructures of: (a) ZA82, (b) ZA102, (c) ZA122, (d) ZA104, (e) ZA74, and (f) ZA75 alloys cast in a permanent mould.

up to 100 h, based on the ASTM E139 standard procedure. The

total creep strain was measured from each elongation-time curve
and averaged over three tests.
3. Results and discussion
3.1. Microstructure
The typical as-cast microstructures of all the studied alloys
are shown in Fig. 3. In these alloys, the microstructure
mainly shows dendritic morphology with the secondary phases
distributed in interdendritic spacings and along grain boundaries. Further detailed observation reveals that these primary
intermetallic particles manifest different crystallographic characteristics, as can be seen from Fig. 4. A typical microstructure
of ZA82 alloy is illustrated in Fig. 3(a), only cramp lump com-

pounds are identified in the boundary areas, which is much like a

set of fish skeleton with stings and bone lumps under higher magnification (Fig. 4(a)). XRD pattern of the ZA82 alloy indicates
this alloy is mainly composed of -Mg phase and a small amount
of Mg32 (Al, Zn)49 , also known as phase (Fig. 5(a)), which has
a = 1.416 nm
a body-centered cubic structure (space group Im3,
[17]). In ZA102 alloy, in addition to -Mg phase and phase,
another phase, MgZn phase, also known as phase, was also
detected by XRD in the as-cast microstructure (Fig. 5(b)). SEM
observation reveals the detailed structure of the phase, as shown
in Fig. 4(b), which consists of many approximately paralleled
white laths. Cramp lumps are also found in the boundary areas.
With increasing Zn content, the boundary areas of ZA122 alloy
are dominated by the wide parallel structured eutectic phases,
which likely make the boundary thicker, as obviously seen from
Fig. 3(c). In addition, the dendritic structure becomes some-

46

J. Zhang et al. / Materials Science and Engineering A 456 (2007) 4351

Fig. 4. SEM Micrographs showing different crystallographic characteristics of the primary phases of (a) , (b) , and (c) Q in the experimental alloys, and (d) SAD
pattern recorded from the particles in the ZA75 alloy shown in (c).

what coarser with the increase of zinc content. The apparent

width boundary indicates that the outer and inner regions of
these boundary phases have different etching behaviors, suggesting different composition with each other. XRD analysis
shows similar result about the microstructure of the ZA122 alloy,
which is mainly composed of -Mg, , and a small amount of
phases (Fig. 5(c)). Whereas with increasing Al content, the
parallel structured eutectic phase present in the ZA102 alloy disappears again in ZA104 alloy, whose typical as-cast structure
is manifested in Fig. 3(d), showing solely cramp lump compounds distributing along grain boundaries, similar with that
of alloy ZA82. XRD result also shows the only intermetallic
phase , besides the matrix -Mg phase, presents in ZA104 alloy
(Fig. 5(d)).

Fig. 5. X-ray diffraction patterns of (a) ZA82, (b) ZA102, (c) ZA122, and (d)
ZA104 alloys.

When the Zn/Al concentration ratio moves further to lower

side than that of ZA104 alloy, alloy ZA74 exhibit a completely
different microstructure characteristics, as shown in Fig. 3(e).
Under optical microscope, the microstructure has a kind of necklace morphology, with more or less regular black and white
particles connecting into a chain. Such feature becomes even
more obvious in ZA75 alloy (Fig. 3(f)), when the ratio goes even
lower. Further detailed observation under SEM reveals a cluster
of corn-like phase marked with eyelike spots (Fig. 4(c)), which
might interpret the necklace feature on the metallographical
cross section of the specimen. However, in spite of a large quantity of compounds existing in the specimens, no second phase
was detected besides the -Mg matrix phase in XRD at the same
condition with the other specimens. Using electron diffraction,
Nie et al. [6] concluded that most primary intermetallic particles
were found to be a ternary icosahedral quasi-crystalline phase
in the as-cast microstructure of a similar permanent mould-cast
alloy with the nominal composition of Mg8 wt.%Zn4 wt.%Al,
a non-equilibrium phase so far have been reported only after
extreme non-equilibrium processing in the MgAlZn system
[18], such as melt-spinning and mechanical alloying. Similar
observations concerning the phase identification by TEM were
reported by Vogel et al. [9,10] in ZA85 alloy samples fabricated by die-casting (cooling rate 20 K/S) as well as after
re-melting (cooling rate 1 K/S), further widening the range of
cooling rates within standard casting conditions. However, the
quasi-crystalline grain boundary phase was deemed by Nie et
al. [6] to be a metastable phase with a composition close to that

J. Zhang et al. / Materials Science and Engineering A 456 (2007) 4351

of the equilibrium phase (Mg5 Zn2 Al2 , primitive orthorhombic structure, space group Pbcm, a = 0.8979 nm, b = 1.6988 nm,
c = 1.9340 nm [17]) and to be gradually replaced by the equilibrium phase during thermal exposure at elevated temperature
for extended periods. Whereas Vogel et al. [9,10], showing by
EDX analysis that the chemical composition is in the homogeneity range of the crystalline phase, considered it representing
a metastable modification of the equilibrium phase, the compound that is conventionally recognized as the 1/1 crystalline
approximant of the icosahedral quasi-crystal in the MgAlZn
system (I phase). The selected area diffraction patterns recorded
from the intermetallic grain-boundary particles in the ZA75
alloy in our case, Fig. 4(d), strongly support the grain boundary
quasi-crystalline phases, but its formation condition, chemical
composition, and the relationship between them, as well as its
connection with the equilibrium phase and the phase remain
to be further elucidated.
According to the ternary phase diagram, two ternary intermetallic compounds are known in the MgZnAl system:
the magnesium-rich phase (Mg5 Zn2 Al2 ) and the phase
(Mg32 (Al, Zn)49 ) [19]. While the phase possesses a wide composition range near the center of the ternary phase diagram,
the phase occupies a slit with a narrow variation of the Mg
concentration. The Mg17 Al12 phase () and the MgZn () are
the other binary intermetallic phases in the Mg-rich MgZnAl
alloys. No phases were found in all the experimental alloys,
indicating that the formation of the eutectic Mg17 Al12 phase is
completely suppressed by the high zinc content. X-ray diffraction peaks of these phases were found to be somewhat shifted
from the standard positions, and from each other for the same
phase in different alloys as well, indicating small changes in the
lattice parameters. These differences can be ascribe to the variable concentrations of solutes in the -Mg or that of the ternary
solute in the binary phase MgZn, as well as the wide range of
compositions of the ternary intermetallic phases.
It should be mentioned that although based on the MgZnAl
phase diagram of liquidus projection, is one of the two ternary
intermetallic phases existing in equilibrium in the system, as
shown in Fig. 6 [19], the stability of phase has not been
determined with certainty up to now [15]. It was reported by
Clark [20] that was found to be stable at 335 and 204 C,
it seems reasonable to assume that its stability extends down
to room temperature. Recently, Zhang et al. [13] confirmed its
existence at ambient temperature based on XRD and EDX studies in permanent-mould cast alloys ZA106 and ZA146. It is
worthy mentioning that all the microstructural data obtained by
Zhang et al. were under near-equilibrium with a cooling rate of
0.05 K/s. Since the two alloys are of similar Zn/Al ratios in the
high-Al concentration side to the ZA74 and ZA75 alloys, the
results seem to suggest a relationship between the metastable
quasi-crystal and the equilibrium phase, viz. the quasi-crystal
might be a replacement of the phase under non-equilibrium
solidification conditions.
In summary, three different kinds of primary compounds with
distinct crystallographic morphologies are formed along with the
change of Zn or Al contents, and the Zn/Al concentration ratio
under normal permanent-mould cast conditions. Accordingly,

47

Fig. 6. Liquidus projection of the MgZnAl (Mg rich) ternary phase diagram
[19].

these alloys within the high zinc castable domain can then be distinguished into three groups by a characteristic compound: Type
I, ZA82 alloy and ZA104 alloy, consist of -Mg and fish-bone
like phase; Type II, including ZA102 alloy and ZA122 alloy, is
composed of -Mg, , and the binary eutectic phase, which is a
predominant phase and characterized by its approximately paralleled lath structure; whereas Type III, locating in comparatively
low Zn/Al ratio and high Al content, such as ZA74 alloy and
ZA75 alloy, consists of -Mg with icosahedral quasi-crystalline
phases.
The trend of microstructural constituent variation with the
Zn/Al ratio and Al content are schematically shown in Fig. 7.
It can be seen that the low-zinc side of the high-zinc domain is
roughly partitioned into three regions, marked with the characteristic compounds of each alloy type, with the high Zn/Al and
low Al section corresponding to Type II, intermediate Zn/Al and
Al to Type I, and the low Zn/Al and high Al to Type III. The two

Fig. 7. Schematic diagram of microstructural constituent as a function of the

Zn/Al ratio and the Al content. Note: Boundary 1 and 2 corresponding to the
two boundaries of the hot cracking area in Fig. 1.

48

J. Zhang et al. / Materials Science and Engineering A 456 (2007) 4351

Fig. 8. DSC curves of the alloy ZA104.

curves represent the boundaries of the hot cracking area in Fig. 1.

To verify the effectiveness of the diagram, a new alloy, ZA73,
was cast and subjected to XRD analysis. The results [21] showed
that the primary compounds in the as-cast microstructure are
phases, falling well into the corresponding phase section of the
composition.
3.2. Thermal analysis
To study the crystallization behaviors during solidification,
DSC measurements were carried out under controlled heating
and cooling programs. Fig. 8 shows a typical DSC curve of
the experimental alloys. There are two (sets of) peaks in each
DSC curve, corresponding to the melting of the -Mg matrix
(peak I) and the second phase transformations (peak II), respectively. The equilibrium phase transformation temperature can be

approximately calculated by averaging the two corresponding

temperatures in heating curve and cooling curve, thus offsetting the temperature lags. All the characteristic temperatures
emerging in the DSC curves, including the peak temperatures,
the starting and ending temperatures of the second phase transformations, as well as the calculated equilibrium temperatures,
of all the experimental alloys are listed in Table 1. Combined
with the ternary phase diagram showing the liquidus surface
(Fig. 6), the sequence of the secondary phase transformations
during solidification could be determined, which are also given
in Table 1.
In ZA102 alloy, the second phase reactions occur over a
temperature range from 344 to 316 C. There are three thermal events involved. The first corresponds to a binary eutectic
reaction L1 -Mg + + L2 at a peak temperature of 341 C.
Following this is a ternary quasi-peritectic reaction L2 + Mg + at 334 C. The third peak is very weak, indicating a
very small enthalpy change, from which solid phase transformations, i.e. secondary precipitate reaction from the primary -Mg
solution, took place at around 316 C, and probably included
-Mg II , and -Mg II . The final microstructure of the
alloy is composed of -Mg, and phases. The corresponding endothermal reactions in the ZA102 alloy were not detected
separately as shown in Table 1, largely because of the variable
contact resistance in heating and cooling processes.
Alloy ZA122 showed similar thermal events to ZA102 alloy.
In the case of ZA82 alloy, the amount of eutectic product
should be very small and is expected to be used up in the ternary
quasi-peritectic reaction. It is noteworthy that the peak II temperature increases slightly with the increase of Zn content, being

Table 1
Characteristic temperatures emerging in the DSC curves and possible events of the second phase transformations
Peak I ( C)

Alloy

Second phase transformations

Peak(s) II ( C)

Starting ( C)

Ending ( C)

Events

ZA82

Heating
Cooling
Equilibrium

611.9
602.3
607.1

341.7
324.4
333.1

352
327
340

335
322
329

L1 + + L2
L2 + + + L3
L3 +

ZA102

Heating
Cooling
Equilibrium

610.5
600.4
605.5

342.7
332.4 325.5 307.8
341 334.1 316.4

355
333
344

327
307
316

L1 + + L2
L2 + +
II , II

ZA122

Heating
Cooling
Equilibrium

627.1
616.9
622.0

342.1
329.9
336

350
330
340

337
320
329

L1 + + L2
L2 + +

ZA104

Heating
Cooling
Equilibrium

595.6
577.6
586.6

361.5
338.2
349.9

390
350
370

340
310
325

L1 + + L2
L2 + + + L3
L3 +

ZA74

Heating
Cooling
Equilibrium

612.3
578.8
595.6

369.3
350.8
360.1

398
358
378

358
320
339

L1 + + L2
L2 + + + L3
L3 +

ZA75

Heating
Cooling
Equilibrium

598.5
580.5
589

372.2
354.8
363.5

398
358
378

358
327
343

L1 + + L2
L2 + + + L3
L3 +

Note: Characteristic temperatures emerging in both the heating process (heating) and the cooling process (cooling) of DSC measurements are given in this
table, as well as the equilibrium temperatures (equilibrium), calculated by averaging the two corresponding temperatures in heating curve and cooling curve. To
determine the solidification path, the starting temperature (starting) and ending temperature (ending) of the second set of peak (peak II), which corresponds to
the second phase transformations, are also listed in the table.

J. Zhang et al. / Materials Science and Engineering A 456 (2007) 4351

333, 334, and 336 C in alloy ZA82, ZA102, and ZA122, respectively. It is most probably due to the different contributions of
each thermal event to the total thermal effect. The higher the temperature, the greater the contribution of the first binary eutectic
reaction is, suggesting more is formed. This can also account
for the different solidification microstructures: the least binary
product in ZA82 alloy results in the sole second phase in this
alloy; whereas the most product in ZA122 alloy makes it a
dominant compound in the final microstructure. It is reasonable
to infer that further increase of the Zn concentration would lead
solely to the secondary phase, which is also indicated in Fig. 7.
Type II alloys can thus be named as -type high zinc magnesium
alloys. Different levels of the reactions in alloy ZA82 and ZA122
from those in ZA102 may also result in inseparable and overlapping of thermal evolutions, accounting for only one detectable
peak in the former two.
With increasing Al content, the second phase transformation
moves to a higher temperature range. In alloy ZA104, it starts
at 370 C with a binary eutectic reaction L1 -Mg + + L2 ,
followed by a ternary quasi-peritectic reaction L2 + Mg + + L3 (343 C in the equilibrium phase diagram), and
ends with another binary eutectic reaction L3 -Mg + . The
amount of eutectic phase should be very small and is expected
to be used up in the following quasi-peritectic reaction. The subsequent solidification microstructure consists of the -Mg +
phases. The overlapping of these thermal events makes the peaking temperature, 350 C, locates between the first and the second
reaction. It is noted that although these have the same phase
constituents, alloy ZA104 and ZA82 underwent different solidification processes. Its possible effects on micro-scale structures
and properties remain unclear at present. These alloys characterized by the compound can also be named the -type high zinc
magnesium alloy, which occupies a wide composition range and
Zn/Al concentration ratios, as seen in Fig. 7.
To alloys with a lower Zn/Al concentration ratio and a higher
Al content, the second phase transformations occur at even
higher temperature ranges. ZA74 and ZA75 alloys have almost
the same temperature range and similar thermal events during
solidification, both of which start at 378 C. According to the
equilibrium phase diagram, the first is a binary eutectic reaction
L1 -Mg + + L2 , which proceeds until the residual liquid
temperature reaches 362 C, where a ternary quasi-peritectic
reaction L2 + -Mg + + L3 occurs, and finally, the binary
eutectic reaction L3 -Mg + takes place. The peak II temperatures, being 360.1 and 363.5 C for ZA74 alloy and ZA75
alloy, respectively, are in good agreement with the equilibrium temperature, which strongly supports the above mentioned
events, suggesting that could be a major intermetallic phase
in the as-cast microstructure. However, this is not the case. It
has been known that the unit cell of the intermetallic phase
contains icosahedral clusters of atoms [22]. It is perhaps not
surprising that the resulting microstructure also contains such
clusters. However, it is unclear how the packing of the icosahedral clusters results in the quasi-crystalline, instead of forming
a crystalline equilibrium phase; and whether there exists a
structural relationship between the icosahedral phase, phase,
and the phase. Given that the icosahedral quasi-crystalline

49

phase may be related with the phase, which usually crystallizes in aluminum-rich alloys, it may not be surprising either that
it only appears in the high-Al experimental alloys. The quasicrystalline phase is denoted by the initial letter Q at this stage
because of its uncertainty, to distinguish it with the I phase in the
MgAlZn system [18]. This type of alloy can also be named as
the Q-type high zinc magnesium alloy. It should be mentioned
that with a higher Al content, ZA75 alloy shows a higher peak II
temperature than ZA74 alloy, indicating a greater contribution
of the first binary eutectic reaction to the total thermal effect,
suggesting more phase was formed. It is reasonable to infer
that with further increase of the Al content, the phase cannot
be consumed in the subsequent ternary quasi-peritectic reaction
and will emerge in the as-microstructure, which is detrimental to
elevated temperature creep resistance. Well-developed commercial MgAlZn alloys are characterized by the phase, which
normally contain about 1%Zn, situated just at the boundary of
the hot cracking area of the low zinc side. For comparison, this
type of alloy, named as the -type magnesium alloy, is also
denoted in the diagram shown in Fig. 7.
3.3. Mechanical properties
Mechanical properties (UTS, YS, elongation) of the experimental alloys are shown in Fig. 9. The Y-axis gives the total
element content in these alloys. For comparison, mechanical
properties of the AZ91 alloy cast in the same condition are
also shown as a reference in Fig. 9. It can be seen that all
the experimental alloys have comparable ultimate strength and
yield strength with the AZ91 alloy. However, the elongations are
somewhat less than that of the AZ91 alloy, with the exception
of the ZA82 alloy, which has the same total element content
as the AZ91 alloy. It is probably because of the larger amount
of intermetallics in the interdendritic spacing, which can hinder deformation. It may also be noted that the ultimate strength
and yield strength increase slowly with the increase of the total
content of the alloying elements in those high zinc magnesium
alloys. It seems that the type of compounds present in the alloys
does not have any particular influence on the ambient temperature mechanical properties.

Fig. 9. Mechanical properties (UTS, YS, elongation) of the experimental alloys.

50

J. Zhang et al. / Materials Science and Engineering A 456 (2007) 4351

resistance, the and Q phases exhibit unbalanced strengthening

effects at ambient temperature: ambient-temperature properties
solely depend on the total element contents, and not on the particular type of the compounds present. It is most desirable that
the second phase could contribute to both. Again further studies
are required to identify the important microstructural parameters that strengthen magnesium alloys, and to develop reasonable
processing techniques, in order to fully utilize the potential of
this promising group of creep-resistant magnesium alloys.
4. Conclusions

Fig. 10. Total creep strains of the experimental alloys at 150 C for 100 h under
60 MPa.

Fig. 10 shows the total creep strain of the alloys tested at

150 C for 100 h under a stress of 60 MPa. As seen from the
figure, the total strains of all the high zinc experimental alloys are
far smaller than that of the AZ91 alloy, indicating a significant
improvement in the creep resistance. Apparently, this is due to
the completely inhibition of the formation of the phase in the
MgZnAl alloys. It can also be seen that the three types of high
zinc magnesium alloys show different levels of creep resistance.
The -type alloy (ZA122, ZA102) exhibits the highest creep
deformation. The creep resistance of the Q-type alloy (ZA74
and ZA75) is comparable with that of the ZA104 alloy, but is far
superior to that of the ZA82, the other -type alloy. The inferior
creep resistance of the ZA82 alloy is probably due to the small
amount of compounds containing in this alloy, compared with
the ZA104 alloy, which cannot effectively pin grain boundaries,
leading to excess creep deformation by gain boundary sliding.
It is well known that second phases and solid-solution
additions impose great effects on creep behavior. Since the
experimental alloys have approximately the same total alloy contents and total quantity of intermetallic compounds, the type of
the second phase is likely to play a key role in creep resistance.
Thermal stability of the second phase is one of the important
factors that affect its capacity to improve creep resistance. The
phase has a much higher melting or decomposition temperature
of 535 C than the phase, which melts at 347 C. It is reported
that the quasi-crystalline grain boundary phases presenting in the
MgZnAl alloy has pronounced thermal stability [6,9]. The
different stabilities of the three characteristic phases can thus
provide a preliminary explanation to the different levels of creep
resistance at elevated temperature. However, other microstructural parameters, such as phase size, shape, distribution, and
coherency with the matrix, as well as possible precipitation
processes during elevated temperature creep testing, may also
interfere with the creep process and affect creep behavior. Systematic investigation is required to clarify the creep mechanism
and the contributions of these factors to the creep resistance of
high-zinc magnesium alloys.
It should be noted that whereas these demonstrate a clear
influence on the improvement of elevated temperature creep

(1) The as-cast microstructural constituents are determined by

both the Zn and Al contents and the Zn/Al ratio in the permanent mould cast MgZnAl alloys. Alloy ZA82 and ZA104
with an intermediate Al content and Zn/Al ratio is characterized by the -Mg and a fish-bone like cramp lump
phase; Alloy ZA102 and ZA122 with a relatively low Al
content and a high Zn/Al ratio is composed of the -Mg,
a small amount of the , and the binary eutectic phase,
which is the dominant second phase and characterized by
its approximately paralleled lath structure; and alloy ZA74
and ZA75, with comparatively low Zn/Al ratios and high
Al contents, consists of the -Mg matrix with icosahedral
quasi-crystalline phases, which display as a cluster of cornlike phases marked with eye-like spots. No Mg17 Al12 ()
phase was found in these ZA series alloys.
(2) The temperature range, in which the second phase transformations occur, increases with the increase of the Al content
and the decrease of the Zn/Al ratio. Further shift of the composition is likely to result in the emergence of the Mg17 Al12
() phase in the as-cast microstructure.
(3) The ZA series alloys possess similar ultimate strength and
yield strength, but a low elongation, compared with the
AZ91 alloy. The total creep strains are far smaller than that
of the AZ91 alloy.
(4) The ambient ultimate strength and yield strength of the
experimental ZA series alloys increase slowly with the
increase of the total element content. The elevated temperature creep resistances of the -type alloys (ZA104) and the
Q-type alloys (ZA74 and ZA75) are superior to that of the
-type alloys (ZA122 and ZA102).
Acknowledgements
The authors are grateful for the financial support of the
National Natural Science Foundation of China (Grant No.
50301018), the Scientific Research Foundation for ROCS, State
Education Ministry of China, and Chongqing Municipal Science and Technology Commission, China. Jing Zhang gratefully
acknowledges the support of the British Royal Society and the
KC Wong Education Foundation for a KC Wong Fellowship.
Thanks are also due to Mr. Qin Ren from Chongqing University,
China, and Mrs. Weina Yang and Mr. Jinhan Yao from Queen
Mary, University of London, UK, for their assistance in some
experiments.

J. Zhang et al. / Materials Science and Engineering A 456 (2007) 4351

References
[1] P. Humble, Mater. Forum. 21 (1997) 4556.
[2] A.A. Luo, M.P. Balogh, B.R. Powell, Metall. Mater. Trans. A 33 (2002)
567.
[3] G.Y. Yuan, Y.S. Sun, W.J. Ding, Mater. Sci. Eng. A 308A (2001) 38.
[4] Y.S. Sun, K.Z. Wen, G.Y. Yuan, Chin. J. Nonferrous Met. 9 (1999) 55.
[5] B.L. Mordike, K.U. Kainer (Eds.), Proceedings of the Magnesium Alloys
and Their Applications, Werkstoff-Informationsge-sellschaft, Frankfurt,
1998.
[6] L. Bourgeois, C.L. Mendis, B.C. Muddle, J.F. Nie, Philos. Mag. Lett. 81
(2001) 709718.
[7] C. Mendis, L. Bourgeois, B. Muddle, J.F. Nie, Magnesium Technology 2003, 2003 TMS Annual Meeting, San Diego, CA, USA, 2003, pp.
183188.
[8] A. Luo, T. Shinoda, J. Mater. Manufact. 107 (1998) 8694.
[9] M. Vogel, O. Kraft, G. Dehm, E. Arzt, Scripta Mater. 45 (2001) 517524.
[10] M. Vogel, O. Kraft, E. Arzt, Scripta Mater. 48 (2003) 985990.
[11] Y.X. Wang, S.K. Guan, X.Q. Zeng, et al., Mater. Sci. Eng. A 416 (2006)
109118.

51

[12] Z. Zhang, R. Tremblay, D. Dube, Mater. Sci. Eng. A 385 (2004) 286
291.
[13] Z. Zhang, R. Tremblay, D. Dube, in: J. Hryn (Ed.), Magnesium Technology,
TMS, 2001, pp. 147151.
[14] I.A. Anyanwu, S. Kamado, T. Honda, et al., Mater. Sci. Forum 350351
(2000) 7378.
[15] H. Liang, S.L. Chen, Y.A. Chang, Metall. Mater. Trans. A 28A (1997)
17251734.
[16] G.S. Foerster, Proceeding of the IMA 33rd Annual Meeting, Montreal,
Quebec, Canada, May 2325, 1976, pp. 3539.
[17] L. Bourgeois, B.C. Muddle, J.F. Nie, Acta Mater. 49 (2001) 27012711.
[18] C. Janot, Quasicrystals, Oxford University Press, Oxford, 1997.
[19] D.V. Petrov, in: G. Petzow, G. Effenberg (Eds.), Ternary Alloys: A Comprehensive Compendium of Evaluated Constitutional Data and Phase
Diagrams, vol. 7, VCM, Weinheim, 1993, pp. 5771.
[20] J.B. Clark, Trans. Am. Soc. Met. 53 (1961) 295306.
[21] J. Zhang, F.S. Pan, Z.X. Guo, Proceedings of the Second International
Conference on Magnesium, Beijing, China, June 2528, 2006.
[22] P. Schobinger-Papamantellos, P. Fisher, Naturwissenschaften 57 (1970)
128.