Intermetallics 67 (2015) 145e155

Contents lists available at ScienceDirect

Intermetallics
journal homepage: www.elsevier.com/locate/intermet

Diffusion kinetics, mechanical properties, and crystallographic

characterization of intermetallic compounds in the MgeZn binary
system
C.C. Kammerer a, *, S. Behdad b, L. Zhou a, F. Betancor a, M. Gonzalez a, B. Boesl b,
Y.H. Sohn a
a

University of Central Florida, Advanced Materials Processing and Analysis Center, Department of Materials Science and Engineering, Orlando, FL 32816,
USA
Florida International University, Department of Mechanical and Materials Engineering, Miami, FL 33174, USA

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 19 May 2015
Received in revised form
5 August 2015
Accepted 8 August 2015
Available online xxx

Pure Mg was diffusion bonded to pure Zn at 315
C for 168 h to produce equilibrium intermetallic
compounds of the MgeZn system. All equilibrium phases at 315
C, Mg21Zn25, Mg4Zn7, MgZn2, Mg2Zn11,
were observed to develop. Concentration proles by electron probe microanalysis, electron diffraction
patterns by transmission electron microscopy, and loadedisplacement curves by nano-indentation were
examined to characterize the phase constituents, crystal structure, diffusion kinetics, and mechanical
properties. Mg21Zn25 with trigonal, Mg4Zn7 with monoclinic, and Mg2Zn11 with cubic structures were
found and their lattice parameters were reported herein. Mg4Zn7 and Mg2Zn11 were observed to have a
range of solubility of approximately 2.4 at% and 1.6 at%, respectively. Interdiffusion in MgZn2 occurred
most rapidly, was an order of magnitude slower in Mg4Zn7 and Mg2Zn11, and was the slowest in
Mg21Zn25. Composition-dependence of interdiffusion within each intermetallic phase was negligible. The
intermetallic phases exhibited insignicant creep, but evidence of discontinuous yielding was observed.
The average hardness and reduced moduli were similar for Mg21Zn25, Mg4Zn7, and MgZn2 phases, ~5 GPa
and ~90 GPa, respectively. However, the Mg2Zn11 phase had lower hardness of 3.76 GPa and higher
modulus of 108.9 GPa. The mechanical properties in the characterized intermetallic phases, exclusive of
Mg21Zn25, were strongly concentration-dependent.
2015 Elsevier Ltd. All rights reserved.

Keywords:
Crystal chemistry
Diffusion
Mechanical properties
Alloy design
Heat treatment
Phase transformations

1. Introduction
The high strength-to-weight ratio makes Mg alloys one of the
most promising structural materials for the automotive, aerospace,
and consumer electronics industries [1,2]. Mg also has exceptional
biocompatibility, with potential application as prosthetic implants
[3,4]. Some of the more common commercial Mg alloys are AZ
series (MgeAleZn), ZE series (MgeZneRE), and ZK series
(MgeZneZr). Clearly, Zn is one of the most important alloying elements in Mg. Therefore, understanding the crystal structure, kinetic behavior during thermal processing, and mechanical
properties of the intermetallic compounds in the MgeZn binary
systems is essential.

* Corresponding author. Department of Materials Science and Engineering, University of Central Florida, 12760 Pegasus Drive, Orlando, FL 32816, USA.
E-mail address: Catherine.Kammerer@ucf.edu (C.C. Kammerer).
http://dx.doi.org/10.1016/j.intermet.2015.08.001
0966-9795/ 2015 Elsevier Ltd. All rights reserved.

As shown in the phase diagram in Fig. 1 [5], there are ve

intermetallic compounds in the MgeZn system e Mg51Zn20,
Mg21Zn25, Mg4Zn7, MgZn2, and Mg2Zn11. The Mg51Zn20 phase is not
stable at room temperature. While Mg21Zn25 is the equilibrium
intermetallic with Mg solid-solution (a-Mg), formation of GuinierPreston (GP) zones and metastable precipitates of MgZn2 and
Mg4Zn7 take place during articial aging in commercial Mg alloys
containing Zn [6,7]. There is considerable debate on the exact
stoichiometry and solubility limits of the intermediate phases in
the MgeZn system. Similarly, there is uncertainty in the specic
details of the crystal structure of the MgeZn intermetallic compounds. In recent years, investigation into the stoichiometry and
crystal structure of the phases has been conducted [5,8e10]. In fact,
the phase diagram given in Fig. 1 is based on the 2012 work of
Ghosh et al. [8] in which all available experimental and computational data for thermodynamic properties and phase boundaries of
the MgeZn system were modeled. Nonetheless, little corroborating
data is available to resolve the precise stoichiometry, extent of

146

C.C. Kammerer et al. / Intermetallics 67 (2015) 145e155

Fig. 1. MgeZn phase diagram taken from the 2013 review by Okamoto [5] and based on the work of Ghosh et al. [8].

homogeneity, and crystal structure of the room-temperature

equilibrium intermetallic phases.
Alloy development and optimization of diffusion controlled
processes such as solidication, homogenization, and precipitation
necessitate knowledge of interdiffusion coefcients. The rst
interdiffusion study of the MgeZn system was conducted by
Sakakura and Sugino [11] via polycrystalline solid-to-solid diffusion
couples annealed between 200
C and 300
C. They observed only
the MgZn2 and Mg2Zn11 phases and calculated interdiffusion coefcients and parabolic growth constants. More recently, Brennan
et al. [12] determined the parabolic growth constants for phases
evolved during diffusion anneals between 295
C and 325
C of
polycrystalline couples of pure Mg and Zn. Under the assumption of
diffusion controlled parabolic growth, the interface velocity is
inversely proportional to the interphase thickness therefore, the
parabolic growth constant is proportional to the interdiffusion coefcient and is often used as an indicator of kinetic behavior. Only
four of the ve phases were reported; as expected, Mg51Zn20 was
observed in only the couple annealed at 325
C, but Mg21Zn25 was
not observed in any couple. Though the absence of certain phases
can be attributed to difculties in nucleation, phases present in very
small quantities may not be observed due to limitations of characterization techniques [13]. In the most comprehensive diffusion
study to date, Das et al. [14] examined the time-dependent,
anisotropic kinetic behavior in the MgeZn system. Couples of single crystal Mg and polycrystalline Zn were annealed at
280
Ce330
C for 4, 10, and 20 days. All intermediate phases were
observed, thicknesses measured, and parabolic growth constants
computed. Interdiffusion coefcients for Mg4Zn7 (reported in [14]
as Mg2Zn3), MgZn2 and Mg2Zn11 were experimentally determined
while those for Mg21Zn25 (reported in [14] as Mg12Zn13) and
Mg51Zn20 were calculated from a multiphase diffusion model. Most
recently Mostafa and Medraj [15] annealed solid-to-solid MgeZn
diffusion couples between 250
C and 300
C. They observed and
measured the thickness of all phases indicated by the phase diagram for the studied temperature range. Still, there is limited data

which can be used to offer consensus on the kinetic behavior in the

intermetallic phases.
Surprisingly very few investigations into the mechanical properties of intermetallic compounds in the MgeZn system exist in the
literature. The lack of mechanical-property data for these compounds is partly due to the requirement for a bulk, homogeneous
specimen for traditional mechanical property evaluation methods
(e.g. tensile testing). While precipitation from supersaturated solidsolutions and solidication from melt of intermetallic compounds
result in metastable phases and microstructural gradients, solidstate synthesis of intermetallic compounds by isobaric,
isothermal diffusion reaction between two metals has long been
recognized as a means of eliciting intermediate phases occurring in
the same sequence of the appropriate equilibrium phase diagram
[16,17]. Thus, solid-to-solid diffusion couples can serve as combinatorial specimens yielding sufciently thick layers of intermetallic
phases to allow for nano-mechanical testing. In addition, when
homogeneity ranges exist within an intermediate phase, variation
in elastic and plastic properties as a function of composition can be
examined through nanoindentation.
In this study, three solid-to-solid diffusion couples consisting of
pure Mg diffusion bonded to pure Zn at 315
C for 7 days were
assembled. Consistency of the diffusion reaction products were
established by comparing layer thicknesses examined with a
Scanning Electron Microscope (SEM). Resulting phases were subjected to Electron Probe Micro Analysis (EPMA), quasi-static nanomechanical testing, and Selected Area Electron Diffraction (SAED)
via Transmission Electron Microscopy (TEM). Crystallographic
characteristics for Mg21Zn25, Mg4Zn7, and Mg2Zn11 are examined
and reported. Parabolic growth constants and interdiffusion coefcients for the phases are computed and compared to literature
values. The elastic and plastic properties for each phase are determined through analysis of the loadedisplacement proles by the
Oliver-Pharr method. To our best knowledge, this study is the rst
measurement of mechanical properties of intermetallic compounds
in polycrystalline MgeZn system.

C.C. Kammerer et al. / Intermetallics 67 (2015) 145e155

2. Materials and methods

The combinatorial samples were fabricated by diffusion bonding
pure, polycrystalline Mg and Zn at 315
C for 7 days. Rods of
polycrystalline Mg (99.9%), 7.9 mm in diameter, and Zn (99.9%),
10 mm in diameter, from Alfa Aesar were sectioned into disks
approximately 3 mm in thickness. These disks were then metallographically polished down to 1 mm using a non-oxidizing lubricant
at each grinding and polishing step. The polished faces of the disks
were clamped in a stainless steel jig. Inert alumina (Al2O3) barrier
material was inserted between the MgeZn couple and clamping
surface of the jig to prevent chemical interaction during heat
treatment. The jig assembly was then placed in a quartz capsule
that was repeatedly evacuated to ~106 torr and ushed with
hydrogen and ultra-high purity argon. The capsule was then
backlled with ultra-high purity argon to a pressure slightly above
1 atm at the annealing temperature and sealed. The encapsulated
couple was placed into furnace (Paragon Bluebird) which had
been preheated to the annealing temperature of 315
C. A type-K
thermocouple was used to independently monitor the furnace
temperature which was stabilized within 2
C. The sample was
held at temperature for 168 h (i.e. 7 days). After the diffusion
anneal, the quartz capsule was quickly removed and quenched in
cold water. The whole jig assembly was mounted in epoxy, crosssectioned to reveal the diffusion reaction zone, and again metallographically prepared down to 1 mm nish. Each of the three
diffusion couples was prepared independently. A eld emissionSEM (Zeiss Ultra55) with an angular selective backscattered
electron detection system for maximum compositional contrast
was utilized to image the diffusion reaction products for subsequent image analysis. Thickness, x, of each phase was measured at
30 random points from multiple SEM elds of view using ImageJ
image processing and analysis software (National Institutes of
Health, http://imagej.nih.gov/ij/). X-ray Energy Dispersive Spectroscopy (XEDS) was used to verify consistency between each of the
diffusion couples of chemical composition of the evolved phases.
One diffusion couple was subjected to EPMA (JEOL 8900R
Superprobe) for precise quantitative chemical analysis of the

~ int
D
AX BY

147

parameters of each phase. At least six orientations within each

phase were obtained by tilting the specimens.
Nanoindentation was conducted with a Hysitron TI Premier
instrumented quasi-static nano-mechanical tester using a 100 nm
diamond Berkovich tip. No less than ten indentations were made on
each phase. Two of the three assembled couples were tested. The
indentation loading prole consisted of 10-s loading to a peak load,
followed by a 3-s hold segment, and then 10-s unloading. When the
phase layer thickness exceeded 50 mm the peak load was 7 mN. For
layers greater than 5 mm but less than 50 mm the peak load was
reduced to 5 mN. Layers less than 5 mm in thickness were subjected
to 3.5 mN peak loading. The reduction in peak load was necessary to
ensure test results were not affected by the presence of an interface.
Indent location on the narrow intermetallic phases were conrmed
after nanoindentation with secondary and backscattered electron
imaging. To reveal any concentration dependencies on the mechanical properties, three indentation line scans were performed on
the central region of the sample subjected to EPMA, perpendicular to
the interface and extending into the solid solution phases. The load
prole remained the same as previously described however the peak
load was 7 mN and the step size was 5 mm.

3. Analytical framework
In diffusion controlled growth in a solid-to-solid diffusion
couple, and under the assumption of negligible nucleation time, the
thickness of a growing phase after time, t, can be described by the
parabolic relationship [17,20] as given by:

kp

x2
2t

[1]

The layer thickness and parabolic growth constant depends on

the composition of neighboring phase, and can be used to determine the integrated interdiffusion coefcient by adjusting for the
chemical effects of the neighboring phases as described by Huang
et al. [21] adopted from Wagner [22]:

3
2
Z


 Zx1
AX BY 
Dx
7
6
A B
NBAX BY 1  NBAX BY kAp X BY
NB dx NBAX BY
1  NB dx5
4 1  NB X Y
2t


concentration proles through the reaction zone. EPMA line scans

were performed on the central region of the sample and orthogonal to the interface extending approximately 150 mm into the
solid solution phases. Measured x-ray intensities were corrected
for matrix effects by applying the Heinrich ZAF correction factors.
The pure Mg and Zn terminal ends were used as standards for
equipment calibration. Operating conditions were 15 keV with a
steady probe current of 20 nA. The step size of the EPMA line scan
was 2 mm.
Crystal structure of the Mg21Zn25, Mg4Zn7, and Mg2Zn11 phases
was examined using an FEI/Tecnai F30 TEM microscope with a
eld-emission source working at 300 kV. Site-specic specimens
were prepared using an FEI TEM200 Focused Ion Beam (FIB) in situ
lift-out (INLO) technique [18,19]. The spacing and orientation of the
diffraction spots in a series of single crystal Selected Area Electron
Diffraction (SAED) patterns obtained through a range of tilt angles
was used to determine the interplanar spacing and lattice

x2

[2]
NB 0;NB 1

Dayananda [23] relates the integrated interdiffusion coefcient

to the interdiffusion coefcient by:
x2

~ int
D

ZCi

~
DdC


Zx2

~Jdx

[3]

x1

x
Ci 1

The interdiffusion coefcient can be calculated with the BoltzmanneMatano approach at any specic composition such that the
interdiffusion coefcient is expressed as:

Z
~ 
D

1
2t

Co

x  xo dC
vC
vx

[4]

where C and C refer to the composition at terminal ends of the

148

C.C. Kammerer et al. / Intermetallics 67 (2015) 145e155

diffusion couple, xo is the position of the Matano plane, and Co is the

composition at the Matano plane. The Matano plane denotes the
location of mass balance across the interdiffusion zone such that:

ZC o

C 
Z

xdC
C

xdC 0

[5]

Co

However, this approach is only valid when the deviation in

molar volume, Vm, from that of ideal solution, e.g. Vegard's law, is
negligible. Non-ideal solutions deviate from this rule. The SauerFreise Method [24] with Wagner's adaptation [22] can be used to
calculate the interdiffusion coefcient when changes in molar
volume occur as in the case of multicomponent diffusion reactions
[25]. This method [20] essentially normalizes the concentration
and the diffusion coefcient is modied as:

3
2
Zx
Zx*
*


 *
V
Y
1

Y
7
6
*
M
~ Y 
dx Y
dx5
D
4 1Y
VM
VM
2tdY=dxx*
x*

x

[6]
where Y (N  N)/(NN) and N denotes the mole fraction
of the component at any position; N denotes at the right terminal end of the couple, N- denotes at the left terminal end of the
couple. The SauereFreise method is not subject to the error
potentially introduced through the determination of the plane of
mass balance in the BoltzmanneMatano method.
The interdiffusion coefcient typically varies over the composition range. Therefore, the effective interdiffusion coefcient
[23,26] can be computed and provides a single nominal coefcient
for the compositional spectrum of a given phase. Integrating the
ux over an interval from x1 to x2, and dividing by the change in
composition over the interval, yields the effective interdiffusion
coefcient as per:

x2

~Jdx
eff
x1
~
D 
Cx2  Cx1

4. Results and discussion

The solid-to-solid diffusion couple technique was used to
fabricate samples which could be subjected to a combination of
characterization methods. To validate this approach, three samples
were independently fabricated and qualitatively compared. The
three samples are ostensibly identical. The following results are
obtained from one or more of the assembled couples.
4.1. Phase constituents and crystallography

[7]

For assessment of mechanical behavior via nanoindentation, the

Oliver-Pharr method [27e29] was utilized. Hardness, H, is the peak
load, Pmax, divided by the contact area, which can be calculated as a
function of contact displacement, hc. The contact area function, A,
for the Berkovich tip can be described by an nth (n 3e6) order
polynomial in which the coefcients are determined through tip
area calibration on a standard fused quartz sample before testing.
Therefore, ideal indenter geometry deviations (i.e. manufacturing
tolerances and wear) can be taken into account prior to testing of
the sample. The reduced modulus, Er, is determined from its relationship to contact area function and the initial unloading stiffness,
S. A schematic of a typical loadedisplacement curve is given in
Fig. 2 in order to further elaborate how reduced modulus and
hardness can be calculated from nanoindentation.
The reduced elastic modulus is a composite value inuenced by
the elastic properties of both the specimen and the indenter [27].
Given knowledge of the indenter material elastic properties (i.e.
Young's modulus, Ei, and Poisson's ratio, yi) and the Poisson's ratio
of the material being tested, ys, the reduced modulus can easily be
converted to the materials Young's modulus, ES, using the
relationship:

1  y2i
1  y2S
1

Er
Ei
ES

Fig. 2. Schematic of a typical load vs. displacement curve with the fundamental
equations of the OliverePharr method [26e28].

[8]

A backscattered electron micrograph of the diffusion reaction

products from the Mg vs. Zn couple annealed at 315
C for 7 days is
presented in Fig. 3. All room-temperature equilibrium compounds
are present. Due to magnesium's strong afnity for oxygen, a natural marker plane from oxidation of the initial end-members is

Fig. 3. Backscattered electron micrograph in the interdiffusion zone from the diffusion
couple Mg vs. Zn annealed at 315
C for 168 h: phases and approximate location of
TEM specimens are annotated.

C.C. Kammerer et al. / Intermetallics 67 (2015) 145e155

Fig. 4. Concentration prole across the interdiffusion zone from the diffusion couple
Mg vs. Zn annealed at 315
C for 168 h. Open symbols are from the measured EPMA
data and solid line is the tted prole. The large scatter in the Mg2Zn11 phase is an
artifact caused by the crack at the marker plane.

observed within the Mg2Zn11 phase [12,15]; as a consequence this

phase was more susceptible to cracking and pull out during subsequent processing. Although the location of this marker plane is
equivalent to a Kirkendall plane, its precise position cannot be
determined due to the phase pullout. As a result, the intrinsic
diffusion coefcients at the plane and microstructural differences
on either side of the plane [30] are not determined. In addition,
there is a slight contrast in the Mg2Zn11 phase on either side of the
marker plane. Composition across the diffusion reaction zone was
determined via EPMA, and the concentration proles were curve t
for each phase. Fig. 4 presents the EPMA data along with the tted
concentration prole. The scatter within the Mg2Zn11 phase is
assumed to be due to pull-out at the marker plane during polishing.
Mg4Zn7, MgZn2, and Mg2Zn11 are all present with denite concentration gradients (vC/vx s 0). With the exception of that for aMg, the concentration proles for each phase were best t with a
simple linear least squares regression. The Zn concentration in the
a-Mg phase was t with an exponential function. Consistent with
the phase diagram shown in Fig. 1, there was no apparent solubility
of Mg in Zn therefore the concentration prole was not t for this
phase.
From EPMA line scans shown in Fig. 4, the line compounds,
Mg4Zn7 and Mg2Zn11, have a homogeneity range of 2.4 at% Zn and
1.6 at%Zn, respectively. The absolute composition for each phase

149

presents with slightly lower Zn concentration than stoichiometry

would suggest. EPMA can precisely determine concentration
gradient which is essential for diffusion and relative concentration
studies, but the technique can have increased error in the absolute
concentration; EPMA is precise to the ppm level but only accurate
within ~2% [31].
To examine the crystal structure of the intermetallic compounds, three samples were extracted from the interdiffusion zone
for TEM. The locations of the extracted specimens are annotated in
Fig. 3 as TEM-1, TEM-2, and TEM-3. A bright eld micrograph from
TEM-1 specimen in Fig. 5a reveals a-Mg, Mg21Zn25, and Mg4Zn7
phases. A bright eld micrograph from TEM-3 showing Mg2Zn11
phase is presented in Fig. 5b. TEM-2 with Mg2Zn11 was indistinguishable from TEM-3 under bright eld imaging and is therefore
not presented.
The intermetallic compound in equilibrium with Mg solidsolution is Mg21Zn25; it has also been described as Mg12Zn13 or
the equiatomic MgZn despite the composition being Zn rich.
Casting further doubt on the exact stoichiometry of this phase is the
fact that the crystal structure of MgZn or Mg12Zn13 has not been
decidedly reported [5,6,32,33]. Tilting series SAED patterns were
obtained for Mg21Zn25 from TEM-1, four of which are shown in
Fig. 6. From the SAED study, the Mg21Zn25 was found to have a
trigonal symmetry (space group R3ch) with lattice parameters of
a 2.5518 nm and c 0.8713 nm. In 2002, Cerny and Renaudin [9]
identied, using single crystal x-ray diffraction (XRD), a MgeZn
compound that is isomorphous with Zr21Re25. From their work, it
was shown Mg21Zn25 nucleates and grows in a trigonal structure
with 6 formula units per unit cell and lattice parameters of
a 2.5776 nm and c 0.8762 nm. In another study, the Mg21Zn25
phase, formed during mechanical alloying and annealing of pure
Mg, Zn and Si powders, was characterized by De Negri et al. [34] via
powder XRD. The crystal structure was likewise identied as
trigonal with cell parameters of a 2.5640 nm and c 0.8714 nm.
The results obtained in this study by electron diffraction for
Mg21Zn25, summarized in Table 1, agree well with those determined by XRD [9,34] and validate the stoichiometry of Mg21Zn25.
Similarly, Mg4Zn7 had generally been accepted as Mg2Zn3 as
originally reported by Takei [35], Laves [36], and later Clark and
Rhines [37]. The crystal structure of Mg2Zn3 was thought to be
triclinic (a 1.724 nm, b 1.445 nm, c 0.520 nm, a 96
, b 89
,
g 138
) [6]. In 1975, Yarmolyuk et al. [10] determined the crystal
structure and ideal stoichiometry of the intermetallic phase as
having monoclinic symmetry with cell parameters of a 2.596 nm,
b 0.524 nm, c 1.428 nm, and b 102.5
and 10 formula units per
unit cell; stoichiometry was identied as Mg4Zn7. Nonetheless,
consensus on the crystal structure and stoichiometry of Mg4Zn7 has

Fig. 5. Bright eld micrographs: a) TEM-1 showing the a-Mg, Mg21Zn25, and Mg4Zn7 phases; b) TEM-3 specimens showing Mg2Zn11 phase.

150

C.C. Kammerer et al. / Intermetallics 67 (2015) 145e155

Fig. 6. Tilting series SAED patterns from the Mg21Zn25 phase in TEM-1 indicating the trigonal symmetry (space group R3ch) and lattice parameters of a 2.5518 nm and
c 0.8713 nm: zone axis a) 421; b) 112; c) 114; d) 225.

Table 1
Experimentally determined roometemperature equilibrium phase crystallographic parameters.
Structure
Mg21Zn25
Mg4Zn7
MgZn2
Mg2Zn11

Trigonal
Monoclinic
C14 Laves
cubic

a (nm)

b (nm)

2.564
2.669
0.511
0.5223
0.8415e0.8462

c (nm)
0.8714
1.411
0.8566

b

90
90

90

90

g

101.85

120
90

120

Formula units

Molar volume (cm3/mol)

6
10
4
3

10.74
10.31
10.15
9.20e9.35

MgZn2 data taken from PDF#04-003-2083 [42].

Fig. 7. SAED patterns for Mg4Zn7 from TEM-1 indicates a monoclinic structure (space group B2/m) with lattice parameters of a 2.669 nm, b 0.511 nm, c 1.411 nm, and
b 101.85
: zone axis a) [130] and b) [140].

C.C. Kammerer et al. / Intermetallics 67 (2015) 145e155

151

Fig. 8. SAED patterns for Mg2Zn11 from TEM-2 (top row: aec) and TEM-3 (bottom row: def) reveal a cubic structure (space group Pm3) with a variation in lattice parameters
a 0.8462 nm (TEM-2) and a 0.8415 nm (TEM-3): zone axis a) [100], b) 110 , c) 120 , d) [100], e) 110 , f) 120 .

remained unresolved [5,38]. In the present study, tilting series

SAED patterns were obtained for Mg4Zn7 as shown in Fig. 7. A
monoclinic structure (space group B2/m) with lattice parameters
a 2.669 nm, b 0.511 nm, c 1.411 nm, and b 101.85
were
determined. These results are in good agreement with those of
Yarmolyuk et al. [10].
Mg2Zn11 has been cited as a line compound with cubic symmetry and a 0.8552 nm [5,39]. To study inuence of concentration on lattice parameter in Mg2Zn11, SAED patterns from TEM-2
and TEM-3 specimens, presented in Fig. 8, were analyzed. Both
specimens exhibit a cubic structure (space group Pm3) with slightly
different lattice parameters a 0.8462 nm (TEM-2) and
a 0.8415 nm (TEM-3). A smaller lattice parameter is observed for
Mg2Zn11 near the Zn side. The molar volume of Mg is 13.99 cm3/mol
while the molar volume of Zn is 9.15 cm3/mol. Thus, consistent with
Vegard's law [40], an increase in Zn content would decrease the
molar volume of the phase and hence the lattice parameter. This
nding is consistent with the homogeneity range determined by
EPMA. The results of SAED and EPMA strongly suggest that the
Mg2Zn11 has a range of solubility with a corresponding variation in
lattice parameter.
The crystal structure of MgZn2 is well established as the prototypical C14 Laves Phase [36,41,42]. Therefore, no TEM was carried
out. Table 1 provides a summary of the crystallographic parameters
determined in this study along with that for MgZn2 from PDF#04003-2083 [42].
4.2. Diffusion kinetics characteristics
The concentration dependent interdiffusion coefcient was
calculated by applying both the BoltzmanneMatano approach,
Equation (4), and the Sauer-Freise approach, Equation (6), using the
molar volumes estimated from the lattice parameters determined
by TEM. The results of the two analytical approaches are compared
in Fig. 9. The observed molar volume variation of Mg2Zn11 has

Fig. 9. Interdiffusion coefcients determined using SauereFreise and BoltzmanneMatano methodologies and plotted as a function of composition.

negligible inuence on the computed interdiffusion coefcients as

demonstrated in Fig. 9. In addition to the congruency of results
obtained, an examination of Fig. 9 reveals that there is no appreciable concentration dependence on the interdiffusion coefcients
for Mg4Zn7 and MgZn2, while there is slight concentration dependence in Mg2Zn11. As previously mentioned, a Kirkendall plane is
present in the Mg2Zn11 phase. While the quantitative intrinsic
diffusion coefcients cannot be determined, a qualitative assessment of the diffusivity of each component can be made. Since the
marker plane occurs on the Zn-rich side of the diffusion couple, the
Zn is generally the faster moving species [43].
The thickness of each phase was measured as previously
described. Under the presumption of parabolic growth, the phasespecic growth constant, kp, was calculated using Equation (1), and

152

C.C. Kammerer et al. / Intermetallics 67 (2015) 145e155

Table 2
Thickness, parabolic growth constant, integrated interdiffusion coefcients, and average effective interdiffusion coefcients of intermetallic phases.

Thickness (mm)
kp(m2/s)
kp (m2/s)
kp (m2/s)
~ int

a-axis
c-axis

D (m /sec)
~ eff (m2/sec)
D
eff
~ (m2/sec)
D
~ eff (m2/sec)
D

a-axis
c-axis

Mg21Zn25

Mg4Zn7

MgZn2

Mg2Zn11

Source

0.94 (0.32)
1.6  1018
e
7.24  1019
3.27  1018
4.74  1017
e
8.13  1015
7.72  1015
e

33.7 (3.3)
1.9  1015
1.1  1015
2.21  1015
2.36  1015
2.31  1015
9.76  1014
1.19  1013
1.12  1013
1.44  1014

237.0 (2.4)
9.3  1014
4.9  1014
6.17  1014
9.21  1014
1.53  1014
1.13  1012
5.07  1013
4.83  1013
2.87  1013

30.8 (1.7)
1.6  1015
1.2  1015
2.12  1015
2.35  1015
9.97  1016
6.02  1014
4.70  1014
4.66  1014
2.55  1013

This Study
This Study
Brennan [12]
Das [14]

reported in Table 2. MgZn2 phase grows the thickest while

Mg21Zn25 does not form a layer of substantial thickness. Mg4Zn7 is
thicker than Mg2Zn11. The integrated interdiffusion coefcient for
Mg21Zn25 was determined using Equation (2), while the integrated
interdiffusion coefcient for Mg4Zn7, MgZn2, and Mg2Zn11 were
calculated by using Equation (3). The effective interdiffusion coefcients was calculated using Equation (7), where the ux was
determined using the BoltzmanneMatano approach. The interdiffusion coefcients are reported in Table 2 as well. Interdiffusion in
MgZn2 is orders of magnitude faster than in the other intermetallic
phases. Mg21Zn25 has the lowest interdiffusion coefcient.
In consideration of ndings by other researchers, Table 2 also
presents comparative data for growth and interdiffusion at 315
C.
Nominal pre-exponential factor and activation energy were used to
~ Mostafa et al. [15]
compute average interdiffusion coefcients, D.
computed interdiffusion coefcient of each interphase boundary
(i.e. Mg/Mg21Zn25, Mg21Zn25/Mg4Zn7, Mg4Zn7/MgZn2, MgZn2/
Mg2Zn11, and Mg2Zn11/Zn) as a xed-frame of reference, and
therefore they are not comparable to the present work and have
been omitted from Table 2. There is good agreement in the parabolic growth constants determined by Brennan et al. [12] and Das et
al. [14] with those calculated in this study. Das et al. has previously
demonstrated that growth of the intermetallic phases is in fact
parabolic and independent of orientation [14], and corroborate the
presumption of parabolic growth used in this study. Slight difference of the interdiffusion coefcients determined in this study and
the ndings of other researchers can be attributed to the different
analytical and computational methods. Das et al. used the
Heumann-Matano method, and Sakakura et al. exercised the
BoltzmanneMatano method; both calculate the numerical average
value of the measured interdiffusion coefcients over a composi~ which is comparable but not equivalent to the
tion range, D,
effective interdiffusion coefcient.
4.3. Assessment of mechanical behavior
The mechanical behavior of the intermetallic compounds was
examined by nanoindentation. For comparison and validation, the
mechanical properties of the pure Mg and Zn were also measured.
The reduced modulus and hardness were calculated from the
multiple loadedisplacement curves. The calculated values along
with their standard deviations are summarized in Table 3. All
intermetallic compounds exhibit higher hardness values compared
to pure Zn and Mg. The increase in reduced modulus is less

This Study
This Study
Das [14]
Sakakura [11]

signicant in consideration of the high modulus possessed by Zn.

Mg2Zn11 has the highest reduced modulus but the lowest hardness
of the intermetallic compounds. The primary strengthening phases
in MgeZn system (MgZn2, Mg4Zn7, and Mg21Zn25) all have similar
hardness and reduced moduli.
Fig. 10 shows the typical nanoindentation loadedisplacement
curves for the Mg21Zn25, Mg4Zn7, MgZn2, and Mg2Zn11. Several
qualitative observations can be made based on the
loadedisplacement curves. Specically, a holding segment is usually included in nanoindentation loading prole to avoid the effect
of creep on the unloading segment, which is used for calculations
based on Oliver-Pharr method. In this study, a 3 s holding segment
was utilized. All four intermetallic phases undergo only 20e30% of
the creep exhibited by the pure Zn and Mg.
Additionally, serrations seen in the loading prole are formed
because of discontinuous yielding during plastic deformation.
These serrations are termed pop-ins. This phenomenon is most
commonly attributed to dynamic strain aging, often called Portevin
e Le Chatelier (PLC) effect, and is due to interactions between
moving dislocations and diffusing solute atoms, which leads to
instantaneous negative strain rate sensitivity and localization of
ow [44,45]. The solute atoms catch and pin the dislocations. When
the dislocations are torn away from the solute atoms, a jump in
displacement is observed. The PLC effect has been well studied in
AleMg alloys and is associated with free ight of mobile dislocations between subsequent blockings at obstacles [46e48]. However, dynamic strain aging is not the only phenomenon that can
cause serrated loadedisplacement curves. Order-disorder transformations and mechanical twinning during deformation as well as
work hardening sensitivity can also produce the same effect
[49,50]. In fact, it is understood that the PLC effect occurs within
specic limits of temperature, strain, strain rate, and impurity
concentration [45,48]. It is interesting to note, on the basis of Fig. 10,
Mg2Zn11 shows very little evidence of pop-ins compared to the
other intermetallic compounds. A dislocation avalanche, associated
with twinning and cross-slip, is observed in the pure Zn as a pop-in
in the loading segment. On the other hand, the pure Mg shows only
subtle indications of pop-ins.
To our best knowledge, there are no previous reports on measurement or theoretical estimation of mechanical properties of
Mg2Zn11 or Mg21Zn25. Xie et al. [51] used rst-principle calculations
and VoigteReusseHill approximation to calculate upper and lower
bounds of elastic properties of polycrystalline MgZn2 and Mg4Zn7.
Wu et al. [52] estimated elastic constants of MgZn2 by rst-

Table 3
Reduced modulus and hardness determined from loadedisplacement curves.

Reduced modulus (GPa)

Hardness (GPa)

Mg

Mg21Zn25

Mg4Zn7

MgZn2

Mg2Zn11

Zn

35.02 (1.69)
0.54 (0.04)

83.25 (9.15)
4.66 (0.92)

94.09 (3.64)
5.11 (0.46)

87.30 (3.10)
5.08 (0.32)

108.94 (6.39)
3.76 (0.38)

76.42 (3.74)
0.80 (0.05)

C.C. Kammerer et al. / Intermetallics 67 (2015) 145e155

153

Fig. 10. Typical loadedisplacement curves for the intermetallic compounds in the MgeZn system including pure Mg and pure Zn. Serrations in loading segment, annotated by
arrows, are indications discontinuous yielding.

Table 4
Calculated and measured elastic moduli and Poisson's ratio of MgZn2 and Mg4Zn7.

MgZn2

Mg4Zn7

Reference

E (GPa)

Er (GPa)

This work
First-principle calc. [51]
First-principle calc. [52]
This work
First-principle calc. [51]

e
60.53
85.91
e
82.06

e
0.34
0.28
e
0.27

87.3 3.10
64.58
86.20
94.09 3.64
82.16

principle calculations and predicted elastic modulus and Poisson's

ratio of MgZn2. Both papers compared their predictions of elastic
constants with Seidenkranz's measurement of elastic constants of
single crystalline MgZn2 [53]. The results are shown in Table 4. For
ease of comparison, reduced modulus, Er, was calculated based on
given values for elastic modulus and Poisson's ratio in the references using Equation (8). As can be seen, rst principle calculations
are comparable with nanoindentation results for the MgZn2 and
Mg4Zn7.
Mechanical properties are strongly dependent on the crystal
structure of a material. For highly ordered intermetallic compounds, as in the case of the MgeZn intermetallic phases, the
structure (i.e. defect structure) is dependent on the composition. In
fact, the deformation properties of the Laves phase MgZn2 have
been reported to be strongly dependent on the compositions
within the range of solubility [54]. By correlating the composition
vs. position curve t EPMA data with the nanoindentation results
vs. position, using interphase boundaries as reference positions, the
reduced modulus and hardness as a function of composition are
obtained and presented in Fig. 11. Solid lines serve as visual guides

Fig. 11. Reduced modulus and hardness as function of composition.

of the composition dependent response in mechanical properties.

The hardness and reduced modulus increase with increasing Zn
concentration for Mg4Zn7 and MgZn2 but decrease with increasing
concentration for Mg2Zn11. There is a notable gradient in reduced
modulus and hardness as a function of zinc concentration in all
intermediate phases of substantial thickness. While the diffusion
behavior was determined to be unaffected by compositional

154

C.C. Kammerer et al. / Intermetallics 67 (2015) 145e155

changes within the homogeneity ranges of the intermetallic compounds, the mechanical properties were observed to be strongly
concentration dependent.
5. Conclusions
Pure Mg was diffusion bonded to pure Zn in order to nucleate
and grow equilibrium intermetallic compounds; all roomtemperature equilibrium phases formed during diffusion anneal.
Analytical techniques yielded concentration proles, electron
diffraction patterns, and loadedisplacement curves which were
used to characterize the phase constituents and crystallography,
diffusion kinetics, and mechanical properties of the intermetallic
compounds.
The phase in equilibrium with a-Mg was found to be Mg21Zn25
with a trigonal structure and lattice parameters, a 2.5518 nm and
c 0.8713 nm. The crystal symmetry and lattice parameters
established in this study agreed well with those determined by the
recent XRD studies and validated the stoichiometry of the phase.
Mg4Zn7 was determined to have a monoclinic symmetry
(a 2.669 nm, b 0.511 nm, c 1.411 nm, and b 101.85
),
consistent with the structure and symmetry of the Mg4Zn7 reported in literature. This study found Mg4Zn7 to have a 2.4 at%Zn
range of solubility rather than being a xed line compound as
indicated on the phase diagram. This study also found that the
Mg2Zn11 phase has a 1.6 at%Zn range of solubility with a corresponding variation in cubic lattice parameter (0.8415e0.8462 nm).
Kinetic properties were examined through conventional BoltzmanneMatano and Sauer-Freise methods. Observed variation in
molar volume did not inuence the magnitude of interdiffusion
coefcients determined for the MgeZn system. Interdiffusion and
growth rates followed the same trend; MgZn2 (kp 9.3  1014)
was greatest while Mg4Zn7 (kp 1.9  1015) and Mg2Zn11
(kp 1.6  1015) were an order of magnitude smaller, and
Mg21Zn25 (kp 1.6  1018) was substantially less than all other
equilibrium phases. Interdiffusion within each intermetallic phase
did no vary signicantly as a function of composition.
Mechanical properties of the equilibrium intermetallic compounds were determined by nanomechanical testing and reported
for the rst time. Results of this study were comparable with rst
principle calculations of MgZn2 and Mg4Zn7 reported in literature.
While Zn and Mg undergo substantial creep, Mg21Zn25, Mg4Zn7,
MgZn2, and Mg2Zn11 showed insignicant creep during the holding
segment. Serrations in the loading prole of the intermetallic
compounds were evidence of the PLC effect; Mg2Zn11 was less
susceptible to this discontinuous yielding during plastic deformation than the other intermetallic phases. The hardness and reduced
modulus for Mg21Zn25, Mg4Zn7, and MgZn2 were all quite similar.
The average hardness for these phases was within 0.5 GPa, and the
reduced modulus was within 11 GPa of each other. The hardness of
Mg2Zn11 was almost 1 GPa lower, and its reduced modulus was
nearly 15 GPa higher than all other intermetallic phases. The mechanical properties in intermetallic phases, exclusive of Mg21Zn25,
were concentration dependent. The hardness and reduced modulus
increased with Zn concentration for Mg4Zn7 and MgZn2, but
decreased with increasing concentration for Mg2Zn11.
Acknowledgments
Authors would like to express their gratitude to the U.S. Army
Research Laboratory for nancial support of the Advanced Materials Design and Innovative Processing Development project,
W911NF1120020. S.B. would like to thank the nancial support
from the Doctoral Evidence Acquisition Fellowship by Florida International University Graduate School.

References
[1] H.E. Friedrich, B.L. Mordike, Magnesium Technology: Metallurgy, Design Data,
Automotive Applications, Springer, 2006.
[2] J. Hirsch, T. Al-Samman, Superior light metals by texture engineering: optimized aluminum and magnesium alloys for automotive applications, Acta
Mater. 61 (3) (2013) 818e843.
[3] K.U. Kainer, F. Kaiser, Magnesium Alloys and Technologies, Wiley, Weinheim,
2003.
[4] M. Deutsche Gesellschaft fr, K.U. Kainer, Magnesium Alloys and their Applications, Wiley-VCH, Weinheim; Chichester, 2000.
[5] H. Okamoto, Supplemental literature review of binary phase diagrams: CseIn,
CseK, CseRb, EueIn, HoeMn, KeRb, LieMg, MgeNd, MgeZn, MneSm, OeSb,
and SieSr, J. Phase Equilib. Diffusion 34 (3) (2013) 251e263.
[6] J.F. Nie, Precipitation and hardening in magnesium alloys, Metall. Mater.
Trans. A Phys. Metall. Mater. Sci. 43A (11) (2012) 3891e3939.
[7] X. Gao, J.F. Nie, Characterization of strengthening precipitate phases in a
MgeZn alloy, Scr. Mater. 8 (2007) 645.
[8] P. Ghosh, M. Mezbahul-Islam, M. Medraj, Critical assessment and thermodynamic modeling of MgeZn, MgeSn, SneZn and MgeSneZn systems, Calphad
(2012) 28.
[9] R. Cern, G. Renaudin, The intermetallic compound Mg21Zn25, Acta Crystallogr. Sect. C. Cryst. Struct. Commun. 58 (Pt 11) (2002) i154ei155.
[10] Y.R. Yarmolyuk, P.I. Kripyakevich, E.V. Melink, Crystal structure of the compound Mg4Zn7, Sov. Phys. Crystallogr. 20 (3) (1975) 329e331.
[11] T. Sakakura, S. Sugino, Fundamental study on interdiffuison in H.C.P. Alloys.
Part 2: magnesiumezinc system, Memories Suzuka Coll. Technol. 10 (1977)
141e153.
[12] S. Brennan, K. Bermudez, N. Kulkarni, Y.H. Sohn, Diffusion Couple Investigation of the MgeZn System, in Magnesium Technology 2012, John Wiley &
Sons, Hoboken, 2012, pp. 323e327.
[13] A.A. Kodentsov, G.F. Bastin, F.J.J. van Loo, Chapter six e application of diffusion
couples in phase diagram determination, in: J.C. Zhao (Ed.), Methods for Phase
Diagram Determination, Elsevier Science Ltd, Oxford, 2007, pp. 222e245.
[14] S.K. Das, Y.-M. Kim, T.K. Ha, I.-H. Jung, Investigation of anisotropic diffusion
behavior of Zn in HCP Mg and interdiffusion coefcients of intermediate
phases in the MgeZn system, Calphad (2013) 51.
[15] A. Mostafa, M. Medraj, On the atomic interdiffusion in Mge{Ce, Nd, Zn} and
Zne{Ce, Nd} binary systems, J. Mater. Res. 29 (13) (2014) 1463e1479.
[16] G. Tammann, H.J. Rocha, Die Diffusion zweier Metalle ineinander unter Bildung intermetallischer Verbindung, Z. fr Anorg. Allg. Chem. 199 (1) (1931)
289.
[17] J.B. Clark, F.N. Rhines, Layer growth in AleMgeZn, Trans. Am. Soc. Metals 51
(1959) 199e221.
[18] L.A. Giannuzzi, F.A. Stevie, A review of focused ion beam milling techniques
for TEM specimen preparation, Micron 30 (3) (1999) 197e204.
[19] B. Kempshall, L.A. Giannuzzi, In-situ lift-out FIB specimen preparation for TEM
of magnetic materials, Microsc. Microanal. 8 (Supplement S02) (2002)
390e391.
[20] J.M. Philibert, Atom Movements e Diffusion and Mass Transport in Solids, EDP
Sciences, 1991.
[21] K. Huang, Y. Park, A. Ewh, B.H. Sencer, J.R. Kennedy, K.R. Coffey, Y.H. Sohn,
Interdiffusion and reaction between uranium and iron, J. Nucl. Mater. 424
(2012) 82e88.
[22] C. Wagner, The evaluation of data obtained with diffusion couples of binary
single-phase and multiphase systems, Acta Metall. 17 (2) (1969) 99e107.
[23] M.A. Dayananda, Average effective interdiffusion coefcients in binary and
multicomponent alloys, Diffusion Defect Forum (1993) 521e536.
[24] F. Sauer, V. Freise, Diffusion in binaren gemischen mit volumenanderung,
Z. fur Elektrochem. 66 (4) (1962) 353e363.
[25] H. Mehrer, Diffusion in Solids: Fundamentals, Methods, Materials, Diffusioncontrolled Processes, Springer-Verlag, , Berlin, 2007.
[26] M.A. Dayananda, C.W. Kim, Metallurgical Mater. Trans. A (1979) 1333e1339.
[27] W.C. Oliver, G.M. Pharr, An improved technique for determining hardness and
elastic modulus using load and displacement sensing indentation experiments, J. Mater. Res. 7 (06) (1992) 1564e1583.
[28] W.C. Oliver, G.M. Pharr, Nanoindentation in materials research: past, present,
and future, MRS Bull. 35 (11) (2010) 897e907.
[29] W.C. Oliver, G.M. Pharr, Measurement of hardness and elastic modulus by
instrumented indentation: advances in understanding and renements to
methodology, J. Mater. Res. 19 (01) (2004) 3e20.
[30] A. Paul, M.J.H. van Dal, A.A. Kodentsov, F.J.J. van Loo, The Kirkendall effect in
multiphase diffusion, Acta Mater. 52 (3) (2004) 623e630.
[31] P. Carpenter, Electron-probe microanalysis (EPMA): an overview for beginners and a status report for experts, Microsc. Microanal. 14 (Suppl. S2) (2008)
1150e1151.
[32] Y. Khan, Dynamic temperature crystallization behavior of amorphous and
liquid Mg70zn30 Alloy, J. Mater. Sci. 24 (3) (1989) 963e973.
[33] X. Gao, J.F. Nie, Structure and thermal stability of primary intermetallic particles in an MgeZn casting alloy, Scr. Mater. 57 (7) (2007) 655e658.
 ska, S. Delno, A. Saccone, The MgeZneSi system:
[34] S. De Negri, M. Skroban
constitutional properties and phase formation during mechanical alloying,
Intermetallics 18 (9) (2010) 1722e1728.
[35] T. Takei, The equilibrium diagram of the system magnesiumezinc, Kinzoku-no

C.C. Kammerer et al. / Intermetallics 67 (2015) 145e155

Kenkyu 6 (1929) 177e183.
[36] F. Laves, Zur Konstitution der MagnesiumeZinkeLegierungen, Naturwissenschaften 27 (26) (1939) 454e455.
[37] J.B. Clark, F.N. Rhines, Central region of the magnesiumezinc phase diagram,
J. Metals 9 (1957) 425e430.
[38] H. Okamoto, Comment on MgeZn (magnesiumezinc), J. Phase Equilib. 15 (1)
(1994) 129.
[39] S. Samson, The crystal structure of Mg2Zn11: isomorphism between Mg2Zn11
and Mg2Cu6Al5, Acta Chem. Scand. 3 (1949) 835e843.
[40] L. Vegard, Die Konstitution der Mischkristalle und die Raumfllung der
Atome, Z. fr Phys. 5 (1) (1921) 17.

[41] K. Eichler, H. Kubsch, T. Mller, P. Pauer, Anderung

von Verformungseitsbergenschaften der intermetallischen Verbindung MgZn2 im Homogenita
eich, Kristall Tech. 11 (11) (1976) 1185.
[42] T. Ohba, Y. Kitano, Y. Komura, The charge-density study of the Laves phases,
MgZn2 and MgCu2, Acta Crystallogr. Sect. C 40 (1) (1984) 1e5.
[43] M.J.H. van Dal, A.A. Kodentsov, F.J.J. van Loo, Formation of CoeSi intermetallics
in bulk diffusion couples. Part II. Manifestations of the Kirkendall effect
accompanying reactive diffusion, Intermetallics 9 (2001) 451e456.
[44] H. Li, A.H.W. Ngan, M.G. Wang, Continuous strain bursts in crystalline and
amorphous metals during plastic deformation by nanoindentation, J. Mater.
Res. 20 (11) (2005) 3072e3081.
[45] A. Yilmaz, The Portevin-Le Chatelier effect: a review of experimental ndings,

155

Sci. Technol. Adv. Mater. 12 (6) (2011) 16.

th, Z. Kova
cs, A. Juha
sz, G. Be
rces, J. Lendvai, Kinematic
[46] N.Q. Chinh, G. Horva
and dynamic characterization of plastic instabilities occurring in nano- and
microindentation tests, Mater. Sci. Eng. A 409 (1e2) (2005) 100e107.
th, N.Q. Chinh, J. Gubicza, J. Lendvai, Plastic instabilities and dislo[47] G. Horva
cation densities during plastic deformation in AleMg alloys, Mater. Sci. Eng. A
445e446 (0) (2007) 186e192.

user, Portevin[48] H. Dierke, F. Krawehl, S. Graff, S. Forest, J. Sachl,
H. Neuha
LeChatelier effect in AleMg alloys: inuence of obstacles e experiments and
modelling, Comput. Mater. Sci. 39 (2007) 106e112.
[49] G.E. Dieter, Mechanical Metallurgy, McGraw Hill, , Boston, 1986.
[50] P. Rodriguez, Serrated plastic ow, Bull. Mater. Sci. 6 (4) (1984) 653e663.
[51] Y.-P. Xie, Z.-Y. Wang, Z.F. Hou, The phase stability and elastic properties of
MgZn2 and Mg4Zn7 in MgeZn alloys, Scr. Mater. 68 (7) (2013) 495e498.
[52] M.-M. Wu, L. Wen, B.-Y. Tang, L.-M. Peng, W.-J. Ding, First-principles study of
elastic and electronic properties of MgZn2 and ScZn2 phases in MgeSceZn
alloy, J. Alloys Compd. 506 (2010) 412e417.
[53] T. Seidenkranz, E. Hegenbarth, Single-crystal elastic constants of MgZn2 in the
temperature range from 4.2 to 300 K, Phys. Status Solidi (A) 33 (1) (1976)
205e210.
[54] T.H. Mller, P. Pauer, Yield strength of the monocrystalline intermetallic
compound MgZn2, Phys. Status Solidi (a) 40 (2) (1977) 471e477.