Journal of Alloys and Compounds 596 (2014) 63–68

Contents lists available at ScienceDirect

Journal of Alloys and Compounds

journal homepage: www.elsevier.com/locate/jalcom

First-principles study of the properties of Li, Al and Cd doped Mg alloys

Lijuan Zhou a, Kehe Su a,⇑, Yanli Wang a, Qingfeng Zeng b, Yulong Li c
a
Key Laboratory of Space Applied Physics and Chemistry of the Ministry of Education, School of Natural and Applied Sciences, Northwestern Polytechnical University, Xi’an,
Shaanxi 710072, People’s Republic of China
b
National Key Laboratory of Thermostructure Composite Materials, Northwestern Polytechnical University, Xi’an, Shaanxi 710072, People’s Republic of China
c
School of Aeronautics, Northwestern Polytechnical University, Xi’an, Shaanxi 710072, People’s Republic of China

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

Article history: The major mechanical and electronic properties of the Li, Al and Cd doped Mg solid solutions were inves-
Received 16 December 2013 tigated by employing the first principles density functional theory methods. The solute atoms were cho-
Received in revised form 24 January 2014 sen since they have a larger range of solubility. Each of the alloy models consists of fifteen Mg atoms and
Accepted 27 January 2014
one solute atom. The elastic parameters Cij were evaluated, and the bulk modulus (B), shear modulus (G),
Available online 2 February 2014
Young’s modulus (E) and Poisson’s ratio (m) were further derived. The results show that the B/G ratio (an
estimation of the metal ductility) of Mg, Mg–Al, Mg–Li and Mg–Cd are 1.92, 1.80, 1.47 and 2.17, respec-
Keywords:
tively, which indicates that the Mg–Al and Mg–Cd alloys are ductile and the Mg–Li alloy is slightly brittle.
First-principles calculation
Mg alloy
For the Mg–Cd alloy, its plasticity might be better than that of the others due to a larger Poisson’s ratio
Mechanical properties and lower values of C11  C12 and E. The tensile test simulation of the Mg metal and Mg–Al, Mg–Li and
Strain–stress Mg–Cd alloys reaches the strengths of 5.05, 5.32, 5.28 and 4.89 GPa, respectively.
Ó 2014 Elsevier B.V. All rights reserved.

1. Introduction the tensile strength of Mg alloys by adding one and three Al atoms
into the super-cell of Mg54 and reported that the tensile strength of
Magnesium alloys, the lightweight metallic structural materi- Mg51Al3 is obviously increased compared with Mg54 and Mg53Al.
als, have been used broadly in the fields of aerospace, automotive, Additionally, Counts et al. [18,19] predicted the structure and
computer, communications and home appliances industry owing properties of the bcc Mg–Li alloys. Since the solute elements of
to their high specific strength, stiffness and low density known Al, Li and Cd have a larger range of solubility [9,20], and have sim-
well as the twenty-first century green engineering materials ilar atomic radii [21] in different rows and columns of the periodic
[1–4]. The improvement of the mechanical properties attracts table, this paper will, therefore, examine some of the mechanical,
great interest mostly by the addition of different alloy elements the electronic properties and the stabilities of the stable hcp Mg al-
[5–7]. Three mechanisms in the increase of the strength of Mg al- loys at a relatively higher concentration of the solute atoms.
loys were proposed including solid solution strengthening, precip-
itation strengthening and dispersion strengthening. Among those, 2. Computational method
the intermetallic phases are believed to be the most significant
[8,9]. Experimental and theoretical investigations mainly on the The first-principles calculations were performed with the Cambridge Serial To-
tal Energy Package (CASTEP) code in the Materials Studio software [22,23]. The den-
intermetallic phases were thus broadly carried out, e.g. in Refs.
sity functional theory (DFT) of the plane-wave pseudopotential method [24,25] was
[8,10–15]. For the solid solution strengthening, Cáceres and Rovera chosen. The generalized gradient approximation (GGA) of Perdew–Burke–Ernzerhof
[5] detected that the hardness of Mg alloys, forming a hexagonal (PBE) [26,27] exchange–correlation functionals were employed combined with the
close-packed (or hcp) structure, would increase along with increas- finite basis set correction [28] and the Pulay scheme of density mixing [29]. The
models were designed as a super-cell containing 16 atoms with three dimensional
ing the Al content in the solid solution within 1–8 wt%. Theoreti-
periodic conditions. The cell structure was full optimized using the Broyden–Fletch-
cally, Ganeshan et al. [16] investigated the elastic properties of er–Goldfarb–Shanno (BFGS) [30] minimization technique, where the cut-off energy
Mg by doping 12 alloying elements at a concentration of for the plane wave expansion was 420 eV, the Brillouin zone sampling was carried
2.77 at%. The lattice structure, bulk and shear modulus, elastic con- out using a 7  7  4 set of the Monkhorst–Pack mesh [31] and the SCF energy tol-
stant and ductility were discussed. Wang et al. [17] investigated erance was set as 5.0  107 eV. For all calculations, the further thresholds were:
the total energy being converged into 5.0  106 eV/atom, the finite atomic dis-
placement [32] being less than 5.0  104 Å and the maximum force being less than
0.01 eV/Å. In the elastic constants evaluations, the additional GGA-PW91 [33]
⇑ Corresponding author. Tel.: +86 29 88431672 (O); fax: +86 29 88493325. method was examined in order to compare the accuracy of the calculations. In
E-mail address: [email protected] (K. Su). the tensile simulations, uniaxial strain was introduced into the super-cell of the

http://dx.doi.org/10.1016/j.jallcom.2014.01.199
0925-8388/Ó 2014 Elsevier B.V. All rights reserved.
64 L. Zhou et al. / Journal of Alloys and Compounds 596 (2014) 63–68

equilibrium configuration with the minimum energy. The tensile test was simu- less than 1%, an error that is acceptable for the theoretical calcula-
lated by gradually increasing the dimension along the z-direction at the tensile
tions [16]. Our results also show that the c/a ratio is 1.6399 and the
strain increment of 2% of the system. The atomic positions of the new configuration
were thus fully relaxed at each strain stage. The next initial strain configuration
volume of the primitive cell is 46.079 Å3, which is consistent with
would start from the former relaxed configuration. The procedure continued until the experimental and calculated results [16,37–41]. It should be
the convergence criterion was unable to meet. noted that the experimental lattice constants are measured at
room temperature and the theoretical results correspond to the
temperature at zero Kelvin. However, as has been examined in
3. Results and discussions
Ref. [42], the computed results with the GGA functionals at 0 K
are also in excellent agreement with the room temperature exper-
3.1. Structure and lattice constant
iments. This might represent the fact that the lattice constants
would not be significantly changed for temperatures rising from
The models of Mg and Mg alloys are shown in Fig. 1. The Mg
0 K to the room temperature.
crystal is a hexagonal close-packed (hcp) structure with the space
Compared with Mg metal, the cell volumes of the alloys de-
group P63/mmc. A primitive cell contains 2 atoms and thus the
creased. This should be mainly due to the smaller atomic radii of
super-cell model contains eight units. The Mg alloy models involve
aluminum and lithium, and the larger electronegativity of cad-
15 Mg atoms and one solute atom that takes the place of a magne-
mium. The volumes of the Mg–Al, Mg–Li and Mg–Cd alloy cells
sium atom. This implies that the concentration is 6.25 at% which is
are 45.002, 45.542 and 45.679 Å3, respectively. It is interesting that
within the range of the solid solution phase for a large region of
the c/a ratios increased slightly by adding Al and Cd atoms at
temperatures according to the equilibrium phase diagrams of the
1.6448 and 1.6430, respectively, but that decreased slightly by Li
Mg–Al, Mg–Li and Mg–Cd alloys [34–36].
at 1.6257.
As shown in Table 1, the lattice constants of Mg crystal have
been widely determined (e.g., in Refs. [37–39]) by a = b = 3.2026–
3.2099 Å and c = 5.1997–5.2108 Å. The theoretical calculations 3.2. Heat of formation and cohesive energy
[16,40,41] reproduced the results by a = b = 3.187–3.195 Å and
c = 5.1777–5.1939 Å. Our results, a = b = 3.1915 Å and c = 5.2338 Å, The alloying ability or stability may be estimated with either
consist well with the experiments [37–39] by a deviation of the heat of formation (DH, or equivalently the internal energy for

Fig. 1. Crystal structure of Mg and Mg–X (X = Al, Li, Cd) alloys.

Table 1
Lattice constant, cohesive energy and the heat of formation of Mg and its 6.25 at% alloys obtained with PBE density functional theory.

Crystal Method a (Å) c (Å) c/a V (Å3) Ecoh (eV/atom) DH (eV/atom)

Mg (this work) GGA-PBE 3.1915 5.2338 1.6399 46.152 1.4915 0.0193
Mg (exp.)a 3.2026 5.1997 1.6236
Mg (exp.)b 3.2099 5.2108 1.6234
Mg (exp.)c 3.2095 5.2106 1.6235
Mg (calc.)d GGA-PW91 3.190 5.1777 1.6231
Mg (calc.)e GGA 3.195 5.178 1.621
Mg (calc.)f GGA-PW91 3.187 5.1939 1.6297
Mg–Al (this work) GGA-PBE 3.1617 5.2004 1.6448 45.002 1.6260 0.0057
Mg–Al (calc.)e 3.183 5.177 1.627
Mg–Li (this work) GGA-PBE 3.1866 5.1804 1.6257 45.542 1.5288 0.0285
Mg–Li (exp.)g 3.2071 5.2011 1.6218
Mg–Li (calc.)e GGA 3.194 5.150 1.613
Mg–Cd (this work) GGA-PBE 3.1787 5.2227 1.6430 45.679 1.4599 0.0352
a
Experimental results at room temperature from Ref. [37].
b
Experimental results at room temperature from Ref. [38].
c
Experimental results at room temperature from Ref. [39].
d
Theoretical value from Ref. [40].
e
Theoretical value from Ref. [16]. The data for alloys are at a concentration of 2.77 at% of the alloying elements.
f
Theoretical value from Ref. [41].
g
Experiments at 4.66 at% from Ref. [43].
 L. Zhou et al. / Journal of Alloys and Compounds 596 (2014) 63–68 65

condensed phases) or the cohesive energy (Ecoh) defined [44,45] thus have a larger structural stability. This result is also consistent
approximately as: with the calculated cohesive energies in this work.

1  AB 
DH ¼ Etot  xEAsolid  yEBsolid ð1Þ
xþy 0.6

Density of States (states / eV)

0.4 Mg 3s
3s
0.2 2p 3p
Mg 2p
1  AB 
Ecoh ¼ Etot  xEAatom  yEBatom ð2Þ 0.0
0.6
xþy
0.4 Al 2s
3s
In the equations, x and y refer to the numbers of the respective 0.2 2p 3p
Al 2p
A and B atoms in the super-cell. EAB tot represents the total energy of 0.0
12
the combined system. EAsolid or EBsolid represents the energy per atom
8
of A or B in its crystal, while EAatom or EBatom represents the energy of Mg-Al
4
the isolated A or B atom. The energy of the isolated atom was cal-
0
culated in a cubic box (super cell) at the lattice constant of 10 Å.
-50 -40 -30 -20 -10 0 10
The total energy convergence is 1.0  105 eV/atom and the
Energy (eV)
8  8  8 k-point is employed in the cohesive energy calculations.
0.6

Density of States (states / eV)

It should also be noted that the present calculations have not in-
0.4
volved the zero point energy and thermal corrections due to the Mg 3s
3s
0.2 Mg 2p
2p 3p
limitation of our computational resources in the vibrational analy-
0.0
ses. This must lead to a significant error in the enthalpy evalua- 0.6
tions, especially for the Mg–Li model that has a larger zero point 0.4 Li 2s
1s 2s
energy due to the small atomic weight of Li atom. 0.2 Li 2p

The results listed in Table 1 shows that the DH for the pure Mg 0.0
12
metal is a negative value 0.0193 eV/atom. This should be consid- 8
ered as the systematic error of the present calculations since the 4
Mg-Li

ideal value should be zero by definition. By comparing the DH val- 0

ues of Mg–Li and Mg–Cd alloys, the data are more negative. This -50 -40 -30 -20 -10 0 10
represents that the alloying process is definitely spontaneous be- Energy (eV)
cause the entropy change is positive. These results are consistent
0.6
with the equilibrium phase diagrams of Mg–Li [34] and Mg–Cd
Density of States (states / eV)

0.4 Mg 3s
3s
[35], in which each of the systems exhibits an alloy of unsaturated 2p 3p
Mg 2p
0.2
solid solution. The DH for the Mg–Al alloy is less negative (or is a
0.0
small positive value 0.0136 eV if the systematic error is taken into 0.6
account). The equilibrium phase diagram of Mg–Al [36] alloy 0.4 Cd 5s
5s
Cd 4p
5p
shows that the solid solution alloy is stable between 310 and 0.2 Cd 4d
550 °C, implying that the Gibbs free energy is negative within this 0.0
12
temperature region but is positive below 310 °C due to the positive 8
Mg-Cd
DH. Qualitatively, our results reproduce well the facts within the 4
Mg–Al alloy phase diagram. 0
It is also shown in Table 1 that the cohesive energies of Mg and -50 -40 -30 -20 -10 0 10
Mg–Al, Mg–Li, Mg–Cd alloys are larger negative values by 1.4915, Energy (eV)
1.6260, 1.5288 and 1.4599 eV/atom, respectively. This indi- 12
cates that the models are much more stable compared with their Mg-Cd

isolated atoms. Among those, Mg–Al alloy is most stable by the 8

lowest cohesive energy. 4

5.75
0
12
Density of States (states / eV)

Mg-Li
3.3. Electronic structure
8

The partial and total density of states (DOS) of Mg and its alloys 4
5.71
near the Fermi level are plotted in Fig. 2. It is shown that the partial 0
12
DOS represent the differences of the bands, especially contributed Mg-Al
from the different valence orbitals of Mg, Al, Li and Cd elements. 8
But the features of the total DOS close to the Fermi level of the
4 6.47
alloys are similar to those of the pure Mg metal mainly from
the valence s and p orbitals of Mg with the dominant concentra- 0
12
Mg
tion. However, the total DOS of Mg, Mg–Al, Mg–Li and Mg–Cd have
different values by 5.81, 6.47, 5.71 and 5.75 states/eV on the Fermi 8
level, consistent qualitatively well with the number of the valence 4
electrons of the system. Naturally, the DOS for Mg–Cd has a sharp 5.81
peak at 9 eV mainly from the contribution of the full occupied 4d 0
-8 -4 0 4 8
orbital of Cd that have not mixed with the valence orbitals. Since
Energy (eV)
the DOS value at the Fermi level may estimate the relative bonding
electron numbers and a higher number of bonding electrons repre- Fig. 2. Partial and total density of states of Mg and its alloys near the Fermi level
sents a stronger charge interaction [46,47], the Mg–Al alloy should obtained with DFT-PBE (the Fermi level is plotted as 0 eV).
66 L. Zhou et al. / Journal of Alloys and Compounds 596 (2014) 63–68

3.4. Elastic properties elastic constants for Mg metal are consistent with the experimen-
tal [43,54] and the reported theoretical values [16,42]. The elastic
The elastic stiffness constants Cij (i,j = 1, 2, 3, 4, 5, 6), forming a constants Cij of Mg obtained with the PBE and PW91 functionals
6  6 symmetric tensor matrix, determine the response of the have the average absolute deviations (AADs) of 1.272 GPa and
strain to external forces [48,49] that can be obtained by applying 1.820 GPa, respectively, compared with the experimental data
a set of finite homogeneous deformations to the unit cell and by [54] at 0 K. However, the respective data have the AADs of
optimizing the internal atomic freedoms to calculate the stresses. 2.016 GPa and 0.764 GPa compared with the experimental data
For the hexagonal crystal of Mg and its solid solution alloys, there [54] at 298 K, and also have the AADs of 2.011 GPa and
are six different elastic coefficients C11, C12, C13, C33, C44 and C66, 0.927 GPa compared with the 298 K experiments of Ref. [43]. Inter-
and five of which are independent due to C66 = (C11  C12)/2. The estingly, the calculation of the constants of Mg metal [42] with an
results obtained with both DFT-PBE and DFT-PW91 are listed in Ta- approach by using the stress-based least-squares fitting method of
ble 2. LePage and Saxe [56] resulted in a larger AAD value of 2.818 GPa
The Voigt’s (GV) [50], Reuss’s (GR) [51] approximations for max- and an even smaller AAD value of 0.598 GPa with the same GGA
imum and minimum values of the shear moduli and the Hill’s (G) PW91 functionals compared with the experiments of Ref. [54] at
approximation by the average value of GV and GR for the hexagonal 0 K and 298 K, respectively. Although the further discussions will
crystal are given as follows [50–53]: emphasize the results from the PW91 functionals, we may not con-
clude here that the PW91 method is superior in reproducing the
GV ¼ ½C 11 þ C 33  2C 13 þ 6C 44 þ 5C 66 =15 ð3Þ
Mg elastic constants because the theoretical computations have
not yet corrected the thermal influence. The much better results
GR ¼ 15=½8S11 þ 4S33  4S12  8S13 þ 6S44 þ 3S66 Þ ð4Þ
for Mg metal by comparing the ‘298 K’ experiments must be an
accidental case that needs to be further investigated.
G ¼ ðGV þ GR Þ=2 ð5Þ
The elastic moduli (B, G, E) of Mg metal and Mg–Al, Mg–Li and
where the compliance constants Sij are: S11 + S22 = C33/C, Mg–Cd alloys are listed in Table 2 and the results from the PW91
S11  S22 = 1/(C11  C12), S13 = C13/C, S33 = (C11 + C12)/C, S44 = 1/C44, functionals are shown in Fig. 3. The results show that the elastic
S66 = 1/C66 and C = (C11 + C12)C33  2(C13)2. moduli of the pure Mg metal consist very well with both of the
The Young’s modulus and Poisson’s ratio can be developed by experimental [43,54] and theoretical results [16,42] within a devi-
the shear modulus G and the bulk modulus B [50–53]: ation of 6%. However, these data have an error as much as 26%
compared with those, for example, from Ref. [41]. It should be
9GB
E¼ ð6Þ noted that the data from Ref. [41], obtained with a different Per-
G þ 3B dew functional, have a maximum error of about 22% compared
with all those from Refs. [16,42,43,54]. Since our results are more
3B  2G
m¼ ð7Þ close to the experiments, the data obtained in the present work
2ð3B þ GÞ
should be more reliable.
It is shown that all the elastic constants in Table 2 satisfy the For the Mg–Al alloy, all the bulk, shear and Young’s moduli are
Born stability criteria C11 > 0, C11  C12 > 0, C44 > 0 and (C11 + C12)- predicted to be improved by 1.58, 2.11 and 4.88 GPa, respectively,
C33  2(C12)2 > 0 for the hexagonal structure [55], representing that compared with those of the pure Mg metal. This is different from
the Mg metal and its alloys are mechanical stable. Our calculated the calculations of Ref. [16] where a smaller concentration of Al

Table 2
Elastic properties of Mg and its Al, Li and Cd alloys calculated with DFT-PBE and DFT-PW91. (Cij are the elastic stiffness constants, B is the bulk modulus, G is the shear modulus, E
is the Young’s modulus, m is the Poisson’s ratio and B/G is the ratio of bulk/shear moduli. All the data are in GPa except m and B/G. The C11, C12, C13, C33, C44 and B are given directly
in the calculations. The G, E, m and underlined data are calculated presently with the equations listed in the text of this paper by using the original elastic constants from
references.)

Crystal Method C11 C33 C44 C12 C13 B G E m B/G

Mg (exp.)a 63.48 66.45 18.42 25.94 21.70 36.89 19.33 49.38 0.28 1.91
Mg (exp.)b 59.43 61.64 16.42 25.6 21.4 35.25 17.36 44.75 0.29 2.03
Mg (exp.)c 59.50 61.55 16.35 26.12 21.805 35.55 17.21 44.45 0.29 2.07
Mg (calc.)d 63.5 66 19.3 24.85 20.0 35.83 18.5 47.4 0.28 1.93
Mg (calc.)e 49.7 59.7 16.1 24.4 26.8 35.0 14.3 37.7 0.32 2.45
Mg (calc.)f GGA 59.3 61.6 14.2 25.8 21.0 35 16 42 0.30 2.19
LDA 66 71.6 16.1 31.2 23.7 40 18 47 0.30 2.22
Mg (this work) PBE 59.89 67.19 19.01 26.39 20.71 35.84 18.77 47.95 0.28 1.91
PW91 59.41 62.58 18.53 25.77 20.82 35.14 18.31 46.80 0.28 1.92
Mg–Al (calc.)d 65.6 69.0 13.6 25.9 19.3 36.6 17.9 46.2 0.28 1.98
Mg–Al (this work) PBE 71.64 74.87 16.62 25.27 17.24 37.50 21.32 53.77 0.26 1.76
PW91 70.24 69.96 15.88 24.92 17.63 36.72 20.42 51.68 0.27 1.80
c
Mg–Li (exp.) 57.95 59.94 15.87 25.39 21.26 34.62 16.74 43.26 0.29 2.07
Mg–Li (calc.)d 58.9 54.0 15.0 24.5 23.2 34.84 16.17 42.00 0.30 2.15
Mg–Li (this work) PBE 70.63 68.02 20.82 16.15 17.11 34.44 24.18 58.79 0.22 1.42
PW91 68.64 62.90 20.16 16.45 17.90 33.84 22.99 56.23 0.22 1.47
Mg–Cd (this work) PBE 70.12 80.98 9.07 25.65 14.01 36.50 16.97 44.08 0.30 2.15
PW91 68.87 76.22 8.91 25.51 14.28 35.78 16.45 42.80 0.30 2.17
a,b
Experiments from Ref. [54]. The data for note ‘a’ are obtained at 0 K and those for ‘b’ are at 298 K.
c
Experiments from Ref. [43] at 298 K. The data for Mg–Li alloy is at a concentration of 7.0 at% of Li atoms.
d
Theoretical values from Ref. [16]. The data for alloys are at a concentration of 2.77 at% of the alloying elements.
e
Theoretical values from Ref. [41].
f
Theoretical values from Ref. [42].
 L. Zhou et al. / Journal of Alloys and Compounds 596 (2014) 63–68 67

at 2.77 at% was examined within which the shear and the Young’s 70
bulk modulus (B)
moduli are slightly decreased. For the Mg–Li alloy, the bulk modu- shear modulus (G)
lus is slightly reduced by 1.3 GPa but both of the shear and Young’s 60 Young's modulus (E )
56.23
moduli are improved. Especially, the Young’s modulus is predicted
significantly increased by 9.43 GPa. This is different from the 51.68
50

Elastic moduli / GPa

experimental [43] and the theoretical [16] results. The reason must 46.80
42.80
be mainly from the different solute concentrations and, probably,
40
from the computational deficiencies (e.g., the accuracy of the cur- 35.13
36.72 35.78
33.84
rent functionals). For the Mg–Cd alloy, the present work predicts
that the solution of Cd element into Mg does not significantly 30
changed the bulk modulus (i.e., 35.78 vs. 35.14 GPa) but both the 22.99
20.42
shear and Young’s moduli are reduced by 1.86 and 4.00 GPa, 20 18.31
16.45
respectively.
The ductility and brittleness of a metal or an alloy can be esti- 10
mated by using the ratio of the bulk/shear moduli (B/G) suggested
by Pugh [57] by a critical value of B/G = 1.75. The data in Table 2 0
show that the B/G value of pure Mg metal is 1.92, consistent well Mg Mg-Al Mg-Li Mg-Cd
with either the experiments [54] or the calculations [16]. The B/G
Fig. 3. Elastic moduli of Mg and its alloys computed with GGA PW91.
ratios for Mg–Al, Mg–Li and Mg–Cd alloys are 1.80, 1.47 and
2.17, respectively, showing that the Mg–Al and Mg–Cd alloys have
the B/G greater than 1.75, which means they are ductile and the
Mg–Li alloy is slightly brittle. This is consistent with the experi- 70
C11-C12 / GPa
ments [43] and the previous theoretical calculations [16]. It should Young's modulus (E) / GPa
be noted that the Mg–Cd alloy has a largest B/G ratio of 2.17, indi- 60
56.23
cating that Mg–Cd may have a good ductility among the three Mg 51.68 52.18
alloys. 50
46.80 45.31
In addition to B/G ratio, it is reported that the C11  C12 and 43.36 42.80
Young’s modulus (E) values are also responsible for the plasticity 40
of a material [40,58] where a smaller value represents a better 33.64
plasticity. The C11  C12 values and E are shown in Fig. 4 that shows
30
the plasticity of the Mg–Al and Mg–Li alloys are decreased by the
addition of the alloy elements with a concentration of 6.25 at%.
20
Since the C11  C12 value is increased while the Young’s modulus
(E) of the Mg–Cd alloy is decreased, the plasticity of which cannot
be predicted solely by these two properties. 10
The Poisson’s ratio (m) reflects the stability of a crystal against
shear usually ranging from 1 to 0.5. The respective Poisson’s 0
Mg Mg-Al Mg-Li Mg-Cd
ratios of Mg, Mg–Al, Mg–Li and Mg–Cd are 0.28, 0.27, 0.22 and
0.30, consistent well with the available data as listed in Table 2. Fig. 4. Values of C11  C12 and Young’s modulus (E) of Mg and Mg alloys.
Among these materials, the Mg–Cd alloy has the largest Poisson’s
ratio, showing that it should be of best plasticity.
It should also be noted that the arithmetic average (Eq. (4)) of
6
the elastic bounds GV (Eq. (2)) and GR (Eq. (3)) obtained with the Mg
data of the single crystalline models have been compared with Mg-Al
Mg-Li
the respective values of the experiments with the randomly ori-
Mg-Cd
ented polycrystalline aggregates. The bridging and validity have,
for example, been presented reasonable in Ref. [18] and reviewed 4
in Ref. [60] at least for the cubic Mg alloys.
Stress / GPa

3.5. Tensile test

The theoretical strain–stress relationships of Mg and its Al, Li 2

and Cd alloys are obtained with the tensile simulations and the re-
sults are illustrated in Fig. 5. It is shown that the stresses rise rap-
idly with the increment of strains at the first stage. For the strains
at about 14%, the stresses increase slowly with the increase of
strains. After the strain of 8%, the stresses of Mg–Al and Mg–Li al- 0
0.0 0.1 0.2 0.3
loys are always higher than that of pure Mg metal but the stress of
Strain
Mg–Cd is lower. The test reaches the tensile strength of Mg, Mg–Al,
Mg–Li and Mg–Cd at 5.80, 5.19, 5.55 and 5.40 GPa where the com- Fig. 5. Calculated stress–strain relation of Mg and Mg alloys with tensile
putational test terminated due to the convergence criterion was simulation.
unable to meet except for Mg metal that reaches a smaller value
after the maximum stress of 5.80 GPa at around 30% strain. This
is different from the previous theoretical calculations of Ref. [17] stressed that the theoretical simulations imply a tensile process
where the stress of Mg54 reaches the maximum value 7.47 GPa at without any defect in the material at zero Kelvin and the plastic
the strain of 18% by using the PW91 functionals. It should be deformation is not considered. This, naturally, results in a much
68 L. Zhou et al. / Journal of Alloys and Compounds 596 (2014) 63–68

higher (at least over an order) stress or strength compared with the [15] J. Wro9 bel, L.G. Hector Jr., W. Wolf, S.L. Shang, Z.K. Liu, K.J. Kurzydłowski, J.
Alloys Comp. 512 (2012) 296–310.
experimental detections [59] since the experiments apply the ten-
[16] S. Ganeshan, S.L. Shang, Y. Wang, Z.K. Liu, Acta Mater. 57 (2009) 3876–3884.
sile force within a limited time and the defects are unavoidable. [17] C. Wang, P.D. Han, L. Zhang, C.L. Zhang, X. Yan, B.S. Xu, J. Alloys Comp. 482
(2009) 540–543.
[18] W.A. Counts, M. Friák, C.C. Battaile, D. Raabe, J. Neugebauer, Phys. Status Solidi
4. Conclusions B 245 (2008) 2630–2635.
[19] W.A. Counts, M. Friák, D. Raabe, J. Neugebauer, Acta Mater. 57 (2009) 69–76.
In this work, the mechanical properties of Mg metal and its Al, [20] S.C. Zhang, H.Q. Duan, O.Z. Cai, B.K. Wei, H.T. Lin, W.C. Chen, Foundry 50 (2001)
310–315.
Li and Cd solute alloys at a concentration of 6.25 at% were calcu- [21] D.R. Lide, CRC Handbook of Chemistry and Physics, 77th ed., CRC Press, New
lated theoretically using first-principles with DFT-PBE/plane-wave York, 1996–1997.
and partially with DFT-PW91/plane-wave functionals. The results [22] M.D. Segall, P.J.D. Lindan, M.J. Probert, C.J. Pickard, P.J. Hasnip, S.J. Clark, M.C.
Payne, J. Phys.: Condens. Matter 14 (2002) 2717–2744.
show that the lattice parameters and volume of Mg tend to de-
[23] J.C. Stewart, D.S. Matthew, J.P. Chris, J.H. Phil, I.J.P. Matt, R. Keith, C.P. Mike,
crease by the addition of the alloying elements. The bulk modulus, Comp. Crystallogr. 220 (2005) 567–570.
shear modulus, and the Poisson’s ratio show that the Mg–Al and [24] P. Hohenberg, W. Kohn, Phys. Rev. B 136 (1964) 864–871.
[25] W. Kohn, L.J. Sham, Phys. Rev. A 140 (1965) 1133–1138.
Mg–Cd alloys are ductile while the Mg–Li alloy is slightly brittle.
[26] J.P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 77 (1996) 3865–3868.
In examination of the C11  C12 and the Young’s modulus values, [27] M. Marlo, V. Millman, Phys. Rev. B 62 (2000) 2899–2907.
we found that the plasticity of Mg would decrease with the addi- [28] G.P. Francis, M.C. Payne, J. Phys.: Condens. Matter 2 (1990) 4395–4404.
tion of the Al and Li elements. The Poisson’s ratio of the Mg–Cd al- [29] B. Hammer, L.B. Hansen, J.K. Nørskov, Phys. Rev. B 59 (1999) 7413–7421.
[30] B.G. Pfrommer, M. Côté, S.G. Louie, M.L. Cohen, J. Computat. Phys. 131 (1997)
loy is the largest, showing that it could have a better plasticity. The 233–240.
stress–strain relationships show that the stress of the Mg–Al and [31] H.J. Monkhorst, J.D. Pack, Phys. Rev. B 13 (1976) 5188.
Mg–Li alloys would be increased compared with the pure Mg metal [32] G.J. Ackland, M.C. Warreny, S.J. Clark, J. Phys.: Condens. Matter 9 (1997) 7861–
7872.
after the strain of 8%. For the larger strains, the Mg–Al alloy has the [33] J.P. Perdew, Y. Wang, Phys. Rev. B 45 (1992) 13244–13249.
largest stress and the stress of the Mg–Cd alloy is slightly less than [34] T.B. Massalski, Binary Alloy Phase Diagrams, second ed., ASM International,
that of the pure Mg metal. Materials Park, OH, 1990.
[35] M. Asta, R. McCormack, D. de. Fontaine, Phys. Rev. B 48 (1993) 748–766.
[36] W. Siller, H. Hoffmeister, Z. Metallkd. 70 (1979) 817–824.
Acknowledgments [37] G.V. Raynor, Proc. R. Soc. London A 180 (1942) 107.
[38] D. Hardie, R.N. Parkinspages, Philos. Mag. 4 (1959) 815–825.
[39] A.A. Nayeb-Hashemi, J.B. Clark, A.D. Pelton, Alloy Phase Diagrams 5 (1984)
Calculations were performed in the High Performance Compu-
365.
tation Center of the Northwestern Polytechnical University. Sup- [40] W.Y. Yu, N. Wang, X.B. Xiao, B.Y. Tang, L.M. Peng, W.J. Ding, Solid State Sci. 11
ports by the National Natural Science Foundation of China (No. (2009) 1400–1407.
[41] C.P. Liang, H.R. Gong, J. Alloys Comp. 489 (2010) 130–135.
10932008) are greatly acknowledged.
[42] L.G. Hector Jr., J.F. Herbst, Phys. Rev. B 76 (014121) (2007) 1–18.
[43] A.R. Wazzan, L.B. Robinson, Phys. Rev. 155 (1967) 586–594.
References [44] B.R. Sahu, Mater. Sci. Eng. B 49 (1997) 74–78.
[45] R.C. King, O.J. Kleppa, Acta Metall. 12 (1964) 87–97.
[1] C. Potzies, K.U. Kainer, Adv. Eng. Mater. 6 (2004) 281–289. [46] C.L. Fu, X.D. Wang, Y.Y. Ye, K.M. Ho, Intermetallics 7 (1999) 179–184.
[2] Z.B. Sajuri, T. Umehara, Y. Miyashita, Y. Mutoh, Adv. Eng. Mater. 5 (2003) 910– [47] J. Nylén, F.J. Garcia, B.D. Mosel, R. Põttgen, U. Hãussermann, Solid State Sci. 6
916. (2004) 147–155.
[3] Z. Liu, Y. Wang, Z.G. Wang, F. Li, Z.Y. Shen, J. Mater. Res. 14 (2000) 449–456. [48] N.W. Ashcroft, N.D. Mermin, Solid State Physics, Saunders College,
[4] B.L. Mordike, T. Ebert, Mater. Sci. Eng. A 302 (2001) 37–45. Philadelphia, 1976.
[5] C.H. Cáceres, D.M. Rovera, J. Light Met. 1 (2001) 151–156. [49] V. Milman, M.C. Warren, J. Phys.: Condens. Matter 13 (2001) 241–251.
[6] L. Gao, R.S. Chen, E.H. Han, J. Alloys Comp. 472 (2009) 234–240. [50] W. Voigt, Lehrbuch der Kristallphysik, Teubner, Leipzig, 1928.
[7] A. Srinivasan, J. Swaminathan, U.T.S. Pillai, K. Guguloth, B.C. Pai, Mater. Sci. Eng. [51] A. Reuss, Z. Angew. Math. Mech. 9 (1929) 49–58.
A 485 (2008) 86–91. [52] R. Hill, Proc. Phys. Soc. A 65 (1952) 349–354.
[8] A. Srinivasan, J. Swaminathan, M.K. Gunjan, U.T.S. Pillai, B.C. Pai, Mater. Sci. [53] M.J. Mehl, J.E. Osburn, D.A. Papaconstantopoulos, B.M. Klein, Phys. Rev. B 41
Eng. A 527 (2010) 1395–1403. (1990) 10311–10323.
[9] H.T. Guo, L.T. Xia, Res. Stud. Foundry Equip. 2 (2007) 40–43. [54] L.J. Slutsky, C.W. Garland, Phys Rev 107 (1957) 972–976.
[10] C. Ravi, C. Wolverton, Acta Mater. 52 (2004) 4213–4227. [55] M. Born, Proc. Cambridge Philos. Soc. 36 (1940) 160–165.
[11] S. Ganeshan, S.L. Shang, H. Zhang, Y. Wang, M. Mantina, Z.K. Liu, Intermetallics [56] Y.L. Page, P. Saxe, Phys. Rev. B 65 (2002) 104104.
17 (2009) 313–318. [57] S.F. Pugh, Philos. Mag. Ser. 45 (1954) 823–843.
[12] Y.F. Wang, W.B. Zhang, Z.Z. Wang, Y.H. Deng, N. Yu, B.Y. Tang, X.Q. Zeng, W.J. [58] A. Sumer, J.F. Smith, J Appl. Phys. 33 (1962) 2283–2286.
Ding, Computat. Mater. Sci. 41 (2007) 78–85. [59] J.R. Davis, Tensile Testing, second ed., ASM International Paddyfield ShopInHK,
[13] Y. Zhong, J.O. Sofo, A.A. Luo, Z.K. Liu, J. Alloys Comp. 421 (2006) 172–178. 2004.
[14] W.A. Counts, M. Friák, D. Raabe, J. Neugebauer, Adv. Eng. Mater. 12 (2010) [60] F. Roters, P. Eisenlohr, L. Hantcherli, D.D. Tjahjanto, T.R. Bieler, D. Raabe, Acta
572–576. Mater. 58 (2010) 1152–1211.