Alloying element effect on elastic constants

(elastic properties) of Al alloys

MECHANICAL BEHAVIOR

Name: Ayman Aydan Mahmoud Abdelhadi

(‫)أيمن أيدن محمود عبد الهادي‬
3rd year of metallurgical and materials engineering department
Section No: 2
 • Abstract
The effects of alloying elements (Co, Cu, Fe, Ge, Hf, Mg, Mn, Ni, Si, Sr, Ti,
V, Y, Zn, and Zr) on elastic properties of Al have been investigated using
first-principles calculations within the generalized gradient approximation. A
supercell consisting of 31 Al atoms and one solute atom is used. A good
agreement is obtained between calculated and available experimental data.
Lattice parameters of the studied Al alloys are found to be depended on
atomic radii of solute atoms. The elastic properties of polycrystalline
aggregates including bulk modulus (B), shear modulus (G), Young’s
modulus (E), and the B/G ratio are also determined based on the
calculated elastic constants (cij , s) It is found that the bulk modulus of Al
alloys decreases with increasing volume due to the addition of alloying
elements and the bulk modulus is also related to the total molar
volume (Vm ) and electron density (nAl31X ) with the relationship of nAl31X =
1.0594 + 0.0207√B/Vm These results are of relevance to tailor the
properties of Al alloys.

1. Introduction

With a density approximately one third of that of steel or copper, aluminum

(Al) alloys with alloying elements Cu, Mg, Si, Zn, and Zr, etc., are widely
used as engineering materials where light weight or corrosion resistance is
required. The properties of Al, which make this metal and its alloys the
most economical and attractive for a wide variety of uses, are appearance,
light weight, fabric ability, physical properties, mechanical properties, and
corrosion resistance.
Therefore, a detailed understanding of the thermodynamic and elastic
properties of Al alloys is crucial for a better realization of its potential in
currently available applications and in developing new ones.
The thermodynamic modeling through integrating first-principles
calculations and CALPHAD (Calculation of Phase Diagram) method has
proven to be efficient and robust and demonstrated for relevant binary,
ternary and multicomponent systems of Al alloys of. Recently, the
enthalpies of formation and elastic properties for binary Al compounds were
systematically predicted by first-principles calculations. However, there are
no theoretical studies addressing the elastic property changes in Al induced
by alloying elements. It is known that the elastic properties of materials can
be used to assess certain mechanical properties such as
ductility/brittleness, hardness, strength and so on The theoretical prediction
for the effect of alloying additions on the elastic constants (cij , s) can
provide essential guidance in identifying materials with desired mechanical
properties. The effects of alloying elements on the elastic properties of AlTi
and AlTi3, AlNi, AlNi3, Mg and Ni were studied via first-principles approach.
These works are important for tailoring the properties of existed alloys and
designing new alloys. In this report, the effects of alloying elements (Co,
Cu, Fe, Ge, Hf, Mg, Mn, Ni, Si, Sr, Ti, V, Y, Zn, and Zr) on the elastic
properties in the Al dilute solid solutions are predicted via first-principles
calculations using the efficient stress-strain method. The present work,
together with the previous work on the elastic constants of compounds,
forms a basis for predicting the elastic properties of Al alloys. It is our
ambition to spark systematic experimental studies of the elastic properties
with this contribution. The rest of the present report is described as follows:
the details of first-principles calculations using Vienna Ab-into Simulation
Package (VASP) are presented in Section 2, including the brief
introduction of equation of state (EOS) and elastic theory used herein. In
Section 3, The investigated equilibrium properties include the
volume(𝑉0 ) energy (𝐸0 ), bulk modulus (𝐵0 ) and its pressure derivative
(𝐵0 ′) of the compounds, determined via EOS fitting, and the single crystal
elastic constants (𝑐𝑖𝑗 's ) together with structural stabilities and the
polycrystalline aggregates are presented and discussed. Finally, in Section
4, the summary of the present work is given.

2. Theory and methodology

First-principles calculations are performed using the VASP code with the
projected augment wave (PAW) method to describe the electron ion
interaction and the generalized gradient approximation (GGA) to depict the
exchange correction functional. All the structures are fully relaxed with
respect to cell shape, volume, and atomic coordinates. For consistency, a
400-eV energy cutoff is used for all the elements. A 2×2×2 fcc
(face centered-cubic) supercell including 31 Al atoms and one alloying
atom (X) is employed in this study. The energy convergence criterion of
electronic self-consistency is chosen as 10-6 eV/atom for all the
calculations. The reciprocal space energy integration is performed by the
Methfessel-Paxton technique for structure relaxations, while the final
calculations of total energies for EOS fittings and stresses for determining
the (𝑐𝑖𝑗 's )are performed by the linear tetrahedron method including Blöchl
corrections. The samplings of k-point are 15×15×15 and 11×11×11 for EOS
and elastic constants calculations in terms of the Monkhorst-Pack scheme,
respectively.
In order to fit the first principles calculated E-V (energy-volume) data points,
the 4-parameter Birch Murnaghan equation of state with its linear form
given by Shang et al.is employed,
2 4
𝐸(𝑉) = 𝑎 + 𝑏𝑉 −3 + 𝑐𝑉 −3 + 𝑑𝑉 −2 (1)
where a, b, c and d are fitting parameters. In the present work, usually 10
data points in the volume range of 0.88 − 1.16𝑉0 are used for the EOS fitting
of each structure. The equilibrium properties
estimated from EQS include the volume (𝑉0 ), energy (𝐸0 ). bulk modulus
(𝐵0 ) and its pressure derivative (𝐵0′ ). It is worth mentioning that the fitting
parameters are representable by the equilibrium properties, and vice
versa.
To calculate the single crystal elastic stiffness constants (𝑐𝑖𝑗 ‘s ), an
efficient strain-stress method proposed by Shang et al.is employed in
the present work. Under this methodology, for a given set of strains
𝝐 = (𝜖1 , 𝜖2 , 𝜖3 , 𝜖4 , 𝜖5 , 𝜖6 )( where 𝜖1 , 𝜖2 and𝜖3 are the normal strains and
𝜖4 , 𝜖5 and 𝜖6 are the shear strains) imposed on a crystal with lattice vectors
L specified in the Cartesian coordinates
𝑎1 𝑎2 𝑎3
𝐿 = (𝑏1 𝑏2 𝑏3 ) (2)
𝑐1 𝑐2 𝑐3
where 𝑎𝑙 , 𝑎2 and 𝑎3 are the 𝑥, 𝑦, 𝑧 components of the lattice vector a,
respectively, and it is the same for lattice vectors b and c.
 After the deformation due to strain ϵ, the deformed lattice vectors are
obtained as follows:
1 + 𝜀1 𝜀6 /2 𝜀5 /2
𝐿𝑑𝑒𝑓 = 𝐿 ( 𝜀6 /2 1 + 𝜀2 𝜀4 /2 ) (3)
𝜀5 /2 𝜀4 /2 1 + 𝜀3
Accordingly, a set of stresses,
𝝈 = (𝜎1 , 𝜎2 , 𝜎3 , 𝜎4 , 𝜎5 ,𝜎6 ), associated with the deformed crystal will be
determined through first-principles calculations in the present work.
Correspondingly, for n sets of strains ϵ (n-by-6 matrix) and the resulting
stresses σ, the elastic constants c (6-by-6 matrix as shown in Voigt’s
notation.) are determined according to the general Hooke’s law as follows:

𝑐11 𝑐12 𝑐13 𝑐14 𝑐15 𝑐16 𝜀1,1 𝜀1,𝑛 −1 𝜎1,1 𝜎1,𝑛
𝑐21 𝑐21 𝑐21 𝑐21 𝑐21 𝑐21 𝜀2,1 𝜀2,𝑛 𝜎2,1 𝜎2,𝑛
𝑐31 𝑐32 𝑐33 𝑐34 𝑐35 𝑐36 𝜀3,1 𝜀3,𝑛 𝜎3,1 𝜎3,𝑛
= (4)
𝑐41 𝑐42 𝑐43 𝑐44 𝑐45 𝑐46 𝜀4,1 𝜀4,𝑛 𝜎4,1 𝜎4,𝑛
𝑐51 𝑐52 𝑐53 𝑐54 𝑐55 𝑐56 𝜀5,1 𝜀5,𝑛 𝜎5,1 𝜎5,𝑛
(𝑐61 𝑐62 𝑐63 𝑐64 𝑐65 𝑐66 ) (𝜀6,1 𝜀6,𝑛 ) (𝜎6,1 𝜎6,𝑛 )
where “-1” represents the pseudo-inverse, which can be solved based on
the singular value decomposition method to get the least square solutions
of elastic constants. Due to the symmetry of crystals, the minimum linearly
independent sets of strains to determine the elastic constants are two for
cubic, three for hexagonal and rhombohedral, four for tetragonal, and six
for orthorhombic, monoclinic, and triclinic structures [13, 23]. In this work,
the following linearly independent sets of strains are selected:

𝑥 0 0 0 0 0
0 𝑥 0 0 0 0
0 0 𝑥 0 0 0
0 0 0 𝑥 0 0
0 0 0 0 𝑥 0
(0 0 0 0 0 𝑥)
with x=±0.007, and ±0.01, which are verified to obey the Hooke’s law,
leading to a sufficient redundancy of the nonzero stresses and in turn
accurate elastic constants. Based on 𝑐𝑖𝑗 , polycrystalline aggregates,
including the bulk (B), shear (G), and Young’s (E) modulus, can be
computed via the Voigt’s approach ,viz ,
(𝑐11 + 2𝑐12 ) (𝑐11 − 𝑐12 + 3𝑐44 )
𝐵= ,𝐺 =
3 5
and𝐸 = 9𝐵𝐺/(𝐺 + 3𝐵) for cubic structures. More details regarding the
calculations of elastic constants and the applications of strain-stress
method can be found elsewhere.

3. Results and discussion

Calculated lattice parameters of pure fcc Al and Al-X alloys along with the
available experimental data are summarized in Table 1. Among the 15 Al-X
alloys studied herein, experimental data are available for Al-Fe, Al-Si, and
Al-Ti from X-ray diffraction (XRD) as shown in Table 1. A good agreement
is obtained between the calculated and experimental data, with the
differences of -0.51%, - 0.23%, and -0.16% for Al-Fe, Al-Si, and Al-Ti,
respectively. It is found in Table 1 that the lattice parameters of Al-X alloys
are proportional to the corresponding lattice parameter of pure element X in
the fcc structure. Figure 1 depicts the change of the lattice parameter of Al
due to the addition of solute elements in comparison with the atomic radii of
solute atoms calculated from their fcc structures. Here, the atomic radii of
pure elements are calculated as half of the nearest-neighbor atomic
distance, being consistent with those calculated by Wang et al. Figure 2
shows the change of the nearest-neighbor distance between Al and X atom
against the atomic radii of solute atoms. The nearest-neighbor distance
between Al and X show a similar relationship against atomic radii of solute
atoms.
For dilute solutions the change of lattice parameter can be treated as a
linear function of composition according to Wang et al. The linear
regression coefficients for each element are calculated using the following
equation:
𝑘𝑥 = 𝑁 (𝑎𝐴𝑙31𝑋 − 𝑎𝐴𝑙32 ) /100[pm/ at. %] (5)
where N is the number of atoms in the supercell (N = 32 for the present
work), 𝑎𝐴𝑙31𝑋 the lattice parameter of the cell with 31 Al atoms and one X
atom, and 𝑎𝐴𝑙32 the lattice parameters of the cell with 32 Al atoms.
The calculated linear regression coefficients are listed in Table 1. The
linear regression coefficient of Ti and Zn are almost zero indicating that the
lattice parameter rarely changes due to their additions, which also can be
seen in Figure 1
Table 1. Lattice parameters of Al-3.125X (X in at. %) fcc dilute solutions and linear regression
coefficients of Eq. 5
The changes of the nearest-neighbor
distance around a solute atom can be
described by local lattice distortion
which is listed in Table 2 along with
available experimental measurements
[33]. The calculated local lattice
distortions in the present work are
within the experimental uncertainties
for Cu, Ge, Mn, and Zn [33], as shown
in Figure 3. The local lattice distortion
due to the addition of Zn is the closest
to experimental data. This also verifies
the zero linear regression coefficient of
Zn shown in Table 1.
The predicted properties for Al-3.125X (at. %)
dilute solid solutions, including the elastic
constants (𝑐𝐼𝐼 , 𝑐12 and 𝑐44 ), the bulk modulus
(𝐵), shear modulus(G), Young’s (E) modulus,
and B/G ratio along with the available
experimental data are shown in Table 3. The
estimated bulk modulus and equilibrium volume
using EOS (Eq. 1) are also shown for
comparison. The fitting error (Eq. 4 in Ref) of
EOS is smaller than 0.1, indicates the high
qualities of first-principles calculations. The bulk
moduli of Al-X dilute solutions calculated using
the two methods ((𝑐𝑖𝑗 's ) and EOS) are very
close to each other. The bulk moduli obtained
from EOS are slightly smaller than those from
(𝑐𝑖𝑗 's ), since the volume ranges used in EOS
fitting are wider [13]. The available experimental data for Al-X dilute
alloys are for Al-Cu [24, 34] and Al-Mg. The reported (𝑐𝑖𝑗 's )for Al-5Cu (at.
%) are 308.22,
262.56 and 27.03 GPa for𝑐𝑙𝑙 , 𝑐𝑗 , and𝑐44 , respectively, where𝑐𝐼𝐼 and 𝑐12 are
much larger than the calculated values at Al-3.125Cu (at. %). The
extremely large 𝑐11 and 𝑐12
thus, result in unreasonable large bulk modulus
(277.8) and B/G ratio (10.9). This is probably
due to the samples were in precipitation
hardened state and texture were existence.
The calculated and experimental elastic
properties data for the Al- 3.125Mg (at. %) solid
solution is compared in Figure 4. It should be
mentioned that the experimental elastic
constants of Al-Mg alloy with composition of
3.125 at % Mg are obtained by linear
interpolation. The calculated values are slightly
larger than experimental data, which is
reasonable since the first principles
calculations are performed at 0 K, while the experimental data were
measured at room temperature, and elastic constants decrease with
increasing temperature. All the alloy systems shown in Table 3 satisfy the
Born criteria for mechanical stability, i.e.,
𝑐11 − |𝑐12 | > 0, 𝑐11 + 2|𝑐12 | > 0 and𝑐44 > 0 for
cubic structure, indicating that the Al alloys
with 3.125%X are within the limit of
mechanical stability.
Figure 5 shows that the calculated bulk
moduli of Al-X alloys decrease linearly with
respect to the increase of nearest-neighbor
distances between Al and X atoms. The
smallest bulk modulus is due to the addition
of Sr, while the largest one is due to the
addition of Fe among all the Al alloys studied
in the present work. It can also be seen from
Figure 6 which shows a strong dependence of the bulk modulus on the
atomic volume of the alloys, due to the addition of the alloying element. The
calculated bulk moduli of Al-X alloys is further plotted in Figure 7 with
respect to the bulk modulus of pure solute atom X. Alloying elements with
higher bulk moduli, such as Co, Mn, Fe, and Mn result in the higher bulk
moduli of Al-X alloys, and vice versa.
According to Pugh criterion, a metal can be
considered to be brittle when its bulk/shear
modulus ratio is smaller than 1.75, otherwise
ductile. All Al-X dilute solid solutions have their
B/G ratios greater than 1.75, as shown in
Table3. This means Al will remain be ductility
with the addition of alloying
element.
To understand which factors are correlated with
the bulk modulus, Kim et al used the
empirical relationship between the bulk
modulus and volume
reported by Miedema et
al. According to Miedema
et al. and Li and
Wu, √𝐵/𝑉𝑚has a linear
relationship with the
change of electron
density, n, at the
boundary of Wigner-Seitz
cell for pure elements,
where𝑉𝑚 is the molar
volume of the element for
alkali metals and non-
transition metals. The
electron density n ofAl31 X
solution can be calculated by the following equation
31 1
𝑛𝐴𝑙3 𝑋 = 𝑁𝐴𝑙1𝑋 /𝑉𝐴𝑙1 ,𝑋 = ( 𝑛𝐴𝑙 𝑉𝐴𝑙 + 𝑛𝑋 𝑉𝑋 )
32 32
The calculated results are listed in Table 4 including the bond
valence 𝑍𝐵 = 𝑛𝑉where V is the volume per atom of the ground state
elemental metal at 0 K. A linear relationship between√𝐵/𝑉𝑚 and 𝑛𝐴𝑙31𝑥 is
shown in Figure 8, 𝑛𝐴𝑙1𝑋 = 1.0594 + 0.0207√𝐵/𝑉𝑚 , which allows us to
predict bulk modulus from electron density and volume.
4.Summary
The effects of alloying elements
(Co, Cu, Fe, Ge, Hf, Mg, Mn, Ni, Si,
Sr, Ti, V, Y, Zn and Zr) on elastic
properties of Al have been
investigated by an efficient first-
principles stress-strain method. A
good agreement is obtained
between calculated and available
experimental data. The elastic
properties of polycrystalline
aggregates including bulk modulus
(B), shear modulus (G), Young’s
modulus (E), and B/G ratio are
determined based on the calculated
elastic stiffness. It is found that (i)
the bulk moduli of Al alloys decrease with increasing volume caused by
alloying elements and (ii) the bulk modulus is also related to the total molar
volume(𝑉𝑚 ) and electron density with the relationship of

𝑛𝐴𝑙1 𝑋 = 1.0594 + 0.0207√𝐵/𝑉𝑚

References
[1] J.R. Davis, R. Joseph, In: ASM Specialty Handbook: ASM International; 1993.
[2] G.E. Totten, M. D.S., In, New York: Marcel Dekker; 2003.
[3] Z.K. Liu, Journal of Phase Equilibria and Diffusion, 30 (2009) 517-534.
[4] Y. Du, S.H. Liu, L.J. Zhang, H.H. Xu, D.D. Zhao, A.J. Wang, L.C. Zhou,
CALPHAD, 35 (2011) 427-445.
[5] J. Wang, S.L. Shang, Y. Wang, Z.G. Mei, Y.F. Liang, Y. Du, Z.K. Liu, CALPHAD,
35 (2011) 562-573.
[6] D.G. Clerc, H.M. Ledbetter, J. Phys. Chem. Solids, 59 (1998) 1071.
[7] Z.K. Liu, H. Zhang, S. Ganeshan, Y. Wang, S.N. Mathaudhu, Scripta Materialia, 63
(2010) 686-691.
[8] D. Music, J.M. Schneider, Phys. Rev. B, 74 (2006).
[9] C.L. Zhang, P.D. Han, J.M. Li, M. Chi, L.Y. Yan, Y.P. Liu, X.G. Liu, B.S. Xu, J.
Phys. D. Appl. Phys., 41 (2008) 095410.
[10] D.E. Kim, S.L. Shang, Z.K. Liu, Intermetallics, 18 (2010) 1163-1171.
[11] S. Ganeshan, S.L. Shang, Y. Wang, Z.K. Liu, Acta Mater., 57 (2009) 3876-3884.
[12] D. Kim, S.L. Shang, Z.K. Liu, Comp. Mater. Sci., 47 (2009) 254-260.
[13] S.L. Shang, Y. Wang, Z.K. Liu, Appl. Phys. Lett., 90 (2007) 101909.
[14] G. Kresse, J. Furthmuller, Phys. Rev. B, 54 (1996) 11169-11186.
[15] G. Kresse, J. Furthmuller, Comp. Mater. Sci., 6 (1996) 15-50.
[16] P.E. Blöchl, Phys. Rev. B, 50 (1994) 17953-17979.
[17] G. Kresse, D. Joubert, Phys. Rev. B, 59 (1999) 1758- 1775.
[18] J.P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett., 77 (1996) 3865-3868.
[19] M. Methfessel, A.T. Paxton, Phys. Rev. B, 40 (1989) 3616-3621.
[20] P.E. Blöchl, O. Jepsen, O.K. Andersen, Phys. Rev. B, 49 (1994) 16223-16233.
[21] H.J. Monkhorst, J.D. Pack, Phys. Rev. B, 13 (1976) 5188-5192.
[22] S.L. Shang, Y. Wang, D. Kim, Z.K. Liu, Comp. Mater. Sci., 47 (2010) 1040-1048.
[23] Y. Le Page, P. Saxe, Phys. Rev. B, 65 (2002) 104104.
[24] G. Simmons, H. Wang, In: M.I.T Press, Cambridge; 1971.
[25] H. Zhang, S.L. Shang, Y. Wang, A. Saengdeejing, L.Q. Chen, Z.K. Liu, Acta
Mater., 58 (2010) 4012-4018.
[26] S.L. Shang, A. Saengdeejing, Z.G. Mei, D.E. Kim, H. Zhang, S. Ganeshan, Y.
Wang, Z.K. Liu, Comp. Mater. Sci., 48 (2010) 813-826.
[27] S. Ganeshan, S.L. Shang, H. Zhang, Y. Wang, M. Mantina, Z.K. Liu, Intermetallics.,
17 (2009) 313-318.
[28] A. Tonejc, A. Bonefaci, J. Appl. Phys., 40 (1969) 419- .
[29] A. Bendijk, R. Delhez, L. Katgerman, T.H. Dekeijser, E.J. Mittemeijer, N.M.
Vanderpers, J. Mater. Sci., 15 (1980) 2803-2810.