Available online at www.sciencedirect.

com

Scripta Materialia 67 (2012) 5760

www.elsevier.com/locate/scriptamat

A study of low Youngs modulus TiNbZr alloys using d electrons

alloy theory
Li You and Xiping Song
State Key Laboratory for Advanced Metals and Materials, University of Science and Technology Beijing, Beijing 100083, China
Received 6 March 2012; revised 18 March 2012; accepted 19 March 2012
Available online 24 March 2012

Six kinds of alloys in TiNbZr alloy system were designed using d electrons alloy theory, and the relations between their
Youngs modulus and electron parameters were investigated. The results show that the Youngs moduli of the designed alloys
are low, with a minimum of 38.8 GPa. A formula was proposed to represent the interatomic bonding force of these alloys, which
showed good agreement with the Youngs modulus. Low Youngs modulus b-Ti alloys can be designed by evaluating their interatomic bonding force.
Crown Copyright 2012 Published by Elsevier Ltd. on behalf of Acta Materialia Inc. All rights reserved.
Keywords: Titanium alloys; Elastic behavior; d electrons alloy theory; X-ray diraction; Tension test

Titanium alloys are now the most widely used

biomaterials for implant applications because of their
excellent biocompatibility and high corrosion resistance
[1,2]. However, the fact that the Youngs moduli of the
titanium alloys currently used are much larger than
those of human bones [3] causes stress shielding in the
vicinity of the implant area, which is harmful to human
health. Therefore, recent research of biomedical titanium alloys has focused on the design of low Youngs
modulus titanium alloys [47].
A new theory proposed by Morinaga et al. [8], named
d electrons alloy theory, has proved to be an eective
method of designing titanium alloys with low Youngs
modulus. For example, Ti29Nb13Ta4Mo (wt.%),
Ti29Nb13Ta2Sn (wt.%) [9] and Ti23Nb0.7Ta
2Zr1.2O (at.%) [10] alloys designed using the theory
have low Youngs moduli, ranging from 45 to 55 GPa.
The theory involves two parameters, Bo and Md [8].
The former is a measure of the covalent bond strength
between Ti and an alloying element M, and the latter
correlates well with the electronegativity and the metallic radius of elements. Each element has a specic Bo
value and a specic Md value for the body-centered
cubic (bcc) cluster and separately for the hexagonal
close-packed (hcp) cluster, based on calculations using

Corresponding

author. Tel.: +86 10 62333213; fax: +86 10

62333447; e-mail: xpsong@skl.ustb.edu.cn

the DV-Xa molecular orbital method (see Table 1). A

titanium alloy can be shown as a point in the binary
BoMd diagram [8] by taking the compositional average
of the Bo and Md values of each element in the titanium
alloy, which are denoted as Bo and Md.
In the BoMd diagram of the bcc-Ti cluster, all the
titanium alloys can be distributed into three regions
(a, a + b and b). Around the a + b/b boundary the bphase is metastable, and will probably transfer into
non-equilibrium-phases (x, a0 and a00 ) under the condition of water quenching. It has been reported that the
metastable b-phase has the lowest modulus among
these-phases, and that the Youngs modulus of the titanium alloys decreases as the Bo and Md increase along
the metastable b-phase boundary [1113], so the low
Youngs modulus of the titanium alloys can be achieved
if we can obtain metastable b-phase with high Bo and
Md.
The TiNbZr alloy system has proved to be a good
substitute for developing absolutely safe Ni-free biomedical Ti alloys because Ti, Nb and Zr are non-toxic
elements and do not cause any adverse reaction in the
human body. However, the Youngs moduli of the Ti
NbZr alloys reported so far are not low enough
[4,14], and there has been little research into the Youngs
moduli of metastable b-phase TiNbZr alloys with
high Bo and Md. Moreover, the quantitative relationship
between the Youngs modulus and the Bo and Md has

1359-6462/$ - see front matter Crown Copyright 2012 Published by Elsevier Ltd. on behalf of Acta Materialia Inc. All rights reserved.
http://dx.doi.org/10.1016/j.scriptamat.2012.03.020

58

L. You, X. Song / Scripta Materialia 67 (2012) 5760

Table 1. List of Bo and Md values for various alloying elements in

hcp-Ti and bcc-Ti clusters.
Element

Bo
hcp-Ti

Bo
bcc-Ti

Md (eV)

Ti
V
Cr
Mn
Fe
Co
Ni
Cu
Zr
Nb
Mo
Hf
Ta
W
Al
Si
Sn

3.513
3.482
3.485
3.462
3.428
3.368
3.280
3.049
3.696
3.767
3.759
3.664
3.720
3.677
3.297
3.254
2.782

2.790
2.805
2.779
2.723
2.651
2.529
2.412
2.114
3.086
3.099
3.063
3.110
3.144
3.125
2.426
2.561
2.283

2.447
1.872
1.478
1.194
0.969
0.807
0.724
0.567
2.934
2.424
1.961
2.975
2.531
2.072
2.200
2.200
2.100

not yet been claried. Therefore, the purpose of this

paper is to design low Youngs modulus TiNbZr alloys by using d electrons alloy theory and to propose a
formula to represent the quantitative relation of the
Youngs modulus and the Bo and Md that could provide
a theoretical guide to the design of low Youngs modulus biomedical Ti alloys.
The Bo and Md values of single b-phase and non-single b-phase TiNbZr ternary alloys [15,16] were rst
calculated; these are shown in Figure 1. It was found
that the boundary between the single b-phase and the
non-single b-phase was almost a straight line (the solid
straight blue line in Fig. 1). We could thus assume that
the straight line was the metastable b-phase boundary of
the TiNbZr alloy system. It can be seen in Figure 1
that the metastable b-phase boundary of the TiNb
Zr alloy system has an oset from that proposed by
Morinaga et al. [8] and Abdel-Hady et al. [13] for alloys
without Zr. This dierence can be explained by the shift
eect of the element Zr [13,17]. Six kinds of alloys in the
TiNbZr alloy system were then designed along the
metastable b-phase boundary in order to obtain a low
Youngs modulus, and to investigate the corresponding
relation of the Youngs modulus and the Bo and Md, as

Figure 1. BoMd diagram of the bcc-Ti cluster. The six selected alloy
locations are numbered 16 along the deduced metastable b-phase
boundary of the TiNbZr alloy system.

indicated by the numbers 16 in Figure 1. The corresponding compositions were calculated to be Ti24Nb,
Ti20Nb12Zr, Ti17Nb21Zr, Ti11Nb38Zr, Ti6Nb53Zr
and Ti70Zr by denition of the Bo and Md. Here, the alloy compositions are expressed in atomic percent.
The alloys in this study were prepared by arc-melting
from pure metals of Ti, Nb and Zr, with a purity of
more than 99 mass% under argon atmosphere. They
were re-melted three times to ensure a homogeneous
composition, then suction cast into a copper mould.
This suction casting is similar to water quenching: it
cools the alloy rapidly and is conducive to obtaining a
single b-phase in the alloy. Tension samples were cut
from the as-cast alloys. The dimensions of the tension
samples are shown in Figure 3. The Youngs moduli
of these alloys were measured from the stressstrain
curve in tensile tests with a strain gauge. For each alloy,
three samples were tested for each experimental Youngs
modulus. The phase structures were also determined by
X-ray diraction with Cu Ka radiation.
Figure 2 shows the XRD proles of the six kinds of
alloys in the TiNbZr alloy system. It can be seen that
Ti24Nb, Ti20Nb12Zr, Ti17Nb21Zr and Ti11Nb38Zr
are composed only of b-phase (bcc), while Ti6Nb53Zr
is composed of both b- and a0 -phases (hcp), and Ti70Zr
is composed only of a0 -phase. From Ti24Nb to
Ti6Nb53Zr, the b-phase diraction peaks move toward
smaller angles, indicating that the b-phase lattice constant becomes larger.
The dimensions of the tension samples and the
Youngs modulus with error bars are shown in Figure 3.
For alloy Nos. 16, the Youngs moduli are 66.7, 54.8,
52.4, 38.8, 48.0 and 74.7 GPa, respectively. These results
indicate that the Youngs moduli of the TiNbZr alloys
designed near the metastable b-phase boundary are indeed very low. The minimum Youngs modulus is
38.8 GPa, which is lower than that of any b-Ti alloy so
far reported. It can also be seen that, as the Bo and Md

Figure 2. XRD proles of the six kinds of alloys in the TiNbZr alloy
system. The-phase structures of the TiNbZr alloys change from
b-phase to a0 -phase as the number of alloys increases.

L. You, X. Song / Scripta Materialia 67 (2012) 5760

Figure 3. The dimensions of the tension samples and the Youngs

modulus of the six kinds of alloys in the TiNbZr alloy system.

of these alloys increase along this boundary, the Youngs

modulus rst decreases and then increases. The increase is
because of the appearance of a0 -phase from the XRD results. When the alloys are composed of a single b-phase,
their Youngs modulus decreases as their Bo and Md increase. This is in agreement with the ndings of AbdelHady et al. [13].
It is well known that Youngs modulus, which is an
intrinsic feature of materials, is determined by the interatomic bonding force. When atoms of metallic elements
come together to form a metal, the electrons and nuclei
of dierent atoms interact and form metallic bonds.
Transition metals have special d electrons, which are
more localized than s and p electrons. The more localized d electrons tend to form covalent bonds. In order
to demonstrate this conveniently, we show an illustration of a transition metallic bonding model in Figure
4. The positive ion cores and valence electrons attract
and conne each other within the metallic crystal.
Essentially, the bonding force of metallic bonds is derived from Coulomb electrostatic interactions among
valence electrons and positive ion cores. It can thus be
estimated approximately by Coulombs law.
Coulombs law shows that the force between two
electric charges is proportional to the multiplication of
the charges divided by the square of the distance between them, which can be given by:
qq
1
F / 12 2
r
where q1 and q2 are the charges, and r is the distance between them.
In the following, we attempt to correlate Coulombs
law with the parameters Bo and Md of d electrons alloy
theory.
In metallic bonding, the charges of positive ion cores
are eective nuclei charges felt by valence electrons.
Hence, they can be regarded as q1. The eective nuclei
charges Ze felt by each valence electron can be calculated by Z  s, where Z is the atomic number and s is
the Slater shielding constant (for details, see Ref. [18]).
Here we take the average values of Ze felt by each valence electron as the q1 for each positive ion core.

59

Figure 4. Illustration of a transition metallic bonding model. The small

red points represent the nuclei; the small yellow point represents the
probability density centre of the valence electrons; the green circles
represent positive ion cores; the dark blue parts represent the overlap
of the probability densities of the valence electrons. (For interpretation
of the references to color in this gure legend, the reader is referred to
the web version of this article.)

As can be seen from the calculation of the Bo in d electrons alloy theory, the parameter Bo corresponds to the
overlap population between atoms of Ti and an alloying
element M [19], which is a measure of their shared electron density. A larger Bo indicates that the valence
electrons have a larger probability density appearing between atoms of Ti and the alloying element M. Hence Bo
can be regarded as q2 charges, which act to neutralize the
charge of the positive ion core and conne each other in
the metallic crystal.
Parameter Md is the energy level of the d orbit of an
alloying element M, which correlates well with the electronegativity and metallic radius of an element. In particular, the Md levels increase linearly with the metallic
radius, independently of whether the elements are 3d,
4d or 5d [20]. Thus, the Md can be regarded as the
approximate value for the metallic radius. On the other
hand, since the valence electrons are in the outermost
shell of a metallic element, the distance between the valence electrons and the ion core can be represented by
the metallic radius. From this discussion, it can be concluded that Md can be used as a parameter to represent
the distance r between valence electrons and the positive
ion core.
Now, combining Coulombs law and the d electrons
alloy theory, the interatomic bonding force of titanium
alloys can be estimated by:
Z eff  Bo
2
Md 2
where Z eff , Bo and Md are the compositional average
values of Ze, Bo and Md in a titanium alloy.
We have calculated the Z eff and bonding force of the
TiNbZr alloys in our study. For single b-phase (bcc)
and a0 -phase (hcp) alloys, the Bo values in the bcc-Ti
cluster and the hcp-Ti cluster (see Table 1), respectively,
are used in the calculations. For the alloy with a b + a0 phase, the average Bo value in the bcc-Ti and hcp-Ti
clusters is used approximately in the calculation. The results are shown in Table 2.
It can be seen that, from alloy No. 1 to alloy No. 6,
the bonding force rst decreases and then increases,
while the Z eff steadily decreases. The variation trend
bonding force /

60

L. You, X. Song / Scripta Materialia 67 (2012) 5760

Table 2. Z eff and bonding force of the six kinds of alloys in the Ti
NbZr alloy system.
Alloy

Z eff

Bonding force

Ti24Nb
Ti20Nb12Zr
Ti17Nb21Zr
Ti11Nb38Zr
Ti6Nb53Zr
Ti70Zr

3.477
3.464
3.454
3.435
3.419
3.400

1.671
1.599
1.549
1.459
1.541
1.593

of the bonding force is the same as that of the Youngs

modulus. The minimum bonding force of 1.459 corresponds to the minimum Youngs modulus of 38.8 GPa.
Bonding force is related not only to atom types and
the distances between atoms, but also to crystal structure. From Table 1, it can be seen that the Bo values
in the hcp-Ti cluster are larger than those in the bcc-Ti
cluster, and the Md values are the same. For the
Ti11Nb38Zr alloy its crystal structure is bcc, while for
the Ti70Zr alloy its crystal structure is hcp, so the Bo
value of the Ti70Zr alloy will be larger than that of
the Ti11Nb38Zr alloy. As a result, the bonding force
of the Ti70Zr alloy calculated by formula (2) will be larger than that of the Ti11Nb38Zr alloy. Therefore the increase in the Youngs modulus of the Ti70Zr alloy can
also be explained by the increase in the bonding force
between atoms.
Good agreement between the bonding force and the
Youngs modulus validates the bonding force and indicates that it can be used as a useful parameter for
designing low Youngs modulus alloys in the TiNb
Zr alloy system.
As an example, we have calculated the bonding force
of the Ti41.1Nb7.1Zr (wt.%) alloy with a bcc-phase
structure reported by Elias et al. [14]. The bonding force
of this alloy is calculated to be 1.661, which is larger than
that of the Ti11Nb38Zr alloy (1.459). Therefore the actual Youngs modulus of the Ti41.1Nb7.1Zr (wt.%) alloy should be larger than that of the Ti11Nb38Zr alloy.
As expected, the actual Youngs modulus of the Ti
41.1Nb7.1Zr (wt.%) alloy is 65 GPa [14], while that of
the Ti11Nb38Zr alloy is 38.8 GPa. On the other hand,
the bonding force of the Ti41.1Nb7.1Zr (wt.%) alloy
is close to that of the Ti24Nb alloy (1.671) and the
Youngs modulus of the Ti24Nb alloy is 66.7 GPa in
our study. This result shows that same bonding force
corresponds to the same Youngs modulus when the
phase structure is bcc crystal. Therefore, the bonding
force proposed in this paper can be used to estimate
the Youngs modulus of the TiNbZr alloy system. Further research should be done to develop this conclusion.
In summary, six low Youngs modulus alloys in the
TiNbZr alloy system, with the nominal compositions

of Ti24Nb, Ti20Nb12Zr, Ti17Nb21Zr, Ti11Nb38Zr,

Ti6Nb53Zr and Ti70Zr, were designed using d electrons
alloy theory. The minimum Youngs modulus was
38.8 GPa. Based on Coulombs law, a formula was proposed to represent the interatomic bonding force of
these alloys, which showed good agreement with the
Youngs modulus. Low Youngs modulus b-Ti alloys
can be designed by evaluating their interatomic bonding
force.
The authors wish to express their thanks for
support from the State Key Laboratory of Advanced
Metals and Materials of the University of Science and
Technology Beijing and from the Natural Science Foundation of China (Project No. 21171018).
[1] Y. Okazaki, S. Rao, Y. Ito, T. Tateishi, Biomaterials 19
(1998) 1197.
[2] H.J. Rack, J.I. Qazi, Mater. Sci. Eng. C 26 (2006) 1269.
[3] M. Niinomi, Mater. Sci. Eng. A 243 (1998) 231.
[4] V. Brailovski, S. Prokoshkin, M. Gauthier, K. Inaekyan,
S. Dubinskiy, M. Petrzhik, et al., Mater. Sci. Eng. C 31
(2011) 643.
[5] Y.L. Hao, S.J. Li, S.Y. Sun, C.Y. Zheng, R. Yang, Acta
Biomater. 3 (2007) 277.
[6] Y.L. Zhou, M. Niinomi, T. Akahori, Mater. Sci. Eng. A
483484 (2008) 153.
[7] Y.L. Hao, S.J. Li, S.Y. Sun, R. Yang, Mater. Sci. Eng. A
441 (2006) 112.
[8] M. Morinaga, N. Yukawa, T. Maya, K. Sone, H. Adachi,
in: The Sixth World Conference on Titanium, France, vol.
1, 1988 p. 1601.
[9] D. Kuroda, M. Niinomi, M. Morinaga, Y. Kato, T.
Yashiro, Mater. Sci. Eng. A 243 (1998) 244.
[10] T. Saito, T. Furuta, J.H. Hwang, S. Kuramoto, K.
Nishino, T. Sakuma, et al., Science 300 (2003) 464.
[11] L.A. Matlakhova, A.N. Matlakhova, S.N. Monteiro,
S.G. Fedotov, B.A. Goncharenko, Mater. Sci. Eng. A
393 (2005) 320.
[12] Y.L. Zhou, M. Niinomi, T. Akahori, Mater. Sci. Eng. A
371 (2004) 283.
[13] M. Abdel-Hady, K. Hinoshita, M. Morinaga, Scr. Mater.
55 (2006) 477.
[14] L.M. Elias, S.G. Schneider, S. Schneider, H.M. Silva, F.
Malvisi, Mater. Sci. Eng. A 432 (2006) 108.
[15] J.I. Kim, H.Y. Kim, T. Inamura, H. Hosoda, S. Miyazaki,
Mater. Sci. Eng. A 403 (2005) 334.
[16] M. Abdel-Hady, H. Fuwa, K. Hinoshita, H. Kimura, Y.
Shinzato, M. Morinaga, Scr. Mater. 57 (2007) 1000.
[17] M. Abdel-Hady, K. Hinoshita, H. Fuwa, Y. Murata, M.
Morinaga, Mater. Sci. Eng. A 480 (2008) 167.
[18] J.C. Slater, Phys. Rev. 36 (1930) 57.
[19] M. Morinaga, H. Adachi, M. Tsukada, J. Phys. Chem.
Solids 44 (1983) 301.
[20] M. Morinaga, N. Yukawa, H. Adachi, J. Phys. Soc. Jpn.
53 (1984) 653.