Sei sulla pagina 1di 8

Biomaterials 20 (1999) 2115}2122

Structure and properties of cast binary Ti}Mo alloys


W.F. Ho, C.P. Ju, J.H. Chern Lin*
Department of Materials Science and Engineering, National Cheng-Kung University, Tainan, Taiwan, ROC
Received 12 January 1999; accepted 26 May 1999

Abstract
Structure and properties of a series of binary Ti}Mo alloys with molybdenum contents ranging from 6 to 20 wt% have been
investigated. Experimental results indicated that crystal structure and morphology of the cast alloys were sensitive to their
molybdenum contents. The hexagonal a phase c.p. Ti exhibited a feather-like morphology. When Mo content was 6 wt%, a "ne,
acicular martensitic structure of orthorhombic a phase was observed. When Mo content was 7.5 wt%, the entire alloy was dominated
by the martensitic a structure. When Mo content was increased to 10 wt% or higher, the retained b phase became the only dominant
phase. Among all Ti}Mo alloys, the a phase Ti}7.5Mo alloy had the lowest hardness. The bending strength of Ti}7.5Mo was similar
to that of Ti}15Mo and Ti}13Nb}13Zr, and higher than c.p. Ti by nearly 60%. The bending modulus of the a-dominated Ti}7.5Mo
alloy was lower than that of Ti}15Mo by 22%, of Ti}6Al}4V by 47%, of Ti}13Nb}13Zr by 17%, and of c.p. Ti by 40%.  1999
Elsevier Science Ltd. All rights reserved.
Keywords: Titanium}molybdenum alloy; Implant material; Structure; Mechanical properties

1. Introduction
There has been a concern from time to time regarding
the stress shielding phenomenon, i.e., insu$cient loading
of bone due to the large di!erence in modulus between
implant device and its surrounding bone. This phenomenon, more often observed in cementless hip and knee
prostheses [1], can potentially lead to bone resorption
[2] and eventual failure of the arthroplasty [1].
Both strain gauge analysis [3] and "nite element analysis [4] have demonstrated that lower modulus (more
#exible) femoral hip implant components result in
stresses and strains that are closer to those of the intact
femur, and a lower modulus hip prosthesis may better
simulate the natural femur in distributing stress to the
adjacent bone tissue [5,6]. Canine and sheep implantation studies have shown signi"cantly reduced bone resorption in animals with low modulus hip implants [7],
and the bone loss commonly experienced by hip prosthesis patients may be reduced by a prosthesis having
lower modulus [7,8].

* Corresponding author. Tel./fax: #886-06-274-8086.


E-mail address: chernlin@mail.ncku.edu.tw (J.H. Chern Lin)

Titanium and titanium alloys have become one of the


most attractive implant materials due to their light
weight, high biocorrosion resistance, biocompatibility
and mechanical properties, including low modulus. For
example, the most widely used titanium alloy, Ti}6Al}
4V, according to Pilliar [9], has an elastic modulus
(108 GPa) only about half of that of 316L stainless steel
(200 GPa) or Co}Cr}Mo alloy (210 GPa) that is
still popularly used today. Although a/b type Ti}6Al}4V
alloy is widely used as an implant material, studies
have reported that the release of Al and V ions from the
alloy might cause some long-term health problems
[10}13]. Moreover, the low wear resistance of Ti}6Al}
4V could accelerate the release of such harmful ions
[14}16].
Recently, much research e!ort was devoted to the
study of more biocompatible, lower modulus, better processability b or near-b Ti alloys, such as Ti}13Nb}13Zr
[17], Ti}11.5Mo}6Zr}2Fe [14] and Ti}15Mo [18]. The
near-b Ti}13Nb}13Zr alloy, developed by Smith and
Nephew Richards Inc. [19], was reported to consist of
hexagonal martensite (a) phase under water-quenched
condition. With subsequent aging, the bcc b phase was
precipitated. This aged Ti}13Nb}13Zr alloy had a lower
(by 30}40%) modulus than mill-annealed Ti}6Al}4V
alloy [17].

0142-9612/99/$ - see front matter  1999 Elsevier Science Ltd. All rights reserved.
PII: S 0 1 4 2 - 9 6 1 2 ( 9 9 ) 0 0 1 1 4 - 3

2116

W.F. Ho et al. / Biomaterials 20 (1999) 2115}2122

The metastable b phase Ti}15Mo alloy is being evaluated for orthopaedic implant applications by Synthes
USA. The rapidly quenched Ti}15Mo alloy was reported
to have a "ne-grained bcc structure with a lower modulus
(77.7 GPa) than that of 316L stainless steel, Grade IV Ti,
Ti}6Al}4V and Ti}6Al}7Nb [18]. According to Bania
[20], to stabilize b phase at room temperature in
a Ti}Mo alloy, a minimum of 10 wt% of this isomorphous b-stabilizing elememt is needed.
In the present study, the structure and properties of
a series of binary Ti}Mo alloys with Mo contents ranging
from 6 to 20 wt% were investigated. Emphasis of the
research was on the comparison of the properties of
b phase alloys with those of the orthorhombic a phase
alloy that seems to have great potential to be a new
candidate for implant material.

2. Experimental procedure
The molybdenum contents selected for this study include 6, 7.5, 9, 10, 12.5, 15, 17.5 and 20 wt%. The series of
the binary alloys were prepared from 99.9% pure titanium and 99.95% pure molybdenum using a commercial
arc-melting vacuum-pressure-type casting system (Castmatic, Iwatani Corp., Japan). The melting chamber was
"rst evacuated and purged with argon. An argon pressure of 1.5 kgf/cm was maintained during melting. Appropriate amounts of metals were melted in a U-shaped
copper hearth with a tungsten electrode. The ingots were
remelted thrice prior to casting to improve chemical
homogeneity.
Prior to casting, the ingots were melted once again in
an open-based copper hearth under an argon pressure of
1.5 kgf/cm. The di!erence in pressure between the two
chambers allowed the molten alloys to instantly drop
into the graphite mold when melted.
The cast alloys were sectioned using a Buehler Isomet low-speed diamond saw to obtain specimens
for various purposes. Surfaces of the alloys for
microstructural study were mechanically polished via
a standard metallographic procedure to a "nal level of
0.3 lm alumina powder and then etched in a solution
of water, nitric acid, and hydro#uoric acid (80 : 15 : 5
in volume). Microstructure of the etched alloys was
examined using an optical microscope (MC80, ZEISS,
Germany).
X-ray di!raction (XRD) for phase analysis was conducted using a Rigaku di!ractometer (Rigaku D-max
IIIV, Rigaku Co., Tokyo, Japan) operated at 30 kV and
20 mA. A Ni-"ltered CuK radiation was used for this
?
study. Phase was identi"ed by matching each characteristic peak with the JCPDS "les. The microhardness of
polished alloys was measured using a Matsuzawa
MXT70 microhardness tester with a load of 200 g for
15 s.

Fig. 1. Schematic diagram of bending test used in this study.

Three-point bending tests were performed using


a desk-top mechanical tester (Shimadzu AGS-500D,
Tokyo, Japan). The bending strengths were determined
using the equation, p"3P/2bh [21], where p is the
bending strength (MPa); P is the load (kg), is the span
length (mm), b is the specimen width (mm), and h is the
specimen thickness (mm) (Fig. 1). The dimensions of the
specimens were: "30 mm, b"5.0 mm and h"
1.0 mm. The modulus of elasticity in bending is
calculated from the load increment and the corresponding de#ection increment between the two points on the
straight line as far apart as possible using the equation,
E"*P/4bh*d, where E is the modulus of elasticity in
bending (Pa), *P is the load increment as measured from
preload (N), and *d is the de#ection increment at midspan as measured from preload. The average bending
strength and modulus of elasticity in bending were taken
from at least "ve tests under each condition. The elastic
recovery (springback) capability for each material was
evaluated from the change in de#ection angle when loading was removed. As schematically shown in Fig. 2, the
springback capability is a measurement of h !h , where


h is the de#ection angle somewhere in the plastic defor
mation regime and h is the angle after loading is re
leased.

3. Results and discussion


3.1. X-ray diwraction
The X-ray di!raction (XRD) results indicated that
crystal structure of the binary Ti}Mo alloy is sensitive to
the composition (molybdenum content) of the alloy. The
results are shown in Fig. 3 and summerized in Table 1. To
save space, &Ti}xMo' will henceforth stand for &Ti}x wt%
Mo' in the text.
As shown in Fig. 3, the c.p. Ti was comprised entirely
of a hexagonal a phase. When Mo content was 6 wt%,
the orthorhombic a phase was observed in the cast alloy,
as indicated in the splitting of the single a (1 0!1 0)
peak into two a peaks [22]. When Mo content was

W.F. Ho et al. / Biomaterials 20 (1999) 2115}2122

2117

Fig. 2. Schematic diagram illustrating elastic recovery measurement.

7.5 wt%, the cast alloy was almost entirely made up of a


phase. When the Mo content increased to 9 wt%, a signi"cant amount of b phase was retained.
When the Mo content increased to 10 wt% or higher,
only the retained b phase was observed in the XRD
patterns. The retention of b phase at higher Mo contents
is consistent with the early result of Davis et al. [23] and
the more recent result of Bania [20]. Bania reported that
a minimum of 10 wt% Mo was needed in a Ti}Mo alloy
to fully stabilize b phase at room temperature.
The sensitivity dependence of phase/crystal structure
on alloy composition in the range between 6 and 10 wt%
Mo can be seen more clearly in a lower scanning speed
(0.53/min) XRD patterns (Fig. 4). The splitting of XRD
peaks was a direct indication of the existence of orthorhombic a phase that distorted the unit cell and decreased its symmetry level. As early as four decades ago
Bagariatskii et al. [22] had found that increasing alloy
content in a number of quenched binary Ti alloys
(Ti}Mo, Ti}Nb, Ti}Ta, etc.) could cause the normally
observed hexagonal a structure to distort and transform
into an orthorhombic a phase. Davis et al. [23] reported
that, in Ti}Mo system, the martensitic structure changed
from hexagonal (a) to orthorhombic (a) at around
6 wt% Mo. As Mf falls below room temperature, b phase
is retained.
3.2. Light microscopy
The microstructure of the series of Ti}Mo alloys, as
shown in Fig. 5, was consistent with the XRD results. The

Fig. 3. X-ray di!raction patterns of c.p. Ti and Ti}Mo alloys.

Table 1
Phases and crystal structure of c.p. Ti and Ti}Mo alloys fabricated in
this study
wt% Mo

Phase

Crystal structure

c.p. Ti
3}5
6
7.5
9
10}20

a
a
a/a
a
a/b
b

Hexagonal
Hexagonal
Hexagonal/orthorhombic
Orthorhombic
Orthorhombic/bcc
bcc

hexagonal c.p. Ti exhibited a typical rapidly-cooled


metastable feather-like microstructure. When Mo content was 6 wt%, the "ne, acicular martensitic structure of
a phase was observed. When Mo content was 7.5 wt%,
the entire alloy was dominated by the martensitic a
structure. When the Mo content increased to 9 wt%,
a signi"cant amount of equi-axed, retained b phase was
observed. When the alloy contained 10 wt% or more
Mo, b phase became the only dominant phase.

2118

W.F. Ho et al. / Biomaterials 20 (1999) 2115}2122

According to Davis et al. [23] and Duerig et al. [24],


the orthorhombic a structure is intermediate between
bcc and hcp structures and hence the phase transition of
bcc b phase to twinned, orthorhombic a involves smaller
strains than that required to produce dislocated, hexagonal a structure. Such smaller strains cause the transformed a martensitic structure to be less hardened. This
might help explain why Ti}7.5Mo alloy had a lower
hardness level than Ti}6Mo alloy. The b phase alloys
generally had a higher hardness level than a and a
probably due to a stronger solid solution e!ect. Since
casting defects, such as porosity/voids, were not observed
under light microscope, their e!ect, if existing, should not
be signi"cant.
3.4. Bending strength and modulus

Fig. 4. X-ray di!raction patterns of Ti}Mo alloys at low scanning


speed.

3.3. Microhardness
As shown in Fig. 6, all the Ti}Mo alloys (containing
6}20 wt% Mo) had much higher microhardness values
(270}340 HV) than that of c.p. Ti (160 HV). Among all
Ti}Mo alloys, the a phase alloy (Ti}7.5Mo) had the
lowest microhardness value (263 HV), while the b phase
alloy (especially those containing 10 and 12.5 wt% Mo)
had the highest hardness. The microhardness value of
Ti}7.5Mo was lower than that of as-cast Ti}6Al}4V
(294 HV) and of Ti}13Nb}13Zr (285 HV) by 10.5 and
7.7%, respectively.
One-way ANOVA statistical analysis and Sche!e multiple comparison indicated that the microhardness values
of all Ti alloys were signi"cantly higher than that of
c.p.Ti (P(0.001). Signi"cant di!erences (P(0.001)
were found among Ti}6Mo, Ti}7.5Mo and Ti}9Mo
alloys. The microhardness values of these three alloys
were signi"cantly lower than those of the alloys containing only b phase (Mo content ranging from 10 to 20
wt%).
Although the measurements were straightforward, the
interpretation of the measured microhardness values
could be quite complex. Such e!ects as solid solution
strengthening, precipitation hardening, strain aging,
grain size and crystal structure/phase (a, a, b) could all
a!ect the hardness of the alloy.

Like microhardness, the bending strengths of all the


Ti}Mo alloys (1400}1750 MPa) were much higher than
that of c.p. Ti (880 MPa), as shown in Fig. 7. Although
a little lower than that of Ti}6Al}4V (1860 MPa), the
bending strength of Ti}10Mo (1750 MPa, the highest
among all Ti}Mo alloys) was twice that of c.p. Ti
(884 MPa). Though lower than that of Ti}6Al}4V, the
bending strength of Ti}7.5Mo (1371 MPa) was similar to
that of Ti}15Mo (1440 MPa) and of Ti}13Nb}13Zr
(1360 MPa), and higher than c.p. Ti by nearly 60%.
The variation in alloy bending strength with Mo content had a trend similar to that in microhardness, except
for the alloys containing 12.5 and 15 wt% Mo. These two
alloys showed an unexpectedly lower strength than other
b phase alloys. The reason is not certain at this moment.
The dependence of modulus was more sensitive on
phase/crystal structure than on other factors. The early
work of Graft et al. [25] indicated that the metastable a
phase in Ti}Mo alloys (with Mo contents close to 3 wt%)
had the lowest elastic modulus among di!erent phases.
The results of the present study, however, strongly suggest that the orthorhombic a phase (with Mo contents
close to 7.5 wt%) has a lower modulus than all other
phases in the binary Ti}Mo system.
It is widely known that b phase Ti alloys generally
have a lower modulus level than that of a or a/b type
alloys [18]. As shown in Fig. 8, all the Ti}Mo alloys
tested in this study, except Ti}10Mo, had lower moduli
than that of c.p. Ti (92 GPa). Among all b phase alloys
(with 10}20 wt% Mo), Ti}10Mo had the highest
modulus (95 GPa), while Ti}15Mo had the lowest
(71 GPa).
It is worth noting that the bending modulus of the
a-dominated Ti}7.5Mo alloy (55 GPa) was even lower
than all b phase Ti}Mo alloys. This bending modulus
was lower than that of Ti}15Mo (with the lowest
modulus among all b phase alloys) by 22%, of Ti}6Al}4V
by 47%, of c.p. Ti by 40%, and of Ti}13Nb}13Zr by
17%. As mentioned in the Introduction, using implant

W.F. Ho et al. / Biomaterials 20 (1999) 2115}2122

materials with lower moduli (closer to that of human


bone) could reduce the stress shielding e!ect.
The advantage in mechanical properties of Ti}7.5Mo
alloy is also demonstrated in their high elastic recovery
capability. High elastic recovery (springback) capability
of a metal is an indication of high strength and low
modulus and is essential for many load-bearing implant
and dental applications. As shown in Fig. 9, the elastic
recovery capability of Ti}7.5Mo was greater than all other
Ti alloys fabricated in this study. For example, the elastically recoverable angle of Ti}7.5Mo was higher than that of
Ti}15Mo by 53%, of Ti}6Al}4V by 46%, of Ti}13Nb}13Zr
by 35%, and of c.p. Ti by as much as 440%.
The phases/crystal structure, hardness and bending
properties of c.p. Ti and a variety of Ti alloys used or

2119

potentially used as implant material are summarized in


Table 2. Such advantages as low hardness, low modulus
and high elastic recovery capability of the a phase
Ti}7.5Mo alloy are clearly demonstrated in the table. In
the current search for a better implant material, the low
hardness, low modulus, excellent elastic recovery capability and reasonably high strength a-dominated
Ti}7.5Mo alloy seems to serve as a promising candidate.

4. Conclusions
1. The hexagonal a phase c.p. Ti exhibited a feather-like
morphology. When 6 wt% Mo was contained, a "ne,
acicular martensitic structure of orthorhombic a

Fig. 5. Light micrographs of c.p. Ti and Ti}Mo alloys.

2120

W.F. Ho et al. / Biomaterials 20 (1999) 2115}2122

Fig. 5. (continued).

Fig. 6. Microhardness of c.p. Ti and Ti alloys.

Fig. 7. Bending strengths of c.p. Ti and Ti alloys.

phase was observed. When 7.5 wt% Mo was contained, the entire alloy was dominated by the martensitic a structure. When the Mo content increased to
10 wt% or higher, retained b phase became the only
dominant phase.
2. Among all Ti}Mo alloys, the a phase Ti}7.5Mo alloy
had the lowest hardness, while the b phase alloys
containing 10 and 12.5 wt% Mo had the highest
hardness. The microhardness value of Ti}7.5Mo was

lower than that of Ti}6Al}4V and of Ti}13Nb}13Zr


by 10.5 and 7.7%, respectively.
3. The bending strength of Ti}10Mo was the highest
among all Ti}Mo alloys and was twice that of c.p. Ti.
The bending strength of Ti}7.5Mo was similar to that
of Ti}15Mo and of Ti}13Nb}13Zr, and higher than
c.p. Ti by nearly 60%.
4. All the Ti}Mo alloys, except Ti}10Mo, had lower
moduli than that of c.p. Ti. Among all b phase alloys

W.F. Ho et al. / Biomaterials 20 (1999) 2115}2122

(with 10}20 wt% Mo), Ti}10Mo had the highest


modulus, while Ti}15Mo had the lowest. The bending
modulus of the a-dominated Ti}7.5Mo alloy was
lower than that of Ti}15Mo by 22%, of Ti}6Al}4V by
47%, of Ti}13Nb}13Zr by 17%, and of c.p. Ti by
40%.
5. The elastically recoverable angle of Ti}7.5Mo was
higher than that of Ti}15Mo by 53%, of Ti}6Al}4V

Fig. 8. Bending moduli of c.p. Ti and Ti alloys.

Fig. 9. Elastic recovery angles of c.p. Ti and Ti alloys.

2121

by 46%, of Ti}13Nb}13Zr by 35%, and of c.p. Ti by as


much as 440%.

References
[1] Sumner DR, Galante JO. Determinants of stress shielding: design
versus materials versus interface. Clin Orthop Relat Res
1992;274:202}12.
[2] Engh CA. Bobyn JD. The in#uence of stem size and extent of
porous coating on femoral bone resorption after primary cemenless hip arthroplasty. Clin Orthop Relat Res 1988;231:7}28.
[3] Lewis JL, Askew MJ, Wixson RL, Kramer GM, Tarr RR. The
in#uence of prosthetic stem sti!ness and of a calcar collar on
stresses in the proximal end of the femur with a cemented femoral
component. J Bone Jt Surg 1984;66-A:280}6.
[4] Koeneman JB, Hansen TM, Toal TR. E!ects of implant geometry
position and boundary conditions on cancellous bone stresses:
a "nite element analysis. Proceedings of Biomechanics Symposium, vol. 120. New York, NY: American Society of Mechanical
Engineers, 1991. p. 117}20.
[5] Cheal E, Spector M, Hayes W. Role of loads and prosthesis
material properties on the mechanics of the proximal femur after
total hip arthroplasty. J Orthop Res 1992;10:405}22.
[6] Prendergast P, Taylor D. Stress analysis of the proximo-medial
femur after total hip replacement. J Biomed Engng 1990;
12:379}82.
[7] Bobyn JD, Mortimer ES, Glassman AH, Engh CA, Miller JE,
Brooks CE. Producing and avoiding stress shielding. Clin Orthop
Relat Res 1992;274:79}96.
[8] Bobyn JD, Glassman AH, Gotto H, Krygier JJ, Miller JE, Brooks
CE. The e!ect of stem sti!ness on femoral bone resorption after
canine porous-coated total hip arthroplasty. Clin Orthop Relat
Res 1990;261:196}213.
[9] Pilliar RM. Modern metal processing for improved load-bearing
surgical implants. Biomaterials 1991;12:95}100.
[10] Rao S, Ushida T, Tateishi T, Okazaki Y, Asao S. E!ect of Ti, Al,
and V ions on the relative growth rate of "broblasts (L929) and
osteoblasts (MC3T3-E1) cells. Bio-Med Mater Engng 1996;
6:79}86.
[11] Yumoto S, Ohashi H, Nagai H, Kakimi S, Ogawa Y, Iwata Y,
Ishii K. Aluminum neurotoxicity in the rat brain. Int J PIXE,
World Scienti"c Publishing Company, 1992;2(4):493}504.
[12] Walker PR, LeBlanc J, Sikorska M. E!ects of aluminium and
other cations on the structure of brain and liver chromatin.
Biochemistry 1989;28(9):3911}5.
[13] McLachlan DRC, Farnell B, Galin H. Aluminum in human brain
disease. In: Sarkar B, editor. Biological aspects of metals and
metal-related disease. New York: Raven Press, 1983. p. 209}18.
[14] Wang K. The use of titanium for medical applications in the USA.
Mater Sci Engng 1996;A213:134}7.

Table 2
Microhardness and bending properties of c.p. Ti and Ti alloys fabricated in this study

c.p. Ti
Ti}6Al}4V
Ti}13Nb}13Zr
Ti}15Mo
Ti}7.5Mo

Microhardness (HV)

Bending strength (MPa)

Bending modulus (GPa)

Elastic recovery angle (deg)

156
294
285
307
263

884
1857
1471
1348
1395

92
105
66
71
55

7
27
29
26
40

2122

W.F. Ho et al. / Biomaterials 20 (1999) 2115}2122

[15] McKellop HA, RoK stlund TV. The wear behavior of ion-implanted
Ti}6Al}4V against UHMW polyethylene. J Biomed Mater Res
1990;24:1413}25.
[16] Rieu J, Pichat A, Rabbe LM, Rambert A, Chabrol C, Robelet M.
Structural modi"cations induced by ion implantation in metals
and polymers used for orthopaedic prostheses. Mater Sci
T 1992;8:589}93.
[17] Mishra AK, Davidson JA, Poggie RA, Kovacs P, FitzGerald TJ.
Mechanical and tribological properties and biocompatibility of
di!usion hardened Ti}13Nb}13Zr*a new titanium alloy for surgical implants. In: Brown SA, Lemons JE, editors. Medical applications of titanium and its alloys: the material and biological
issues, ASTM STP, vol. 1272. West Conshohocken, PA: ASTM,
1996. p. 96}113.
[18] Zardiackas LD, Mitchell DW, Disegi JA. Characterization of
Ti}15Mo beta titanium alloy for orthopaedic implant applications. In: Brown SA, Lemons JE, editors. Medical applications of
titanium and its alloys: the material and biological issues, ASTM
STP, vol. 1272. West Conshohocken, PA: ASTM, 1996. p. 60}75.

[19] Davidson J, Kovacs P. U.S. Patent no. 5169597. Biocompatible


low-modulus titanium alloy for medical implants, 1992.
[20] Bania PJ. Beta titanium alloys and their role in the titanium
industry. In: Eylon D, Boyer R, Koss D, editors. Beta titanium
alloys in the 1990's. Warrendale, PA: TMS, 1993. p. 3}14.
[21] Guha A. Metals handbook, 9th ed., vol. 8. Ohio: ASM International, 1985. p. 133}6.
[22] Bagariaskii IA, Nosova GI, Tagunova TV. Factors in the formation of metastable phases in titanium-base alloys. Sov Phys Dokl
(Engl Transl) 1959;3:1014}8.
[23] Davis R, Flower HM, West DRF. Martensitic transformations in
Ti}Mo alloys. J Mater Sci 1979;14:712}22.
[24] Duerig TW, Albrecht J, Richter D, Fischer P. Formation and
reversion of stress induced martensite in Ti}10V}2Fe}3Al. Acta
Metall 1982;30:2161}72.
[25] Graft WH, Rostoker W. The measurement of elastic modulus of titanium alloys. Symposium on titanium: presented
at the Second Paci"c Area National Meeting, ASTM, 1957.
p. 130}44.