Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Acta Biomaterialia
journal homepage: www.elsevier.com/locate/actabiomat
a r t i c l e i n f o a b s t r a c t
Article history: Ti6Al4V reticulated meshes with different elements (cubic, G7 and rhombic dodecahedron) in Materi-
Received 13 February 2014 alise software were fabricated by additive manufacturing using the electron beam melting (EBM)
Received in revised form 13 May 2014 method, and the effects of cell shape on the mechanical properties of these samples were studied. The
Accepted 9 June 2014
results showed that these cellular structures with porosities of 8858% had compressive strength and
Available online 24 June 2014
elastic modulus in the range 10300 MPa and 0.515 GPa, respectively. The compressive strength and
deformation behavior of these meshes were determined by the coupling of the buckling and bending
Keywords:
deformation of struts. Meshes that were dominated by buckling deformation showed relatively high col-
Electron beam melting
Titanium alloys
lapse strength and were prone to exhibit brittle characteristics in their stressstrain curves. For meshes
Reticulated mesh dominated by bending deformation, the elastic deformation corresponded well to the GibsonAshby
Compressive deformation behavior model. By enhancing the effect of bending deformation, the stressstrain curve characteristics can change
Cell shape effect from brittle to ductile (the smooth plateau area). Therefore, Ti6Al4V cellular solids with high strength,
low modulus and desirable deformation behavior could be fabricated through the cell shape design using
the EBM technique.
2014 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
1. Introduction 0.10.2, which are much lower than that of bulk alloy (0.6)
[11]. Since the above limits can be improved by optimizing cell
Porous titanium and its alloys, which possess low elastic mod- geometry such as shape and size [12], many studies have been per-
ulus matching to bone tissues and are capable of providing space formed. For examples, Nieh et al. [13] investigated open-cell Al
for in-growth of bony tissue to achieve better xation, have been foams fabricated by the directional solidication technique, and
considered as promising replacements for dense implants [1]. their results showed that both elastic modulus and yield stress
Recently, additive manufacturing (AM) using the electron beam were higher along the longitudinal direction than along the trans-
melting (EBM) method has been applied successfully to fabricating verse direction, but the cell size had a negligible effect. Li et al.
titanium cellular meshes and foams [24]. In comparison with [14,15] conducted nite element (FE) simulations, and their results
previous methods, it offers the advantages of accurate control of indicated that the elastic modulus increased as cell shape became
internal pore architectures and complex cell shapes, thus receiving irregular, but decreased with increase in variation of cell wall
extensive attention. thickness. Luxner et al. [16] investigated the spatial deformation
Until now, most of the reported EBM titanium-based open cel- distribution of uniform and disordered cellular structures by the
lular structures and foams have been fabricated using Ti6Al4V FE method and concluded that the catastrophic failure of highly
powders. Systematic and various geometrical arrays of biocompat- porous biological materials would be retarded by a certain degree
ible reticulated meshes and foams with interconnected porous of disorder. Yamada et al. [17] investigated the mechanical proper-
structures have been manufactured with complex monolithic ties of three types of open-cell Mg alloys and explained their
structures [57]. Previous studies have found the following limits: results from the viewpoint of bending, buckling and yielding of
one is their brittle deformation characteristics if they have an the struts. Vesenjak et al. [18] reported that the circular cells
elastic modulus close to that of human bone tissues [3,810]; favored energy absorption. Cote et al. [19] found the sandwich
and another is their low fatigue endurance ratios in the range panels containing corrugated and diamond lattice cores were
weaker than these containing pyramidal and square-honeycomb
Corresponding author. Tel.: +86 24 83978841; fax: +86 24 23972021. cores. Fazekas et al. [20] and An et al. [21] studied the effects of cell
E-mail address: ylhao@imr.ac.cn (Y.L. Hao). shape on the elastic modulus and mechanical behavior of
http://dx.doi.org/10.1016/j.actbio.2014.06.010
1742-7061/ 2014 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
4538 S.J. Li et al. / Acta Biomaterialia 10 (2014) 45374547
two-dimensional and three-dimensional cellular solids and also <1 mm exhibits better biocompatibility than larger cells
found strong dependence on cell geometry. These theoretical and [2325]. Lastly, cells >1 mm can be used in scaffolds of bioactive
experimental results provide helpful insight into the mechanical coatings to overcome the bio-inert nature of titanium alloys [25].
properties and deformation behavior of reticulated meshes. The apparent density of the open cellular structure was
In this study, three types of reticulated meshes with cubic, G7 estimated from its mass and apparent volume. The porosity P
and rhombic dodecahedron cell shapes were fabricated by the was calculated as
EBM method using Ti6Al4V (ELI) alloy powders. The compres-
sion tests demonstrate that the meshes prone to buckling deforma- P 1 q=q0 100% 1
tion tend to have high strength and be brittle, while these prone to
where q and q0 are the densities of cellular and bulk alloys,
bending deformation exhibit a contrasting tendency. These results
respectively.
suggest that cell structure design is an effective way to balance the
The microstructures of these meshes were characterized by
strength, elastic modulus and deformation behavior of the reticu-
optical microscopy (OM) and scanning electron microscopy
late meshes.
(SEM). The specimens for OM observation were mounted, mechan-
ically polished and etched in a solution consisting of 8 vol.% HF, 15
2. Experimental vol.% HNO3 and 77 vol.% H2O. The cell structure and size were
studied by SEM using a Hitachi S-3400N microscope primarily
Reticulated meshes of Ti6Al4V alloy with extra low intersti- operated at 20 kV. The Youngs moduli of these cellular specimens
tials (ELI) were fabricated by an EBM system (Arcam A1, Arcam, of 30 30 60 mm were measured by means of a damping ana-
Sweden). These monolithic structures were built layer-by-layer lyzer (RFDAHTVP1750-C), which measures the resonant frequency
using precursor, medical-graded Ti6Al4V (ELI) powders, with of a specimen arising from a mechanically induced vibration.
an average particle diameter of 50 lm. Each powder layer was Youngs modulus (E) is generally given by [26]
created by raking powder gravity fed from two cassettes, and then
2 3
heated to 730 C by multiple pre-heat scans, followed by melting E 0:9465 mf r =w l =t3 T 2
of the selected areas controlled by a CAD program. In the study,
CAD programs were used to design meshes with unit cells of cubic, where fr is the measured resonant frequency, m, w, l and t are the
G7 (Materialise/Magics cell unit) and rhombic dodecahedron, as mass, width, length and thickness of the specimen, respectively,
shown in Fig. 1ac, respectively. To obtain different densities and T is a factor related to l, t and the Poissons ratio of the material.
(0.52.0 g cm3), porosities (5888%) and cell sizes (0.53.0 mm), Specimens for the compression test were prepared with dimensions
the built units are dimensionally scaled under the condition of 20 20 30 mm to avoid free-surface effects. Static compression
constant strut thickness 0.5 mm. The above cell parameters were tests were carried out using an Instron 5582 instrument at room
selected on the basis of previous studies [2225]. First, higher temperature under a strain rate of 103 s1. To measure the static
porosity favors better bone ingrowth [22]. Second, a cell size modulus, the calibrated strain extensometer was used to record
Fig. 1. (ac) Cubic, G7 and rhombic dodecahedron element in the Materialise software and (df) the corresponding Ti6Al4V prototype blocks fabricated by the EBM
method.
S.J. Li et al. / Acta Biomaterialia 10 (2014) 45374547 4539
(a) (b)
500m 500m
(c) (d)
500m 50m
Fig. 3. SEM images of the meshes with (a) cubic, (b) G7 and (c) rhombic dodecahedron cells and (d) typical OM image of mesh strut.
4540 S.J. Li et al. / Acta Biomaterialia 10 (2014) 45374547
102 35
Young's modulus, GPa Cubic 0.62 g/cm3
G7
(a)
30 0.73 g/cm3
rhombic dodecahedron
rhombic dodecahedron 0.91 g/cm3
101 25
Stress, MPa
20
100 15
10
5
10-1
0.0 0.5 1.0 1.5 2.0
0
Density, g/cm3 0 2 4 6 8
Compressive strength, MPa
3 Strain,%
10
Cubic
G7 (b) Fig. 5. Loading and unloading stressstrain curves recorded by the strain exten-
rhombic dodecahedron someter to estimate the static modulus of three meshes with rhombic dodecahe-
2 dron structure.
10
300 250
1.6g/cm3 (b)
0.9g/cm3
(a) 1.8g/cm3
1.6g/cm3
0.7g/cm3
200
1.0g/cm3
Stress, MPa
Stress, MPa
200 0.6g/cm3 0.8g/cm3
0.5g/cm3 150
100
100
50
0 0
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8
Strain, mm/mm Strain,mm/mm
140
0.62g/cm3 (c)
120 0.73g/cm3
0.91g/cm3
100
Stress, MPa
1.18g/cm3
80 1.68g/cm3
60
40
20
0
0.0 0.2 0.4 0.6 0.8 1.0
Strain, mm/mm
Fig. 6. Nominal compressive stressstrain curves of the meshes with (a) cubic, (b) G7 and (c) rhombic dodecahedron structures.
the deformation behavior of cell wall/strut (referred to as cell wall to a higher modulus. Therefore the meshes have higher and lower E
and strut for cellular foam and mesh, respectively, in general). The if their structures tend to buckling (larger P2) and bending (larger
early studies failed to establish accurate relations because the P1), respectively. The analysis of strength also has a similar
dominated deformation mechanisms of the cell walls had not been tendency.
revealed [3234]. Later investigations have shown that, for the cel- It should be noted that, for stochastic foams of ductile materials
lular solids, their elastic moduli are related primarily to bending such as polymer and aluminum foams, previous studies have
deformation of cell walls, elastic collapse is caused by elastic buck- shown their deformations dominated mainly by the bending
ling, and plastic collapse is initiated by plastic hinges [3538]. mechanism and neglected generally for the buckling mechanism.
Based on these deformation modes, a simple model was estab- The study of reticulate meshes of Ti6Al4V alloy, however, found
lished to estimate accurately mechanical properties such as elastic that their deformations were controlled by both mechanisms
modulus and collapse strength using the properties of cell wall and (Fig. 9). Since the increase in bending contribution tends to
the cell geometry as input [37,38]. The following investigations on ductility of the meshes (Fig. 6), the following experiments were
the mechanical properties of cellular solids are based mainly on conducted to give a solid conrmation: for the rhombic dodecahe-
this model. dron meshes presented in Fig. 10, the a angles were designed as 36,
The videos and in situ observations presented in Section 3.3 30 and 23, corresponding to the increase in bending component
show that the cell struts are deformed mainly by buckling and/or P1. The compression tests (Fig. 10) found that the stress plateau
bending in compression. Thus, different mechanical behavior of gradually became smooth with the decrease in the angle, while
the meshes should be related to the interaction of two deformation the strength and modulus decreased correspondingly. The above
mechanisms. To characterize their contributions, the load P applied analyses suggest that the brittle deformation behavior of the
to a strut in Fig. 9 was divided into two components P1 and P2 along EBM meshes studied can be avoided by promoting the bending
the bending and buckling deformation directions, respectively. It is deformation mechanism through structure design.
clear that the P2 component decreases in the order cubic, rhombic
dodecahedron and G7, whereas P1 follows the reverse order. 4.2. Comparison with the classical model
The above variation tendencies of P1 and P2 components with
cell shapes are in agreement with those of the strength, modulus Over several decades, many models have been established to
and deformation behavior described in Sections 3.2 and 3.3. That relate the porous structure and mechanical properties. The most
is, under identical density, higher P1 (lower P2) correspond to lower successful one is the GibsonAshby model [12]. According to the
strength and modulus, and their stressstrain curves are easily model, the relative modulus (E/Es), the relative collapse strength
shown to have a ductile character. For example, Fig. 4 shows that (rp/rs), the relative buckling strength (rp/Es) and the relative den-
strength and modulus decrease in turn for the cubic, rhombic sity (q/qs) of cellular materials can be described as
dodecahedron and G7 cells, while their deformation behavior
changes from brittle to ductile characteristics. The above empirical
E=Es q=qs 2 3
relation can be explained by the simple model presented in Fig. 9.
As for the elastic modulus E of a mesh, the analyses in the Appen- rp =rs Cq=qs 1:5 4
dix (Eq. (A19)) shows that it is proportional to sin a= cos2 a, in
which a is the angle between P1 and P. As a result, the larger a leads rp =Es C 0 q=qs 2 5
4542 S.J. Li et al. / Acta Biomaterialia 10 (2014) 45374547
(a)
(b)
(e) (f)
(c)
(a)
P P P
A A A
B B B
500m
P P P
D D D
C C C
500m
P (P2)
A D F
P2 P2
P1 P1
P P
B C E
Fig. 9. Schematic to depict the buckling and bending vectors of the load applied on (a) cubic, (b) G7 and (c) rhombic dodecahedron cell struts.
In this study, the reticulated meshes with different cell shapes irregular struts can be analyzed with aid of the FE method [39], the
were manufactured by the AM-EBM method. The relative modulus, consistency of theoretical and experimental n exponent can be
strength and density of these meshes are measured and plotted in achieved.
Fig. 11. For the relative modulus and density, the experimental The relationship between the relative strength and density is
exponent n is 2.4, 2.0 and 2.2 for cubic, G7 and rhombic dodecahe- also linear, with the exponential factor n0 of 1.7, 1.9 and 2.2 for
dron structures, respectively (Fig. 11a). These results indicate that cubic, G7 and rhombic dodecahedron meshes, respectively
the factors for G7 and rhombic dodecahedron meshes dominated (Fig. 11b, c), which shows the discrepancy with the ideal value of
by the bending deformation match well with the theoretical value the GibsonAshby model in the Appendix. This discrepancy can
of the model in the Appendix. As for the cubic meshes dominated be explained as follows. One is the microstructure, irregular
by the buckling deformation, however, the experimental factor surfaces of the struts, which is crucially important to determine
(2.4) is much higher than the theoretical one (1). The inconsis- the plastic strength. The other is the cell shape. For the deforma-
tency would be caused by the unreasonable assumption that the tion of G7 and rhombic dodecahedron meshes in the Appendix,
struts are uniform. Actually, they are very irregular (Fig. 8), owing only the bending component was considered. In order to obtain
to the limitation of the EBM method. Since the deformation of such an accurate theoretical estimate of the strength of the cellular
4544 S.J. Li et al. / Acta Biomaterialia 10 (2014) 45374547
16 cubic (a)
b (a) G7
c 0.1 rhombic dodecahedron
12 d
E/Es
Stress,MPa
8 0.01
4
1E-3
0.1 0.2 0.3 0.4 0.5
/s
0
0 30 60 90 1
Strain,% (b) G7
rhombic dodecahedron
(b) F (c) F
(d)
F
p/s
E E E 0.1
F F F 0.01
0.1 0.2 0.3 0.4 0.5
E E E /s
10
Cubic (c)
F F
P2 P2 F
P2
p/Es
P1 P1 1
P P1
P P
E E E
Fig. 10. (a) Stressstrain curves of the reticulated meshes with different rhombic
dodecahedron unit cells; (b) unit cell designed by the Materialise software; (c) and 0.1
(d) unit cells designed to increase the bending component of the load applied on the 0.1 0.2 0.3 0.4 0.5
struts. /s
Fig. 11. Plots of (a) the relative modulus (E/Es), (b) the relative strength (rp/rs) and
(c) (rp/Es) vs. relative density (q/qs) for the Ti6Al4V meshes studied.
structure, the coupling effect of the bending and buckling compo-
nent should be taken into account in the theoretical analysis, espe-
and desirable deformation behavior can be fabricated by the
cially for the structures with a high buckling component.
EBM technique.
5. Conclusions
Acknowledgements
In this study, the reticulated meshes of Ti6Al4V alloys with
cubic, G7 and rhombic dodecahedron cells were fabricated by This work was partially supported by MOST grants
AM using the EBM method. The effects of the cell shape on the 2012CB933902, 2012CB619103 and 2012CB933901, and NSFC
compressive deformation behavior were investigated. The results grants 51271182 and 51271180.
are summarized as follows.
Appendix A. Mechanistic model of the studied mesh arrays
1. The compressive deformation behavior and strength of the
studied meshes are determined by the coupled mechanisms
Cubic-cell mesh
of the buckling and bending.
2. Meshes dominated by the buckling mechanism possess a
A representative cubic unit cell and schematic drawing for load-
high collapse strength and exhibit brittle characteristics in
ing analysis are shown in Fig. 1Aa. The compressive load P applied
the stressstrain curves.
on the strut can be expressed as
3. Meshes dominated by the bending mechanism obey the the-
oretical GibsonAshby model. By enhancing the bending mech- rl2
P A1
anism through the cellular structure design, the stressstrain 4
curve exhibits ductile characteristic, whereas the strength where r is the stress applied on the meshes, and l is the length of
decreases correspondingly. the strut. Assuming the strut has a uniform cross section, its linear
4. Youngs modulus and the compressive strength of these elastic deformation Dl is [40]
meshes are in the range 0.515 GPa and 10300 MPa, respec- 3
tively. By optimizing the buckling and bending deformation, Pl rl
Dl A2
Ti6Al4V cellular solids with high strength, low modulus Es A 4Es t 2
S.J. Li et al. / Acta Biomaterialia 10 (2014) 45374547 4545
q t2 G7-cell meshes
/ A4
qs l2
Then, Eq. (A3) becomes A representative G7 unit cell and the loading analysis are shown
in Fig. 1Ab Based on the SEM observation, the strut was dominated
E q by the bending mechanism. The deformation coordination equa-
C1 A5
Es qs tions are
The collapse stress rp, can be obtained by the same method. d11 X 1 d12 X 2 D1P 0 A9
SEM observation indicates that buckling is the major deformation
mode for the struts of cubic-cell meshes, hence, according to
d21 X 1 d22 X 2 D2P 0 A10
Eluthers theory, the collapse force carried on the strut section
Pcr, is obtained by the following formula where d11, d12 and D1P are the unit displacements along the X axis
p2 Es I 4p2 Es I caused by X1, X2 and P, respectively; d21, d22 and D2P are the unit dis-
Pcr 2
2
A6 placements along the Y axis caused by X1, X2 and P, respectively.
ll l
Based on the Moore theorem, these coefcients can be calculated as
where I is the two moment of inertia, l is the degree of constraint at Z 2
the ends of the column. Then, the collapse stress of the meshes rp is 1 L
sin aL3
d11 M 1 2 dx A11
given by EI 0 3EI
4546 S.J. Li et al. / Acta Biomaterialia 10 (2014) 45374547
Z 2
1 L
sin aL2 E q
d12 M 1 M 2 dx A12 C5 A25
EI 0 2EI Es qs
Z L 1:5
1 L
d22 M 2 2 dx A13 rp q
EI EI C6 A26
0 ry qs
Z L
1 1 SEM observation indicates that bending is the dominant defor-
D1P M1 MP dx P sin a cos aL3 A14 mation mechanism for rhombic dodecahedron struts in this work.
EI 0 3EI
The loading direction is consistent with direction 1, as shown in
Z Ref. [41]. Thus, the elastic and plastic deformation behavior of
1 L
PL2
D2P M2 MP dx cos a A15 the rhombic dodecahedron open cellular structures studied are
EI 0 2EI
suggested to obey Eqs. (A25) and (A26).
Take Eqs. (A11)(A15) into Eqs. (A9) and (A10) to obtain
[21] An Y, Wen C, Hodgson PD, Yang CH. Investigation of cell shape effect on the [31] Hao YL, Niinomi M, Kuroda D, Fukunaga K, Zhou YL, Yang R, et al. Youngs
mechanical behaviour of open-cell metal foams. Comput Mater Sci modulus and mechanical properties of Ti29Nb13Ta4.6Zr in relation to a00
2012;55:19. martensite. Metall Mater Trans A 2002;33:313744.
[22] Li JP, Habibovic P, Van den Doel M, Wilson CE, De Wijn JR, Van Blitterswijk CA, [32] Lederman JM. The prediction of the tensile properties of exible foams. J Appl
et al. Bone ingrowth in porous titanium implants produced by 3D ber Polym Sci 1971;15:693703.
deposition. Biomaterials 2007;28:281020. [33] Chan R, Nakamura M. Mechanical properties of plastic foams: the dependence
[23] Wu SH, Li Y, Zhang YQ, Li XK, Yuan CF, Hao YL, et al. Porous Titanium- of yieldstress and modulus on the structural variables of closed-cell and open-
6Vanadium cage has better osseointegration and less micromotion than a cell foams. J Cell Plast 1969;5:1128.
poly-ether-ketone cage in sheep vertebral fusion. Artif Organs [34] Barma P, Rhodes MB, Salovey R. Mechanical properties of particulate-lled
2013;37(12):E191201. polyurethane foams. J Appl Phys 1978;49:498591.
[24] Ponader S, Vairaktaris E, Heinl P, Wilmowsky C, Rottmair A, Korner C, et al. [35] Menges G, Knipschild F. Estimation of mechanical properties for rigid
Effects of topographical surface modications of electron beam melted Ti polyurethane foams. Polym Eng Sci 1975;15:6237.
6Al4V titanium on human fetal osteoblasts. J Biomed Mater Res A [36] Ko WL. Deformation of foamed elastomers. J Cell Plast 1965;1:4550.
2008;84:11119. [37] Gibson LJ, Ashby MF, Schajer GS, Robertson CI. The mechanics of two-
[25] B.D. Liu, Effect of pore sizes of porous titanium alloys as bone material on bone dimensional cellular materials. Proc R Soc Lond A 1982;382:2542.
formation (Master thesis), Fourth Military Medical University, 2012. [38] Gibson LJ, Ashby MF. The mechanics of three-dimensional cellular materials.
[26] Manual of RFDA system 23 (RFDA-MF, version 6.3.0), p. 14. Proc R Soc Lond A 1982;382:4359.
[27] Gu DD, Meiners W, Wissenbach K, Poprawe R. Laser additive manufacturing of [39] Campoil G, Borleffs MS, Yavari SA, Wauthle R, Weinans H, Zadpoor AA.
metallic components: materials, processes and mechanisms. Int Mater Rev Mechanical properties of open-cell metallic biomaterials manufactured using
2012;57:13364. additive manufacturing. Mater Des. 2013;49:95765.
[28] Goldstein SA. The mechanical properties of trabecular bone: dependence on [40] Beer FP, Johnston ER, Dewolf JT. Mechanics of Materials. 3rd ed. New
anatomic location and function. J Biomech 1987;20:105561. York: McGraw-Hill; 2001.
[29] Reilly DT, Burstein AH. The mechanical properties of cortical bone. J Bone Joint [41] Babaee S, Jahromi BH, Ajdari A, Nayeb-Hashemi H, Vaziri A. Mechanical pro
Surg Am 1974;56:100122. perties of open-cell rhombic dodecahedron cellular structures. Acta Mater
[30] Ashby MF, Evans A, Fleck NA, Gibson LJ, Hutchinson JW, Wadley HNG. Metal 2012;60:287385.
Foams: A Design Guide. Waltham, MA: Butterworth-Heinemann; 2000.