Biomaterials 29 (2008) 36253635

Contents lists available at ScienceDirect

Biomaterials
journal homepage: www.elsevier.com/locate/biomaterials

Porous titanium scaffolds fabricated using a rapid prototyping and powder

metallurgy technique
Garrett E. Ryan a, b, Abhay S. Pandit a, *, Dimitrios P. Apatsidis b
a
b

National Centre for Biomedical Engineering and Science, National University of Ireland, Galway, Ireland
Department of Mechanical and Biomedical Engineering, National University of Ireland, Galway, Ireland

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 31 March 2008
Accepted 19 May 2008
Available online 16 June 2008

One of the main issues in orthopaedic implant design is the fabrication of scaffolds that closely mimic the
biomechanical properties of the surrounding bone. This research reports on a multi-stage rapid prototyping technique that was successfully developed to produce porous titanium scaffolds with fully interconnected pore networks and reproducible porosity and pore size. The scaffolds porous
characteristics were governed by a sacricial wax template, fabricated using a commercial 3D-printer.
Powder metallurgy processes were employed to generate the titanium scaffolds by lling around the wax
template with titanium slurry. In the attempt to optimise the powder metallurgy technique, variations in
slurry concentration, compaction pressure and sintering temperature were investigated. By altering the
wax design template, pore sizes ranging from 200 to 400 mm were achieved. Scaffolds with porosities of
66.8  3.6% revealed compression strengths of 104.4  22.5 MPa in the axial direction and 23.5  9.6 MPa
in the transverse direction demonstrating their anisotropic nature. Scaffold topography was characterised using scanning electron microscopy and microcomputed tomography. Three-dimensional reconstruction enabled the main architectural parameters such as pore size, interconnecting porosity, level
of anisotropy and level of structural disorder to be determined. The titanium scaffolds were compared to
their intended designs, as governed by their sacricial wax templates. Although discrepancies in architectural parameters existed between the intended and the actual scaffolds, overall the results indicate
that the porous titanium scaffolds have the properties to be potentially employed in orthopaedic
applications.
2008 Elsevier Ltd. All rights reserved.

Keywords:
Titanium
Scaffold
Rapid prototyping
Three-dimensional printing
Microstructure

1. Introduction
Scaffolds are of signicance for orthopaedic tissue engineering
applications as they allow biological anchorage to the surrounding
bone tissue through the ingrowth of mineralized tissue into the
porous network [1]. Although porous ceramics and polymers have
been studied as potential bone graft scaffolds [24], they cannot
reliably satisfy the mechanical demands of load-bearing applications, such as stand alone interbody spinal fusion devices [5,6]. For
this reason, porous metals including titanium [7,8], titaniumnickel
[9,10], and tantalum [1113] have been investigated. Titanium
metal is widely chosen due to its corrosion resistance, high
strength-to-weight ratio, and proven biological acceptance [14,15].
Several techniques have been developed to introduce a degree of
porosity in titanium and titanium alloy scaffolds including

* Corresponding author. Tel.: 353 91 492758; fax: 353 91 563991.

E-mail addresses: garrett.ryan@nuigalway.ie (G.E. Ryan), abhay.pandit@nuigalway.ie (A.S. Pandit), dapa@nuigalway.ie (D.P. Apatsidis).
0142-9612/$ see front matter 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.biomaterials.2008.05.032

compression and sintering of beads and bres [16,17], combustion

synthesis [18], electron beam melting [19,20], solid-state foaming
by expansion of argon-lled pores [21] and polymeric sponge
replication [22].
Notwithstanding the advancements that have attained in scaffold fabrication, the control over scaffold architecture using these
conventional techniques is highly process dependent. Rapid prototyping (RP) techniques are considered as a viable alternative for
achieving extensive and detailed control over scaffold architecture
[23,24], by combining computer-aided design (CAD) with computer-aided manufacturing (CAM). RP makes it possible to build
objects with predened microstructure and macrostructure and
provides the potential for making scaffolds with controlled hierarchical structures [25]. The transfer of RP technologies to metallic
materials for tissue engineering and orthopaedic implants poses
a signicant challenge, although a number of investigators have
made substantial progress in this regard. Curodeau et al. produced
porous surfaced CoCr implants using a variation of the lost wax
process [26]. By pouring molten CoCr into porous ceramic moulds
that had been fabricated using a three-dimensional printing

3626

G.E. Ryan et al. / Biomaterials 29 (2008) 36253635

technique, textures with ve layers of distinct geometric denition

were created. Using a similar investment casting technique, LopezHeredia et al. created a porous titanium scaffold with 100% interconnected structure and porosity of around 60% [27,28]. Abundant
bone formation was found inside the rapid prototyped titanium
scaffolds after 3 and 8 weeks of implantation in the femoral
epiphysis of New Zealand White rabbits [27]. Using a technique
known as three-dimensional ber deposition, Li et al. produced
porous Ti6Al4V scaffolds by depositing Ti6Al4V slurry through
a computer controlled syringe followed by sintering of the titanium
powders [29]. Hollander et al. used a technique called direct laser
forming to create porous scaffolds from Ti6Al4V powder followed by
sand blasting of the scaffold to remove loose powder particles [30].
A key requirement for such RP technologies is control over the
scaffolds pore structure, including pore size, shape, volume and
interconnectivity [31]. It is generally accepted that pore sizes between 100 and 400 mm are optimal for bone ingrowth [32]. High
pore interconnectivity greatly affects mass transport through the
scaffold and is necessary to ensure adequate delivery of cells and
nutrient supply during subsequent culture throughout the complete porous scaffold [33,34]. Porosity and pore geometry also inuences the scaffolds mechanical properties and affect stress
shielding and fatigue strength [35,36]. Tight control over the scaffolds porous architecture requires that reliable methods are essential for its characterisation following the design and fabrication
processes. This enables an accurate assessment of the level of
precision that the fabrication process can deliver. Although scanning electron microscopy (SEM) images provide high resolution
and detailed views of the surface topology an inherent weakness is
its limitation to two-dimensional measurements on relatively small

elds of view. State of the art three-dimensional imaging techniques enable researchers to describe the complex structure of
materials more accurately. Microcomputed tomography (mCT),
a non-destructive technique, is being increasingly used to provide
structural information within an object by mathematically reconstructing its three-dimensional images from a consecutive series of
x-ray images [3739]. This technique provides the opportunity to
experimentally measure the complex morphology of the pore
space of scaffolds in three dimensions at resolutions in the micrometre range.
The objective of this study was to develop a porous titanium
scaffold with controlled architecture using a multi-stage RP technique. The major parameters involved in the fabrication process
were examined, in order to evaluate the signicance of each parameter on the scaffolds mechanical properties. Subsequently,
scaffolds were created to identify the inuence of increasing porosity on the mechanical and structural properties. These scaffolds
were scanned using a mCT scanner and their structural characteristics were calculated using the 3D digitized images. In this way
a comparison with the initial CAD designs was possible and any
differences in topography could be assumed to be due to inaccuracies of the fabrication process. Finally, in vitro studies were conducted to assess the cellular response to the porous scaffold.
2. Materials and methods
2.1. Scaffold fabrication
In this section we outline the method used to create the porous titanium scaffolds.
A porous wax model was fabricated using a Thermojet (3DSystems Corporation,
Valencia, CA) 3D-printer according to the CAD template shown in Fig. 1(A). The

Fig. 1. (A) Schematic of the CAD design template used to create the sacricial wax model, and (B) schematic demonstrating the pore space reconstruction and centreline generation
from serial mCT scans.

G.E. Ryan et al. / Biomaterials 29 (2008) 36253635

3627

Table 1
Process parameters used for optimising the powder metallurgy fabrication process
Process parameter
Powder metallurgy parameters

Variables
Slurry concentration
(mg Ti/ml ethylene glycol)
Pressure (MPa)
Sintering temp. ( C)

3D-printed sacricial
wax template

template consists of a repeating array of unit-cells in the three principal directions. A

restriction to the wax template design was presented by the 3D-printer. A dispensing
angle equal to or less than 7 needs to be adhered to, because the printer will lay down
automatic support structures, if this angle is exceeded. The wax templates were
designed to accommodate this limitation, according to the schematic shown in
Fig. 1(A). The wax model was placed in a compaction die and a layer of absorptive
paper was placed underneath the wax. Titanium powder (350 mesh, grade 2) (Alfa
Aesar GmbH & Co KG, Karlsruhe, Germany) was mixed with ethylene glycol solution
(SigmaAldrich, MO) using a vortex mixer to form a titanium slurry and was immediately poured into the die under gravity. Over a period of 12 h the tissue paper
absorbed almost all of the ethylene glycol from the slurry by capillary action, drying
out the wax-titanium construct. The absorptive paper was removed and the construct
was then compacted using a hydraulic press. During the compaction process, the
press was switched from load control to displacement control to prevent further
compaction of the construct. A cartridge heater, built into the die, was employed to
heat the die at a rate of approx. 5  C/min to approx. 90  C. The heating step was
monitored using a thermocouple. The temperature was maintained at 90  C for
15 min to completely melt the wax template. At this point, the load from the hydraulic
press was removed. Once the construct had cooled to room temperature the waxtitanium construct was removed from the die and placed in xylene solution at 60  C
for 15 min. This procedure was repeated three times to ensure complete dissolution
of the wax. The resulting porous titanium scaffold was placed in a high vacuum
furnace (105 mbar) at temperatures ranging from 1100 to 1300  C for 1 h to sinter
the titanium powder particles. The sintering step produced a functionally stable titanium scaffold with a porous structure that approximated the inverse morphology
of the sacricial wax template. By modifying the design parameters of the sacricial
wax template (Fig. 1(A): 1, 2, L1L4), and by adjusting powder metallurgy (PM)
parameters, scaffolds with various pore sizes and porosities could be fabricated.
Scanning electron microscopy (SEM) pictures were taken with a Hitachi S-4700
(HitachiHisco Europe GmbH, Berkshire, UK) with the following settings: 15 kV
Voltage and 10 mA emission current. Energy dispersive analysis (EDX, Oxford Instruments, Oxfordshire, UK) was performed at a working distance of 15 mm and
15 kV to distinguish the elements present at the surface of the scaffold.
2.2. Process optimisation
The major parameters involved in the fabrication process were examined to
determine their inuence on the mechanical strength of the titanium scaffolds. A
summary of these parameters is presented in Table 1. Initial testing was conducted
to examine the inuence of the PM parameters, which involved investigating the
effect of slurry concentration, compaction pressure and sintering temperature. For
these tests, a common wax template (L1 L2 4.0 mm, L3 L4 1.0 mm,
1 350 mm, 2 700 mm) was chosen and the titanium scaffolds were fabricated
using this template. To determine the signicance of a single variable within a given
parameter, all other parameters were held constant. The highest value of each parameter was chosen as the constant value. The resulting titanium scaffolds were
subjected to uniaxial compression tests (n 3) along their transverse direction at
a rate of 1 mm/min using a universal testing machine (Instron 8874; Instron Corporation, Norwood, MA, USA). The compression strengths of the scaffolds were
compared to assess the inuence of each parameter. All parameter variables were
compared against the variables that yielded highest scaffold strength. Statistical
analyses were carried out using statistical software (Minitab, v. 13.32). Statistical
variance between parameter variables was determined using one-way analysis of
variance (ANOVA). Tukeys honesty signicant difference test was used for post hoc
evaluation of differences between groups. A P value of <0.05 was considered to be
statistically signicant.
To study the inuence of the 3D-printed sacricial template, three different wax
templates were prepared using CAD software (AutoCAD 2002; Autodesk, Inc., CA).
The design variables for the three wax templates are presented in Table 1. These
values were chosen to create scaffolds with increasing levels of porosity based on
initial testing using various designs. The PM parameters, which resulted in highest
mean strength values, were chosen to create the three titanium scaffolds. Uniaxial
compression tests (n 3) were performed along the axial and transverse directions
to examine the mechanical properties of the resulting scaffolds.

5/7

5/5

5/3

50
1100
Template 1

150
1200
Template 2

250
1300
Template 3

1 350 mm
2 700 mm
L1 L4 4 mm
L2 L3 1 mm

1 350 mm
2 700 mm
L1 L4 3.6 mm
L2 L3 1 mm

1 400 mm
2 800 mm
L1 L4 4 mm
L2 L3 1 mm

2.3. Characterisation of scaffold morphology

Three cylindrical porous titanium samples, approximately 14 mm in diameter
and 15 mm in height, from each scaffold design described in Table 1 were fabricated.
Total scaffold porosity was determined by measuring the apparent density of the
scaffold using volume and weight measurements and the known solid density of
titanium 4507 kg/m3. The samples were scanned using a mCT desk scanner (SkyScan 1072; e2v Scientic Instruments Ltd., UK). The output format for each sample
was 600 serial 1024  1024 bitmap images with an isotropic voxel size of 15 mm.
Using MatLab v. 7.0.1 (The Mathworks, Inc., MA), sequential 500  500 pixel images
were cropped from the serial images of each sample. The images were converted to
black and white and the black-to-white ratio was used to determine the porosity for
each consecutive image. In this way, porosity could be determined as a function of
height for each of the scaffolds. Based on these images and using a reconstruction
software (Mimics v. 10.1; Materialise, Leuwen, Belgium) 3D models of each scaffold
were created. Thresholds of the grey scale images were inverted to allow measurements of the volume of all pore spaces. Subsequently, a region-growing operation was performed, creating a mask consisting only of interconnected pore spaces.
Volume for this region-grown mask was also determined, and the ratio of regiongrown volume to total volume was calculated. This percentage is described as the
interconnecting porosity. To analyse the distribution of pore size throughout the
scaffold, the pore space over the height of three unit-cells was isolated as demonstrated in Fig. 1(B). Using Mimics, centre lines were constructed for the pore space
and the best-t diameter was obtained. This process was performed at three random
locations within each scaffold and the average pore size, as a function of height, was
determined for scaffolds from each design. The degree of anisotropy, dened as unitcell horizontal length/axial length, was determined for each of the three scaffolds.
Three unit-cells were isolated from three scaffolds at each porosity level to produce
an average of these measurements. The level of structural disorder within each
scaffold was evaluated by counting the number of non-interconnecting struts in
a given cross-section and dividing by the total possible number of interconnecting
struts. Three random cross-sections were chosen from three scaffolds at each porosity level for this process.
To assess the closed-cell microporosity formed through sintering of the titanium
powder particles as would be found in the struts of the nal porous scaffold
cylindrical titanium billets were prepared without the use of the wax template but
using identical PM parameters. Scanning electron microscopy (SEM) was used to
assess the surface topography of the billets, and the billet porosity was evaluated
using the identical technique used for the open-cell porous scaffolds.
2.4. In vitro experiments
SAOS-2 pre-osteoblast cells were cultured on porous titanium scaffolds over
a period of 3 weeks. Cylindrical scaffold samples with a diameter of 14 mm and
height of 8 mm were used for this purpose. The scaffolds had a mean pore size of
465  170 mm and porosity of 45.1 1.7%. Standard cell culture plastic was used as
the control surface. The culture medium was McCoys-5A supplemented with 10%
fetal bovine serum (SigmaAldrich) and 1% penicillin and streptomycin (Sigma
Aldrich). The samples were placed in a 24-well plate (Sarstedt, Numbrecht, Germany) and seeded at 5 x 104 cells/per sample in 1 ml of medium, and cultured at
37  C in a humidied atmosphere with 5% CO2 concentration. The medium was
changed every 23 days. After 14 days of culture, the medium was changed to
complete McCoys medium supplemented with ascorbic acid (50 mg/L; Sigma
Aldrich, Ireland), dexamethasone (10 ng/L; SigmaAldrich, Ireland) and b-glycerophosphate (10 mM; SigmaAldrich, Ireland) [4042]. All assays were performed after
1, 7, 14 and 21 days in culture.
Cell viability was assessed using the AlamarBlue assay (BioSource). The cellseeded scaffolds were removed from the original wells and placed in fresh wells
containing 1 ml of 10% (v/v) AlamarBlue solution in Hanks balanced salt solution
(SigmaAldrich). After 1 h of incubation, the uorescence was measured using an
FLx800 microplate uorescence reader (Bio-Tek Instruments, Inc.) at 520 and
590 nm excitation and emission wavelengths, respectively.
Cell proliferation on the titanium scaffolds was determined by DNA analysis. The
substrates were placed in a new well containing 1 ml deionized distilled water and

3628

G.E. Ryan et al. / Biomaterials 29 (2008) 36253635

Fig. 2. (A) Top and side proles of a porous titanium scaffold with 59.1% porosity and (B) scanning electron micrographs of porous titanium scaffolds with pore sizes of 200, 300, and
400 mm.
frozen at 80  C for 2 h followed by thawing to room temperature. After three
freeze-thaw cycles, the samples were sonicated briey. A volume of 100 ml of the
DNA sample was added to 100 ml Picogreen stain (Sarstedt). The samples were incubated for 5 min in the dark at room temperature. Fluorescence was measured
using an FLx800 microplate uorescence reader (lem 520 nm and lex 480 nm) and
the DNA content was calculated from a standard curve (calf thymus DNA (Sigma
Aldrich)).
The cell colonization of the scaffold was analysed by scanning electron microscopy (SEM). The samples were xed with 2.5% glutaraldehyde in phosphate
buffer (0.1 M (NaH2PO4 0.18 M, Na2HPO4 0.17 M)) (SigmaAldrich, Ireland) for 2 h at
room temperature. After xation, the samples were dehydrated through graded
ethanol solutions (SigmaAldrich, Ireland). The samples were then immersed in
hexamethyldisilazane for 30 min, dried at room temperature and gold coated
(Emitech K550 Sputter Coater, Emitech Ltd., Ashford, Kent, UK). SEM was then
performed on the samples.

3. Results
3.1. Scaffold fabrication
Fig. 2(A) shows a porous titanium scaffold that has been created using the present RP fabrication process. The scaffold

approximates the inverse morphology of the wax template and

results in a fully interconnected porous network. For the 3Dprinter used in this fabrication process the minimum wax feature
size was approximately 200 mm and thereafter the wax models
were approximately 35 mm greater than the CAD design models
for the three principal directions. The wax template bestowed the
titanium scaffold with non-uniform architectural and mechanical
properties in the axial and transverse directions and ensured
greater strength of the samples in the axial direction. The repeating unit-cell geometry of the wax template produced a uniform distribution of pore size throughout the scaffold. The
template was modied to produce specic pore sizes, as shown in
Fig. 2(B). The predominant factor governing the pore size in the
titanium scaffolds was the diameter of the struts (1 and 2) in the
sacricial wax template unit-cell. The main factor inuencing
porosity in the titanium scaffolds was the height of the unit-cell
(L1) and the unit-cell spacing (L2L4). Both macro and micropores
exist in the scaffold. The sizes of titanium powder particles used in
the fabrication process ranged from 40 to 63 mm. Typical titanium
powder topographies for the different stages of sintering in this

G.E. Ryan et al. / Biomaterials 29 (2008) 36253635

3629

Fig. 3. Scanning electron micrographs of (A) as received CP2 titanium powder, (B) compacted titanium powder, (C and D) compacted and sintered titanium powder, (E) struts of the
titanium scaffold showing the sintered powders, and (F) the EDX spectra of the scaffold showing the titanium peaks.

PM process are shown in Fig. 3. During sintering, particle bonding

is achieved by neck growth through a solid-state diffusion process.
Sintering produced micropores in the size range approx. 115 mm
in the titanium struts. Except for the outer strut surfaces, the
sintered particles created an almost completely closed-cell porous
structure. Signicant scaffold shrinkage was evident during sintering. At 1300  C the horizontal shrinkage was 11.9  1.1% and the
vertical shrinkage 9.2  0.6%, corresponding to a volumetric decrease of 29.7  2.2%. As shown in Fig. 3(F), the chemical analysis
of the scaffold did not show any element other than titanium.
Although an oxide layer is created on all titanium surfaces, the
EDX did not detect any oxygen present. The chemical analysis was
identical to that of the titanium powders. This suggests that no
additional impurities or oxides were formed during the sintering
process.

3.2. Process optimisation

The inuence of the PM parameters on scaffold strength is
presented in Fig. 4. The highest values of compaction pressure
(250 MPa), sintering temperature (1300  C), and titanium slurry
concentration (5 g/3 ml) produced the scaffold with the highest
strength. These parameters produced scaffolds with yield strengths
of 72.2  7.6 MPa in the transverse direction. Decreasing the compaction pressure, while holding the other parameters constant,
resulted in a reduction in scaffold strength. Scaffolds produced
using a compaction pressure of 50 MPa had yield strengths of
48.3  7.8 MPa. However, there was no statistical difference found
between the compaction pressure of scaffolds compacted at 150
and 250 MPa (P > 0.05). Sintering temperature was also found to
inuence scaffold strength. Reducing the sintering temperature to

3630

G.E. Ryan et al. / Biomaterials 29 (2008) 36253635

Fig. 4. The effect of different PM processes on the mechanical strength of the titanium
scaffolds. Values are reported as mean  Std. Dev; *P < 0.05, n 3.

1100  C, while holding other parameters constant, produced scaffolds with yield strengths of 56.33  5.85 MPa. However, no statistical difference was found between samples sintered at 1200 and
1300  C (P > 0.05). Slurry concentration was found to have the
greatest inuence on scaffold strength. Slurry consisting of 5 g of
titanium powder in 7 ml of ethylene glycol produced scaffolds with
mean yield strengths of 42.0  8.46 MPa. A limit to the concentration of the slurry was reached at 5 g titanium powder in 3 ml
ethylene glycol, as above this concentration the slurry became too
viscous to completely penetrate the sacricial wax template.
The design of the sacricial wax template was found to greatly
inuence the scaffolds morphological and mechanical properties.
Given that this wax model was lost in the fabrication process,
a decrease in the wax template porosity resulted in an increase in
titanium scaffold porosity. The mechanical properties of three titanium scaffolds created using the wax templates described above
are presented in Fig. 5. The morphological characteristics of these
three scaffolds are presented in greater detail in Table 2. Youngs
modulus and scaffold strength decreased with increasing porosity.
In the axial direction the relationship between Youngs modulus
and porosity was almost linear. Structural disorder increased with
increasing porosity. The scaffolds were found to be anisotropic in
nature. The mean Youngs modulus was approximately 69.3% lower
in the transverse direction compared to the axial direction for all
three scaffolds. Also, mean ultimate compression strength was on
average 56.7% lower in the transverse direction compared to the
axial direction.
3.3. Characterisation of scaffold morphology
Three-dimensional computer simulations of titanium scaffolds
were successfully constructed from serial mCT data as shown in
Fig. 6. These were used for comparing the porous morphology of
the nal samples to that of the initial CAD drawing that they have
been derived from, in order to identify any discrepancies. Visual
inspection of the models revealed that the level of anisotropy increased as the porosity of the scaffolds increased. These values are
summarised in Table 2, along with the overall scaffold porosity and
total interconnecting porosity as calculated using Mimics. The
closed-cell microporosity was found to be 9.5  1.1%, through
evaluation of the sintered titanium billets. For the titanium scaffolds, the mean difference in total porosity as measured using apparent density calculations and interconnecting porosity as
measured using Mimics was approximately 5.7%.
The variation in porosity and pore size over the height of the
three titanium scaffolds is presented in Fig. 6(A) and (B). Both

Fig. 5. The mechanical properties of three titanium scaffolds created using different
sacricial wax templates showing (A) Youngs modulus and (B) Yield strength (n 3).
Porosity values for the three templates are presented in Table 2.

porosity and pore size showed a repeating pattern, governed by the

unit-cell geometry. In general, an increase in pore size over the
height of the unit-cell corresponded to an increase in scaffold porosity. The distribution of pore sizes for the three scaffolds is presented in Fig. 6(C). For all scaffolds the dominant pore size ranged
between 400 and 550 mm. However, Template 3 possessed a greater
proportion of pores at and above 1000 mm and this contributed to
a scaffold with overall greater porosity.
There was a notable difference in morphology between the
intended scaffold design and the physical titanium models. This is
evident in Fig. 7(A), which shows the idealised unit-cell for the
sacricial Template 2 and a unit-cell isolated from its corresponding titanium scaffold. The graph shows the change in porosity over
the height of the scaffold. It is evident that the architecture was
signicantly altered due to the fabrication process. The porosities of
the titanium scaffolds were on average 41.2% greater than the porosities of their corresponding idealised models. Also, the physical
titanium scaffolds were on average 29.2% shorter than their idealised models. Fig. 7(B) shows the change in pore size over the height
Table 2
Intended and achieved porosities and structural characteristics of three porous
titanium scaffolds created using different design templates
Template 1

Template 2

Template 3

Idealised model porosity (%)

Total porosity (%)
Interconnecting porosity (%)
Level of structural
disorder (%)
Degree of anisotropy

36.9
51.4  1.2
46.5  1.1
1.8  0.4

40.8
59.1  1.7
53.4  2.0
3.2  1.7

47.8
66.8  3.6
60.4  2.6
5.2  2.1

0.51  0.01

0.34  0.01

0.35  0.02

Closed-cell microporosity (%)

9.5  1.1

G.E. Ryan et al. / Biomaterials 29 (2008) 36253635

3631

Fig. 6. Typical models of porous titanium scaffolds created using three different sacricial templates and reconstructed using 3D reconstruction software (Mimics; Materialise) are
presented. The graphs show (A) porosity and (B) pore size as a function of height for the porous titanium scaffolds, and (C) distribution of pore size for the three scaffolds.

of the scaffold. Again, the repeating unit-cell structure is evident.

However, pore sizes for titanium scaffolds ranged from 300 to
1000 mm, which is contrast to the idealised models, which possessed only two possible pore sizes.

microporous surface and some migrated inside the micropores. As

shown in Fig. 9, AlamarBlue and total DNA analysis showed osteoblast proliferation with increasing cellular activity from day 1 to
14. The highest cell growth period was seen between day 7 and 14.
After day 14, cell growth decreased, as the two assays conrm.

3.4. In vitro experiments

4. Discussion
Fig. 8 shows SEM images of SAOS-2 cells on the surface of porous
titanium samples after 1, 7, 14 and 21 days in culture. At day 14
a conuent cell layer was observed on the titanium surface. Polygonal and spindle-shaped cells attached and spread on the

We have developed a new process of fabricating porous titanium scaffolds that possess high levels of interconnecting porosity
combined with high levels of mechanical strength. In the

3632

G.E. Ryan et al. / Biomaterials 29 (2008) 36253635

Fig. 7. The difference between the idealised and actual scaffold morphology over the
height of the scaffold in terms of (A) porosity and (B) pore size for the titanium scaffold
created using Template 2.

fabrication process, a wax template acts as a sacricial polymeric

space holder, and the resulting titanium scaffold takes the inverse
morphology of this template. Wen et al. used a similar technique
based on a space holder material in the form of ammonium hydrogen carbonate particles that were randomly mixed with titanium powders and subsequently compacted [43]. The ammonium
hydrogen carbonate was thermally removed using a heating step
followed by sintering of the titanium powders at elevated temperatures. Using this technique scaffolds with 80% porosity and
compression strengths of 40.0 MPa were created. Greater control
over pore interconnectivity was achieved by Li et al., using a technique they pioneered, which is known as the polymeric sponge
replication method [22,44]. In this technique, commercially available polyurethane foams were dipped into slurry containing
Ti6Al4V powder particles causing the powder particles to be impregnated onto the struts of the sponge. The polyurethane material
was subsequently pyrolyzed followed by sintering of the titanium
alloy powders. With this method they were able to create scaffolds
with 89  1.6% porosity and compression strengths of
10.3  3.3 MPa. Using an RP technique known as 3D ber deposition, Li et al. created porous Ti6Al4V scaffolds with fully
interconnected porous networks [29,45]. By depositing Ti6Al4V
slurry in a process similar to 3D-printing, followed by sintering of
the Ti6Al4V powder particles, scaffolds with approx. 60% porosity
and compressive strengths of around 520 MPa were created. LopezHeredia et al. used a reverse engineering technique to create
macroporous titanium implants with controlled shape and porosity
[27,28]. The rst step in this technique was to create a porous wax
scaffold template using an RP machine. The wax template was
embedded in a refractory material and the wax was subsequently
removed by pyrolysis. Molten titanium was then poured into the

mould and once cooled, the refractory material was removed by

grit-blasting. This process resulted in titanium scaffolds with
approx. 60% porosity and compressive strength of around 80 MPa.
The fabrication process developed in the present research has
similarities with the aforementioned techniques. An RP technique
was used to create the sacricial wax template allowing the porous
characteristics to be tailored to a high degree and ensure complete
interconnectivity of the porous architecture. A compaction step was
employed to bring the powder particles into close proximity and
ensure optimum strength of the nal sintered powders. The pore
cell walls microporosity of 9.5  1.1% was formed using a compaction pressure of 250 MPa and sintering temperature of 1300  C. The
microporosity was analysed under these conditions as these values
of compaction pressure and sintering temperature yielded the
highest strength scaffolds during process optimisation. An increase
in scaffold porosity corresponded with a decrease in scaffold
compressive strength, which is in agreement with Gibson and
Ashbys model of cellular solids [46]. Scaffolds with porosities of
66.8  3.6% possessed compression strengths of 104.4  22.5 MPa
in their axial direction and 23.5  9.6 MPa in their transverse direction. An important implication of the compaction step is that
titanium particles are less likely to break away from the body of the
porous material. Loose particles could potentially cause inammation of the surrounding implant and can cause aseptic
loosening of the implant [47,48].
Several factors were found to inuence the nal architectural
and mechanical properties of the titanium scaffolds. For process
optimisation, a single wax template design was chosen to evaluate
the effects of compaction pressure, sintering temperature and
slurry concentration. Although the results from testing these parameters would change when using different wax templates, it was
assumed that the general trend would remain the same within
a single parameter. The design of the sacricial wax model had the
greatest effect on scaffold properties. Due to the high computational demands of making these models, they were designed as
hexagonal in cross-section, to reduce the size of the stereolithography CAD le that is used for 3D-printing. Nevertheless, the
resulting wax models were circular in cross-section due to the accuracy limitations of the machine. Due to 3D-printer software
constraints, the maximum dispensing angle for wax printing is set
to 7. For this reason, the design of the titanium scaffold was governed by this limitation. However, with relevant software modications to the current 3D-printer or by using a 3D-printer that does
not have this limitation we intend to generate wax models with
a greater choice of possible design characteristics. By removing the
7 limitation the level of anisotropy for the scaffold can be adjusted
or completely removed, if desired. Removing the 7 limitation will
also permit scaffolds with much higher porosity to be developed.
The wax template structure can be tailored to optimise the slurry
inltration process and eliminate strut defects and structural
inhomogeneity.
Another factor that ultimately limits the porosity of the scaffold
is the size of the titanium powder particles (approximately 45 mm
were used in this work). Through initial testing of different wax
structures, it was found that the smallest aperture size that will
allow titanium powder slurry inltration into the wax template
was approx. 60 mm. The wax models are required to be of sufcient
porosity to allow an even distribution of titanium powder to ll
from the top to the bottom of the wax template. For this reason, the
composition and concentration of the slurry are critical. It was also
found that increasing the concentration of ethylene glycol increased the viscosity of the slurry but this also meant that less titanium powder penetrated the scaffold. This produced an increase
in scaffold porosity compared to the idealised model, due to the
increased microporosity that resulted in the struts of the porous
structure. There was also a weakening of titanium struts, as they did

G.E. Ryan et al. / Biomaterials 29 (2008) 36253635

3633

Fig. 8. Appearance of SAOS-2 cells on the surface of the porous titanium scaffold after (A) 1, (B) 7, (C) 14, and (D) 21 days in culture.

not conform to the original template geometry. Possible ways to

overcome this are to use ner titanium powders or to add additional substances that might increase the viscosity of the slurry as
demonstrated in previous work by Li et al. [44].
The reduction in scaffold height was a prominent feature in all
scaffolds that were created, and mainly resulted from the uniaxial
compaction step that condenses the titanium powder and wax
template. Future research may incorporate a hydrostatic compaction press so that alterations to the intended design can be minimized. The compaction process owes its success to the concurrent
heating step that simultaneously melts the wax inside in the die
cavity. During this process the hydraulic press is changed from load
control to displacement control, which prevents further deformation of the titanium structure while the wax is melted. The
melting step is necessary, as without this multiple cracks developed
in the scaffold struts prior to sintering due to the difference in
compressibility between the wax material and the compacted titanium powder. If not heated, the wax springs back to its original
shape once force from the press is removed, and this movement
breaks the adhesion bonds formed between the titanium powder
particles. It should be noted that during this process the wax is not
removed from the die but retains its position, embedded within the
titanium powders. After the die has cooled the wax/titanium construct can be removed and is quite robust. The use of xylene to
remove the wax material by dissolution prior to sintering was
chosen over pyrolysis to avoid contamination of the titanium with
carbon residues from the wax space holder that can be left behind
[49]. Several rounds of rinsing are necessary to completely remove
the wax, after which the titanium powders are held together by the
adhesion forces formed in the compaction process. With increasing
porosity, and once the wax llers have been removed, the scaffolds
become very delicate and care needs to be exercised when moving

the scaffolds to the furnace. Sintering temperature did not seem to

affect scaffolds mechanical strength greatly once the temperature
exceeded 1200  C. However, a furnace-vacuum of greater than
105 mbar was necessary for complete sintering to occur, without
any oxides interfering with the particles fusion.
Although the elongated pores in the titanium scaffolds are
a result of limitations due to the 3D-printer, they create a material
that more closely resembles natural bone than fully dense, isotropic
titanium. The anisotropy of bone is well documented [50,51]. A
number of authors have utilised this attribute in the design and
optimisation of biomaterial scaffolds with elongated pores [52,53].
Scaffolds with a level of anisotropy have the potential to address
issues of reduction in implant weight and stiffness while minimising strength reduction in the loading direction. In this research,
scaffolds with 66.8  3.6% porosity possessed a Youngs modulus of
20.5  2.0 GPa in the axial direction and 4.35  0.5 GPa in the
transverse direction. Compared to bone, these values are quite high.
The mechanical properties of trabecular bone are widely studied,
but still reported values for Youngs modulus range vary from 0.1 to
10.4 GPa depending on the source and method of analysis [54].
Future research will aim to increase the porosity of the scaffolds to
further reduce the scaffolds modulus. Finding appropriate mechanical properties for a prospective implant will be a signicant
challenge where levels of strength are sacriced for increasing
porosity and reduced stiffness. In vivo tests will be necessary to
identify appropriate designs that will best match loading in such
a way that will prevent stress shielding as well as early failure. A
possible way to further reduce the elastic modulus would be to use
a titanium alloy powder. Beta titanium alloys such as Ti35Nb7Zr5Ta
and Ti13Nb13Zr have lower Youngs moduli, compared to pure titanium and have received considerable attention as orthopaedic
implants in recent years [5557].

3634

G.E. Ryan et al. / Biomaterials 29 (2008) 36253635

Promising cellular responses to the porous titanium scaffolds

were seen in this study. Both AlamarBlue and DNA tests showed an
increase in cell number in the rst 2 weeks of culture. Thereafter,
differentiation media was added after 14 days of culture, which
explains the reduction in cellular growth until day 21, as this would
have coincided with cell differentiation. We conclude that the in
vitro experiments show that the titanium scaffold allows the
growth of vital pre-osteoblast cells on its surface, which is comparable to the cellular behaviour on other porous titanium scaffolds
[30,45].
5. Conclusion
A porous titanium scaffold with controlled porous structure and
high strength has been created using a combination of RP and PM
techniques. A wax model, created using a commercial 3D-printer,
provided the template for the porous titanium scaffold. Characteristics of the PM process, such as titanium slurry concentration,
compaction pressure, and sintering temperature, were found to
inuence the structure and strength of the scaffolds. At present,
a signicant level of variability exists between scaffolds and their
intended designs due to factors present in the PM process. Future
research will involve gaining a greater control over these factors so
that this mismatch can be minimized. The in vitro experiments
showed that the osteoblasts maintain their metabolic activity on
the surface of the porous titanium scaffolds. In conclusion, the results of this study demonstrate the potential of such scaffolds for
use in orthopaedic tissue engineering applications.
Fig. 9. (A) Metabolic activity and (B) change in density of SAOS-2 cells on porous
titanium scaffolds for up to 21 days in culture. *P < 0.05, tissue culture plastic vs. porous titanium (n 3).

A range of structural properties were quantied using mCT.

Scaffold reconstruction using serial mCT images enabled structural
parameters that were previously unavailable experimentally to be
measured, for example the change in porosity and pore size
throughout the height of the scaffold. Analysis of these parameters
enabled us to determine the relationship between the intended
scaffold design and the physical titanium scaffold. The height of the
titanium scaffolds is signicantly shorter compared to their projected design. This difference is largely due to the uniaxial compaction process although sintering of the green titanium scaffolds
following xylene dissolution also causes a reduction in scaffold
height through coalescence of the powder particles. The level of
structural disorder in the scaffolds increased with increasing levels
of porosity. In the fabrication process, increasing the porosity requires increasing the ratio of sacricial wax template to titanium
powder. In doing so, the titanium powder is less likely to completely ll the porous space around the sacricial template and
more anomalies in the scaffold architecture are likely.
In the 3D reconstruction process, the accuracy of the scaffold
geometries is dependent on the spatial resolution of the scans and
image segmentation [58]. In this research, a voxel size of 15 mm was
sufcient to describe the main structural features of the scaffold,
because the average powder particle size was 45 mm. Using
Mimics, image thresholding was executed to segment titanium
and pore phases prior to 3D modelling. This is a critical step as it
determines the success of subsequent analysis and visualization.
The microporosity generated between the sintered titanium powders (9.5  1.1%) was lost during this process due to resolution
limitations. The difference in total porosity as measured using
Archimedes principle and interconnecting porosity as measured
using Mimics should equal this nding but it is lower by 3.8%. This
slight divergence may be attributed inaccuracies in the 3D reconstruction process and in particular during segmentation.

Acknowledgements
This work was supported by Enterprise Ireland (PC/2005/012).
References
[1] Karageorgiou V, Kaplan D. Porosity of 3D biomaterial scaffolds and osteogenesis. Biomaterials 2005;26(27):547491.
[2] Giannoudis PV, Dinopoulos H, Tsiridis E. Bone substitutes: an update. Injury
2005;36(Suppl. 3):S207.
[3] Al Ruhaimi KA. Bone graft substitutes: a comparative qualitative histologic
review of current osteoconductive grafting materials. Int J Oral Maxillofac
Implants 2001;16(1):10514.
[4] Laurencin C, Khan Y, El-Amin SF. Bone graft substitutes. Expert Rev Med Dev
2006;3(1):4957.
[5] Fantigrossi A, Galbusera F, Raimondi MT, Sassi M, Fornari M. Biomechanical
analysis of cages for posterior lumbar interbody fusion. Med Eng Phys 2007;
29(1):1019.
[6] Agazzi S, Reverdin A, May D. Posterior lumbar interbody fusion with cages: an
independent review of 71 cases. J Neurosurg 1999;91(Suppl. 2):18692.
[7] Lin CY, Wirtz T, Lamarca F, Hollister SJ. Structural and mechanical evaluations
of a topology optimized titanium interbody fusion cage fabricated by selective
laser melting process. J Biomed Mater Res A 2007.
[8] Takemoto M, Fujibayashi S, Neo M, So K, Akiyama N, Matsushita T, et al. A
porous bioactive titanium implant for spinal interbody fusion: an experimental study using a canine model. J Neurosurg Spine 2007;7(4):43543.
[9] Assad M, Jarzem P, Leroux MA, Coillard C, Chernyshov AV, Charette S, et al.
Porous titaniumnickel for intervertebral fusion in a sheep model: part 1.
Histomorphometric and radiological analysis. J Biomed Mater Res B Appl Biomater 2003;64B(2):10720.
[10] Likibi F, Assad M, Coillard C, Chabot G, Rivard CH. Bone integration and apposition of porous and non porous metallic orthopaedic biomaterials. Ann Chir
2005;130(4):23541.
[11] Levi AD, Choi WG, Keller PJ, Heiserman JE, Sonntag VK, Dickman CA. The radiographic and imaging characteristics of porous tantalum implants within
the human cervical spine. Spine 1998;23(11):124550. Discussion 1251.
[12] Wigeld C, Robertson J, Gill S, Nelson R. Clinical experience with porous
tantalum cervical interbody implants in a prospective randomized controlled
trial. Br J Neurosurg 2003;17(5):41825.
[13] Levine B, Sporer S, Della Valle CJ, Jacobs JJ, Paprosky W. Porous tantalum in
reconstructive surgery of the knee: a review. J Knee Surg 2007;20(3):18594.
[14] Head WC, Bauk DJ, Emerson Jr RH. Titanium as the material of choice for cementless femoral components in total hip arthroplasty. Clin Orthop 1995;
(311):8590.
[15] Hirschhorn JS, McBeath AA, Dustoor MR. Porous titanium surgical implant
materials. J Biomed Mater Res Symp 1971:4967.

G.E. Ryan et al. / Biomaterials 29 (2008) 36253635

[16] Oh IH, Nomura N, Masahashi N, Hanada S. Mechanical properties of porous
titanium compacts prepared by powder sintering. Script Mat 2003;49:1197
202.
[17] Weiss MB. Titanium ber-mesh metal implant. J Oral Implantol 1986;12(3):
498507.
[18] Zhang X, Ayers RA, Thorne K, Moore JJ, Schowengerdt F. Combustion synthesis
of porous materials for bone replacement. Biomed Sci Instrum 2001;37:4638.
[19] Heinl P, Muller L, Korner C, Singer RF, Muller FA. Cellular Ti-6Al-4V structures
with interconnected macro porosity for bone implants fabricated by selective
electron beam melting. Acta Biomater 2008.
[20] Ponader S, Vairaktaris E, Heinl P, Wilmowsky CV, Rottmair A, Korner C, et al.
Effects of topographical surface modications of electron beam melted Ti-6Al4V titanium on human fetal osteoblasts. J Biomed Mater Res A 2008;84(4):
11119.
[21] Davis NG, Teisen J, Schuh C, Dunand DC. Solid-state foaming of titanium by
superplastic expansion of argon-lled pores. J Mater Res 2001;16(5):150819.
[22] Li JP, Li SH, van Blitterswijk CA, de Groot K. A novel porous Ti6Al4V: characterization and cell attachment. J Biomed Mater Res 2005;73A(2):22333.
[23] Yeong WY, Chua CK, Leong KF, Chandrasekaran M. Rapid prototyping in
tissue engineering: challenges and potential. Trends Biotechnol 2004;22(12):
64352.
[24] Hutmacher DW, Sittinger M, Risbud MV. Scaffold-based tissue engineering:
rationale for computer-aided design and solid free-form fabrication systems.
Trends Biotechnol 2004;22(7):35462.
[25] Sun W, Darling A, Starly B, Nam J. Computer-aided tissue engineering: overview, scope and challenges. Biotechnol Appl Biochem 2004;39(Pt 1):2947.
[26] Curodeau A, Sachs E, Caldarise S. Design and fabrication of cast orthopedic
implants with freeform surface textures from 3-D printed ceramic shell.
J Biomed Mater Res 2000;53(5):52535.
[27] Lopez-Heredia MA, Goyenvalle E, Aguado E, Pilet P, Leroux C, Dorget M, et al.
Bone growth in rapid prototyped porous titanium implants. J Biomed Mater
Res A 2008;85(3):66473.
[28] Lopez-Heredia MA, Sohier J, Gaillard C, Quillard S, Dorget M, Layrolle P. Rapid
prototyped porous titanium coated with calcium phosphate as a scaffold for
bone tissue engineering. Biomaterials 2008;29(17):260815.
[29] Li JP, de Wijn JR, van Blitterswijk CA, de Groot K. Porous Ti6Al4V scaffolds
directly fabricated by 3D bre deposition technique: effect of nozzle diameter.
J Mater Sci Mater Med 2005;16(12):115963.
[30] Hollander DA, von Walter M, Wirtz T, Sellei R, Schmidt-Rohlng B, Paar O,
et al. Structural, mechanical and in vitro characterization of individually
structured Ti-6Al-4V produced by direct laser forming. Biomaterials 2006;
27(7):95563.
[31] Sachlos E, Czernuszka JT. Making tissue engineering scaffolds work. Review:
the application of solid freeform fabrication technology to the production of
tissue engineering scaffolds. Eur Cell Mater 2003;5:2939. Discussion 3940.
[32] Cameron HU, Pilliar RM, Macnab I. The rate of bone ingrowth into porous
metal. J Biomed Mater Res 1976;10(2):295302.
[33] Gomes ME, Sikavitsas VI, Behravesh E, Reis RL, Mikos AG. Effect of ow perfusion on the osteogenic differentiation of bone marrow stromal cells cultured
on starch-based three-dimensional scaffolds. J Biomed Mater Res A 2003;
67(1):8795.
[34] Cartmell SH, Porter BD, Garcia AJ, Guldberg RE. Effects of medium perfusion
rate on cell-seeded three-dimensional bone constructs in vitro. Tissue Eng
2003;9(6):1197203.
[35] Hollister SJ, Maddox RD, Taboas JM. Optimal design and fabrication of scaffolds to mimic tissue properties and satisfy biological constraints. Biomaterials
2002;23(20):4095103.
[36] Fang Z, Starly B, Sun W. Computer-aided characterization for effective mechanical properties of porous tissue scaffolds. Comput-Aided Des 2005;37(1):
6572.

3635

[37] Gupta G, Zbib A, El-Ghannam A, Khraisheh M, Zbib H. Characterization of

a novel bioactive composite using advanced X-ray computed tomography.
Compos Struct 2005;71(34):4238.
[38] Knackstedt MA, Arns CH, Senden TJ, Gross K. Structure and properties of
clinical coralline implants measured via 3D imaging and analysis. Biomaterials
2006;27(13):277686.
[39] Jones JR, Poologasundarampillai G, Atwood RC, Bernard D, Lee PD. Non-destructive quantitative 3D analysis for the optimisation of tissue scaffolds.
Biomaterials 2007;28(7):140413.
[40] Meretoja VV, De Ruijter AE, Peltola TO, Jansen JA, Narhi TO. Osteoblast differentiation with titania and titaniasilica-coated titanium ber meshes. Tissue Eng 2005;11(910):148997.
[41] Beloti MM, Rosa AL. Osteoblast differentiation of human bone marrow cells
under continuous and discontinuous treatment with dexamethasone. Braz
Dent J 2005;16(2):15661.
[42] Schutze N, Noth U, Schneidereit J, Hendrich C, Jakob F. Differential expression
of CCN-family members in primary human bone marrow-derived mesenchymal stem cells during osteogenic, chondrogenic and adipogenic differentiation. Cell Commun Signal 2005;3(1):5.
[43] Wen CE, Yamada Y, Shimojima K, Chino Y, Asahina T, Mabuchi M. Processing
and mechanical properties of autogenous titanium implant materials. J Mater
Sci Mater Med 2002;13(4):397401.
[44] Li JP, Li SH, de Groot K, Layrolle P. Preparation and characterization of porous
titanium. Key Eng Mater 2002;218:514.
[45] Li JP, de Wijn JR, Van Blitterswijk CA, de Groot K. Porous Ti6Al4V scaffold
directly fabricating by rapid prototyping: preparation and in vitro experiment.
Biomaterials 2006;27(8):122335.
[46] Gibson L, Ashby M. Cellular solids: structure and properties. 2nd ed. Cambridge: Cambridge University Press; 1999.
[47] Cunningham BW, Orbegoso CM, Dmitriev AE, Hallab NJ, Sefter JC, Asdourian P,
et al. The effect of spinal instrumentation particulate wear debris. An in vivo
rabbit model and applied clinical study of retrieved instrumentation cases.
Spine J 2003;3(1):1932.
[48] Wang JC, Yu WD, Sandhu HS, Betts F, Bhuta S, Delamarter RB. Metal debris
from titanium spinal implants. Spine 1999;24(9):899903.
[49] Robinson SK, Paul MR. Debinding and sintering solutions for metals and ceramics. Met Powder Rep 2001;56(6):246.
[50] Jacobs CR, Simo JC, Beaupre GS, Carter DR. Adaptive bone remodeling incorporating simultaneous density and anisotropy considerations. J Biomech
1997;30(6):60313.
[51] Smit TH, Odgaard A, Schneider E. Structure and function of vertebral trabecular bone. Spine 1997;22(24):282333.
[52] Lin CY, Kikuchi N, Hollister SJ. A novel method for biomaterial scaffold internal
architecture design to match bone elastic properties with desired porosity.
J Biomech 2004;37(5):62336.
[53] Spoerke ED, Murray NG, Li H, Brinson LC, Dunand DC, Stupp SI. Titanium with
aligned, elongated pores for orthopedic tissue engineering applications. J Biomed Mater Res A 2007.
[54] Kim HS, Al-Hassani ST. A morphological model of vertebral trabecular bone.
J Biomech 2002;35(8):110114.
[55] Banerjee R, Nag S, Fraser HL. A novel combinatorial approach to the
development of beta titanium alloys for orthopaedic implants. Mater Sci Eng
C Biomim Supramol Syst 2005;25(3):2829.
[56] Eisenbarth E, Velten D, Muller M, Thull R, Breme J. Biocompatibility of betastabilizing elements of titanium alloys. Biomaterials 2004;25(26):570513.
[57] Taddei EB, Henriques VAR, Silva CRM, Cairo CAA. Production of new titanium
alloy for orthopedic implants. Mater Sci Eng C Biomim Supramol Syst 2004;
24(5):6837.
[58] Ho ST, Hutmacher DW. A comparison of micro CT with other techniques used
in the characterization of scaffolds. Biomaterials 2006;27(8):136276.