Accepted Manuscript

Title: Control of Structural and Mechanical Properties in

Bioceramic Bone Substitutes via Additive Manufacturing
Layer Stacking Orientation
Author: Mihaela Vlasea Robert Pilliar Ehsan Toyserkani
PII:
DOI:
Reference:

S2214-8604(15)00014-7
http://dx.doi.org/doi:10.1016/j.addma.2015.03.001
ADDMA 31

To appear in:
Received date:
Revised date:
Accepted date:

21-9-2014
2-2-2015
5-3-2015

Please cite this article as: Vlasea M, Pilliar R, Toyserkani E, Control

of Structural and Mechanical Properties in Bioceramic Bone Substitutes
via Additive Manufacturing Layer Stacking Orientation, Addit Manuf (2015),
http://dx.doi.org/10.1016/j.addma.2015.03.001
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Control of Structural and Mechanical Properties in Bioceramic Bone

Substitutes via Additive Manufacturing Layer Stacking Orientation
Robert Pilliar2,3

Ehsan Toyserkani1

ip
t

Mihaela Vlasea1

cr

1-University of Waterloo, Department of Mechanical and Mechatronics Engineering, Waterloo, Ontario, N2L 3G1

us

2-Institute of Biomaterials and Biomedical Engineering, University of Toronto, Toronto, Ontario, Canada M5S 3G9

Journal of Additive Manufacturing

Submitted to:

an

3-Faculty of Dentistry, University of Toronto, Toronto, Ontario, M5G 1G6, Canada

September 21, 2014

Number of Pages:

21

Ac
ce
p

te

Submission Date:

Number of Figures:

Number of Tables:

Contact Author:

Mihaela Vlasea, PhD

Address:

Department of Mechanical and Mechatronics Engineering,

University of Waterloo, Waterloo, Ontario, N2L 3G1 Canada

Email:

mlvlasea@uwaterloo.ca

Phone:

1-519-722-1368

1
Page 1 of 22

Control of Structural and Mechanical Properties in Bioceramic Bone

ip
t

Substitutes via Additive Manufacturing Layer Stacking Orientation

Abstract: Additive manufacturing (AM) is a promising approach for fabricating structures to

serve as bone substitutes, or as biomaterial components in biphasic implants for repair of

cr

osteochondral defects. In this study, the three dimensional printing (3DP) AM process was

investigated to determine the effect of powder layer orientation on mechanical and structural

us

properties of fabricated parts. Five types of standard cylindrical parts were manufactured via AM
with 0, 30, 45, 60 and 90stacking layer orientations relative to the vertical z-axis of the print

an

bed, using amorphous calcium polyphosphate (CPP) powder of irregular particle shape, average
aspect ratio 1.70 and particle size between 75-150 m. It was concluded that layer orientation

had an effect on porosity and compressive strength, based on induced powder particle orientation
in the green part during powder layering. The resulting bulk porosity values ranged between 30.0
2.4% to 38.2 2.7%, while the compressive strength ranged between 13.50 1.95 MPa to

45.13 6.82 MPa. The orientation with the highest compressive strength was 90, while

te

orientations with the weakest compressive strength were 0 and 45. Based on these results, it
was established that AM-made parts are structurally and mechanically anisotropic. The stacking

Ac
ce
p

layer orientation which results in the highest strength performance along a preferred loading
orientation can be implemented to further optimize mechanical strength of constructs along the
maximum loading direction.

Keywords: Additive manufacturing, 3D printing, oriented layering, calcium polyphosphate,

bioceramic bone substitutes

2
Page 2 of 22

Introduction

Porous bone substitutes serve as an artificial matrix providing the mechanical and structural
template for new bone formation. Such porous structures must be biocompatible and ideally

ip
t

should promote osteogenesis by being osteoconductive and osteoinductive while degrading invivo at an appropriate rate to allow their replacement by newly-formed bone [1,2]. The porous
structure must also be designed to have an anatomically accurate three dimensional (3D) shape

cr

in order to maintain a natural contact load distribution post implantation [3] and initial internal
structure in terms of micro- and macro-interconnected porosity to promote cell proliferation,

us

metabolic exchange and vascularisation [4]. Furthermore, the mechanical strength and porous
architecture of the bone substitute should ideally be designed to match the load bearing

an

requirements at the substitution site and to promote the appropriate bone growth cues as dictated
by mechanostat theory [5]. Ideally, this would suggest anisotropic structural and mechanical

characteristics throughout the construct, depending on the implantation site.

Powder-based additive manufacturing (AM) utilizing three dimensional printing (3DP) is a very

promising fabrication method for making scaffolds or porous constructs in the field of tissue
engineering and regenerative medicine, and specifically for bone substitute fabrication [2,610].

te

Using this approach, the anatomical shape and internal porous configuration of the implant is

Ac
ce
p

first designed in a computer-aided design (CAD) environment. Subsequently, the CAD model is
converted into image slices and the scaffold is manufactured in a layer-by-layer fashion by
repeated stacking powder layers and subsequently injecting a binder solution at locations dictated
by the cross-sectional image of the part layer to be formed as shown in Figure 1. The resulting
product is referred to as a green part. For ceramic structures, such as the calcium polyphosphate
(CPP) used in the present study, further post-green part processing, usually involving thermal
annealing, is necessary to achieve required strength properties and structural characteristics (i.e.
% porosity, pore size and interconnectedness) of the final part [4,8].

3
Page 3 of 22

ip
t

cr

Figure 1 Process description for conventional AM via powder-based 3DP

Prior investigations by a number of groups have focused on fine-tuning 3DP process parameters

us

to examine their effect on mechanical and structural properties of the final product. The
appropriate binder-powder material systems compatible with both the additive manufacturing

an

process and the biological requirements of the bone substitute were investigated [2,11]. Other
studies have examined the effect of the liquid binder used during powder lay-up and green part

formation with regard to its chemical composition [12,13], concentration [10,12,13] and
saturation levels [1416] to determine appropriate binder-powder interactions that would
produce samples with a desired compressive and flexural strength and porosity [10,12,13,15].

Other studies have focused on defining powder composition and blends to yield better powder

te

flow characteristics [17,18] and improved mechanical performance of the final structure [12].
The effect of powder particle size and its effect on physical, structural and mechanical properties

Ac
ce
p

of AM-formed constructs has also been studied [19]. Layer thickness is another parameter that
can be controlled during the 3DP process, with a range of layer thicknesses having been used for
preparation of samples in order to study the effect on mechanical properties [15]. These studies
concluded that, in general, the flexural and compressive strength performance is inversely
proportional to layer thickness [15,20]. The effect of including open or closed macro-channels
within the porous structures during the 3DP processing and its effect on mechanical strength and
biological response of porous constructs has also been studied [12,19,2124]. In this context, a
new type of 3DP platform has been investigated, capable of creating interconnected macrochannels with a feature size below 500 m within the part while avoiding the risk of having
particles trapped within the macro-channels [25,26].
In 3DP, due to the nature of the layer-by-layer manufacturing process, the effect of layer
stacking orientation within the part may influence the physical, structural and mechanical
4
Page 4 of 22

properties of constructs so formed. Shanjani et al. [27] and Zhang et al. [20] studied the effect of
layer orientation along the direction of the printing axes and concluded that mechanical strength
characteristics were related to orientations used in forming parts. This effect has not been
explored in detail, as the two previously reported studies focused only on orientations along the

ip
t

printing axes z, y, z, without considering intermediate orientations. In the present study, to better
understand the correlation between layer orientation and mechanical properties, standard

cr

cylindrical parts with 0, 30, 45, 60, and 90 layer stacking orientations with respect to the vertical
axis (z-axis) in the build chamber were fabricated and characterized in terms of porosity, bulk

us

density, and compressive strength. It is proposed that the stacking layer orientation within a part
which results in the highest strength can be aligned during part fabrication in the direction of

an

anticipated maximum loading, if this is known during the design stage. Or, contrarily, the
orientation resulting in the lowest strength can be avoided from coinciding with expected high

Materials

te

2.1

Materials and Methods

load-carrying directions.

In this study, the powder material used in the AM process was amorphous calcium

Ac
ce
p

polyphosphate (CPP) powder formed as reported in earlier studies [28]. The powder is
characterized by an irregular particle shape, an aspect ratio 1.70 and particle size range
between 75-150 m. This powder was mixed with polyvinyl alcohol (PVA) powder (Alfa Aesar,
Ward Hill, MA) of particle size < 63 m at a composition ratio of 90 wt% CPP and 10 wt%
PVA. To ensure a homogeneous blending, the CPP and PVA powders were mixed for 4 hours
using a rotating jar mill (US Stoneware, OH). The PVA powder served as an additional binding
agent in combination with the liquid binder (ZbTM58) (3D Systems, Burlington, MA) which also
served as an aqueous solvent for the PVA particles to give acceptable green strength to samples
formed by 3DP.
2.2

Sample Fabrication

Previous studies reported on preparation of porous CPP samples and the effect of CPP powder
size and method of porous construct preparation have been reported for conventional gravity
5
Page 5 of 22

sintering and AM using 3DP [11,27]. In vivo studies of samples so-made have been reported
[29,30].The focus of the present study is the determination of how the layer orientation affects
mechanical strength characteristics of samples. To achieve this, cylindrical test samples of 4 mm
diameter and 6 mm height were manufactured with layer orientations at 0, 30, 45, 60, and 90

ip
t

with respect to the vertical axis in the build compartment as seen in Figure 2. A 3DP machine
(ZPrint 310 Plus, 3D Systems, Burlington, MA) was used to manufacture the parts in a layer-by-

cr

layer fashion. The layer thickness was selected to be 175 m with the AM layer-by-layer powder
spread process being undertaken at a 38C. The cylindrical test parts were designed using

us

computer-aided design (CAD) software (SolidWorks Corp., Concord, MA) and imported into the
3D printing software (ZPrintTM) in stereolithography (STL) file format. The cylindrical test parts

an

were then oriented within the build bed, as seen in Figure 2, with n=10 parts prepared for each
orientation. The green parts were then air-annealed (Lindberg/Blue M, ThermoScientific) with a
50% R.H. in-furnace environment using a pre-established heat treatment protocol [31]. The

annealing cycle used has been reported elsewhere [11,31], with a heat-uprate of 10C/min from
room temperature to 400C, 2 h dwell at 400C to burn off organic binder constituents, continued

heat-up at 10C/min to 500C and then5C/min to 630C, hold for 1 h (this is the so-called Step-

te

1 sinter [31]),and then increasing the temperature to 950C at 10C/min and holding at 950C for
1 h (to achieve complete CPP crystallization and final microstructural development [31]).

Ac
ce
p

Samples were then allowed to furnace cool to room temperature.

Figure 2Parts printed with 0, 30, 45, 60, 90 layer orientation, arranged in the build compartment.

6
Page 6 of 22

2.3

Porosity Characterization

The bulk porosity of the sintered cylindrical samples was determined as described previously
[11] by ethanol displacement using Archimedes principle (ASTM C373 standard). An ethanol

ip
t

bath kit (Sartorius YDK01 Density Determination Kit, Sartorius AG, Goettingen, Germany) and
a precision micro-scale balance (APX-203, Denver Instrument, Bohemia, NY, US) were used to
of each specimen. Each specimen was then immersed in

cr

first determine the dry weight

ethanol and sonicated (VWR Ultrasonics Cleaner B2500A-DTH, VWR International, West

us

Chester, PA, US) for one hour at 30C and soaked for another hour. Subsequently, the weight of
the specimen suspended in ethanol

an

was measured. Each specimen was then removed from

ethanol, dabbed with a lint free cloth to remove excess ethanol and weighed immediately after in
. The bulk porosity

order to determine the wet weight,

and bulk density

were

determined based on the formulae below, (Equations 1 and 2), where the density of ethanol

te

at room temperature equals 0.785 g/cm3 and the theoretical density of non-porous CPP

Ac
ce
p

equals 2.850 g/cm3[28]. A population of n=10 samples was used for this data.

2.4

(1)

(2)

Structural Characterization

The microstructure of final sinter-annealed cylindrical CPP samples with different layer
orientations was examined using secondary electron emission scanning electron microscopy
(SEM, JSM-6460, Jeol, Akishima, Tokyo) operating at 20 kV accelerating voltage. In
preparation for SEM examination, the samples were sputter-coated with a 10 nm thick gold layer
7
Page 7 of 22

to make them electrically conductive (Desk II, Denton Vacuum, LCC, Moorestown, NJ, USA).
One representative sample from each category was considered for SEM analysis.
2.5

Powder Size Characterization

ip
t

The powder size and aspect ratio were determined by placing a thin layer of powder on a

conductive substrate and using secondary electron emission scanning electron microscopy (SEM,

cr

JSM-6460, Jeol, Akishima, Tokyo) at 20 kV accelerating voltage to view particle images. Five
images were captured at 100x magnification and three at 200x magnification. The particle

us

dimension, across the longest orientation (length) and along the dimension perpendicular to this
direction (width), were recorded using the SEM AnalySIS tool in order to estimate particle

2.6

an

aspect ratios.
Uniaxial Compression Characterization

Eight samples (n=8) for each different layer orientation were tested in uniaxial compression.
Testing was conducted on dry samples at room temperature using a 1-kN load cell at a loading

Statistical Analysis

te

2.7

rate of 0.2 mm/min (Instron 5548 Micro-Testing, MA).

Volume % porosity (denoted hereafter as bulk porosity), bulk density of the porous structure, and

Ac
ce
p

compression strength data are reported as means and standard deviation. A one-way ANOVA
single factor analysis of variance was used to evaluate the statistical significance of
measurements, followed by Tukey-Kramer post hoc pairwise comparisons to identify the socalled honest significant differences (HSD) between classes of samples using STATISTICA
V12 (StatSoft, Tulsa, OK) with p<0.05 identifying a significant difference. The ASTM C123907 standard was used to analyse and report the performance of ceramics under uniaxial
compression that failed in a brittle fashion. The Weibull distribution described in Equation 3
was used to predict the probability of failure of the ceramic part [32] as well as to provide a
statistical means of evaluating the quality of the measured data set:

(3)

8
Page 8 of 22

where

is the failure probability,

characteristic strength and

is the central value or scale parameter showing the Weibull

is the shape parameter, also known as the Weibull modulus. The

ip
t

Weibull modulus is a measure of reproducibility of samples [33]. In this work, a population of

eight samples (n = 8) was used for all sample categories for compressive strength data. Equation

cr

4 was used as a probability estimator for the Weibull linear regression, where is the

us

corresponding rank of the sample measurement.

3.1

Results
Structural Characterization Results

an

(4)

Figure 3 illustrates an SEM image of a sample prepared at a 60 layer stacking orientation. In

te

Figure 3a), the parallel layer orientations are highlighted by the dashed white lines. Figure 3b)
and Figure 3c) show examples of sinter necks resulting from the post-AM thermal sinter/anneal

Ac
ce
p

treatments used for sample preparation. It can be seen that the particles are well bonded together,
forming an open-pored structure. The SEM images obtained for the other layer stacking
orientations were very similar in nature.

9
Page 9 of 22

ip
t
cr
us
an
M
d

Powder Size Characterization

Ac
ce
p

3.2

te

Figure 3 Images of a CPP sample with 60 layer stacking orientation as viewed under SEM with a
magnification of a)x22, b) x100, and c) x300

The SEM image analysis revealed the distribution of powder size along the length and width of
particles as shown in Figure 4. The powder had an irregular shape, with a mean aspect ratio
greater than one. The bulk of particles had the longest axis above 150 m (average particle size
of 177 42 m), while the smallest axis normal to this long dimension was on average below
120 m (average particle size of 104 28 m). The average particle aspect ratio was determined
to be 1.70.

10
Page 10 of 22

ip
t
cr
us
an
M
d
te

3.3

Ac
ce
p

Figure 4 a) The particle size distribution for particle size sieved between 75-150 m, (n=105) b)
Scanning electron microscope (SEM) view of powder particles at magnification x250

Bulk Porosity (i.e. volume % porosity) and Density

The bulk density and porosity results of the cylindrical samples with oriented layers at 0, 30, 45,
60, 90 were calculated using the Archimedes method for density determination. The results are
summarized in Table 1. Figure 5 illustrates the bulk porosity measurements for the different
layer orientation samples indicating statistically significant differences. The maximum bulk
porosity occurred at an orientation of 0 and 45, with values of 38.22.7% and 37.6 3.1%
respectively. The minimum bulk porosity corresponded to samples prepared at an orientation of
90, where the bulk porosity value was equal to 30.0 2.4%.

11
Page 11 of 22

Table 1 Bulk porosity and bulk density characteristics for cylindrical CPP samples printed with
layer orientations of 0, 30, 45, 60, and 90 respectively, (n=10)

(p < 0.05)
38.2 2.7
32.2 2.8
37.6 2.1
34.0 3.2
30.0 2.4

(p < 0.05)
1.76 0.08
1.93 0.08
1.78 0.09
1.88 0.09
2.00 0.07

te

an

us

cr

0
30
45
60
90

ip
t

3DP
Layer Orient.

3.4

Ac
ce
p

Figure 5 Bulk porosity characteristics of cylindrical samples with layer stacking orientations 0, 30,
45, 60, and 90 respectively (n=10). *(p<0.05)

Mechanical Strength Uniaxial Compression

The results listed in Table 2 indicate that the 90 layering orientation in samples results in the
maximum compressive strength of 45.13 6.82 MPa, while the samples prepared at 45 and 0
layer orientations displayed minimum compressive strengths (13.43 4.60 MPa and13.50 1.95
MPa respectively).

Table 2 - Uniaxial compression results and Weibull moduli of cylindrical samples printed with layer
stacking orientations of 0, 30, 45, 60, and 90 respectively, (n=8)

AM
Layer Orient.
0
30
45
60

(p < 0.05)
13.50 1.95
20.62 6.23
13.43 4.60
28.19 2.46

(p < 0.05)
6.95 (0.91)
3.14 (0.93)
3.18 (0.80)
11.98 (0.95)

12
Page 12 of 22

90

45.13 6.82

7.02 (0.90)

Figure 6 illustrates the compressive strength measurements showing statistically significant

differences. Figure 7 shows the linear regression computed for the Weibull distribution used to

an

us

cr

ip
t

predict the probability of failure of ceramic parts.

Ac
ce
p

te

Figure 6 Compression strength characteristics of cylindrical samples with layer stacking

orientations 0, 30, 45, 60, and 90 respectively, (n=8), (p<0.05) except(*)

Figure 7 Weibull failure probability functions and linear interpolations for compressive strength of
cylindrical samples printed with layer stacking orientations 0, 30, 45, 60, and 90 respectively

13
Page 13 of 22

ip
t

Figure 8 Fracture surface propagation of samples with orientations 0, 30, 45, 60, and 90
respectively. These images are representative of more than 50% of the samples in each category

cr

The samples were analyzed qualitatively under the microscope to view the orientation of the
fracture path after compression testing to failure. From representative illustrations shown in

us

Figure 8, for samples with 30, 45, 60, and 90, the failure occurred, not surprisingly, along
planes parallel to the stacked powder layers. For the 0 orientation, fracture resulted in an

Discussion

an

irregular v-shaped pattern.

In this study, the optimization of the additive manufacturing process focused on establishing an

orientation that would yield improved mechanical strength characteristics under given loading
conditions, while providing the interconnected porosity required for implant stabilization by

te

bone ingrowth throughout a bone substitute. Five categories of test samples, 4 mm in diameter
and 6 mm in height, were fabricated, with different layer orientations at 0, 30, 45, 60, and 90

Ac
ce
p

with respect to the vertical z -axis in the build compartment (Figure 2). Experimental results
showed that layer orientation had a statistically significant impact on porosity and compressive
strength of samples. The results showed that the 45-oriented samples had the lowest
compressive strength, (13.43 4.60 MPa), while the 90-oriented samples showed the highest
compressive strength, (45.13 6.82 MPa). This considerable difference in strength corresponded
to differences in porosity, whereby the 45-oriented samples had a high bulk porosity value (37.6
2.1%), while the 90 orientation samples, displaying the highest strength, had the lowest
porosity (30.0 2.4%). It is expected that higher porosity would reduce the overall mechanical
strength of a sample.
The difference in compressive strength can be attributed to the additive manufacturing process,
where the orientation of the irregular-shaped CPP particles within each powder layer is
14
Page 14 of 22

influenced by the action of the counter-rotating roller as described by Shanjani et al. [27]. The
results shown in Figure 4 illustrate that the CPP powder used in this work has an irregular shape,
with one distinct longer axis. The counter-rotating roller compacts the powder and aligns the
irregular-shaped particles having an aspect ratio larger than 1 with the longest axis generally

ip
t

parallel with the build plane. As reported elsewhere [27], the resulting particle-particle packing
favors inter-particle contacts at particle ends within the build planes (i.e. at smaller radius of

cr

curvature particle profile contacts, with sharper contact zone). The sinter necks are stronger

when the contact profile between particles has a sharper contact zone, where the contact angle

us

between particles is larger, therefore the sinter necks between particles within each build plane
(parallel to xy) are expected to be stronger than the sinter necks between build planes (along z

an

axis) [27,34].

For compressive loading of homogeneous samples displaying isotropic properties, maximum

shear stresses will act at 45 to the applied force direction resulting in shear fracture along this
direction characterized by an oblique fracture path of failed samples. The observed fracture path
for the 45-oriented samples follows this direction which also corresponds to the weakest sinter

neck direction (i.e. that corresponding to larger radius of curvature contact points between

te

particles in general). These are also the lowest fracture strength samples. The 30- and 60oriented samples approach this condition and in view of the irregular particle shapes, their

Ac
ce
p

measured lower strengths may be similarly explained. Fracture path directions shown in Figure 8
correspond closely to these weakest sample plane directions. This is further expanded on as
discussed below using stress tensor equations summarized in Table 3. In contrast, the 90
orientation corresponds to the weakest direction being parallel to the applied force direction, so
that crack propagation and failure is likely along this direction. The 0-oriented samples develop
highest normal stresses along the weakest planes (Table 3) resulting in transverse fractures (or
more likely, complex fracture paths with both transverse and oblique fracture path segments).
This is reflected in the observed fractures displayed by these samples. In porous ceramic
structures such as CPP, sinter necks represent stress concentration sites, therefore the stronger
the sinter neck, the better the resistance to crack propagation. This means that structurally, the
parts fabricated via additive manufacturing are orthotropic parallel and normal to the z- build
directions.

15
Page 15 of 22

Stress tensor theory for uniaxial compression [35] allows shear stress and normal stress at an
orientation

to be computed based on Equation 5 and Equation 6, where the only load acting on
, also denoted by

for compressive loading

ip
t

the body in uniaxial compression is

cr

us

(5)

an

(6)

The shear stress and normal stress computed for each print stacking orientation

are

summarized in Table 3. These theoretical results, along with the structural anisotropy introduced

by additive manufacturing, explain the experimental behavior of the oriented parts shown in

te

Figure 6 and Figure 7, and summarized in Table 2. The parts manufactured in the 0 and 45
orientations had the lowest compressive strength, 13.50 1.95 MPa and 13.43 4.60 MPa,

Ac
ce
p

respectively. This occurs because the layer orientations within those parts coincide with the
direction perpendicular to the principal normal stress for
principal shear stress for

, and along the direction of the

, respectively as seen in Table 3. The 90 plane, based on the

stress tensor theory, does not experience shear or normal stress, therefore the loading is
distributed along the parallel stacked build planes, where the inter-particle contact results in the
strongest sinter neck formation, therefore these parts show the highest compression strength of
45.13 6.82 MPa. When comparing the compressive strength of the 30 orientation, 20.60
6.23 MPa, and 60 orientation, 28.19 2.46 MPa, it can be seen that the 30 orientation can
sustain a lower compressive strength, as the stacked build planes along the 30 orientation
experienced a higher normal stress than compared to the 60 orientation. The experimental
16
Page 16 of 22

results, supported by the stress tensor theory and the structural anisotropy hypothesis,
demonstrate that the 3DP process can significantly influence the mechanical strength of parts.

30

45

Stress plane angle ()

90

120

135

Shear stress along

orientation

an

90

150

180

Normal stress to
orientation

60

cr

us

Print orientation ()

ip
t

Table 3 Normal and shear stress distribution relative to layering planes for samples manufactured
with stacked layers of orientation 0, 30, 45, 60, 90 deg respectively

te

In this study, there is an inversely proportional correlation between the measured porosity and
compressive strength, where the lowest porosity of 30% corresponded to the highest mechanical

Ac
ce
p

strength of 45 MPa, and the highest porosity of 38% and 37% corresponded to the lowest
mechanical strength of 13.5 MPa and 13.4 MPa respectively. This finding is in accordance
with the literature, where an increased porosity is ideal for promoting bone ingrowth into bone
substitutes, however this comes at the cost of reducing mechanical properties [33,36,37]. The
reason for the observed volume % porosity difference between sample categories is not clear but
may be related to the level of binder deposited per volume of sample during the powder layer
build-up.

The porosity of all samples is between the measured porosity of trabecular bone (50-90%) and
cortical bone (3-12%) [33]. In addition, the compressive strength of each category of samples is
between the measured ultimate strength of trabecular bone (4-12 MPa) and cortical bone (130180 MPa) [7]. From a previous study [27], the pore size for CPP samples appears independent of
layer orientation, and ranges between 20-150 m in size, with a mean of ~56 m, which is in an
17
Page 17 of 22

acceptable range for bone substitutes [33,38,39]. The range of compressive strength obtained in
this study for CPP samples at 90 orientation (45MPa) is similar in value with values obtained
in other studies of 3D printing of CPP at the same orientation (50 MPa) [27]. Furthermore, the
reported for the CPP samples fabricated in this work (3-12) are

ip
t

values of Weibull modulus

comparable with reported Weibull moduli reported for calcium phosphate samples in the

cr

literature (3-9)[40] and (5-10)[27], and indicate good reliability in measurements presented in

us

this work, as the Weibull modulus is inversely proportional with data scatter. A higher
correlates with a good distribution of porous distribution within the part and good repeatability

an

[41].

Based on the results shown in this study, the powder-based additive manufacturing process using

CPP as the raw material is a promising approach in manufacturing implants as bone substitutes
or as components of biphasic tissue-engineered constructs for osteochondral defect repair.
Furthermore, the AM-made structures are anisotropic in nature, offering the possibility of

aligning the layer stacking orientation during part fabrication perpendicular to the path of

te

maximum anticipated compressive loading based on kinetic and kinematic data. This benefit will
be explored in a future work. The layer orientation within the part is an important design

Ac
ce
p

parameter in manufacturing bone substitutes, as the loading kinetics and kinematics on the part
can be estimated and layer orientation can be tuned to ensure greater probability of implant
survival under peak loading conditions.

Conclusion

In this work, the powder-based additive manufacturing process was studied to quantify the effect
of powder layer orientation on mechanical properties of the generated parts. It has been shown
that the layer orientation within the part has a significant influence on the compressive strength
of the resulting structure. Thus layer orientation is an important optimization parameter in the
additive manufacturing design cycle. Furthermore, the results shown in this study can be used to
tune the mechanical strength of an implant along the orientation of maximum loading, if this is
known during the implant design stage. It was concluded that samples made with the 90
18
Page 18 of 22

stacking orientation have the highest compressive strength (45.13 6.82 MPa), whereas those
made with the 0 and 45 stacking orientations exhibit the weakest compressive strength (13.50
1.95 MPa).

ip
t

Acknowledgment
The authors appreciate the funding support received from The Natural Sciences and Engineering

References

an

us

cr

Research Council of Canada (NSERC), grant # RGPIN312074 37063.

Szpalski C, Wetterau M, Barr J, Warren SM. Bone tissue engineering: current strategies
and techniques Part I: Scaffolds. Tissue Eng Part B-Reviews 2012;18:24657.
doi:10.1089/ten.teb.2011.0427.

[2]

Butscher A, Bohner M, Hofmann S, Gauckler L, Mller R. Structural and material

approaches to bone tissue engineering in powder-based three-dimensional printing. Acta
Biomater 2011;7:90720. doi:10.1016/j.actbio.2010.09.039.

[3]

Koh JL. The effect of graft height mismatch on contact pressure following osteochondral
grafting: a biomechanical study. Am J Sports Med 2004;32:31720.
doi:10.1177/0363546503261730.

[4]

Hutmacher DW. Scaffolds in tissue engineering bone and cartilage. Biomaterials

2000;21:252943.

[5]

Frost HM. Bones mechanostat: a 2003 update. Anat Rec A Discov Mol Cell Evol Biol
2003;275:1081101. doi:10.1002/ar.a.10119.

[6]

Leong K, Cheah CM, Chua CK. Solid freeform fabrication of three-dimensional scaffolds
for engineering replacement tissues and organs. Biomaterials 2003;24:236378.
doi:10.1016/S0142-9612(03)00030-9.

[7]

Yang S, Leong KF, Du Z, Chua CK. The design of scaffolds for use in tissue engineering.
Part I. Traditional factors. Tissue Eng 2001;7:67989. doi:10.1089/107632701753337645.

[8]

Yang S, Leong K-F, Du Z, Chua C-K. The design of scaffolds for use in tissue
engineering. Part II. Rapid prototyping techniques. Tissue Eng 2002;8:111.
doi:10.1089/107632702753503009.

Ac
ce
p

te

[1]

19
Page 19 of 22

Yeong W-Y, Chua C-K, Leong K-F, Chandrasekaran M, Lee M-W. Indirect fabrication of
collagen scaffold based on inkjet printing technique. Rapid Prototyp J 2006;12:22937.
doi:10.1108/13552540610682741.

[10]

Gbureck U, Hlzel T, Biermann I, Barralet JE, Grover LM. Preparation of tricalcium

phosphate/calcium pyrophosphate structures via rapid prototyping. J Mater Sci Mater Med
2008;19:155963. doi:10.1007/s10856-008-3373-x.

[11]

Shanjani Y, De Croos JNA, Pilliar RM, Kandel RA, Toyserkani E. Solid freeform
fabrication and characterization of porous calcium polyphosphate structures for tissue
engineering purposes. J Biomed Mater Res B Appl Biomater 2010;93:5109.
doi:10.1002/jbm.b.31610.

[12]

Khalyfa A, Vogt S, Weisser J, Grimm G, Rechtenbach A, Meyer W, et al. Development of

a new calcium phosphate powder-binder system for the 3D printing of patient specific
implants. J Mater Sci Mater Med 2007;18:90916. doi:10.1007/s10856-006-0073-2.

[13]

Vorndran E, Klarner M, Klammert U, Grover LM, Patel S, Barralet JE, et al. 3D Powder
Printing of -Tricalcium Phosphate Ceramics Using Different Strategies. Adv Eng Mater
2008;10:B6771. doi:10.1002/adem.200800179.

[14]

Lu K, Reynolds WT. 3DP process for fine mesh structure printing. Powder Technol
2008;187:118. doi:10.1016/j.powtec.2007.12.017.

[15]

Suwanprateeb J. Influence of printing parameters on the transformation efficiency of 3Dprinted plaster of paris to hydroxyapatite and its properties. Rapid Prototyp J
2012;18:4909. doi:10.1108/13552541211272036.

[16]

Hsu TJ, Lai WH. Manufacturing parts optimization in the three-dimensional printing
process by the taguchi method. J Chinese Inst Eng 2010;33:12130.

[17]

Chumnanklang R, Panyathanmaporn T, Sitthiseripratip K, Suwanprateeb J. 3D printing of

hydroxyapatite: Effect of binder concentration in pre-coated particle on part strength.
Mater Sci Eng C 2007;27:91421. doi:10.1016/j.msec.2006.11.004.

[18]

Suwanprateeb J, Chumnanklang R. Three-dimensional printing of porous polyethylene

structure using water-based binders. J Biomed Mater Res Part B-Applied Bomaterials
2006;78B:13845. doi:10.1002/jbmb.

[19]

Will J, Melcher R, Treul C, Travitzky N, Kneser U, Polykandriotis E, et al. Porous

ceramic bone scaffolds for vascularized bone tissue regeneration. J Mater Sci Mater Med
2008;19:278190. doi:10.1007/s10856-007-3346-5.

[20]

Zhang W, Melcher R, Travitzky N, Bordia RK, Greil P. Three-dimensional printing of

complex-shaped alumina/glass composites. Adv Eng Mater 2009;11:103943.
doi:10.1002/adem.200900213.

Ac
ce
p

te

an

us

cr

ip
t

[9]

20
Page 20 of 22

Gbureck U, Hlzel T, Doillon CJ, Mller FA, Barralet JE. Direct printing of bioceramic
implants with spatially localized angiogenic factors. Adv Mater 2007;19:795800.
doi:10.1002/adma.200601370.

[22]

Habibovic P, Gbureck U, Doillon CJ, Bassett DC, Blitterswijk CA, Barralet JE.
Osteoconduction and osteoinduction of low-temperature 3D printed bioceramic implants.
Biomaterials 2008;29:94453. doi:10.1016/j.biomaterials.2007.10.023.

[23]

Igawa K, Mochizuki M, Sugimori O, Shimizu K, Yamazawa K, Kawaguchi H, et al.

Tailor-made tricalcium phosphate bone implant directly fabricated by a three-dimensional
ink-jet printer. J Atificial Organs Off J Japanese Soc Artif Organs 2006;9:23440.
doi:10.1007/s10047-006-0347-y.

[24]

Choi S, Lee J, Igawa K, Sugimori O, Suzuki S, Mochizuki M, et al. Bone regeneration

within a tailor-made tricalcium phosphate bone implant with both horizontal and vertical
cylindrical holes transplanted into the skull of dogs. J Artif Organs Off J Japanese Soc
Artif Organs 2009;12:2747. doi:10.1007/s10047-009-0474-3.

[25]

Vlasea M, Toyserkani E. Experimental characterization and numerical modeling of a

micro-syringe deposition system for dispensing sacrificial photopolymers on particulate
ceramic substrates. J Mater Process Technol 2013;213:19707.
doi:http://dx.doi.org/10.1016/j.jmatprotec.2013.05.011.

[26]

Vlasea M, Shanjani Y, Bothe A, Kandel R, Toyserkani E. A combined additive

manufacturing and micro-syringe deposition technique for realization of bio-ceramic
structures with micro-scale channels. Int J Adv Manuf Technol 2013;68:22619.
doi:10.1007/s00170-013-4839-7.

[27]

Shanjani Y, Hu Y, Pilliar RM, Toyserkani E. Mechanical characteristics of solidfreeform-fabricated porous calcium polyphosphate structures with oriented stacked layers.
Acta Biomater 2011;7:178896. doi:10.1016/j.actbio.2010.12.017.

[28]

Pilliar RM, Filiaggi MJ, Wells JD, Grynpas MD, Kandel RA. Porous calcium
polyphosphate scaffolds for bone substitute applications - in vitro characterization.
Biomaterials 2001;22:96372.

[29]

Pilliar RM, Kandel R a, Grynpas MD, Hu Y. Porous calcium polyphosphate as loadbearing bone substitutes: in vivo study. J Biomed Mater Res B Appl Biomater
2013;101:18. doi:10.1002/jbm.b.32832.

[30]

Shanjani Y, Hu Y, Toyserkani E, Grynpas M, Kandel R a, Pilliar RM. Solid freeform

fabrication of porous calcium polyphosphate structures for bone substitute applications: in
vivo studies. J Biomed Mater Res B Appl Biomater 2013;101:97280.
doi:10.1002/jbm.b.32905.

Ac
ce
p

te

an

us

cr

ip
t

[21]

21
Page 21 of 22

Pilliar RM, Hong J, Santerre PJ. Method of manufacture of porous inorganic structures,
Patent Number 7494614. 7494614, 2009.

[32]

Absi J, Fournier P. Influence of experimental parameters on the estimated value of

Weibulls modulus. J Mater Sci 1999;34:121927.

[33]

Karageorgiou V, Kaplan D. Porosity of 3D biomaterial scaffolds and osteogenesis.

Biomaterials 2005;26:547491. doi:10.1016/j.biomaterials.2005.02.002.

[34]

Bannister M. Shape sensitivity of initial sintering equations. J Am Ceram Soc

1968;51:54853. doi:10.1111/j.1151-2916.1968.tb13319.x.

[35]

Drucker D. Introduction to Mechanics of Deformable Solids. New York: McGraw-Hill;

1967.

[36]

Barralet JE, Grover L, Gaunt T, Wright a J, Gibson IR. Preparation of macroporous

calcium phosphate cement tissue engineering scaffold. Biomaterials 2002;23:306372.

[37]

Lin ASP, Barrows TH, Cartmell SH, Guldberg RE. Microarchitectural and mechanical
characterization of oriented porous polymer scaffolds. Biomaterials 2003;24:4819.

[38]

Porter JR, Ruckh TT, Popat KC. Bone tissue engineering: a review in bone biomimetics
and drug delivery strategies. Biotechnol Prog 2009;25:153960. doi:10.1002/btpr.246.

[39]

Stevens B, Yang Y, Mohandas A, Stucker B, Nguyen KT. A review of materials,

fabrication methods, and strategies used to enhance bone regeneration in engineered bone
tissues. J Biomed Mater Res B Appl Biomater 2008;85:57382. doi:10.1002/jbm.b.30962.

[40]

Miranda P, Pajares A, Saiz E, Tomsia AP, Guiberteau F. Mechanical properties of calcium

phosphate scaffolds fabricated by robocasting. J Biomed Mater Res A 2008;85:21827.
doi:10.1002/jbm.a.31587.

[41]

Montanaro L, Jorand OY, Fantozzib G, Negroa A. Ceramic foams by powder processing.

J Eur Ceram Soc 1998;18:133950.

Ac
ce
p

te

an

us

cr

ip
t

[31]

22
Page 22 of 22