This content has been downloaded from IOPscience. Please scroll down to see the full text.
IP Address: 188.8.131.52
This content was downloaded on 05/03/2015 at 12:40
Please note that terms and conditions apply.
Fabrication of interconnected microporous biomaterials with high hydroxyapatite nanoparticle
View the table of contents for this issue, or go to the journal homepage for more
2010 Biofabrication 2 035006
HomeSearch CollectionsJournalsAbout Contact usMy IOPscience
Biofabrication 2 (2010) 035006 (10pp)
Fabrication of interconnected
microporous biomaterials with high
hydroxyapatite nanoparticle loading
Wei Zhang1, Donggang Yao1,5, Qingwei Zhang2,3,4, Jack G Zhou3and
Peter I Lelkes2,4
1School of Polymer Textile and Fiber Engineering, Georgia Institute of Technology, Atlanta, GA, USA
2School of Biomedical Engineering, Science, and Health Systems, Drexel University, Philadelphia,
PA 19104, USA
3Department of Mechanical Engineering and Mechanics, Drexel University, Philadelphia, PA 19104,
4Surgical Engineering Enterprise, Drexel University College of Medicine, Philadelphia, PA, 19102, USA
Received 10 May 2010
Accepted for publication 12 August 2010
Published 8 September 2010
Online at stacks.iop.org/BF/2/035006
Hydroxyapatite (HA) is known to promote osteogenicity and enhance the mechanical
properties of biopolymers. However, incorporating a large amount of HA into a porous
biopolymer still remains a challenge. In the present work, a new method was developed to
produce interconnected microporous poly(glycolic-co-lactic acid) (PLGA) with high HA
nanoparticle loading. First, a ternary blend comprising PLGA/PS (polystyrene)/HA
(40/40/20 wt%) was prepared by melt blending under conditions for formation of a
co-continuous phase structure. Next, a dynamic annealing stage under small-strain oscillation
was applied to the blend to facilitate nanoparticle redistribution. Finally, the PS phase was
sacrificially extracted, leaving a porous matrix. The results from different characterizations
suggested that the applied small-strain oscillation substantially accelerated the migration of
HA nanoparticles during annealing from the PS phase to the PLGA phase; nearly all HA
particles were uniformly presented in the PLGA phase after a short period of annealing. After
dissolution of the PS phase, a PLGA material with interconnected microporous structure was
successfully produced, with a high HA loading above 30 wt%. The mechanisms beneath the
experimental observations, particularly on the enhanced particle migration process, were
discussed, and strategies for producing highly particle loaded biopolymers with interconnected
microporous structures were proposed.
Biodegradable and biocompatible polymers, due to some
advantageous properties over traditional metallic materials,
have increasing application in biomedical implanting devices
(Higashi et al 1986, Hollinger and Battistone 1986, Bos et al
1991). Extensive research has also been conducted to further
improve the mechanical properties, as well as bioactivity, of
these ‘new’ implanting materials. In particular, combining
5Author to whom any correspondence should be addressed.
hydroxyapatite with a biopolymer such as poly(glycolic-
co-lactic acid) (PLGA) was demonstrated to be able to
generate the synergistic result by taking advantages from
each component (Higashi et al 1986, Hollinger and Battistone
Recently, the importance of interconnected microporous
structures was also emphasized by researchers in this field,
and many intelligent techniques were invented to make such
structures. Existing techniques able to realize this design
concept include fiber bonding (nonwoven meshes) (Mooney
et al 1996b, Mikos et al 1993, Freed et al 1993), solvent
© 2010 IOP Publishing LtdPrinted in the UK
Biofabrication 2 (2010) 035006 W Zhang et al
casting/particulate leaching (Thomson et al 1998, Goldstein
et al 1999, Mikos et al 1993, Shastri et al 2000), phase
separation/emulsification (Whang et al 1995, Schugens et al
1996, Nam and Park 1999, Lo et al 1995), gas foaming
Beckman 1997, Harris et al1998, Wang et al 2006, Singh et al
2004), co-continuous melt blending (Potschke and Paul 2003,
Yuan and Favis 2004, 2005, Yao et al 2009, Zhang et al 2010,
Virgilio et al 2010) and solid freeform fabrication techniques
et al 2005, Taboas et al 2003). These techniques have been
found to be especially useful when dealing with polymers.
In particular, co-continuous melt blending has been shown to
be an extremely versatile technique; interconnected porous
structures with controllable porosity, pore size and even pore
size distribution can be generated with this method (Potschke
and Paul 2003, Yuan and Favis 2004, 2005, Yao et al 2009,
Zhang et al 2010, Virgilio et al 2010).
However, most of these existing methods for porous
materials are not suitable for filled/reinforced polymers,
and the lack in mechanical strength of the resulting porous
materials hinders their applications in load bearing, either
primary or secondary, biomedical devices. Particularly for
methods dealing with a polymer solution, such as solvent
casting and phase separation, one primary difficulty is caused
by the segregation of the particles over the required long
processing period. For other methods coping with a polymer
melt, such as gas forming and melt blending, the extremely
high viscosity of the filled polymer, especially at high particle
loading, causes a processing difficulty.
freeform fabrication techniques such as fusion deposition
modeling and laser sintering can be used to include inorganic
fillers, but they have a low resolution, prohibiting the
production of finely structured materials.
In the present work, we developed and investigated a
technique for processing porous biopolymer composites with
mixing small-strain oscillation was found to be effective in
obtaining the required interpenetrating microporous material.
As a case study, this technique was used to generate a highly
HA-loaded PLGA nanocomposite with an interconnected
microporous structure. The mechanisms underneath the
experimental observations, particularly on the enhanced
particle migration under oscillatory shear, were discussed, and
strategies for producing highly particle-loaded biopolymers
with interconnected microporous structures were proposed.
2. Fabrication method
Polymer melt blending was identified as a potential technique
for generating the required microporous material with high
nanoparticle loading. Due to its high level of process
controllability, melt blending has recently been reported as
a versatile protocol for processing porous materials (Potschke
and Paul 2003, Yuan and Favis 2004, 2005, Yao et al 2009).
In this method, a binary polymer blend with a co-continuous
phase structure was prepared first and, after extraction of the
sacrificial polymer component, a porous material is produced.
To incorporate high loading of nanoparticles in the porous
contains a large amount of particles, channel blockage can
result after sacrificial polymer dissolution. Second, inclusion
of a high concentration of nanoparticles in the blend can result
in extremely high viscosity and therefore a difficult process.
Alternatively, one may elect to premix the biopolymer with
the particles and then blend the resulting mixture with the
sacrificial polymer. However, this would result in a large
difference in viscosity between the two polymer components,
leading to the generation of an unwanted droplet phase
The new technique is aimed at reducing the nanoparticle
usage during blending, resolving the channel blockage
problem after sacrificial extraction, and yet producing
an interpenetrating porous material with the desired high
nanoparticle loading. The following steps are involved in this
process. First, a biodegradable polymer, a sacrificial polymer
and inorganic fillers are melt-blended under appropriate
conditions for forming a co-continuous phase structure. Then,
at the melt state, the blend is subjected to a small-strain
oscillation to facilitate migration of the filler particles from
final step, the sacrificial phase is selectively dissolved to leave
biodegradable polymer and the inorganic fillers. The salient
feature of this technique is the inclusion of an accelerated
blending), but also avoids unwanted accumulation of particles
in the open channels.
3. Experimental details
PLGA (product name: PURASORB PLG 8531) was acquired
from Purac (a division of CSM, The Netherlands).
polymer is a GMP (good manufacturing practice) grade
copolymer of L-lactide and glycolide in an 85/15 molar
mass ratio and with an inherent viscosity midpoint of
3.1 dl g−1. The as-received PLGA was amorphous with a
glass transition temperature at 60◦C, as measured by DSC
(differential scanning calorimetry) at 10◦C min−1ramping
rate. PS (polystyrene) from Dow Chemical (Dow StyronR ?
675 polystyrene) with a Vicat softening point at 108
was chosen as the sacrificial polymer. Both polymers were
obtained in a pellet form from the suppliers. HA nanoparticles
in a National Formulary (NF)/Food Chemical Codex (FCC)
grade with CAS code 1306-06-5 were obtained from Fisher
Scientific (Calcium Phosphate C133-500). The size of the
HA particles was around 100 nm or smaller, as assessed by
scanning electron microscopy (SEM).
A Brabender Intelli-Torque rheometer equipped with two
roller blades capable of tumbling and distributive mixing
was used to mix the raw materials. For all blends, with or
Biofabrication 2 (2010) 035006 W Zhang et al
Figure 1. SEM morphology of the as-mixed blends after extraction of the PS phase: (a) PLGA/PS 50/50 wt% blend and (b) PLGA/PS/
HA 40/40/20 wt% blend.
without HA nanoparticles, the concentration ratio between PS
and PLGA was kept at 1:1 to promote the development of a
phase structure with dual continuity (co-continuity) of both
polymer phases. To reduce hydrolysis caused degradation, all
the materials involved in this study were dried in a vacuum
oven overnight at 90◦C to remove the absorbed moisture
and stored in a desiccator until ready for the next step of
the experiment. A compounding temperature of 200◦C and
a mixing speed of 60 rpm were used during the blending
process. The resulting shear rate was approximately 100 s−1.
Ultrahigh purity nitrogen was used as a purging gas during
the whole mixing process to reduce degradation caused by
oxygen and moisture. A homogeneous blend morphology
inside the chamber was achieved by the development of a
constant blending torque.
In this study, two mixing procedures were compared. For
the first procedure, a single-step mixing procedure, all three
components were fed into the mixer simultaneously. For the
second procedure, a two-step mixing procedure, a uniform
blend of PLGA and HA was premixed first, and then the PS
component was added. After mixing, the molten blend was
immediately scraped out of the chamber and quenched in tap
water, followed by the same drying and storage procedure
The dried blends were compression molded into 1.2 mm
thick sheets at 200◦C and cut into disks of 25 mm diameter.
The disk sample was then reheated to 200◦C and subjected to
oscillatory shear deformation under an isothermal condition at
varied frequency and varied processing time. The oscillatory
shear deformation was applied to the disk sample using a
parallel-plate rheometer (AR2000ex from TA Instrument),
with a gap size set at 1 mm. After the oscillatory shear, the
sample was cooled down to room temperature with a cooling
rate of 10◦C min−1.
To produce a porous material, the solidified ternary
blend after compression molding and dynamic annealing was
submerged in cyclohexane at 75◦C for 3 days to extract the
PS phase, the sacrificial phase. The complete extraction of PS
was verified by comparing the weight of the dissolved sample
with that of the original sample. Completely etched samples
were dried in a vacuum oven at 90◦C.
Different characterization techniques, including XRT (x-ray
tomography), SEM and TGA (thermogravimetric analysis)
were used to investigate the structure and morphology of the
materials produced in this study. For XRT, the disks after
dynamic annealing and cooling but without solvent extraction
were directly inspected on Dage x-ray XD7600NT XRT. SEM
(LEO SEM 1550 at 5 kV) was used to examine the structure
of the porous materials after PS extraction. The SEM samples
were sputtered with a gold/platinum alloy.
materials without sputtering were also examined with TGA
of HA nanoparticles.The TGA was operated under a
pure nitrogen environment at a 10◦C min−1heating rate.
Additionally, the rheological properties of the tertiary blend
were characterized on a parallel-plate rotational rheometer
(AR2000ex from TA Instrument).
4. Results and discussion
4.1. Morphology from the SEM
The SEM was used to observe the phase structure and
nanoparticle distribution in the samples fabricated under
different conditions. The phase structures of the as-mixed
PLGA/PS 50/50 wt% and PLGA/PS/HA 40/40/20 wt%
blends are compared in figure 1. In this comparison, both
blends were prepared using the single-step mixing procedure.
The phase size of the PLGA/PS blend was around 50 μm.
With the ratio of PLGA to PS kept at 1:1, the addition of
20 wt% of HA nanoparticles in the blend yielded a reduced
phase size of approximately 5 μm. This ten times reduction in
phase size may be partly attributed to the increased viscosity
due to the nanoparticle addition.
Figure 2 shows the morphological evolution of the
PLGA/PS/HA 40/40/20 wt% blend during dynamic
annealing under small-strain oscillation.
shown in the figure, the PS phase was extracted.
seen that, with longer annealing time, more HA nanoparticles
entered into the PLGA phase. For the as-mixed sample, HA
For all samples
Biofabrication 2 (2010) 035006 W Zhang et al
Figure 2. SEM morphology of the tertiary blend (PLGA/PS/HA 40/40/20 wt%) after extraction of the PS phase. The blend was prepared
using the single-step mixing approach and after mixing was subjected to oscillatory shear deformation with two levels of oscillation time:
(a) 5 min and (b)–(d) 20 min with different magnifications.
and 1 rad s−1, a considerable number of HA nanoparticles still
remained in the PS phase; after dissolution of the PS phase,
these HA particles resided in the channels inside the PLGA
(figure 2(a), arrow heads). In contrast, after a sufficiently
long oscillation time, e.g. 20 min, the channel became rather
clean, with only a limited number of relatively large HA
particles presented (figures 2(b) and (c), arrow heads). This
indicated that most small nanoparticles had migrated into
the PLGA phase. The SEM image of this sample further
indicated that these nanoparticles were uniformly distributed
in the remaining PLGA phase (figure 2(d), arrow heads). The
resulting material is a porous PLGA/HA composite, with
an interconnected 3D microchannel network running inside
the whole material. For comparison purposes, blends with
only quiescent annealing, i.e. without small-strain oscillation,
were also examined.Without oscillation, a large number
of nanoparticles were presented inside the channels (with
morphology similar to figure 2(a)), even with a prolonged
annealing time over 20 min.
small-strain oscillation effectively assisted in the migration
of the nanoparticles from the PS phase to the PLGA phase.
Complete migration of the nanoparticles yielded a PLGA/HA
composite at a high HA loading level reaching 33%. The
composite material prepared by this method further possessed
interconnected, clean microchannels.
number of HA particles were presented in the PS phase, the
channels after dissolution of the PS phase could be blocked by
These results suggested that
Note that if a large
the residual particles. This could reduce the connectivity of
the channels and hinder the use of the composite material in
tissue engineering applications.
The blends used for the SEM study shown in figure 2
were all prepared using the single-step mixing approach. In
this case, all three components, i.e. PLGA, PS and HA, were
fed into the batch mixer simultaneously during mixing. For
comparison, the morphology of the blends made by premixing
PLGA and HA and then mixing with PS was obtained, as
shown in figure 3. The motivation was that this second
approach, if successful, could be a more effective method for
forming highly HA-loaded PLGA materials/devices because
in this case particle migration is no longer needed. However,
the experimental results shown in figure 3 indicated that the
blend did not form a co-continuous morphology; instead,
a droplet in matrix morphology was formed. This second
method, therefore, did not yield a desirable porous PLGA
composite material with interpenetrating microchannels.
4.2. Thermal analysis
The concentration of HA nanoparticles in the porous
A heating rate of 10◦C min−1was used in the test. The TGA
data of the sample processed at 200◦C and under 1 rad s−1
oscillation for 20 min are shown in figure 4.
testing data, one can find that the residual solid material at
700◦C was approximately 33 wt%. It is known that at such
Biofabrication 2 (2010) 035006W Zhang et al
Figure 3. SEM morphology of the tertiary blend (PLGA/PS/HA 40/40/20 wt%) prepared using the two-step approach: (a) at low
magnification and (b) a zoomed-in view.
0 200400 600800
Weight loss (mg/mg)
Figure 4. TGA of the PLGA/PS/HA blend sample oscillated with
1 rad s−1angular frequency at 200◦C for 2000 s. The PS phase was
extracted by solvent before TGA.
a high temperature, PLGA should be degraded into gasses.
Therefore, the solid content should be HA particles, and the
concentration of the HA in the final porous PLGA composite
should be 33%. Since microchannels inside the porous PLGA
33% HA particles were mainly included in the PLGA matrix
4.3. XRT observation
The contrast in XRT imaging comes from local density
differences, or, more specifically, from local electron density
differences. This technique has been reported to be useful for
examining composite or blend structures with a large density
contrast of constituents, including porous structures (Delisee
et al 2010, Feng et al 2010). The density contrast between the
two polymers involved in this study is much lower that that
PS and HA are 1.2, 1.05 and 3.16 g cm−3, respectively; the
resulting density contrast between PLGA and PS is around 1.1
while the value is 3 for the HA/matrix. Therefore, XRT was
the highly dense component, in the tertiary PLGA/PS/HA
The samples used for XRT imaging in this study were
all prepared using the single-step mixing procedure with a
post-annealing stage under oscillatory deformation.
different levels of annealing time were used during sample
preparation: 5, 10 and 20 min. The results are shown in
figure 5. A number of light-colored dots/regions can be seen
in the XRT images for all samples. From the known density
contrast between HA and the polymers, these light dots were
agreed with the SEM results previously described; when HA
particles migrated from the PS phase to the PLGA phase, the
PS phase had a reduced density, showing a light shade in
XRT. Some kinetics on morphological development was also
shown in the XRT results. For some reason, the light dots
grew initially (from 5 min to 10 min) but reduced in size in a
later stage (i.e. at 20 min). This result is a bit surprising and
yet indicated the possible rearrangement of the nanoparticles
in the polymer phases. At the initial stage of annealing,
it is possible that a large number of highly concentrated
nanoparticles were presented only at the interfacial region
between the PS phase and the PLGA phase. The size of the
PLGA phase would increase when the nanoparticles further
diffused into and became evenly distributed in the PLGA
phase. This rearrangement would cause the relative reduction
of the PS phase.
It should also be noted that the size of the light dots
appeared to be much larger than the size of the actual
microchannels observed in SEM. This would be understood
considering how XRT builds up the final image.
transmission mode techniques, XRT averages the information
normal plane of the incident beam. The projected features in
a cross-sectional view. This could be particularly true for a 3D
size over 0.5 mm and obviously larger than other surrounding
dots, were also observed in the XRT images.
were believed to be air bubbles introduced to the blend during
Biofabrication 2 (2010) 035006W Zhang et al
Figure 5. XRT of the PLGA/PS/HA blend (40/40/20 wt%) subjected to oscillatory shear at 1 rad s−1and 200◦C for different shearing
times: (a) 5 min, (b) 10 min and (c) 20 min.
0.11 10 100
Angular freqency (rad/s)
Complex viscosity (Pa*s)
Figure 6. Complex viscosity versus angular frequency at 200◦C for
PS and PLGA.
The experimental observations suggest that the final structure
of the porous material is strongly dependent on the process
conditions, particularly on the processing temperature and
dynamic loading. The phase evolution under these influences
is considered to strongly correlate with the deformation
properties (or rheological properties) of the blend system.
To understand the underlying mechanism for the structural
development in this fabrication process, rheological tests were
conducted to collect real-time information for interpreting
the observed phaseevolution
All parallel-plate rheological tests were carried out at
1 mm gap size and under small-strain oscillatory shear
in the linear viscoelastic range.
minimized during the test. The relations of complex viscosity
and angular frequency at 200◦C for the pure polymers are
given in figure 6.Using the well-known Cox–Merz rule
(Cox and Merz 1958) for shear viscosity of pure polymers,
one can convert angular frequency equivalent to shear rate.
From the resulting shear rate dependence, it is seen that both
polymers are highly pseudoplastic, with viscosity decreasing
with the increase of shear rate.
and particle migration
In order to preserve the
Two types of viscosity
were important in the current study.
and their ratio at the processing shear rate during batch
mixing were major factors affecting the phase structure of
the blend. For example, the phase inversion could be related
to the viscosity ratio (Miles and Zurek 1988). Second, for
a particulate blend, the viscosity is also expected to affect
the particle migration/diffusion process.
such as the Stoke–Einstein equation (Einstein 1908) predicts
an inverse relation between the diffusion coefficient and the
The shear rate inside the batch mixer was estimated to
be 100 s−1based on the size and geometry of the mixer
components and the RPM of the mixing rotor. The viscosity
ratio of PLGA over PS in this case was around 3, calculated
from the data in figure 6. For a 50/50 wt% PLGA/PS blend
used in this study, this viscosity ratio could result in a co-
continuous phase morphology. However, for the procedure in
form the tertiary blend, the viscosity of the PLGA/HA blend
should be used for final blend morphology prediction instead
of the viscosity of PLGA. It is known that the addition of
inorganic particles into polymers typically causes a viscosity
concentration. This can be used to explain the difficulty in
forming a co-continuous phase structure using the two-step
mixing approach (cf figure 3).
To prepare samples for parallel-plate rheometry, the
cut into 25 mm diameter disks. During time- and frequency-
sweep tests, the gap size and strain amplitude were fixed at
1 mm and 0.01, respectively. In figure 7, time sweep testing
results of PLGA/PS 50/50 wt% blend at 1 and 100 rad s−1
angular frequencies were summarized. For this binary blend,
a steady decrease in complex viscosity was found for both
angular frequencies.It is well known that the complex
viscosity consists of two parts, i.e. elastic contribution and
viscous contribution. Since the testing temperature and shear
rate were kept constant for these time sweep tests, the viscous
part was not expected to cause a large change in complex
viscosity. Therefore, the decrease in complex viscosity as a
function of time can more likely be attributed to the decrease
in the total interfacial area due to phase coarsening during
annealing of the co-continuous blend.
Figure 8 shows different rheological behaviors of HA
ternary blends prepared by different mixing procedures. The
First, the viscosities
Biofabrication 2 (2010) 035006W Zhang et al
0 2000 40006000 8000
Complex viscosity (Pa*s)
0 2000 40006000 8000
Complex viscosity (Pa*s)
Figure 7. Small-strain time sweep data of a PLGA/PS blend
(50/50 wt%) at 200◦C and two different angular frequencies:
(a) 1 rad s−1and (b) 100 rad s−1.
time-sweep data of the blend prepared using single-step
mixing are shown as curve A in the figure.
frequency applied in this test was 1 rad s−1and the testing
temperature was 200◦C. For comparison, the blend with the
same composition but made through two-step mixing was
tested under the same conditions, and the data are shown as
curve B in the figure. Despite the similar complex viscosity
values at the start, the two blends showed a different time
in viscosity over a long period of time (∼20 min) before
a viscosity peak was reached. The increase of viscosity in
this period of time exceeded 100%.
only showed a mild increase in viscosity, approximately 8%,
and the viscosity peak was reached in a much shorter time,
about 6 min. The different responses of the two different
blends during testing may be accredited to different structural
evolution processes. For blend A, there existed a significant
particle migration process; that is, the HA particles migrated
from the PS phase to the PLGA phase during testing, causing
an increase in viscosity of the blend.
consistent with the morphological observations obtained in
SEM and XRT. By comparison, for blend B, the particle
migration process is expected to be a weak process during
testing since the HA particles were already in the PLGA
phase at the beginning of testing due to premixing. PLGA/PS
under the identical conditions as those for the tertiary blends.
In contrast, blend B
This explanation is
0 2000 4000 60008000
Complex viscosity (Pa*s)
A: One-step mixing
B: Two-step mixing
Figure 8. Small-strain time sweep data of PLGA/PS/HA blends
(40/40/20 wt%) at 200◦C and 1 rad s−1angular frequency. Blend
A was prepared by single-step mixing, and blend B was prepared by
monotonically. This provides further evidence that particle
migration was the major mechanism causing the viscosity
increase in the tertiary blends.
alsorevealed inprevious publications, butmostlydealing with
a dispersed morphology (Potschke et al 2004, Potschke et al
2007, Zhao and Huang 2008, Hong et al 2008). Possible
driving forces for particle migration in such systems can
be derived from the difference in chemical affinity between
the nanoparticles and the different polymer components. The
particles tend to migrate to the polymer phase exhibiting
stronger chemical/physical interactions with the additive.
Utilizing this mechanism, one can control the migration of
nanoparticles in immiscible polymer blends. In particular,
one can move nanoparticles from a sacrificial phase to the
desired polymer phase.
Parametric experiments were conducted on blend A to
study the effects of oscillation frequency and processing
temperature on the migration kinetics.
time of HA nanoparticles under different conditions were
measured. From the morphological observations and
rheological behaviors of the PLGA/PS/HA blends (as
described in the previous sections), it appeared reasonable
to use the peak viscosity time as the characteristic time for
particle migration. During testing, two levels of temperature,
200 and 220◦C, and two levels of angular frequency, 1 and
100 rad s−1, were employed, producing four different testing
data sets. The viscosity data are given in figure 9. Note that
the data for 200◦C and 1 rad s−1have already been given in
figure 8 and thus not reproduced.
in all these tests experienced a considerable increase in the
initial stage of testing. However, the time scale for reaching
the peak viscosity differed under different testing conditions.
Furthermore, the magnitude of viscosity was significantly
dependent on the testing conditions.
dependence and the shear rate dependence of viscosity can
The viscosity obtained
Biofabrication 2 (2010) 035006W Zhang et al
Table 1. Shear viscosity of PLGA and PS at different angular frequencies and temperatures.
combined viscosity at
@ 200◦C Material@ 220◦C@ 220◦C
1 rad s−1
100 rad s−1
0 10002000 3000
(A, B) Complex viscosity (Pa*s)
(C) Complex viscosity (Pa*s)
Figure 9. Small strain time sweep data of PLGA/PS/HA blends
(40/40/20 wt%) under different testing conditions: (A) at 200◦C
and 100 rad s−1angular frequency; (B) at 220◦C and 100 rad s−1
angular frequency; (C) at 220◦C and 1 rad s−1angular frequency.
The peak times are indicated in the figure as t1= 140 s, t2= 360 s
and t3= 500 s.
be clearly seen in the data.
expected to be lower at a higher temperature.
the increase in testing frequency can cause a reduction in
the viscosity magnitude due to the shear-thinning behavior
or pseudoplasticity of the polymeric material.
From the classical diffusion theory (Einstein 1908, Elias
et al 2008), temperature and viscosity are expected to be
two major parameters affecting the diffusion coefficient. The
increase in temperature and the decrease in viscosity would
cause an increase in the diffusion coefficient. Based on this
theory, it would be interesting to define a new parameter,
ξ = T/η, and analyze the data in figure 9 in terms of this
ξ parameter. Here, η is the equivalent viscosity of the blend
and T is the testing temperature. For a complex blend system
as the one involved in this study, the arithmetically combined
equivalent viscosity affecting the particle migration process.
Table 1 summarizes the testing conditions, the complex
viscosities and the arithmetically combined viscosities. From
these data, normalized ξ parameters (ξ/ξr) and normalized
peak times (t/tr) were calculated and summarized in table 2.
the reference time equal to 1500 s obtained from figure 8. The
test at 200◦C and 1 rad s−1was chosen as a reference state.
Particularly, the viscosity is
Table 2. Comparison between experimental peek times with the
predicted migration time from the ξ parameter (ξ = T/η). The data
are all normalized with those at the reference state, 200◦C and
1 rad s−1. The subscript r refers to the reference state. The
experimental data listed in table 1 were used in the calculation.
100 rad s−1
1 rad s−1
100 rad s−1
Normalized ξ parameter, ξ/ξr 0.20
Normalized peek time, t/tr
From table 2, it can be seen that the normalized peak times are
least from the first-order approximation. This simple method
for predicting the viscosity peak time, and consequently the
characteristic particle migration time and the kinetics, could
be useful for complex blend systems as the one employed in
In this work, a new technique for processing interpenetrating
microporous material with high nanoparticle loading was
developed and investigated.
PLGA/PS/HA (40/40/20 wt%) was chosen as a model
system. The blend prepared under conditions for formation
of a co-continuous structure was annealed under small-strain
oscillatory shear between two parallel plates.
from different characterizations suggested that the applied
small-strain oscillation substantially accelerated the migration
of HA nanoparticles during annealing from the PS phase to
the PLGA phase; nearly all HA particles were uniformly
presented in the PLGA phase after 20 min annealing at
200◦C. After dissolution of the PS phase, a PLGA/HA
compositematerialwithinterconnected microporous structure
was successfully produced.
the PLGA phase exceeded 30 wt%. Two different mixing
strategies, i.e. single-step mixing and two-step mixing, were
yielded a desirable porous PLGA composite material with
interpenetrating microchannels, while the two-step approach
produced an undesired droplet-in-matrix phase structure.
Parametric studies on the effect of temperature and the
oscillation frequency on the particle migration process were
also conducted. Based on the experimental data, absolute
A tertiary polymer blend
The HA particle loading in
Biofabrication 2 (2010) 035006W Zhang et al
temperature divided by arithmetically combined viscosity was
suggested as the dominant parameter depicting the kinetics of
the HA nanoparticle migration process during post-annealing
under small-strain oscillation.
The authors acknowledge funding from the National
Science Foundation under awards CMMI-0800016, CMMI-
0800735 and CMMI-0927697, as well as support from the
Strategic Initiative of Drexel College of Medicine (Surgical
Engineering Enterprise (SEE)), to QZ, JZ, PIL and from
the Nanotechnology Institute of Southeastern Pennysylvania
(NTI), to PIL.
Ang T H, Sultana F S A, Hutmacher D W, Wong Y S, Fuh J Y H,
Mo X M, Loh H T, Burdet E and Teoh S H 2002 Fabrication of
3D chitosan-hydroxyapatite scaffolds using a robotic
dispensing system Mater. Sci. Eng. C 20 35–42
Bos R R M, Rozema F R, Boering G, Nijenhuis A J, Pennings A J,
Verwey A B, Nieuwenhuis P and Jansen H W B 1991
Degradation of and tissue reaction to biodegradable
poly(L-lactide) for use as internal fixation of fractures—a study
in rats Biomaterials 12 32–6
Cox W P and Merz E H 1958 Correlation of dynamic and steady
flow viscosities J. Polym. Sci. 28 619–22
Delisee C, Lux J and Malvestio J 2010 3D morphology and
permeability of highly porous cellulosic fibrous material
Transp. Porous Med. 83 623–36
Einstein A 1908 Elementary theory of the Brownian motion
Z. Elektrochemie und Angew. Phys. Chem. 14 235–9
Elias L, Fenouillot F, Majeste J C, Alcouffe P and Cassagnau P
2008 Immiscible polymer blends stabilized with nano-silica
particles: rheology and effective interfacial tension Polymer
Feng C, Wang R, Wu Y and Li G 2010 Preliminary analysis of a
linear pore pattern formed on poly(vinylidene fluoride-co-
hexafluoro propylene) porous membrane surfaces J. Membr.
Sci. 352 255–61
Freed L E, Marquis J C, Nohria A, Emmanual J, Mikos A G
and Langer R 1993 Neocartilage formation in vitro and in vivo
using cells cultured on synthetic biodegradable polymers
J. Biomed. Mater. Res. 27 11–23
Goldstein A S, Zhu G, Morris G E, Meszlenyi R K and Mikos A G
1999 Effect of osteoblastic culture conditions on the structure
of poly(D,L-lactic-co-glycolic acid) foam scaffolds Tissue Eng.
Harris L D, Kim B S and Mooney D J 1998 Open pore
biodegradable matrices formed with gas foaming J. Biomed.
Mater. Res. 42 396–402
Higashi S, Yamamuro T, Nakamura T, Ikada Y, Hyon S H
and Jamshidi K 1986 Polymer hydroxyapatite composites for
biodegradable bone fillers Biomaterials 7 183–7
Hollinger J O and Battistone G C 1986 Biodegradable bone repair
materials-synthetic-polymers and ceramics Clin. Orthop. Relat.
Res. 207 290–305
Hong J K, Kim Y K, Ahn K H and Lee S J 2008 Shear-induced
migration of nanoclay during morphology evolution of
PBT/PS blend J. Appl. Polym. Sci. 108 565–75
Hoque M E, Hutmacher E W, Feng W, Li S, Huang M H, Vert M
and Wong Y S 2005 Fabrication using a rapid prototyping
system and in vitro characterization of PEG-PCL-PLA
scaffolds for tissue engineering J. Biomater. Sci., Polym. Edn
Landers R and Mlhaupt R 2000 Desktop manufacturing of complex
objects, prototypes and biomedical scaffolds by means of
computer-assisted design combined with computer-guided 3D
plotting of polymers and reactive oligomers Macromol. Mater.
Eng. 282 17–21
Lo H, Kadiyala S, Guggino S E and Leong K W 1996 Poly(L-lactic
acid) foams with cell seeding and controlled-release capacity
J. Biomed. Mater. Res. 30 475–84
Lo H, Ponticiello M S and Leong K W 1995 Fabrication of
controlled release biodegradable foams by phase separation
Tissue Eng. 1 15–28
Mikos A G, Bao Y, Cima L G, Ingber D E, Vacanti J P and
Langer R 1993 Preparation of poly (glycolic acid) bonded fiber
structures for cell attachment J. Biomed. Mater. Res. 27 183–9
Mikos A G, Thorsen A J, Czerwonka L A, Bao Y, Langer R,
Winslow D N and Vacanti J P 1994 Preparation and
characterization of poly(L-lactic acid) foams Polymer
Miles I S and Zurek A 1988 Preparation, structure, and properties of
two-phase co-continuous polymer blends Polym. Eng. Sci.
Mooney D J, Baldwin D F, Suh N P, Vacanti J P and Langer R
1996a Novel approach to fabricate porous sponges of
poly(D,L-lactic-co-glycolic acid) without the use of organic
solvents Biomaterials 17 1417–22
Mooney D J, Mazzoni C L, Breuer C, McNamara K, Hern D
and Vacanti J P 1996b Stabilized poly glycolic acid fibre-based
tubes for tissue engineering Biomaterials 17 115–24
Nam Y S and Park T G 1999 Biodegradable polymeric
microcellular foams by modified thermally induced phase
separation method Biomaterials 20 1783–90
Nam Y S, Yoon J J and Park T G 2000 A novel fabrication method
of macroporous biodegradable polymer scaffolds using gas
foaming and salt as a porogen additive J. Biomed. Mater. Res.,
Appl. Biomater. 53 1–7
Park A, Wu B and Griffith L G 1998 Integration of surface
modification and 3D fabrication techniques to prepare
patterned poly(L-lactide) substrates allowing regionally
selective cell adhesion J. Biomater. Sci., Polym. Ed. 9 89–110
Park T G 1999 New approaches to fabricate highly porous tissue
scaffolds 4th Asia-Pacific Conf. on Medical and Biological
Engineering (Seoul, Korea)
Potschke P, Bhattacharyya A R and Janke A 2004 Carbon
nanotube-filled polycarbonate composites produced by melt
mixing and their use in blends with polyethylene Carbon
Potschke P, Kretzschmar B and Janke A 2007 Use of carbon
nanotube filled polycarbonate in blends with montmorillonite
filled polypropylene Compos. Sci. Technol. 67 855–60
Potschke P and Paul DR 2003 Formation of co-continuous
structures in melt-mixed immiscible polymer blends
J. Macromol. Sci., Polym. Rev. C 43 87–141
Schugens C, Maquet V, Grandfils C, Jerome R and Teyssie P 1996
Polylactide macroporous biodegradable implants for cell
transplantation: II. Preparation of polylactide foams for
liquid–liquid phase separation J. Biomed. Mater. Res.
Shastri VP, Martin I and Langer R 2000 Macroporous polymer
foams by hydrocarbon templating Proc. Natl Acad. Sci. USA
Singh L, Kumar V and Ratner B D 2004 Generation of porous
microcellular 85/15 poly(D,L-lactide-co glycolide) foams for
biomedical applications Biomaterials 25 2611–7
Sparacio D and Beckman E J 1997 Generation of microcellular
biodegradable polymers in supercritical carbon dioxide Polym.
Preprints Div. Polym. Chem. Am. Chem. Soc. 38 420–3
Taboas J M, Maddox R D, Krebsbach P H and Hollister S J 2003
Indirect solid free form fabrication of local and global porous,
biomimetic and composite 3D polymer-ceramic scaffolds
Biomaterials 24 818–94
Biofabrication 2 (2010) 035006W Zhang et al Download full-text
Thomson R C, Yaszemski M J, Powers J M and Mikos A G 1998
Hydroxyapatite fiber reinforced poly(a-hydroxy ester) foams
for bone regeneration Biomaterials 19 1935–43
Virgilio N, Sarazin P and Favis B D 2010 Towards ultrapoorous
poly(L-lactide) scaffolds from quaternary immiscible polymer
blends Biomaterials 31 5719–28
Wang X X, Li W and Kumar V 2006 A method for solvent-free
fabrication of porous polymer using solid-state foaming and
ultrasound for tissue engineering applications Biomaterials
Whang K, Thomas C H, Healy K E and Nuber G 1995 A novel
method to fabricate bioabsorbable scaffolds Polymer
Yao D, Zhang W and Zhou J G 2009 Controllable growth of
gradient porous structures Biomacromolecules 10 1282–6
Yuan Z H and Favis B D 2004 Macroporous poly(L-lactide) of
controlled pore size derived from the annealing of
co-continuous polystyrene/poly(L-lactide) blends Biomaterials
Yuan Z H and Favis B D 2005 Coarsening of immiscible
co-continuous blends during quiescent annealing AIChE J.
Zhang W, Deodhar S and Yao D 2010 Geometrical confining
effects in compression molding of co-continuous polymer
blends for scaffold applications Ann. Biomed. Eng.
Zhao J F and Huang H X 2008 Migration of nanoclay in PP/PS
blend and effect of its localization on cell structure Proc.
ASME Int. Mechanical Engineering Congress and Exposition
(IMECE 2008) 15 105–9