1
TISSUE ENGINEERING
Volume 8, Number 1, 2002
© Mary Ann Liebert, Inc.
Review
The Design of Scaffolds for Use in Tissue Engineering.
Part II. Rapid Prototyping Techniques
SHOUFENG YANG, Ph.D., KAH-FAI LEONG, M.S.E., M.S.M.E., ZHAOHUI DU, Ph.D.,
and CHEE-KAI CHUA, Ph.D.
ABSTRACT
Tissue engineering (TE) is an important emerging area in biomedical engineering for cre-
ating biological alternatives for harvested tissues, implants, and prostheses. In TE, a highly
porous artificial extracellular matrix or scaffold is required to accommodate mammalian
cells and guide their growth and tissue regeneration in three-dimension (3D). However, ex-
isting 3D scaffolds for TE proved less than ideal for actual applications because they lack
mechanical strength, interconnected channels, and controlled porosity or pores distribution.
In this paper, the authors review the application and advancement of rapid prototyping (RP)
techniques in the design and creation of synthetic scaffolds for use in TE. We also review
the advantages and benefits, and limitations and shortcomings of current RP techniques as
well as the future direction of RP development in TE scaffold fabrication.
INTRODUCTION
A
S DISCUSSED in Part I (
Tissue Engineering
7, 679–689, 2001) of this two-part paper, approximately one-
quarter of patients in need of organ transplants in the United States die while waiting for a suitable
donor.
1,2
The current demands for transplant organs and tissues far outpace supply, and all manner of pro-
jections indicate that this gap will continue to widen.
1,3
Cell transplantation is recently proposed as an al-
ternative treatment to whole organ transplantation for failing or malfunctioning organs.
4–6
For the creation
of an autologous implant, donor tissue is harvested and dissociated into individual cells, and the cells are
attached and cultured onto a proper substrate that is ultimately implanted back at the desired site of the
functioning tissue. Because many isolated cell populations can be expanded
in vitro
using cell culture tech-
niques, only a very small number of donor cells may be needed to prepare such implants. However, it is
believed that isolated cells cannot form new tissues by themselves. Most primary organ cells are believed
to be anchorage-dependent and require specific environments that very often include the presence of a sup-
Design Research Center, School of Mechanical and Production Engineering, Nanyang Technological University, Sin-
gapore.
porting material to act as a template for growth. The success of any cell transplantation therapy relies on
the development of suitable substrates for both
in vitro
and
in vivo
tissue culture. Existing substrates, mainly
in the form of tissue engineering scaffold, are considered to be less than ideal for applications because they
lack mechanical strength, interconnected channels, and controlled porosity or pores distribution.
Rapid prototyping (RP), also termed “solid freeform fabrication (SFF),” is a recent technology based on
the advanced development of computer and manufacturing. The main advantage of these techniques is their
ability to produce complex products rapidly directly from a computer model. RP has been used in the med-
ical field primarily as a means of guiding surgical procedures using tactile models derived from patient
computerized tomography (CT) data.
7
These models have also been used to cast custom titanium orbital
implants. Direct fabrication of custom implants is promising in offering simpler and more rapid surgical
implementations. The potential to intimately control the microstructure of porous channels and the overall
macroscopic shape of the implants makes RP an ideal process for fabricating implant and tissue engineer-
ing scaffold as well. In this paper, we review the applications and advancement of RP techniques in the de-
sign and creation of synthetic scaffolds for use in tissue engineering (TE). Part I of this paper
8
analyzed
the factors necessary to enhance the design and creation of scaffolds for use in TE in terms of materials,
structure, and mechanical properties. Based on these requirements, the authors will now further discuss the
advantages and benefits, and limitations and shortcomings of current RP techniques, as well as the future
direction of application of RP in fabricating TE scaffolds.
EXISTING APPLICATIONS OF RAPID PROTOTYPING IN TISSUE ENGINEERING
Traditional fabrication methods cannot build parts with predefined or controlled microstructure as well
as macrostructure. RP, or SFF, has the distinct advantage of being able to build objects with predefined
macrostructures as well as microstructures. This distinct advantage makes RP a technique with excellent
potential for fabricating scaffold with controlled hierarchical structures for use in TE. The macroscopic
shape of the scaffold, on a scale of up to tens of millimeters, will determine the external appearance and
structure of the final product. For example, it may be desirable to reconstruct an ear or a jaw exactly to
meet patient contours based on images acquired using magnetic resonance imaging (MRI) or CT scans di-
rectly from the patient. The size, orientation, and surface chemistry of pores and channels, on a scale of
hundreds of microns, have a direct impact on the extent and nature of tissue in-growth. On the scale of tens
of microns, local surface texture and porosity become important.
There are also two particular characteristics of RP systems that limits their performance. One is the over-
all resolution of the processes achievable in current systems and the second is the ranging materials from
which prototypes are made. Several different RP processes are described in the following sections. Their
typical features, characteristics and processing limitations with regard to TE are highlighted.
Sheet lamination
Sheet lamination fabrication
7,9
such as the patented laminated object manufacturing (LOM) process builds
three-dimensional (3D) cross-sections out of a roll of sheets lined with thermoplastic adhesive. Layered
cross-sectional profiles are cut with a CO
2
laser and by bonding each cross-sectional layer to the previous
one by applying heat and pressure. The prototype is sequentially made layer by layer, and the area outside
the layer outline and inside any internal closed areas that are not part of the profile are cut into small sec-
tions called “tiles.” These tiles are removed in a postprocessing phase after the completion of all the lay-
ers that comprise the object. Small features, in particular small inner holes, cannot be fabricated by this
method.
Steidle et al.
10,11
have reported the use of a nonresorbable bioceramic composite system consisting of
hydroxyapatite particles bonded together by a calcium phosphate glass phase to build a biocompatible bone
for implant. The implant built is almost completely dense, which renders it unsuitable for use in its intended
application in bone tissue engineering. Although this kind of process has an advantage in its ability of recre-
ating the external shape of the scaffold, it unfortunately still suffers the limitation of a lack of microstruc-
ture control, because the microstructure of each sheet is uniform.
YANG ET AL.
2
Adhesion bonding
In this method, 3D parts are created by a layered printing process with adhesive bonding, using powder
as the base material, according to sliced cross-sectional computer-assisted design (CAD) data of the object.
Each layer of powder is selectively joined where the part is to be formed by ink-jet printing of a binder
material. The process is repeated layer by layer until the part is complete. An example of such technology
is three-dimensional printing (3DP).
Cima et al.
12,13
demonstrated using 3DP to build drug delivery devices and tissue regeneration devices.
Although only polyethylene oxide (PEO) and polycaprolactone (PCL) powders were used in those experi-
ments, theoretically and virtually any materials that can be processed into powdered form can be used for
3D printing.
13
TheriForm™,
14
one of six licensees of 3DP, developed pharmaceutical dosage forms and
medical products, including drug delivery and TE, based on 3DP fabrication process. Cima
15
produced 200-
m
m lines by printing polymer solutions rather than using pure solvent, while the later will produce a line
primitive of 500
m
m in width.
In this method, adapting a appropriate binder for the different materials used can pose problems for bio-
medical applications. Some organic solvents that are in use now as binders, such as chloroform and meth-
ylene chloride, are harmful to the human body and are difficult to remove completely. Even after 1 week,
during which samples of implants are placed in a vacuum to remove excess solvent, the amount of chlo-
roform managed only to be reduced from 10wt% to 0.5wt%.
16
Another difficulty is that excess powder that
is trapped in small channels is difficult to remove. Also, channels of less than 1 mm in diameter have not
been successfully built.
Laser sintering
In laser sintering methods, parts are built by sintering of powder on a powder bed, when an infrared laser
beam hits a thin layer of powdered material such as wax, polycarbonate, nylon, or even metal. The inter-
action of the laser beam with the powder raises the local surface temperature to the glass transition tem-
perature of the powder. This is just below the melting temperature. It results in particle bonding—fusing
the particles onto each other and to the previous layer to form a solid. Limited by the power of the laser
and thermal diffusion, the glass transition temperature and melting point of the powder cannot be too high.
For ceramic powder, the polymer–ceramic mixture is always used, in which case the polymer is used as a
low melting point binder.
Lee et al.
17,18
coated calcium phosphate powder with polymer by spray drying slurry of particulate and
emulsion binder. The coated powder was sintered by selective laser sintering (SLS) to fabricate artificial
calcium phosphate bone implant. Postprocessing included infiltrating the sintered part with calcium phos-
phate solution or phosphoric acid–based inorganic cement. This was used to improve the density of green
part to prevented collapse in the subsequently polymer binder burn out stage. However, the compressive
strengths
19
of the infiltrated sintered part without macropores was only 36.0
6
7.5 MPa. Implant model
with macropores of about 2 mm in diameter was built for preclinical trials to assess the biocompatibility.
It is believed that macropores in the implant should be smaller to increase the surface area for cell attach-
ment and to improve the mechanical strength for load-bearing requirements.
The laser beam diameter of the Sinter Station 2500 (DTM, USA) is about 400
m
m (0.016 inches). Due
to the Gaussian distribution of the laser energy and the nature of powder bonding, it becomes relatively dif-
ficult to form sharp corners and clear boundaries. The conduction and diffusion of laser heat cause neigh-
boring powder of scan vector unwanted bonding, which can sometimes be serious (Fig. 1). These make it
impossible to build small features of less than 400
m
m in size. The fuzzy boundary and the bonding pow-
der induce a coarse inner surface on the macropores, and this makes it difficult to clean out the trapped un-
sintered powder.
Improvements on SLS process are expected to produce the desired scaffold for TE. These include ac-
quiring the ability to create smaller features by using a smaller laser spot size, powder size, and thinner
layer thickness. New ways have to be found to remove trapped loose powder. Potential solutions include
using ultrasonic vibration, compressive air, bead blaster, and/or appropriate solvent. In addition, better sin-
DESIGN OF SCAFFOLDS. PART II. RAPID PROTOTYPING
3
tering environment is necessary so that the humidity, inertness of surrounding atmosphere, and vacuum can
be controlled within a smaller working chamber.
Photopolymerization
The principle of this method is based on the polymerization of photopolymer resins that is initiated by
radiant energy from electromagnetic radiation.
7
Photopolymer resins are mixtures of simple low-molecu-
lar-weight monomers capable of chain-reacting to form solid long-chain polymers when activated by radi-
ant energy within specific wavelength range. There are two basic types of liquid-based commercial RP ma-
chines that use photopolymerization. One uses a laser, while the other uses a masked lamp to cure the
photopolymers.
7,9
In the first system, a deflected laser beam is used to irradiate a thin polymer layer at the
surface of a vat filled with liquid photopolymer resin. The irradiated areas of photopolymer react chemi-
cally to become solid. For example, 3D system’s stereolithography apparatus (SLA) uses an ultraviolet laser
to solidify an epoxy resin. The second system uses masked illumination, instead of a point-by-point method
used by laser system. It irradiates a complete layer of polymer every single time. An example of this masked
lamp technique is Cubital’s solid ground curing (SGC). Rapid micro product development (RMPD),
20
a mi-
crostereolithography method developed by MicroTEC, and mentioned by Chua et al.,
9
is a promising RP
approach. This technology, based on combination of masked lamp technology and laser curing photopoly-
merization, emerged in the middle of 1999.
Langton et al.
21
described a method for the development of a user-defined structural model simulating
cancellous bone of the human calcaneus using SLA. The potential for SLA-produced samples to be used
as a structurally controlled cancellous bone mimic was investigated by producing a 3D rod lattice of 3 mm
center-to-center distance and rod diameters of 1 mm (70% porosity) representing healthy bone and 0.4 mm
(95% porosity) representing osteoporotic bone.
Chu et al.
22
built hydroxyapatite prototypes for bone tissue scaffolds from Image-Based Design files, fea-
turing an interior architecture of void passages. Direct ceramic SLA
22
is done using UV-curable suspen-
sions of ceramic powders in acrylates in a conventional SLA machine. Viscosity control for these highly
concentrated suspensions and cure depth behavior are the main issues for fabricating a ceramic part with
the stereolithography techniques.
Molecular Geodesics, Inc. (MGI, Boston, MA)
23,24
developed a new class of biomimitic materials that
mimic the microstructural organization, mechanical responsiveness, and biocatalytic activities of living cells
and tissues. MGI’s strategy is to study the underlying structures that provide living cells and tissues with
their strength, flexibility, and porosity, and develop ways to apply these structures to synthetic products. A
small-spot laser stereolithography system from 3D System is used. It has been demonstrated to be able to
fabricate features as small as 70
m
m, but the ideal feature required is less than half that size (Figs. 2–4).
The common laser spot size used in SLA is about 250
m
m in diameter. This is a result of a compromise
between the demands of achieving precision and attaining maximum speed. However, new research has
been focused on a small-spot SLA. Representative of such efforts is the RMPD, a new manufacturing tech-
nology in micro-engineering developed by microTEC.
20
Similar to SLA, a controlled laser beam moves
YANG ET AL.
4
FIG. 1. Serious bonding of neighboring powder by laser heat diffusion in SLS (W, wall of channel, sintered part; C,
channel, unsintered part). The arrows indicate the serious bonding of neighboring powder. Laser power: 5 W. Scan
speed: 200 m/sec.
across a specific area structures point by point and, in this manner, hardens the liquid photopolymer by pho-
topolymerization. Typical materials used are acrylics and epoxies, but it is expected that the range of ma-
terials will be extended to include metals, ceramics, and other suitable composites. Components are built
up in steps of less than 1
m
m thick and a resolution finer than 10
m
m. Current technologies limit the prod-
uct size to a maximum size of 35 mm
3
.
Droplet deposition
In droplet deposition,
25–27
molten droplets deposited on the working area will soften the material of the
previous layer and then solidify, joining the droplets to the previous layer. Unfilled areas may be filled with
a soluble molten wax to the same thickness to act as the support material for next layer. When all layers
have been deposited, the object is removed from the platform and support materials are removed by dis-
solving them using appropriate techniques. Examples of machines using such method include Fraunhofer’s
multi-phase jet solidification (MJS), Stratasys’s fused deposition modeling (FDM), 3 D System’s multi-jet
modeling (MJM), and Sanders’s model maker.
7
3D honeycomb porous alumina ceramic structures for bone implants have been fabricated using indirect
route where a polymeric mold is first created using FDM.
27
The mold was then infiltrated with ceramic
slurry, dried, and subjected to a binder burn-out and sintering cycle. 3D honeycomb structures with 33%
DESIGN OF SCAFFOLDS. PART II. RAPID PROTOTYPING
5
FIG. 2. MGI is using stereolithography to build concept-verification models of its tensegrity structures.
24
FIG. 3. The struts of the MGI biomimetic scaffold, which was built using a small-spot laser stereolithography sys-
tem, measure less than 70 mm in diameter.
24
