You are on page 1of 11


Biomaterials 25 (2004) 19912001

Surface modication of ultra thin poly (e-caprolactone) lms using

acrylic acid and collagen
Ziyuan Cheng, Swee-Hin Teoh*
Department of Mechanical Engineering, Centre for Biomedical Materials Applications and Technology (BIOMAT), Division of Bioengineering,
National University of Singapore, Kent Ridge, Singapore 119260, Singapore
Received 20 March 2003; accepted 11 August 2003


Poly (e-caprolactone) (PCL) has been used as a bioresorbable polymer in numerous medical devices as well as for tissue
engineering applications. Its main advantage is its biocompatibility and slow degradation rate. PCL surface, however, is
hydrophobic and cell-biomaterial interaction is not the best. We attempt for the rst time to modify an ultra thin PCL surface with
collagen. The PCL lm was prepared using solvent casting and biaxial stretching technique developed in our laboratory. This biaxial
stretching produced an ultra thin PCL 37 mm thick, ideal for membrane tissue engineering applications. The PCL lm was
pretreated using Argon plasma, and then UV polymerized with acrylic acid (AAc). Collagen immobilization was then carried out.
The modied lm surface was characterized by Fourier Transform Infrared (FT-IR) and X-ray Photoelectron Spectroscopy (XPS).
Water contact angles were also measured to evaluate the hydrophilicity of the modied surface. Results showed that the
hydrophilicity of the surface has improved signicantly after surface modication. The water contact angle dropped from 66 to 32 .
Atomic Force Microscopy (AFM) showed an increase in roughness of the lm. A change from 46 to 60 nm in the surface
morphology was also observed. The effect of cells attachment on the PCL lm was studied. Human dermal broblasts and
myoblasts attachment and proliferation were improved remarkably on the modied surface. The lms showed excellent cell
attachment and proliferation rate.
r 2003 Elsevier Ltd. All rights reserved.

Keywords: Poly (e-caprolactone); Surface modication; Human dermal broblasts; Myoblasts; Collagen immobilization

1. Introduction ability and cell adhesion [6]. In general, however, they

do not have good mechanical strength. Synthetic
Development of biodegradable polymer has been polymers are commonly employed in tissue engineering
considered to be great interest in biomedical applica- because of their remarkable mechanical ability to be
tions, especially tissue engineering and controlled drug shaped easily [7,8]. But their surfaces are not favorable
delivery [1,2]. It is known that the surface properties of a to cells adhesion. Many researchers have focused on
biomaterial dominate the interactions between the understanding the relationship between polymer sur-
material and the biological environment. In the eld of faces and cell afnity. Hydrophilic and protein contain-
tissue engineering, polymeric materials provide surfaces ing surfaces are known to be good for cell growth [911].
for the immobilization of biologically active molecules Synthetic polymers often require selective modication
and living cells [3,4]. Many natural materials, such as to promote hydrophilicity, cell adhesion and biocom-
bronectin, laminnin or collagen, are used as biomater- patibility via introducing special functional groups onto
ials in tissue replacement and wound healing [5]. These the polymer surface [12,13].
materials show excellent biocompatibility, biodegrad- Surface modication is a widely adopted method
because it can enhance the biocompatibility of a
material surface while keeping the bulk properties intact
*Corresponding author. Laboratory for Biomedical Engineering
(LBME), Division Bioengineering, 9 Engineering Drive 1, Singapore
[14,15]. A common approach to polymer surface
117576, Singapore. modication is coating of materials having appropriate
E-mail address: (S.-H. Teoh). biological properties via physical adsorption. However,

0142-9612/$ - see front matter r 2003 Elsevier Ltd. All rights reserved.
1992 Z. Cheng, S.-H. Teoh / Biomaterials 25 (2004) 19912001

the physical adsorped layer is easily removed, 2. Materials and methods

especially by changes in pH of the solution and by
high shear forces. Although a variety of technologies 2.1. Ultra-thin films preparation
have been used for improving surface characteristics,
grafting is a widely applied technology in surface PCL pellets (Mn =80 000, Sigma-Aldrich Company)
modication of polymers. Graft polymerization is an were dissolved in methylene chloride (6% w/w) and cast
attractive way in which a desired monomer can be over glass sheets. The solvent was removed by slow
grafted onto the surface with covalent bond, which evaporation. Films with initial thickness of 100110 mm
renders good stability [16]. One of the simplest were obtained and were further dried in vacuum at 51 C
approaches for grafting is to introduce functional for 24 h. Dried PCL lms were pressed (Cavar Heat
groups to the surface for chemical reaction between Press) with a platen temperature of 55 C to remove
the grafted layer and the substrate. The carboxyl surface defects. All pressed lms were cut to 6  6 cm2
group is one of the best functional groups. Chemical for biaxial drawing. Films were preheated for 30 min at
bonds are produced between amino-groups from 5371 C before drawing. Heated PCL lms were
proteins and carboxyl functional groups. Acrylic acid biaxially drawn 2.5 times its original size to a nal
(AAc) is commonly used in such grafting process. AAc dimension of 15  15 cm2. Ultra-thin PCL lms (37 mm)
acts as a spacer to bind proteins and the substrate. The were obtained. Details of the biaxial stretching have
carboncarbon double bond in the structure of AAc been described previous [25]. The lms were then cut
binds easily with polymers. In addition, the base into rectangles of 1.8 mm  3.0 mm and carefully washed
material can retain their own chemical and physical with ethanol, and dried, prior to surface modication.
properties after graft polymerization. All of these
advantages make AAc a popular chemical in graft 2.2. Argon plasma pretreatment of the PCL film surface
polymerization [1719].
Surface modication of polymers is normally In order to produce the desired active radicals on the
achieved by plasma of gases such as oxygen, nitrogen, material surface, plasma pretreatment was carried out.
inert gases and reactive gases. For example, the blood In this study, argon was used in the plasma pretreat-
compatibility of silk broin lm was improved after ment. This avoids the production of unwanted active
treatment with SO2 gas plasma [20]. Monomers with radicals. Argon plasma pretreatment was carried out in
desirable functional groups have also been polymerized a quartz cylindrical-type glow discharge cell (Model SP
onto the polymer surface [21]. However, these polymers 100, manufactured by Anatech Ltd. of USA). The
have been merely deposited on the surface and could be plasma power applied was kept at 37 W at a radio
easily removed by physical extraction. Another common frequency of 40 kHz. The PCL lm was placed between
approach for graft polymerization on the surface is via the two parallel plate electrodes and subjected to the
free radicals peroxides generated by plasma or ozone glow discharge for 10 s (Fig. 1) at an Argon pressure of
pretreatment [18,19,22]. Since plasma surface treatment
causes changes to a limited depth (several molecular
layers), bulk properties of even the most delicate
Ar Plasma
materials remain unchanged. Treated polymer can be In air O

surface modied through graft polymerization under O

relatively mild conditions. PCL Film
Poly (e-caprolactone) (PCL), an aliphatic polyester
which is bioresorbable and biocompatible, is a ideal
Acrylic acid (AAc) O-(CH2CH)n WSC, 1h
material for pharmaceutical products and wound dres-
sings [29]. PCL is a semicrystalline polymer having a UV, Room temperature 4 C
melting temperature of 60 C and a glass transition
temperature of 60 C. However PCL has poor hydro-
philicity, which results in poor cell attachment and PCL-AAc
proliferation rate. Modications of PCL by crosslink-
ing, or by copolymerization with different monomers O-(CH2CH)n O-(CH2CH)n
and proteins have been reported [23,24]. C=O 0.5mg/ml Collagen
In this study, we report the surface modication of an O 4 C, 5h
ultra-thin PCL lm produced by biaxially stretching C=N-R NH
with collagen. In order to evaluate the effect of grafting NH-R Collagen
of collagen on the surface of the PCL lm, the response
of human dermal broblasts and myoblasts to the Fig. 1. Schematic representation of the process of UV introduced AAc
material surface was also studied. grafting and collagen immobilization.
Z. Cheng, S.-H. Teoh / Biomaterials 25 (2004) 19912001 1993

0.5 Torr. The Argon plasma-pretreated PCL surfaces 2.5.2. XPS measurements
were then exposed to the atmosphere for about The collagen immobilized on the PCL lm surface
10 min to effect the formation of surface peroxides was analyzed by X-ray photoelectron spectroscopy
and hydroperoxides, which were used for the sub- (XPS). The XPS N 1 s core-level signal was used as a
sequent UV-induced graft polymerization experiment marker for the analysis of the relative amount of
[26]. collagen immobilized on the surface.
XPS measurements were made on a VG ESCALAB
2.3. AAc graft polymerization of the PCL film surface MkII spectrometer with a Mg Ka X-ray source
(1253.6 eV photons) at a constant retard ratio of 40. A
For the UV-induced graft polymerization with AAc, survey scan spectrum was taken and surface elemental
the concentration of AAc aqueous solution was varied stoichiometries were determined from peak-area rations.
from 0.03 to 0.11 g/ml. Vitamin B2 (0.0152 mol/l) was
added into each solution, at ratio of 1:20, to reduce the 2.5.3. Water contact angle measurements
dissolved oxygen, which could inhibit the subsequent Water contact angles of the pristine and modied
radical-initiated polymerization. The PCL lm with PCL lms were measured at room temperature and 60%
AAc solution was clipped between two quartzes, and relative humidity, using sessile drop method on a
then put into a Pyrexs tube. The reaction set was telescopic goniometer (Rame-Hart model 100-00(230))
exposed to UV illumination for 30 min, in a Riko [28]. More than ve measurements were carried out and
Rotory, Model RH 40010 W, photochemical reactor the resulting values were averaged.
manufactured by Riko Denki Kogyo of Chiba, Japan.
The reactor was equipped with a 1000 W high-pressure 2.5.4. AFM measurements
Hg lamp and a constant temperature water bath at room The surface topography of unmodied PCL, PCL
temperature of 2771 C. After graft polymerization, the AAc and PCLCol lms was examined in a Dimension
lm was removed from the reaction set and washed with 3100 Scanning Probe Microscope (SPM) (Digital Instru-
owing doubly distilled water, until the residual homo- ments, Veeco Metrology Group) in air. Atomic force
polymer on the lm surface had been removed microscopy (AFM) images were obtained by scanning
absolutely. The lm grafted AAc (PCLAAc) was dried surface in a contact mode (scan size 10  10 mm2, scan
in a vacuum desiccator. rate 0.951.00 Hz) using a silicon nitride probe (model
DNP) [29]. The spring constant was 0.06 N/m. An
arithmetic mean of the surface average roughness (Ra )
2.4. Immobilization of collagen on the PCL film was evaluated directly from the AFM images.

PCLAAc lms were immersed in the phosphate- 2.6. Cell culture

buffered saline (PBS) containing 5 mg/ml of water-
soluble carbodiimide (WSC) [27]. 1-Ethyl-3-(3-dimethy- 2.6.1. Human dermal fibroblasts seeding and culture
laminopropyl) carbodiimide hydrochloride was used in Human dermal broblasts were derived from explant
this experiment. The lm was doused in this solution for cultures of human ear skin samples obtained with
1 h at 4 C to preactivate carboxyls of AAc grafted on patient consent from a local hospital (National Uni-
the PCL lm surface. Then lms were then treated with versity Hospital). The second passage was used in this
Collagen I (calf skin, Sigma)-PBS solution at a experiment. 10% fetal bovine serum (FBS; Hyclone,
concentration of 0.5 mg/ml for 5 h at the same tempera- Logan, UT) and 1% penicillinstreptomycin solution
ture. After immobilization, immersing and washing (Sigma, St. Louis, MO) were supplied into the Dulbec-
the lms in PBS for 1 h at room temperature removed cos Modied Eagle Medium (DMEM; Gibco, Grand
collagen, which was adsorbed physically on the surface. Island, NY) for cell culture. Human dermal broblasts
PCL collagen immobilized (PCLCol) lms were were seeded on PCL and PCLCol lms with 50,000
then dried under reduced pressure and stored in a cells each sample (18 mm  20 mm), and placed in a self-
refrigerator. sterilizable incubator (WTB Binder, Tuttlingen, Ger-
many) at 37 C in 5% CO2, 95% air, and 99% relative
2.5. Surface characterization humidity, with medium change every 2 days.

2.5.1. FTIR measurements 2.6.2. Human myoblasts seeding and culture

Infrared absorption spectra of the PCL, PCLAAc Human myoblasts were puried from rectus femoris
and PCLCol lms were obtained from a Bio-Rad FTS biopsies taken from human donors after their consent.
135 FT-IR spectrophotometer. For each spectrum Super medium (Cell Transplant Singapore) was
obtained, a total of 16 scans were accumulated at a supplemented with 1% penicillin-streptomycin solution
resolution of 8 wavenumbers. (Sigma, St. Louis, MO) in the experiment. The second
1994 Z. Cheng, S.-H. Teoh / Biomaterials 25 (2004) 19912001

passage of human myoblasts were seeded on PCLCol

and PCL lms with density of 10,000 cells each sample (a)
(18 mm  20 mm). Samples seeded with cells were placed
in a self-sterilizable incubator (WTB Binder, Tuttlinger
germany) at 37 C in 5% CO2, 95% air, and 99%
relative humidity, with medium change every 2 days. 1730 cm-1
2.6.3. Phase contrast light microscopy (PCLM)
PCLM of each group lms was taken by an inverted
modulation contrast light microcopy (Olympus IX70)

using original magnication  100 (Human Modulation
Contrast). Films seeded with human dermal broblasts
were viewed on day 1, day 4 and day 8. And lms seeded
with human myoblasts were viewed on day 1, day 3 and
day 6.

2.6.4. Confocal laser scanning microcopy (CLSM)

CLSM of each group was carried out. The samples
seeded with cells were rinsed by PBS to remove
nonattached cells. The viable cells were stained green 1566 cm-1

with uorescein diacetate (FDA; Molecular Probes,

Inc., OR) at concentration of 2 mg/ml in PBS. Samples
2600 2400 2200 2000 1800 1600 1400 1200 1000 800
were incubated at 37 C with FDA for 15 min. The -1
samples were washed for twice in PBS. To stain dead Wavenumber (cm )
cells red, each sample was then placed in 0.1 mg/ml Fig. 2. FT-IR spectra of the (a) pristine PCL lm, (b) PCLAAc lm
propidium iodide (PI; Molecular Probes, Eugene, OR) and (c) PCLCol lm.
for 2 min at room temperature. Residuary PI was
removed by rinsing samples for twice in PBS before
viewing the sample under confocal laser microscope 3.2. XPS spectra
(Olympus IX70-HLSH100 Fluoview) [29].
The XPS wide scan spectrum of the pristine PCL
surface is showed in Fig. 3a. It reveals that carbon and
3. Results oxygen signals are present. The spectrum of PCLAAc
lm surface (Fig. 3b) also shows the same peaks as
Collagen immobilization onto the PCL lm surface pristine PCL surface corresponding to C 1 s and O 1 s.
was achieved by four steps (Fig. 1): First, the PCL lm However, the relative intensity of carbon and oxygen
was pretreated by argon plasma and exposed in air. This peaks varied after grafting with acrylic acid. In contrast,
engendered peroxides on the surface. Second, graft a new peak corresponding N 1 s appeared on the
polymerization of AAc was carried out under the effect spectrum of PCLCol lm surface (Fig. 3c).
of polymerization initiator (peroxides) and UV energy. Fig. 4 shows the respective C 1 s core-level spectra of
Third, WSC reactivated the carboxyl by replacing the the pristine PCL (a), PCLAAc (b) and PCLCol (c)
H of carboxyl groups. Lastly, a chemical bond was lm surfaces. In the case of pristine PCL and PCLAAc
produced between collagen and AAc because WSC was lm surface, the C 1 s core level spectrum contains three
replaced by the collagen. major peak components, with binding energy at
284.6 eV for the C2H species, at 286.2 eV for the
3.1. FT-IR spectra C2O species and at% 288.6 eV for the O2C O species.
After collagen immobilization onto the lm% surface, the
The structure of the modied lm surface was studied C 1 s core level can be curve-tted with ve peak
by FT-IR. Fig. 2 shows the respective FT-IR spectra of components, with binding energy at 284.6 eV for the
the pristine PCL lm (a), PCLAAc (b) lm and PCL C2H species, at 286.2 eV for the C2O species, at
Col lm (c). In the PCL spectrum, a major absorption %
288.6 eV for the O2C O species, at 285.8 eV for the
peak appeared at 1730 cm1 according to the functional %
C2N species and at 287.4 eV for the O2C2N species.
group, COO, in PCL [15]. The same peak was also % The graft concentration on the surface% is expressed
found in the spectrum of PCLAAc and PCLCol lm. simply as the XPS-derived atomic percent ratio of
For the PCLCol spectrum, however, a new absorption nitrogen to carbon. The atomic percent of surfaces was
peak is observed at 1566 cm1. calculated based on the XPS scan spectrum. Fig. 5
Z. Cheng, S.-H. Teoh / Biomaterials 25 (2004) 19912001 1995




Intensity (Arb. Units)





Fig. 3. XPS wide scan spectra of (a) pristine PCL lm surface, 290 288 286 284 282
(b) PCLAAc lm surface and (c) PCLCol lm surface.
Binding Energy (eV)
Fig. 4. XPS C 1 s core-level spectra of (a) pristine PCL lm surface,
(b) PCLAAc lm surface and (c) PCLCol lm surface.

shows the change of concentration of AAc grafted on 0.55

PCL lm surface, represented by the [O]/[C], as the
change of concentration of AAc solution used in 0.50
Graft Conc., [O]/[C]

grafting process. The amount of AAc grafted on PCL

lm surface increases as the concentration of AAc
solution increases at the range of 39%.
Fig. 6 shows the dependence of graft collagen
concentration on the surface, expressed as the [N]/[C] 0.40
ratio, on the concentration of AAc solution used for
grafting acrylic acid onto the PCL lm surface. The 0.35
concentration of immobilized collagen on the surface
increases as the concentration of AAc solution increases. 0.30
2 4 6 8 10 12
3.3. Water contact angle of the modified surface Concentration of AAc Solution (%)
Fig. 5. Effect of the concentration of AAc monomer solution to the
To compare the hydrophilicity of the modied and graft concentration on the PCLAAc lm surface.
unmodied lm surfaces, water contact angles were
measured. The water contact angle of pristine PCL lm
was 66 . With the introduction of AAc onto the lm 3.4. AFM images
surface, the water contact angles dropped to 32 at 9%
(Fig. 7). The water contact angle decreased dramatically In order to further evaluate the effect of grafting AAc
as the concentration of AAc solution increased. After and collagen on the surface structure of PCL lm, AFM
collagen immobilization, the water contact angle of the was used to examine the surface structure of unmodied
lm initially dropped to 45 at 9% AAc. and modied PCL lms. Fig. 8 shows the AFM images
1996 Z. Cheng, S.-H. Teoh / Biomaterials 25 (2004) 19912001

Immobilization Conc., [N]/[C]





2 4 6 8 10 12
Concentration of AAc Solution (%)
Fig. 6. Effect of the concentration of monomer AAc solution to the
immobilization concentration on PCLCol lm surface.

Water Contact Angle ()








1 3 5 7 9
Concentration of AAc Solution (%)
Fig. 7. Effect of the concentration of monomer AAc solution to water
contact angles on the PCLAAc and PCLCol lm surfaces.

of unmodied PCL lm (a), PCLAAc lm (b) and

PCLCol lm (c). Fig. 8a shows dense brils on the
surface of unmodied PCL lm. Fibrils oriented mostly
in a uniaxial direction. After grafting acrylic acid onto
the surface, the PCL brils appeared different (Fig. 8b).
The surface of PCLCol lm (Fig. 8c) showed a
remarkable change in surface texture. The original
PCL brils are now absent and covered with collagen.
The Ra value was calculated from the roughness
prole determined by AFM. The Ra of unmodied Fig. 8. Atomic force microscopy images of (a) PCL, (b) PCLAAc and
PCL lm surface was 46 nm. After modication, the Ra (c) PCLCol lms (contact mode).
increased to 50 and 60 nm for PCLAAc and PCLCol,
scope. After seeding for 2 h, the cells began to attach
3.5. Cell culture onto the lm surface as the cell spread and attened. At
day 1, most of cells attached onto the PCLCol lms
Films seeded with human dermal broblasts were were elongated, exhibiting the spindle morphology that
observed under an inverted phase contrast light micro- is characteristic of dermal broblast cultures. On the
Z. Cheng, S.-H. Teoh / Biomaterials 25 (2004) 19912001 1997

Fig. 9. Human dermal broblasts culture, phase contrast light microscopies (PCLM) and confocal laser scanning microscopies (CLSM) of PCL and
PCLCol surfaces, taken at day 1, day 4 and day 8 (original magnication,  100).

PCL lm, many cells still kept round shapes although with human myoblasts. Cells spread and attened within
the PCL lm also showed good cell attachment. At day 1 day on both the PCLCol and PCL lms. At day 3,
4, round cells disappeared and there was an increase in there was an obvious increase in the cell density on the
cell densities across all samples. PCLCol lm surface. However, there is not much
At day 8, large quantities of cells were observed in change in cell number on the PCL lm surface compared
PCLM images of the PCLCol lm (Fig. 9). The whole to day 1. The PCL lm surface was fully covered by cells
PCLCol lm surface was covered with cells. On the at day 6. The cells appeared to align in a certain
PCL lm surface, the cell density was also increased. In direction. A monolayer of cells was formed on the
comparison with PCLCol lm, however, much fewer PCLCol lm surface. On the PCL lm surface, there
cells were observed on the PCLM image. was only a small increase in cell density. The CLSM
The viable and dead cells were observed under a images showed similar results as PCLM images. More
confocal laser microscope. From Fig. 9, the increase of cells were observed on PCLCol lm surfaces than PCL
cells with culture time was observed clearly on the PCL lm surfaces.
Col lm surface. Viable cells attached uniformly on the
whole PCLCol surface. However, not only was there
no increase of cell number on the unmodied PCL lm 4. Discussion
surface, but also only a few cells were observed.
Fig. 10 shows the phase contrast light microscopies, Plasma treatment technique is a unique and useful
confocal laser microscopies and confocal laser scanning method for the modication of polymeric materials
microscopies of PCL and PCLCol lm surfaces seeded without altering their bulk properties. Argon plasma
1998 Z. Cheng, S.-H. Teoh / Biomaterials 25 (2004) 19912001

Fig. 10. Human myoblasts culture, phase contrast light microscopies (PCLM) and confocal laser scanning microscopies (CLSM) of PCL and PCL
Col surfaces, taken at day 1, day 3 and day 6 (original magnication,  100).

produced peroxides on the surface. No undesired absorption peaks in the spectrum for PCLAAc lm
functional groups were induced because argon is an appears separately at the same position as that of PCL
inert gas (Fig. 1). Under the effect of UV energy, these lm. It can be explained by the chemical structure of
peroxides initiated AAc graft polymerization. The time PCL and AAc. AAc has the same stretching bonds as
of plasma treatment plays an important role in the PCL in chemical structure. This may explain why there
quantity of peroxides. Since PCL is a biodegradable is no new adsorption peak, which corresponds to the
polymer, a relative short treatment time was chosen. stretching bond at a certain position, appears on the
The concentration of AAc aqueous solution deter- spectrum of PCLAAc lm. However, after collagen
mined results of graft polymerization on the lm surface immobilization, the amino group of collagen is intro-
as same argon plasma pretreatment was carried out. duced onto the surface. CNH stretching bond
Similarly, the concentration of carboxyl functional adsorption peak appeared at 1566 cm1 in the FT-IR
groups in the surface plays a critical role in collagen spectrum of PCLCol lm. The FT-IR spectrum
immobilization. suggested that collagen has been successfully immobi-
lized onto the lm.
4.1. FT-IR spectra
4.2. XPS spectra
The adsorption peak at 1730 cm1 corresponds to the
carboxyl functional group of side chain AAc. But the Similarly, no new peak appears in the XPS spectrum
adsorption peak overlapped with that of PCL, which is of PCLAAc lm surface (Fig. 3b) compared with that
the backbone chain. Thus, as shown in Fig. 2, of pristine PCL lm surface (Fig. 3a). However, the
Z. Cheng, S.-H. Teoh / Biomaterials 25 (2004) 19912001 1999

relative intensity of carbon and oxygen peak changed. AAc and immobilizing with collagen. Fig. 7 shows that
AAc with higher oxygen content than PCL was the trend of water contact angles of PCLAAc lm
introduced onto the surface. As a result, the oxygen surfaces is the same as that of the relative concentrations
content on the surface increased. This is also indicated of AAc grafted onto lm surfaces. Carboxyl is a
by the ratio of atomic percent of oxygen to carbon. The hydrophilic functional group. Increasing the concentra-
more the amount of AAc is introduced onto the lm tion of AAc on the lm surface results in increase of
surface, the higher the ratio of atomic percent of oxygen carboxyl functional groups. This leads to a decrease of
to carbon is (Fig. 5). On the other hand, N 1 s peak was the water contact angle. After collagen immobilization,
found in the spectrum of PCLCol surface (Fig. 3c). a part of carboxyl functional groups were blocked with
This is due to the amino-group of collagen, which has collagen. The lm surface, as a result, becomes less
been immobilized on the surface. hydrophilic. The water contact angle decreased with
The C 1 s core-level spectrums of the pristine PCL and increasing concentration of AAc.
PCLAAc lm surface (Figs. 4a and b), show only a
major neutral carbon (C2H) component at binding 4.4. AFM measurements
energy of 284.6 eV and two minor components at
binding energy of 286.2 eV (C2O) and at binding Dense brils were produced in PCL lm (Fig. 8a)
energy 288.6 eV (O2C O), % respectively. For the during biaxial stretching process. This is has been
spectrum of PCLCol %lm surface (c), two new peaks reported previous [25]. The surface topography of lms
associated with C2N and O C2N appeared, at undergoes signicant changes as a result of the plasma
binding energy of %285.8 and 287.4 eV,
% respectively [25]. treatment and subsequent grafting process. This could
The XPS results show that collagen has been immobi- be due to the mechanism of the grafting, which is
lized on the lm surface with engender of covalent initiated by radicals generated at the lm surface [15].
bonds. Covalent immobilization is the direct result of These radicals initiated the grafting of the acrylic acid.
the reaction between the WSC-activated carboxyl Polyacrylic acid grafted on the surface has a thickness in
groups of the grafted AAc polymer with the NH2 sub-micron range. Grafted polyacrylic acid chains
groups of the protein. formed their own domains and morphology on the lm
Fig. 5 shows the relative concentration of AAc grafted surface. As a result, the roughness of the PCLAAc lm
on the lm surface, represented by [O]/[C] ratio, surface increased in comparison with the unmodied
increases with increasing AAc monomer concentration PCL lm. Collagen was immobilized via the covalent
used for graft reaction. Large amount of AAc mono- bond with polyacrylic acid by losing a water molecule.
mers in the environment accelerate grafting polymeriza- Collagen immobilization broadened the polyacrylic acid
tion action. As a result, grafting concentration increases domains on the lm surface, which led to increase of the
as the concentration of AAc solution raises. The highest roughness of the surface. Those domains obscured brils
grafting concentration was observed at 9% AAc on the surface of unmodied PCL lm.
The relative amount of immobilized collagen on the 4.5. Cell culture
surface, similarly, can be expressed by the [N]/[C] ratio.
Fig. 6 shows the amount of collagen immobilized on the Hydrophilicity of the material surface plays a critical
surface as a function of the concentration of AAc role in cell attachment and cell proliferation [30]. Results
monomer solution for grafting. The [N]/[C] ratio showed that the modied lm surfaces, PCLCol lms,
increases rapidly with increasing the concentration of supported cell attachment and proliferation more than
AAc monomer solution less than 9%. Since AAc owns unmodied lm surfaces. This could be due to the more
the active COOH group, which acts as a spacer to bond hydrophilic surface of modied lm and the preferred
with collagen, the quantity of collagen immobilized on cell adhesion to collagen I. Figs. 11 and 12 show relative
the surface is affected by the amount of grafted AAc on occupation percent of viable human dermal broblast
the PCL lm surface. In other words, the concentration and human myoblast cells, according to confocal
of immobilized collagen is dependent on the concentra- images, on the substrate separately. Both human dermal
tion of carboxyl functional group, which is proportion broblast and human myoblast cells showed good cell
with the graft concentration of AAc side chains, on the proliferation on the collagen modied lms. For
lm surface. unmodied PCL lms, poor occupation was shown in
most of the time points except for the human myoblasts
4.3. Water contact angle of the modified surface at day 6 while the increase of cell density was observed
in PCLM images. The most possible explanation was
Hydrophilicity is evaluated by water contact angle that many cells on the PCL lm surface were removed
measurement. The water contact angle of pristine PCL during the staining process before CLSM analysis.
lm surface decreases dramatically after grafting with Better occupation of human myoblasts on unmodied
2000 Z. Cheng, S.-H. Teoh / Biomaterials 25 (2004) 19912001

5. Conclusion
Area of cells / area of the film surface (%)

45 PCL lms, prepared by using solvent casting and
biaxial stretching technique, were successfully modied
by plasma pretreatment, UV-induced graft polymeriza-
tion with AAc and collagen immobilization. AAc graft
30 polymerization and collagen immobilization onto the
25 surface were conrmed using FT-IR and XPS. The
20 concentration of AAc grafted on the lm surface is
15 strongly dependent on the concentration of AAc
monomer solution used for grafting. The amount of
immobilized collagen increases as that of grafted AAc
on the lm surface increases. Water contact angles of
0 lms were reduced signicantly with lm surface
Day 1 Day 4 Day 8
modication. The roughness of the lm surface in-
Fig. 11. Occupation percent of viable human dermal broblast cells on creased after surface modication, and the morphology
PCLCol and PCL lm surfaces at day 1, day 4 and day 8.
of the surface also changed after acrylic acid grafting
and collagen immobilization. PCLCol lms showed
excellent human dermal broblast and myoblast cell
12 attachment and proliferation rate. The present techni-
Area of cells / Area of the film surface (%)

PCL-Col que has posed the way for future membrane tissue
10 engineering layer by layer as rst proposed by Teoh [31]
on PCL.

6 Acknowledgements

The authors acknowledgment the contribution of
Professor Kang En-Tang and Dr. Ying Lei (Chemical
Engineering Department, National University of Singa-
pore) in surface modication of PCL lms.
Day 1 Day 3 Day 6
Fig. 12. Occupation percent of viable human myoblast cells on PCL
Col and PCL lm surfaces at day 1, day 3 and day 6. [1] Choi EJ, Kim CH, Park JK. Synthesis and characterization of
starch-g-polycaprolactone copolymer. Macromolecules
[2] Hutmacher DW. Scaffolds in tissue engineering bone and
PCL lm surface was found at day 6, but it was only cartilage. Biomaterials 2000;21:252943.
amount to 18% of that on the PCLCol lm surface. [3] Schantz JT, Hutmacher DW, Ng KW, Khor HL, Lim TC, Teoh
This might be implied that cell adhesion was improved SH. Evaluation of a tissue engineering membrane-cell construct
remarkably after collagen modication of PCL lms, for guided bone regeneration. Int J Oral Max Implants
especially for human dermal broblasts. Better cell 2002;17:16174.
[4] Khor HL, Ng KW, Schantz JT, Phan TT, Lim TC, Teoh SH,
attachment, in addition, could be suggested that Hutmacher DW. Poly (e-caprolactone) lms as a potential
N-containing groups play an important role for cell substrate for tissue engineering an epidermal equivalent. Mater
adhesion and proliferation. Another advantage of Sci Eng C 2002;20:715.
placing N-containing groups such as amine on the [5] Miralles G, Baudoin R, Dumas D, Baptiste D, Hubert P, Stoltz
surface was that a fraction of them might be positively JF, Dellacherie E, Mainard D, Netter P, Payan E. Sodium
alginate sponges with or without sodium hyaluronate: in vitro
charged at physiological pH because of the protonation engineering of cartilage. J Biomed Mater Res 2001;57:26878.
in the culture medium, which would enhance the [6] Zacchi V, Soranzo C, Cortivo R, Radice M, Brun P, Abatangelo
interaction between the surface and the cells which G. In Vitro engineering of human skin-like tissue. J Biomed Mater
carried negative charge. The roughness of the lm Res 1998;40:18794.
[7] Suggs LJ, West JL, Mikos AG. Platelet adhesion on a
surface may also play an important role in cell
bioresorbable poly (propylene fumarate-co-ethylene glycol) copo-
attachment and cell proliferation. In this study, the lymer. Biomaterials 1999;20:68390.
higher roughness of the lm surface showed the better [8] Hutmacher DW, Schantz T, Zein I, Ng KW, Teoh SH, Tan KC.
cell attachment and proliferation. Mechanical properties and cell cultural response of polycaprolactone
Z. Cheng, S.-H. Teoh / Biomaterials 25 (2004) 19912001 2001

scaffolds designed and fabricated via fused deposition modeling. effect of sulphur dioxide and hexamethyl-disiloxane plasmas.
J Biomed Mater Res 2001;55:20316. Biomaterials 1995;16:101723.
[9] Boyce ST. Skin substitutes from cultured cells and collagen-GAG [21] Haddow DB, France RM, Short RD, MacNeil S, Dawson RA,
polymers. Med Biol Eng Comput 1998;36:7918. Leggett GJ, Cooper E. Comparison of proliferation and growth
[10] Yang J, Bei J, Wang S. Improving cell afnity of poly (d,l-lactide) of human keratinocytes on plasma copolymers of acrylic acid/1,7-
Film modied by anhydrous ammonia plasma treatment. Polym octadiene and self-assembled monolayers. Inc J Biomed Mater
Adv Technol 2002;13:2206. Res 1999;47:37987.
[11] Ayhan H, Piskin E. Collagen immobilization onto P (EGDMA/ [22] Teng MY, Lee KR, Liaw DJ, Lin YS, Lai JY. Plasma deposition
HEMA) microbeads for cell afnity systems. J Bioactive of acrylamide onto novel aromatic polyamide membrane for
Compatible Polym 2000;15:2742. pervaporation. Eur Polym J 2000;36:66372.
[12] Kang IK, Choi SH, Shin DS, Yoon SC. Surface modication of [23] Narkis M, Chaouat SS, Siegmann A, Shkolnik S, Bell J.
polyhydroxyalkanoate lms and their interaction with human Irradiation effects on polycaprolactone. Polymer 1985;26:507.
broblasts. Int J Biol Macromol 2001;28:20512. [24] Lindberg T, Wirsen A, Albertsson AC. Graft polymerisation of
[13] Zhang F, Kang ET, Neoh KG, Wang P, Tan KL. Surface acrylamide onto PCL lm by electron beam pre-irradiation in air
modication of stainless steel by grafting of poly (ethylene glycol) or argon. Morphology in the nal grafted state. Polymer
for reduction in protein adsorption. Biomaterials 2001;22: 2000;41:4099111.
15418. [25] Ng CS, Teoh SH, Chung TS, Hutmacher DW. Simultaneous
[14] Sun YH, Feng LX, Zheng XX. New-type sulphonated polymer biaxial drawing of poly (e-caprolactone) lms. Polymer
surfaces for improving blood compatibility. J Appl Polym Sci 2000;41:585564.
1999;74:282631. [26] Suzuki M, Kishida A, Iwata H, Ikada Y. Graft copolymerization
[15] Gupta B, Plummer C, Bisson I, Frey P, Hilborn J. Plasma- of acrylamide onto a polyethylene surface pretreated with a glow
induced graft polymerization of acrylic acid onto poly (ethylene discharge. Macromolecules 1986;19:18048.
terephthalate) lms: characterization and human smooth muscle [27] Bae JS, Seo EJ, Kang IK. Synthesis and characterization of
cell growth on grafted lms. Biomaterials 2002;23:86371. heparinized polyurethanes using plasma glow discharge. Bioma-
[16] Kang ET, Neoh KG, Huang SW. Surface-functionalized polyani- terials 1999;20:52937.
line lms. J Phys Chem B 1997;101:1074450. [28] Ying L, Wang P, Kang ET, Neoh KG. Synthesis and
[17] Sodergard A. Preparation of poly (e-caprolactone)-co-poly characterization of poly (acrylic acid)-graft-poly (vinylidene
(acrylic acid) by radiation-induced grafting. J Poly Sci Part A uoride) copolymers and pH-sensitive membrances. Macromole-
1998;36:180512. cules 2001;35:6739.
[18] Lee SD, Haiue GH, Chang PCT, Kao CY. Plasma-induced [29] Ng KW, Hutmacher DW, Schantz JT, Ng CS, Too HP, Lim TC,
grafted polymerization of acrylic acid and subsequent grafting of Phan TP, Teoh SH. Evaluation of ultra-thin poly (e-caprolactone)
collagen onto polymer lm as biomaterials. Biomaterials lms for tissue-engineered skin. Tissue Eng 2001;7:44155.
1996;17:1599608. [30] Lin JC, Ko TM, Cooper SL. Polyethylene surface sulfonation:
[19] Matsumura K, Hyon SH, Nakajima N, Peng C, Tsutsumi S. surface characterization and platelet adhesion studies. J Colloid
Surface modication of poly (ethylene-co-vinyl alcohol) (EVA). Interface Sci 1994;164:99106.
Part I. Introduction of carboxyl groups and immobilization of [31] Teoh SH. Scaffolds in cardiovascular tissue engineering-ultra thin
collagen. Int J Biomed Mater Res 2000;50:5127. layer by layer. Cardiovascular Tissue Engineering Symposium,
[20] Lin JC, Cooper SL. Surface characterization and ex vivo blood World Congress Biomechanics Calgary, Canada, 49 August
compatibility study of plasma-modied small diameter tubing: 2002.