An investigation of siloxane cross-linked hydroxyapatite-gelatin/copolymer composites for potential orthopedic applications

Article abstract of DOI:10.1039/c2jm32466k

Causes of bone deficiency are numerous, but biomimetic alloplastic grafts provide an alternative to repair tissue naturally. Previously, a hydroxyapatite-gelatin modified siloxane (HAp-Gemosil) composite was prepared by cross-linking N,N'-bis[(3-trimethoxysilyl)propyl]ethylene diamine (enTMOS) around the HAp-gel nanocomposite particles, to mimic the natural composition and properties of bone. However, the tensile strength remained too low for many orthopedic applications. It was hypothesized that incorporating a polymer chain into the composite could help improve long range interaction. Furthermore, designing this polymer to interact with the enTMOS siloxane cross-linked matrix would provide improved adhesion between the polymer and the ceramic composite, and improve mechanical properties. To this end, copolymers of l-lactide (LLA), and a novel alkyne derivatized trimethylene carbonate, propargyl carbonate (PC), were synthesized. Incorporation of PC during copolymerization affects properties of copolymers such as molecular weight, T_g, and %PC incorporation. More importantly, PC monomers bear a synthetic handle, allowing copolymers to undergo post-polymerization functionalization with graft monomers to specifically tailor the properties of the final composite. For our investigation, P(LLA-co-PC) copolymers were functionalized by an azido-silane (AS) via copper catalyzed azide-alkyne cycloaddition (CuAAC) through terminal alkyne on PC monomers. The new functionalized polymer, P(LLA-co-PC)(AS) was blended with HAp-Gemosil, with the azido-silane linking the copolymer to the silsesquioxane matrix within the final composite. These HAp-Gemosil-P(LLA-co-PC) (AS) composites were subjected to mechanical and biological testing, and the results were compared with those from the HAp-Gemosil composites. This study revealed that incorporating a cross-linkable polymer served to increase the flexural strength of the composite by 50%, while maintaining the biocompatibility of HAp-Gemosil ceramics.

Full text of DOI:10.1039/c2jm32466k

View Online / Journal Homepage
Dynamic Article Links <
C
Journal of
Materials Chemistry
Cite this: DOI: 10.1039/c2jm32466k
www.rsc.org/materials
PAPER
An investigation of siloxane cross-linked hydroxyapatite–gelatin/copolymer
composites for potential orthopedic applications†
a
Jason Christopher Dyke, Kelly Jane Knight, Huaxing Zhou, Chi-Kai Chiu, Ching-Chang Ko
a
a
b
b
*
a
*
and Wei You
Received 19th April 2012, Accepted 6th September 2012
DOI: 10.1039/c2jm32466k
Causes of bone deficiency are numerous, but biomimetic alloplastic grafts provide an alternative to
repair tissue naturally. Previously, a hydroxyapatite–gelatin modified siloxane (HAp–Gemosil)
composite was prepared by cross-linking N,N0-bis[(3-trimethoxysilyl)propyl]ethylene diamine
(enTMOS) around the HAp–gel nanocomposite particles, to mimic the natural composition and
properties of bone. However, the tensile strength remained too low for many orthopedic applications. It
was hypothesized that incorporating a polymer chain into the composite could help improve long range
interaction. Furthermore, designing this polymer to interact with the enTMOS siloxane cross-linked
matrix would provide improved adhesion between the polymer and the ceramic composite, and
improve mechanical properties. To this end, copolymers of L-lactide (LLA), and a novel alkyne
derivatized trimethylene carbonate, propargyl carbonate (PC), were synthesized. Incorporation of PC
during copolymerization affects properties of copolymers such as molecular weight, Tg, and %PC
incorporation. More importantly, PC monomers bear a synthetic handle, allowing copolymers to
undergo post-polymerization functionalization with graft monomers to specifically tailor the properties
of the final composite. For our investigation, P(LLA-co-PC) copolymers were functionalized by an
azido-silane (AS) via copper catalyzed azide–alkyne cycloaddition (CuAAC) through terminal alkyne
on PC monomers. The new functionalized polymer, P(LLA-co-PC)(AS) was blended with HAp–
Gemosil, with the azido-silane linking the copolymer to the silsesquioxane matrix within the final
composite. These HAp–Gemosil–P(LLA-co-PC)(AS) composites were subjected to mechanical and
biological testing, and the results were compared with those from the HAp–Gemosil composites. This
study revealed that incorporating a cross-linkable polymer served to increase the flexural strength of the
composite by 50%, while maintaining the biocompatibility of HAp–Gemosil ceramics.
Due to the drawbacks (e.g., donor site morbidity, shortage of
resources) of autografts, the need for alternate alloplastic mate-
rials is clear. Orthopedic biomaterials, in particular, have been
heavily studied, and comprehensively reviewed by Puppi et al.2
and Shoichet3 in greater detail. In particular, significant progress
has been achieved in engineering materials capable of degrading
in vivo, either by hydrolytic or enzymatic activity to promote
formation of natural osseous tissue, through growth of tissue
into the composite material.4 This biodegradable approach allows
for the body to heal itself more gradually, with the scaffold
serving as a temporary matrix until sufficiently strong osseous
tissue can assume a physiological load.5,6
1. Introduction
Natural bone is a lightweight mineral composite consisting of
inorganic apatites, mainly hydroxyapatite (HAp), within a dense
matrix of organic collagens. The long fibrous collagen makes the
normally brittle HAp more resilient, helping to improve the
flexural strength in natural bone.1 The hierarchy HAp–collagen
structure, however, cannot be reproduced easily using engineering
principles. Sequentially, autografts (tissues from the host) have
become the gold standard for replacement of damaged tissues.
aDepartment of Chemistry, University of North Carolina at Chapel Hill,
Chapel Hill, NC 27599-3290, USA. E-mail: wyou@unc.edu
In response to the needs, several classes of biocompatible and
biodegradable polymers have been established for numerous
medical applications. Polymers such as poly(L-lactide)
(PLLA),7–9 poly(glycolic acid) (PLGA),10 poly(3-caprolactone)
(PCL)11,12 and poly(trimethylene carbonate) (PTMC)13–15 have
been investigated as native or as in blends16,17 for a variety of
medical applications. Each of these polymers has unique
bDepartment of Orthodontics and Applied Materials Sciences Program,
University of North Carolina at Chapel Hill, Chapel Hill, NC 27599-
7450, USA. E-mail: koc@dentistry.unc.edu
† Electronic supplementary information (ESI) available: 1H NMR is
available for all previously unpublished monomers and polymers. 13C
NMR is provided for PC and AS. GPC traces are provided for all
polymers. DSC analysis is given for PLLA, P(LLA-co-PC)6.4% and
P(LLA-co-PC)21.05%. See DOI: 10.1039/c2jm32466k
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem.

View Online
mechanical and degradative properties allowing them to be
utilized in a wide range of biomaterials.2,3,18–21 Though homo-
polymers have good properties for in vivo applications, they are
often limited by their diversity in function. Therefore, copoly-
mers present a useful way to obtain tunable properties such as
molecular weight, crystallinity, glass transition temperature (Tg),
modulus, degradation behavior and tensile strength, all of which
can be specifically optimized for use in preparing scaffolds.10,13
Furthermore, the structure of many of these monomers can be
synthetically altered to tailor their properties. These monomers
can be combined in nearly endless ways to form functional
materials with application specific properties. Because of this, it
is important for synthetic chemists to formulate new monomers
and design new monomers and polymers in an attempt to
improve the utility of materials engineered for specific
applications.
ring-opening polymerizations (ROPs), whereby byproducts are
avoided and polymers of high Mw can be readily obtained.31–33
Since both monomers share a common method of polymeriza-
tion, copolymer composition is more easily controlled. More
importantly, cyclic carbonates can be easily derived34 and the
modifications present on PC would allow the incorporation of a
pendant silane graft monomers onto the polymer backbone (vide
infra), while also imparting similar properties to PTMC.
Second, we designed the chosen copolymer to cross-link within
the HAp–Gemosil composite because this would lead to an
improved tensile strength via a better long range interaction when
compared with physically blending polymers into the composites.
Specifically, this approach – designing polymers with cross-
linkable grafts – would provide two advantages: (a) increase
interfacial adhesion over polymer blends, and (b) enhance long
range interactions when compared with composites that are only
cross-linked by small molecules (e.g., enTMOS). Consequently,
the composite would more effectively mimic the short and long
range chemical interactions seen within bone, thereby improving
the tensile strength of the composite.35 This should in turn help to
resist tensile loading by distributing forces more evenly through
the composite, rather than at the point of application.36 Since
HAp–Gemosil composites were originally cross-linked using an
amino-silane (enTMOS), it would be ideal to design the copoly-
mer to bear a similar cross-linkable silane group.
Recently, Ko and co-workers created
a composite of
hydroxyapatite–gelatin (HAp–gel)22 that mimicked the natural
composition and properties of bone23 and was able to demon-
strate promise for in vivo and in vitro biocompatibility.24
However, challenges remained in developing useful grafts from
these composites because they demonstrated poor processability
and insufficient strength when porous. These problems were
ultimately improved by incorporation of an additional cross-
linking agent, N,N0-bis [(3-trimethoxysilyl)propyl]ethylene
diamine (enTMOS).25 This small molecule is capable of under-
going hydrolysis–condensation of alkoxy-silanes to produce a
silsesquioxane matrix within the hydroxyapatite–gelatin modi-
fied siloxane (HAp–Gemosil) composite. This helps give an
additional structural support to the composite and can help
impart the network strength of the silane matrix into the
composite, leading to enhanced mechanical properties and
molding ability. While this matrix did improve the compressive
strength and processability of the composite, the short chain
siloxane based matrix was brittle and still susceptible to tensile
failure. It was clear that a more robust composite was needed in
order to further advance this system for potential scaffolding
applications. One possible solution is to design and incorporate a
biocompatible and cross-linkable polymer of sufficient chain
length into the composite.
However, the sensitivity of the graft monomer’s silane groups
precluded their direct incorporation prior to polymerization.
Therefore, we employed the CuAAC chemistry (‘‘Click’’ reac-
tions) to impart the silane functionality to the polymer via post-
polymerization functionalization by quantitatively coupling the
graft monomer azide to the main chain alkyne of the polymer
backbone.37,38 Thus the PC monomer was synthesized to bear a
pendant acetylene group, while the azide functionality was linked
with the silane (e.g., 5-azido-N-(3-(trimethoxysilyl) propyl) pen-
tanamide). After the copolymerization of LLA and PC, the
terminal alkyne group of the copolymer would easily react with
the azido-silane (AS) graft monomer via CuAAC chemistry. This
post-polymerization functionalization approach allows the
grafting of the silane functionality to occur after polymerization,
ensuring the fidelity of the silane groups is maintained. P(LLA-
co-PC)(AS) can then be blended with HAp–gel and cross-linked
in the presence of enTMOS to produce a fully cross-linked
composite through hydrolysis–condensation of trialkoxysilyl
groups present on both the copolymer and enTMOS. Such a
programmed composite would improve adhesion through coor-
dination of grafted amide and triazole groups to free carboxylate
groups on gelatin and also through silane cross-linking to HAp.
Based on this rationale, we have synthesized and characterized
a series of P(LLA-co-PC) polymers via ring-opening polymeri-
zation (ROP) and functionalized the polymers with pendant
silanes. Composites formed by incorporating these polymers
with HAp–Gemosil were easily molded and set quickly. To
determine the impact of blending P(LLA-co-PC)(AS) with HAp–
Gemosil, transwell plates were filled with this new composite
material and preosteoblasts MC3T3-E1 were cultured in the
bottom of these plates for 21 days. It was observed that
throughout this period, the growth curves of cells in the presence
of either HAp–Gemosil or HAp–Gemosil–P(LLA-co-PC)(AS)
composites were very similar, suggesting that the synthesized
First, we chose a copolymer of PLLA and PTMC to blend into
HAp–Gemosil composites. PLLA has demonstrated previous
success in use with hydroxyapatite ceramic composites,16,26,27 and
PLLA’s biocompatibility has long been established. However, at
physiological temperature, PLLA is brittle and can contribute to
low composite strength.28 Therefore, an amorphous TMC
derivative, propargyl carbonate (PC), was synthesized and
utilized as the co-monomer. This monomer can act to soften
PLLA and toughen brittle composites,29 as demonstrated in
previous studies where P(LLA-co-TMC) copolymers showed a
decrease in Tg with increasing TMC incorporation.30–32 The
combination of both properties from PLLA and PPC (a deriv-
atized PTMC) polymers could help P(LLA-co-PC) copolymers
to exhibit good flexural strength and elongation, ideal for
creating a more robust ceramic composite. In addition, both
PLLA and PTMC polymers have unique degradative properties,
allowing the degradation behavior of future composites of this
formulation to be controlled. Finally, from the synthetic
perspective, both of these monomers are ideal candidates for
J. Mater. Chem.
This journal is ª The Royal Society of Chemistry 2012

View Online
polymer can help improve mechanical properties without nega-
tively impacting the biocompatibility. In addition, both materials
showed similar cellular growth curves to those of the control
samples, suggesting these materials provide a suitable substrate
for cellular attachment and proliferation. Furthermore, biaxial
bending tests were undertaken to determine the impact that
incorporation of P(LLA-co-PC)(AS) into a ceramic composite
has on the flexural strength. It was also observed that this
polymer helped increase fiber bridging within the composite,
leading to higher flexural strength than that of non-polymeric
HAp–gel composites. These results suggest that the design of this
copolymer, and the use of graft monomers as cross-linking
agents possess the merit for future study in expanding their
applications for bioceramic composites.
Scheme 1 Synthesis of C4 (PC) from THME. Reaction conditions for
synthesis of PC from THME: (i) benzaldehyde, TsOH and THF; (ii)
NaH, propargyl bromide and THF; (iii) MeOH: 1 M HCl (1 : 1, v/v) and
(iv) THF, ethyl chloroformate and TEA.
benzaldehyde (23.2 mL, 0.23 mmol) was added dropwise and
allowed to react for 16 hours. The reaction was then neutralized
with aqueous ammonia and concentrated by rotary evaporation.
The product was then dissolved in DCM and washed 3 times with
water before again concentrating under rotary evaporation to
yield 39 g (94%) of C1 as a colorless solid. 1H NMR (400 MHz,
CDCl3): d 0.81 (s, 3H), 3.66 (d, 2H), 3.91 (s, 2H), 4.06 (d, 2H),
5.44 (s, 1H), 7.35–7.49 (m, 5H).
2. Experimental details
2.1 General methods
All moisture sensitive reactions were performed in flame-dried
glassware under an atmosphere of Ar. Reaction temperatures
were recorded as external bath temperatures. The phrase
‘‘concentrated under reduced pressure’’ refers to the removal of
C1 (35 g, 168 mmol) in THF (50 mL) was added dropwise to a
cold stirring solution of NaH (60% in mineral oil, 12 g,
302 mmol) at 0 ꢀC to yield a milky white solution. After 30
minutes of cold stirring, the solution was heated on an oil bath to
60 ꢀC for 2 hours, yielding a pale yellow solution. Propargyl
bromide (25 g, 211 mmol) was then added dropwise and allowed
to stir for 16 hours. This solution was then quenched with water
to afford a deep red solution with precipitate. This solution was
extracted 3 times with brine and concentrated under high
vacuum to yield crude C2 as a red oil.
€
volatile materials by distillation using a Buchi rotary evaporator
at water aspirator pressure (<20 Torr) followed by removal of
residual volatile materials under high vacuum (<1 Torr). The
term ‘‘high vacuum’’ refers to vacuum achieved by a standard
belt-drive oil pump (<1 Torr).
1H and 13C NMR spectra were recorded on Bruker AC-400
(400 MHz) spectrometers. Chemical shifts are reported in parts
per million (ppm) relative to residual solvent peaks (CHCl3: 1H:
d 7.26). Peak multiplicity is reported as: singlet (s), doublet (d),
triplet (t), quartet (q), multiplet (m), and broad (br).
Crude C2 (12 g, 48.7 mmol) was stirred in 400 mL 1 : 1 v/v
MeOH : 1 M HCl for 2 hours. 1 M NaOH was then used to raise
the pH to 7 before MeOH was removed by rotary evaporation.
The product was extracted using ethyl acetate to yield crude C3.
Purification via flash chromatography using 2 : 1 ethyl aceta-
te : hexanes yielded 6.1 g (54% over 2 steps) C3 as an orange oil.
1H NMR (400 MHz, CDCl3): d 0.84 (s, 3H), 1.65 (b, 2H), 2.45 (s,
1H), 3.49 (s, 2H), 3.58 (d, 2H), 3.67 (d, 2H), 4.15 (s, 2H).
C3 was dissolved (5.18 g, 32.7 mmol) with ethyl chloroformate
(6.68 g, 65 mmol) in THF (300 mL). Triethyl amine (7.1 g,
65 mmol) was then added to this stirring solution dropwise and
allowed to react for 4 hours. Water was then used to quench the
reaction slowly and the reaction was washed with brine. The
organic layer was then concentrated by rotary evaporation and
then purified via flash chromatography with 2 : 1 hexanes : ethyl
acetate to yield 4.52 g (75%) propargyl carbonate (PC). 1H NMR
(400 MHz, CDCl3): d 1.11 (s, 3H), 2.46 (s, 1H), 3.49 (s, 2H), 4.07
(d, 2H), 4.17 (s, 2H), 4.33 (d, 2H). 13C NMR (400 MHz, CDCl3)
d 17.26, 32.8, 58.7, 70.66, 73.81, 75.25, 78.83, 148.17.
2.2 Materials
Ethyl chloroformate (99%) and CaH2 (60% in mineral oil), and 4-
tert butyl benzyl alcohol (98%) were obtained from Acros
Organics and used as received. 1,1,1-Tris(hydroxyl methyl)ethane
(THME, 99%), triethylamine (TEA, 99%), 5-bromovaleryl
chloride (98%) and CuBr (98%) were obtained from Alfa Aesar
and used as received. Benzaldehyde (Aldrich 98%), p-toluene
sulfonic acid (TsOH, Aldrich, 98.5%), N,N0-bis [(3-trimethox-
ysilyl)propyl]ethylene diamine (enTMOS, 95% in MeOH, Gel-
est), tin(II) 2-ethyl hexanoate (SnOct2 98% MP Biomedicals),
propargyl bromide (80% in toluene, TCI) and 3-aminopropyl
trimethoxy silane (96% TCI) were used as received. L-Lactide was
generously donated by Purac and used without further purifica-
tion. Hexanes, acetone, chloroform, dichloromethane (DCM),
anhydrous toluene, anhydrous methanol (MeOH), ethyl acetate
and tetrahydrofuran (THF) were obtained from Fisher. THF was
freshly distilled over sodium before use.
2.4 Synthesis of graft monomer: 5-azido-N-(3-(trimethoxysilyl)
propyl)pentanamide (AS)
2.3 Synthesis of (propargyl carbonate (PC)) monomer
Synthesis of PC monomer is adapted from previously reported
work by Chen et al.39 A typical synthesis of PC is presented and
outlined in Scheme 1. THME (26 g, 0.217 mmol) and TsOH
(1.2 g, 6.3 mmol) were dissolved in anhydrous THF (400 mL) and
stirred at room temperature. To this stirring solution,
Synthesis of AS is shown in Scheme 2. To a dry flask, (3-ami-
nopropyl)trimethoxysilane (1.44 g, 8 mmol) was dissolved in
THF with TEA (1 g, 10 mmol) and set to stir (clear solution). 5-
Bromovaleryl chloride (2 g, 10 mmol) was then added dropwise,
forming a white precipitate. This solution was allowed to react
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem.

View Online
Scheme 2 Synthesis of AS. Reaction conditions for synthesis of AS: (i) 3-aminopropyl trimethoxy silane, TEA and THF and (ii) DMF and NaN3.
for 16 hours under argon. The solution was subsequently filtered
to remove the salt and concentrated by rotary evaporation to
yield 2.73 g (99%) G1 without further purification. 1H NMR
(400 MHz, CDCl3): d 0.65 (q, 2H), 1.63 (p, 2H), 1.80 (m, 4H),
2.19 (t, 2H), 3.24 (t, 2H), 3.4 (t, 2H), 3.57 (s, 9H), 5.65 (b, 1H).
G1 (2.73 g, 8 mmol) was then dissolved in anhydrous DMF
and added to NaN3 under dry conditions. This was allowed to
react for 16 hours under argon atmosphere. DMF was then
removed under vacuum for 36 hours before dry EtOAc was used
to extract AS and the compound was again concentrated to yield
pure AS 2.4 g (98%) which was stored under dry conditions in a
glovebox. 1H NMR (400 MHz, CDCl3): d 0.65 (t, 2H), 1.78 (m,
6H), 2.19 (t, 2H), 3.26 (m, 4H), 3.41 (s, 9H), 5.69 (s, 1H). 13C
NMR (400 MHz, CDCl3) d 6.46, 22.68, 22.81, 28.34, 35.83,
41.77, 50.45, 51.09, 172.27.
was 100 : 1 : 1 [M] : [C] : [I] for all polymerizations. The
following is a typical synthesis for the copolymer with 10 mol%
PC loading. In a glovebox under argon atmosphere, LLA (7 g,
48 mmol) and PC (0.884 g, 4.8 mmol) were added to a high
pressure flask with 10 mL anhydrous toluene. To this solution,
Sn(Oct)2 and 4-tert-butylbenzyl alcohol (0.075 M, 7 mL) were
added quickly and the vessel was then sealed. The reaction
container was quickly transferred to an oil bath outside of the
glovebox and allowed to react for 20 hours before quenching
with methanol. The resulting polymer P(LLA-co-PC) was iso-
lated by precipitation into cold MeOH, and further purified by
precipitation from the DCM solution of the polymer into cold
methanol 3 times. NMR analysis revealed this polymer to
contain 6.4 mol% PC in the backbone and thus was denoted as
P(LLA-co-PC)6.4% for clarity.
2.5 Synthesis of copolymers by Sn(Oct)2 catalyzed ROP
2.6 Characterization of polymers
PC monomer was purified by flash chromatography and its
purity was verified by NMR analysis. The monomer was dried
over CaH2 before use. Polymerization was carried out in toluene
at 120 ꢀC for 20 hours using 4-tert-butylbenzyl alcohol as the
initiator and stannous-2-ethyl hexanoate (Sn(Oct)2) as the cata-
lyst (Scheme 3). The ratio of monomers to catalyst to initiator
The monomer incorporation ratio of all copolymers was deter-
mined using a Bruker spectrometer (400 MHz). CDCl3 was used
as the solvent and the chemical shifts were calibrated against
residual solvent signals. The molecular weight and polydispersity
of the copolymers were determined by a Waters 1515 gel
permeation chromatograph (GPC) using THF as the eluent at a
Scheme 3 ROP of LLA and PC, CuAAC modifications and amalgamation. Scheme 3 shows ROP polymer formation, followed by CuAAC post-
polymerization modification before processing into the HAp–Gemosil composite cement. This cement is stabilized by polymer bridging and cross-
linking through the AS trialkoxysilyl moiety. This silane network will bind HAp, enTMOS, and neighboring polymer chains to increase long range
adhesion and the strength.
J. Mater. Chem.
This journal is ª The Royal Society of Chemistry 2012

View Online
flow rate of 1.0 mL minꢁ1 at 30 ꢀC. A series of narrow poly-
styrene standards were used for the calibration of the columns.
Thermal behavior was investigated using a TA Q100 differential
scanning calorimeter (DSC) under nitrogen atmosphere. All
ratio (n) of 0.3 was used for both materials. The flexure stress at
failure (s in MPa) was calculated using the following expressions:
s ¼ AP/t2
ꢁ1
ꢀ
ꢀ
ꢀ
samples were scanned from ꢁ10 C to 200 C at 10 C min
and
before being quenched with liquid nitrogen. The samples were
then rescanned from ꢁ10 ꢀC to 200 ꢀC and data were collected.
Glass transition (Tg) was taken as the midpoint of heat capacity
change and crystallization (Tc) and melting (Tm) were shown by
exo- and endo-thermal peaks, respectively.
A ¼ (3/4p)[2(1 + n)ln(rs/ro) + (1 ꢁ n)(2rs2 ꢁ ro2)/2(d/2)2 + (1 + n)]
2
where ro ¼ (1.6rp + t2)1/2 ꢁ 0.675t.
2.7 Post-polymerization modification of polymers via CuAAC
2.10 Viability test – preosteoblast cell proliferation by MTS
assay
CuAAC reactions were performed in anhydrous DMF at room
temperature under argon atmosphere. A typical synthesis of
P(LLA-co-PC)6.4% follows. The azide containing linking
monomer (AS), was dissolved with P(LLA-co-PC)6.4% and
CuBr in DMF. Load ratios of AS to acetylene groups of the
polymer were 1.1 : 1 (slight excess of AS) and CuBr was loaded at
2 mol%. Elemental copper was added in trace amounts to help
stabilize Cu(I) ions in solution. This reaction is allowed to run for
2 hours before precipitation in cold MeOH to yield white powder
of the AS functionalized polymer P(LLA-co-PC)(AS)6.4%.
Polymers were treated with EDTA prior to precipitation in
methanol to help remove the excess copper catalyst.
The 6-well transwell plate (Corning #3414 Transwell with 3.0 mm
polycarbonate membrane insert) was used for cell proliferation
testing. The material disc (diameter 15 mm by thickness 1 mm)
was placed on the permeable membrane support of the transwell,
which allowed materials immersed in the culture medium. The
1 ꢂ 10ꢁ5 preosteoblasts MC3T3-E1 were seeded to each well.
Every three days, the medium was replenished with 2 mL fresh
growth medium (a-minimum essential medium with 10% fetal
bovine serum and 1% penicillin and streptomycin). At the end of
cultivation (1, 7 and 21 days in culture), the disk and the
permeable support were removed. 40 mL of 3-(4,5-dimethylth-
iazol-2-yl)-5-(3-carboxymethoxy-phenyl)-2-(4-sulfophenyl)-2H-
tetrazolium salt (MTS) reagent (Promega, Madison, WI USA)
was added to each well containing 400 mL of a-MEM, and the
plate was incubated for 1 hour at 37 ꢀC under a humidified
atmosphere of 5% CO2. From each well, 100 mL of the mixed
solution was transferred into a well of a 96-well plate. Each well
was triplicated. The absorbance of each well at 490 nm was
2.8 Amalgamation of AS functionalized polymer
Composite cement materials were prepared for biaxial flexure
and biocompatibility testing using P(LLA-co-PC)(AS)13.5%
loaded at 10 wt% into the composite. A typical composite
formation is presented here with 10 wt% copolymer in the
composite. HAp–gel (250 mg), prepared as previously reported,25
was mixed with P(LLA-co-PC)(AS)13.5% (25 mg) and the blend
was ground into a fine powder. Next, 100 mL of enTMOS (95% in
MeOH) was added to the powder and mixed thoroughly. To this
mixture, a 10% acetone in PBS solution (420 mL) was added while
gently kneading the composite into a clay. This clay was then
formed into a mold and pressed to remove excess solvent. It was
left for a minimum of 72 hours at room temperature to dry. This
procedure was repeated for all composites. After drying, a solid
material was obtained which could be trimmed and used to test
the material’s flexural strength.
2.9 Biaxial flexure strength
The testing procedure for the biaxial flexure strength was per-
formed according to Ban and Anusavice.40 Four sample disks
(diameter 10 mm by thickness 1 mm) of each group (HAp–
Gemosil and HAp–Gemosil–P(LLA-co-PC)AS13.5%) were
prepared in Teflon molds. The upper and lower surfaces were
polished in order to obtain parallel surfaces with no apparent
defects. After measuring the sample diameter (d) and thickness
(t), the disk was supported on three stainless steel balls (3 mm in
diameter), which were equally spaced along a 3.25 mm radius
(rs). Prior to testing, a stainless steel piston (radius ¼ rp
¼
1.5 mm) was aligned concentrically with the three balls (Chart 1).
A crosshead speed of 0.5 mm minꢁ1 was used, and the maximum
force at failure (P) was determined using an Instron 4411
Machine (model 4411, Instron Co., Norwood, MA). A Poisson’s
Chart 1 Biaxial flexure apparatus.
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem.

View Online
measured using a microplate reader (Microplate Reader 550,
Bio-Rad laboratories, Philadelphia, USA). Relative cell numbers
were quantified on the basis of the concentration of the formazan
product of MTS. Three samples will be used at each time point
for each material group. Three experimental groups of materials
were investigated, including Gemosil, Gemosil–P(LLA-co-PC)
AS13.5%, and dishes as received without coating (control).
This information would assist in future composite planning by
helping elucidate the underlying chemistry that dictates polymer
properties, since these polymer properties will determine how
polymers interact with a ceramic composite. The copolymeriza-
tion was carried out at 120 ꢀC in toluene and for 20 hours using
an Sn(Oct)2 catalyst and 4-tert-butylbenzyl alcohol as the initi-
ator (Scheme 3). Sn(Oct)2 was chosen because of its versatility in
ROP catalysis and ability to run at high temperature. 4-tert-
Butylbenzyl alcohol initiator was employed since its steric bulk
can help inhibit intramolecular chain trans-esterification during
polymerization. 1H NMR was used to determine the ratio of
incorporated carbonate to lactide in the polymer by comparing
integrations of carbonate methyl (d ¼ 0.995 ppm) peaks and
lactide methine (d ¼ 5.18 ppm) peaks, while GPC traces were
taken from a THF solution of the polymer to determine the
molecular weight. The number averaged molecular weight (Mn)
and %PC incorporation are plotted against the mol% PC loading
in Fig. 1A and B respectively. The most notable feature from
Fig. 1A is that the Mn decreases with the increased loading of the
PC monomer, similarly observed by Gu et al. in a recent study.31
Interestingly, the mol% incorporation of PC in the copolymer,
shown in Fig. 1B, is consistently lower than the mol% PC
loading. Both of these observations can be attributed to the faster
rate of polymerization of LLA compared with that of PC as
previously reported for similar TMC–LLA copolymers.30
Previous reports of lactide–carbonate copolymerizations showed
that reaction times of >48 hours are needed to obtain high
molecular weight copolymers that incorporated a high molar
fraction of carbonate monomers.31 However, at these elevated
temperatures and reaction times, PLLA segments could ther-
mally degrade more readily than polycarbonate segments.30 In
our cases, LLA is consumed faster than PC, resulting in a portion
of PC monomers not being incorporated into the copolymer
chain in the chosen reaction time (20 h). Instead, these uncon-
sumed PC monomers form low molecular weight (Mw) chains (2–
4 kDa) consisting primarily of poly-propargyl carbonate, or
remain as unreacted PC monomer. Due to the low MW and
rubbery nature of PC, both remain soluble in methanol and are
washed away during precipitation.
3. Results and discussion
3.1 Monomer and cross-linker synthesis
The hydrolysable cyclic carbonate monomer PC, was synthesized
in four steps from established methods39 as shown in Scheme 1.
Due to the sensitive nature of both the azide and the silane
groups on the graft monomer molecule, azido-silane (AS), it is
important to utilize a synthesis that would allow for highly pure
products in nearly quantitative yields over all steps under mild
conditions. To accomplish this, 5-bromovaleryl chloride was first
reacted with 3-aminopropyl trimethoxy silane to yield 5-bromo-
N-(3-(trimethoxysilyl)propyl)pentanamide (G1 in Scheme 2). An
SN2 reaction with sodium azide was then performed to yield
5-azido-N-(3-(trimethoxysilyl)propyl)pentanamide (AS, G2 in
Scheme 2). Both steps offered products in nearly quantitative
yield with no need for purification as confirmed by NMR.
3.2. Synthesis and characterization of copolymers from
Sn(Oct)2 catalyzed ROP
TMC homopolymer and PLLA have different physical proper-
ties,18,19 therefore the percent incorporation of the PC unit (a
TMC derivative) in the copolymer would impact important
polymer properties such as molecular weight and Tg.28,30–32,34
More importantly, these properties would determine whether or
not the newly designed copolymers are suitable for specific
applications.17 Therefore, understanding the polymerization
behavior and related properties of the copolymers precedes the
development of composites. To accomplish this, we systemati-
cally varied the mol% PC in the load (0–100%) to investigate its
effect on the polymerization and properties of the copolymer.
Fig. 1 (A) Polymer molecular weight (Mn) as a function of increasing %PC load. (B) %PC present in the polymer chain as a function of mol fraction
loaded before polymerization.
J. Mater. Chem.
This journal is ª The Royal Society of Chemistry 2012

View Online
Table
copolymers
1
Summarized polymerization data for P(LLA-co-PC)
Detailed NMR analyses of all polymers elucidate further
structural information of these polymers and the random nature
of the copolymerization. To identify the chemical origin of each
shift in the copolymers, a transition from PLLA homopolymer to
PPC homopolymer with a P(LLA-co-PC)44.6% copolymer is
shown in Fig. 2A. The most interesting and diagnostic feature
comes from the lactide protons (Fig. 2B). In low mol% PC
samples, a single peak is observed for the methine protons at d ¼
5.18 ppm (proton a in Fig. 2B). As the mol% PC increases, this
methine signal begins to split, with a second peak appearing at
d ¼ 5.02 ppm (proton b in Fig. 2B). This secondary methine peak
arises as a result of inductive effects on those methine peaks that
neighbor a carbonate unit. These protons would feel a weaker de-
shielding effect due to the lower electron withdrawing nature of
the carbonate when compared with the ester in LLA, and thus
will be shifted slightly upfield. This effect is observed only in LLA
methine protons that are adjacent to a carbonate in the copoly-
mer. During the copolymerization, if a propagating polymer
chain end belongs to a lactide monomer and this ‘‘lactide’’ chain
end opens up another lactide monomer, then all methine peaks
are equivalent and no alternate shift is observed. When the
propagating ‘‘lactide’’ chain end attacks a carbonate, however,
the additional oxygen on newly incorporated carbonate carbonyl
helps slightly shield the a-methine proton that is next to the
carbonate and leads to the appearance of a second peak upfield
of the first (Fig. 2A). This splitting effect is highlighted in Fig. 2B.
The relative ratio of these methine protons at different chemical
shifts (5.18 ppm vs. 5.02 ppm) gradually decreases as the mol%
PC increases in the copolymer, indicating the ‘‘random’’ nature
of the copolymerization. A summary of polymer composition is
given below in Table 1.
PC loading (%PC)a
%PC incorporationb
Mn (kDa)c
Mw (kDa)
0
0
6.4
66.7
53.0
28.7
23.0
18.6
15.5
13.5
10.5
9.1
93.7
57.8
46.6
35.8
33.5
23.3
22.6
12.4
11.3
9.4
10
20
25
30
40
50
75
90
100
13.5
21.05
25.9
32.1
44.6
64.1
80
100
5.2
a
Copolymers of LLA and PC, denoted by the % loading of PC during
polymerization. Incorporation measured by 1H NMR. Measured by
GPC with the THF eluent.
b
c
into its backbone can help reduce crystallinity and lower the Tg
of the resultant copolymer. This provides a route for altering the
crystallinity and helping to make the copolymer less brittle. This
will allow for improved mechanical properties to be observed
under physiological conditions and in turn, helping to raise the
flexural strength of future composites in vivo.31,41 To demonstrate
the impact on the Tg of the copolymer by the introduction of PC
into PLLA, DSC traces of three polymers with different mol%
incorporation of PC were obtained and compared (Fig. 3). It is
clearly observed that the Tg decreases with the increased PC
content. Specifically, the Tg drops from 57.9 ꢀC for PLLA, to
ꢀ
ꢀ
53.7 C for P(LLA-co-PC)(AS)6.4%, and finally to 52.8 C for
P(LLA-co-PC)(AS)21%. The amorphous nature of the PC
monomer helps to influence Tg by altering chain rigidity and
hindering the chain’s ability to pack effectively. Since both the
HAp and enTMOS portions of the composite are very brittle,
addition of a rubbery copolymer can help improve the polymer
An important implication of employing the copolymer of LLA
and PC is to lower the Tg of the copolymer. PLLA is below its Tg
at the physiological temperature and the incorporation of PC
Fig. 2 (A) The NMR spectra of homopolymers with a P(LLA-co-PC)44.6% copolymer. (B) The splitting observed from lactide methine peaks for
copolymers of varied %PC loadings.
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem.

View Online
tensile strength by increasing flexibility and elongation at break
within the composite.14
after setting. To remedy this, 10% acetone in PBS buffer was used
to facilitate dissolving P(LLA-co-PC)(AS) and subsequent
blending of the polymer with HAp–gel to create a more homo-
geneous composite. More importantly, using this two-solvent
processing, the partially soluble white powder formed after
CuAAC can now be further reacted through these un-cross-
linked, free alkoxysilyl groups with another silane containing
cross-linking reagent, enTMOS, for better setting. The use of
enTMOS allows rapid condensation of trialkoxy silanes, to
rapidly form a strong cross-linked composite. In our investiga-
tion, this functionalized polymer powder is finely ground with
HAp–gel and blended with enTMOS, allowing enTMOS to bond
to free siloxane groups of P(LLA-co-PC)(AS). This reaction
incorporates the newly prepared copolymer into the composite
siloxane matrix through enTMOS. This multiple crosslinking via
enTMOS creates a fully linked gel which can be easily formed
and allows for chemical linking of the polymer to HAp–gel to
enTMOS, increasing long range adhesion and the strength.
After successfully forming these new composites with our
designed copolymers incorporated, we then carried out the
cellular and mechanical studies to determine the effect that
polymer blending has on composites when compared with
previously studied HAp–Gemosil samples. Fig. 4 presents the
results of the 21 day MTS assay. It can be seen from these data
that, when compared with the control sample, both previous
HAp–Gemosil and new HAp–Gemosil–P(LLA-co-PC)(AS)
composites showed similar biocompatibility over a 21 day
period. Furthermore, the resultant growth curves of MC3T3-E1
cells were similar for all three groups. Cells grew up to 7 days
and, then, leveled out. There were no differences in absorbance
between the materials and the control, showing no difference in
cell viability among the tested substrates. These data suggest that
the incorporation of P(LLA-co-PC)(AS) had little to no negative
effect on the biocompatibility of these composites, and minimal
toxic byproducts leaching out from this composite by 21 days.
As previously mentioned, HAp–Gemosil composites lacked
the sufficient flexural strength for use in orthopedic applications.
The incorporation of a cross-linkable polymer was expected to
PLLA and high PLLA content polymers appear as white
fibrous solids at room temperature (0–10% PC incorporation)
due to the high LLA content. As the mol% PC increases in the
polymer, MW decreases and polymers becomes slightly more
yellow in appearance, less fibrous and softer. Above 50 mol% PC
incorporation, the polymers appear as viscous yellow/orange
liquids. These polymers are largely amorphous due to the high
PC content in the backbone, which serves to add steric bulk,
reduce symmetry, and lower rigidity when compared with LLA
segments. The short chain length of these polymers also obstructs
effective packing and crystallization of adjacent chains. Exam-
ples of the physical appearance of several polymers are also given
in Fig. 3 for reference.
3.3 Post-polymerization CuAAC Click functionalization and
amalgamation
After determining that PC can successfully copolymerize with
LLA to give polymers with controlled composition and proper-
ties, it was important to functionalize the pendant acetylene of
PC to help understand how this functionalized polymer can be
processed into HAp–Gemosil composites. CuAAC (a ‘‘Click’’
reaction) allows nearly quantitative coupling of terminal azides
to alkynes via Cu(I) catalysis, with few byproducts and little
purification needed.42 This approach was attempted for several
PC functionalized copolymers and coupling was observed to be
successful.10,43 However, after the CuAAC reaction, the recovery
of the polymer after attempted removal of the residual copper
catalyst proved difficult without initiating minor cross-linking of
the grafted AS backbone groups. This cross-linking and subse-
quent loss of solubility confirmed the coupling reactions were
successful. Fortunately, complete gelation was not observed for
polymers with low mol% PC (<45%), due to the low density of
AS on the backbone. However, the AS functionalized polymers
showed poor solubility in PBS buffer and as-formed composites
(i.e., copolymer mixed with HAp–gel) showed poor properties
Fig. 3 Left: images of several polymers to demonstrate changes in physical appearance caused by changes in polymer chain length and distribution.
Right: DSC traces for three of these polymers are given to demonstrate the tunability of Tg as a function of increased %PC content on the polymer
backbone.
J. Mater. Chem.
This journal is ª The Royal Society of Chemistry 2012

View Online
biocompatibility. Improvements in cross-linking chemistry are
required to allow for better processability and material perfor-
mance. These improvements would allow this system to be more
rigorously studied in vivo to determine the efficacy of this poly-
mer–ceramic system for future scaffolding applications.
4. Conclusions
In summary, we have synthesized a derivatized TMC monomer,
PC, which is capable of undergoing ROP with L-lactide to afford
copolymers with tunable MW, mol% PC incorporation and Tg.
The ability of these monomers to copolymerize and yield
potentially biodegradable and biocompatible polymers of
tunable properties makes this an attractive system for biological
applications. In the current demonstration, we coupled the
copolymer with AS graft agents inspired by enTMOS, rendering
the copolymer ‘‘cross-linkable’’ via these pendant silane groups.
After being processed into the original HAp–Gemosil cement
composite facilitated by the amino-silane enTMOS, these AS
functionalized polymers were capable of bridging the new
composite and providing enhanced long range adhesion, while
still maintaining the biocompatibility of the new composite.
The current grafting approach did have some key limitations,
however. Notably, the sensitivity of polymer bound silanes pre-
vented extensive purification after CuAAC coupling. As a result,
some residual copper from CuAAC was generally trapped in the
composite after coupling, which could lead to the possibility of
increased cell morbidity. Furthermore, the premature cross-
linking could contribute to an inability of these polymers to fully
cross-link into the enTMOS silsesquioxane matrix, leading to
poor adhesion between the hydrophobic polymer and the
hydrophilic HAp–gel moieties in the composite. Therefore,
further work remains to be done, especially regarding the graft
monomer and cross-linking. Fortunately, the PC monomer
introduces a pendant acetylene group on the copolymer, which
provides a synthetic handle for post-polymerization modification
to give more synthetic freedom. This ‘acetylene handle’ allows
various pendant groups to be attached to the copolymer, thus
one can further alter the composite properties to obtain unique,
application specific properties. Future composites will be
synthesized using similar P(LLA-co-PC) polymers as the chem-
istry and properties of these copolymer have been elucidated in
this study, but emphasis will be placed on utilizing alternate
‘‘Click’’ reactions which can preclude the use of potentially toxic
catalysis. Alternate graft monomers will also be investigated to
determine a method for cross-linking which can be easily
degraded. This will eliminate potential problems caused by the
residual material left after degradation. Additionally, a less
sensitive method of cross-linking would be ideal as to allow
better control of cross-linking reactions and thereby improve
processing of the final composite. If these issues can be suffi-
ciently addressed, this HAp–Gemosil–P(LLA-co-PC)(AS)
copolymer system will provide a new springboard to undertake
further scaffolding composite work.
Fig. 4 Cell viability data for cells plated on both HAp–Gemosil and
polymer(AS)13.5%. Cells were plated in 6-well transwell plates and
viability was measured by formazan absorbance at days 1, 7 and 21.
help improve the long range interaction within the composite. As
presented in Fig. 5, the higher flexure strength of HAp–Gemosil–
P(LLA-co-PC)(AS) composites than that of the original HAp–
Gemosil indicates an effect of fiber bridging coming from the
blended long chain copolymers. The stiffness of the force–
displacement curve recorded from the biaxial bending test did
not differ (P ¼ 0.08) between the original (2.06 N mmꢁ1) and the
new (2.15 N mmꢁ1) composites. In our in-house data, the HAp–
Gemosil had a compressive modulus around 862 ꢃ 129 MPa and
a plane strain modulus around 18.0 ꢃ 4.9 GPa measured by the
nanoindentation tester (Hysitron Inc.). Based on the stiffness
data, we expect that the new composite might have similar
modulus values although future tests are required.
Though the results from initial biocompatibility and
mechanical tests are promising, this polymer system presents
several key limitations. Primarily, the sensitivity of the P(LLA-
co-PC)(AS) leads to premature cross-linking of polymer bound
silanes. This in turn made complete blending of composite
cements with the polymer more difficult. This limitation can
influence reproducibility and utility of composites like those
tested in this study. The use of a second, less polar solvent
(acetone) during blending helped improve composite formation,
making it possible to study the interaction of this polymer within
HAp–Gemosil composites. Unfortunately, the use of organic
solvents removes our ability to dope this composite with cells for
scaffolding applications, and these processing issues hinder
the ability to study composite interactions in vivo. Despite the
drawbacks, this system still showed merit in improving the
properties of HAp–Gemosil composites, while preserving their
Acknowledgements
The authors acknowledge funding from NC Biotech Grant
#2008-MRG-1108, NIH/NIDCR, K08DE018695 American
Fig. 5 Changes in UTS between HAp–Gemosil and HAp–Gemosil
doped with P(LLA-co-PC)(AS) polymers.
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem.

View Online
18 J. C. Middleton and A. J. Tipton, Synthetic biodegradable polymers
as orthopedic devices, Biomaterials, 2000, 21(23), 2335–2346.
19 L. L. Hench, Prosthetic implant materials, Annu. Rev. Mater. Sci.,
1975, 5(1), 279–300.
20 S. Mukherjee, C. Gualandi, M. Focarete, R. Ravichandran,
J. Venugopal, M. Raghunath and S. Ramakrishna, Elastomeric
electrospun scaffolds of poly(L-lactide-co-trimethylene carbonate)
for myocardial tissue engineering, J. Mater. Sci.: Mater. Med.,
2011, 22(7), 1689–1699.
Association of Orthodontist Foundation, and UNC University
Research Council. We thank Wonhee Jeong for many useful
discussions regarding our work and John Whitley for his assis-
tance in characterizing and testing polymers and composites.
Additionally we thank Purac Biomaterials for their generous
donation of L-lactide for this project.
21 L. Zheng, F. Yang, H. Shen, X. Hu, C. Mochizuki, M. Sato, S. Wang
and Y. Zhang, The effect of composition of calcium phosphate
composite scaffolds on the formation of tooth tissue from human
dental pulp stem cells, Biomaterials, 2011, 32(29), 7053–7059.
22 M. C. Chang, C.-C. Ko and W. H. Douglas, Preparation of
hydroxyapatite–gelatin nanocomposite, Biomaterials, 2003, 24(17),
2853–2862.
23 C. C. Ko, M. Oyen, A. M. Fallgatter and W.-S. Hu, Mechanical
properties and cytocompatibility of biomimetic hydroxyapatite-
gelatin nanocomposites, J. Mater. Res., 2006, 21(12), 8.
24 C. C. Ko, Y.-L. Wu, W. H. Douglas and R. Narayanan, In vitro and
In vivo Tests of Hydroxyapatite-Gelatin Nanocomposites for Bone
Regeneration, A Preliminary Report in Biological and Bioinspired
Materials and Devices, Symposium Proceedings Material Research
Society Spring Meeting, ed. J. Aizenberg, W. J. Landis, C. Orme,
R. Wang and W.-S. Hu, 2004, 823, 5.
25 T.-J. Luo, C.-C. Ko, C.-K. Chiu, J. Llyod and U. Huh, Aminosilane
as an effective binder for hydroxyapatite–gelatin nanocomposites,
J. Sol-Gel Sci. Technol., 2010, 53(2), 459–465.
26 X. Deng, J. Hao and C. Wang, Preparation and mechanical
properties of nanocomposites of poly(D, L-lactide) with Ca-deficient
hydroxyapatite nanocrystals, Biomaterials, 2001, 22(21), 2867–2873.
27 M. Todo and T. Kagawa, Improvement of fracture energy of HA–
PLLA biocomposite material due to press processing, J. Mater.
Sci., 2008, 43(2), 799–801.
28 L. Zhang, S. H. Goh and S. Y. Lee, Miscibility and crystallization
behaviour of poly(L-lactide)/poly(p-vinylphenol) blends, Polymer,
1998, 39(20), 4841–4847.
29 C. Zhang, H. Subramanian, J. J. Grailer, A. Tiwari, S. Pilla,
D. A. Steeber and S. Gong, Fabrication of biodegradable
poly(trimethylene carbonate) networks for potential tissue
engineering scaffold applications, Polym. Adv. Technol., 2009, 20(9),
742–747.
30 E. Ruckenstein and Y. Yuan, Molten ring-open copolymerization of
L-lactide and cyclic trimethylene carbonate, J. Appl. Polym. Sci., 1998,
69(7), 1429–1434.
31 L.-J. Ji, K.-L. Lai, B. He, G. Wang, L.-Q. Song, Y. Wu and Z.-W. Gu,
Study on poly(L-lactide-co-trimethylene carbonate): synthesis and
cell compatibility of electrospun film, Biomed. Mater., 2010, 5(4),
045009.
References
1 R. A. Field, M. L. Riley, F. C. Mello, J. H. Corbridge and
A. W. Kotula, Bone composition in cattle, pigs, sheep and poultry,
J. Anim. Sci., 1974, 39(3), 493–499.
2 D. Puppi, F. Chiellini, A. M. Piras and E. Chiellini, Polymeric
materials for bone and cartilage repair, Prog. Polym. Sci., 2010,
35(4), 403–440.
3 M. S. Shoichet, Polymer scaffolds for biomaterials applications,
Macromolecules, 2009, 43(2), 581–591.
4 M. Bongio, J. J. J. P. van den Beucken, S. C. G. Leeuwenburgh and
J. A. Jansen, Development of bone substitute materials: from
‘biocompatible’ to ‘instructive’, J. Mater. Chem., 2010, 20(40),
8747–8759.
5 H. R. Imam Khasim, S. Henning, G. H. Michler and J. Brand,
Development of nanocomposite scaffolds for bone tissue
engineering, Macromol. Symp., 2010, 294(1), 144–152.
6 K. Rezwan, Q. Z. Chen, J. J. Blaker and A. R. Boccaccini,
Biodegradable and bioactive porous polymer–inorganic composite
scaffolds for bone tissue engineering, Biomaterials, 2006, 27(18),
3413–3431.
7 J. A. Roether, A. R. Boccaccini, L. L. Hench, V. Maquet, S. Gautier
ꢀ ^
and R. Jerome, Development and in vitro characterisation of novel
bioresorbable and bioactive composite materials based on
polylactide foams and Bioglassꢀ for tissue engineering applications,
Biomaterials, 2002, 23(18), 3871–3878.
8 H. Declercq, M. Cornelissen, T. Gorskiy and E. Schacht, Osteoblast
behaviour on in situ photopolymerizable three-dimensional scaffolds
based on D, L-lactide, 3-caprolactone and trimethylene carbonate,
J. Mater. Sci.: Mater. Med., 2006, 17(2), 113–122.
9 J. K. Oh, Polylactide (PLA)-based amphiphilic block copolymers:
synthesis, self-assembly, and biomedical applications, Soft Matter,
2011, 7(11), 5096–5108.
10 X. Jiang, E. B. Vogel, M. R. Smith and G. L. Baker, ‘‘Clickable’’
polyglycolides: tunable synthons for thermoresponsive, degradable
polymers, Macromolecules, 2008, 41(6), 1937–1944.
11 P. Shokrollahi, H. Mirzadeh, O. A. Scherman and W. T. S. Huck,
Biological and mechanical properties of novel composites based
on
supramolecular
polycaprolactone
and
functionalized
32 T. Tyson, A. Finne-Wistrand and A.-C. Albertsson, Degradable
porous scaffolds from various L-lactide and trimethylene carbonate
hydroxyapatite, J. Biomed. Mater. Res., Part A, 2010, 95(1), 209–221.
12 Y. Zhou, D. W. Hutmacher, S.-L. Varawan and T. M. Lim, In vitro
bone engineering based on polycaprolactone and polycaprolactone–
tricalcium phosphate composites, Polym. Int., 2007, 56(3), 333–342.
13 E. Bat, T. G. van Kooten, J. Feijen and D. W. Grijpma, Crosslinking
of trimethylene carbonate and D, L-lactide (co-) polymers by gamma
irradiation in the presence of pentaerythritol triacrylate, Macromol.
Biosci., 2011, 11(7), 952–961.
copolymers obtained by
Biomacromolecules, 2008, 10(1), 149–154.
a
simple and effective method,
33 F. Nederberg, B. G. G. Lohmeijer, F. Leibfarth, R. C. Pratt, J. Choi,
A. P. Dove, R. M. Waymouth and J. L. Hedrick, Organocatalytic ring
opening
polymerization
of
Biomacromolecules, 2006, 8(1), 153–160.
trimethylene
carbonate,
34 E. J. Vandenberg and D. Tian, A new, crystalline high melting
bis(hydroxymethyl)polycarbonate and its acetone ketal for
biomaterial applications, Macromolecules, 1999, 32(11), 3613–3619.
35 C. Durucan and P. W. Brown, Biodegradable hydroxyapatite–
polymer composites, Adv. Eng. Mater., 2001, 3(4), 227–231.
14 N. Andronova and A.-C. Albertsson, Resilient bioresorbable
copolymers based on trimethylene carbonate, L-lactide, and 1,5-
dioxepan-2-one, Biomacromolecules, 2006, 7(5), 1489–1495.
15 B. L. Dargaville, C. Vaquette, H. Peng, F. Rasoul, Y. Q. Chau,
J. J. Cooper-White, J. H. Campbell and A. K. Whittaker, Cross-
linked poly(trimethylene carbonate-co-L-lactide) as a biodegradable,
elastomeric scaffold for vascular engineering applications,
Biomacromolecules, 2011, 12(11), 3856–3869.
€
36 F. Hoffmann, M. Cornelius, J. Morell and M. Froba, Silica-based
mesoporous organic–inorganic hybrid materials, Angew. Chem., Int.
Ed., 2006, 45(20), 3216–3251.
37 J.-F. Lutz, 1,3-Dipolar cycloadditions of azides and alkynes: a
universal ligation tool in polymer and materials Science, Angew.
Chem., Int. Ed., 2007, 46(7), 1018–1025.
38 W. H. Binder and R. Sachsenhofer, ‘Click’ chemistry in polymer and
materials science, Macromol. Rapid Commun., 2007, 28(1), 15–54.
39 W. Chen, H. Yang, R. Wang, R. Cheng, F. Meng, W. Wei and
Z. Zhong, Versatile synthesis of functional biodegradable polymers
by combining ring-opening polymerization and postpolymerization
modification via Michael-type addition reaction, Macromolecules,
2010, 43(1), 201–207.
16 T. Jiang, S. P. Nukavarapu, M. Deng, E. Jabbarzadeh, M. D. Kofron,
S. B. Doty, W. I. Abdel-Fattah and C. T. Laurencin, Chitosan-
poly(lactide-co-glycolide) microsphere-based scaffolds for bone
tissue engineering: in vitro degradation and in vivo bone
regeneration studies, Acta Biomater., 2010, 6(9), 3457–3470.
17 E. Bat, J. e. A. Plantinga, M. C. Harmsen, M. J. A. van Luyn,
Z. Zhang, D. W. Grijpma and J. Feijen, Trimethylene carbonate
and 3-caprolactone based (co)polymer networks: mechanical
properties and enzymatic degradation, Biomacromolecules, 2008,
9(11), 3208–3215.
J. Mater. Chem.
This journal is ª The Royal Society of Chemistry 2012

View Online
40 S. Ban and K. J. Anusavice, Influence of test method on failure
stress of brittle dental materials, J. Dent. Res., 1990, 69(12), 1791–
1799.
42 H. C. Kolb, M. G. Finn and K. B. Sharpless, Click chemistry: diverse
chemical function from a few good reactions, Angew. Chem., Int. Ed.,
2001, 40(11), 2004–2021.
41 E. M. Valliant and J. R. Jones, Softening bioactive glass for bone
regeneration: sol–gel hybrid materials, Soft Matter, 2011, 7(11),
5083–5095.
43 B. Parrish, R. B. Breitenkamp and T. Emrick, PEG- and peptide-
grafted aliphatic polyesters by click chemistry, J. Am. Chem. Soc.,
2005, 127(20), 7404–7410.
This journal is ª The Royal Society of Chemistry 2012
J. Mater. Chem.