Biodegradable core-shell materials via RAFT and ROP: Characterization and comparison of hyperbranched and microgel particles

Article abstract of DOI:10.1021/ma1027092

Two methodologies for synthesizing novel, degradable, core cross-linked copolymer particles were investigated and the molecular properties of the resultant polymers were compared. The first approach was to synthesize hyperbranched poly(ε-caprolactone-co-N,N-dimethylamino-2-ethyl methacrylate (PCL-co-PDMAEMA) by combining metal-catalyzed ring-opening polymerization (ROP) of ε-caprolactone (ε-CL) and reversible addition-fragmentation chain transfer polymerization (RAFT) of N,N-dimethylamino-2-ethyl methacrylate. First, the hyperbranched core was prepared via ROP copolymerization of ε-CL and branching agent 4,4-bioxepanyl-7,7-dione (BOD). This polymerization was initiated using the hydroxyl moiety of the bifunctional initiator 4-cyano-1-hydroxypent-4-yl dithiobenzoate (ACP-RAFT) which resulted in reactive pendent RAFT groups located on the polymer chains. The hyperbranched structure was confirmed by GPC-MALLS and NMR. Subsequent chain extension of this hyperbranched macromolecule with DMAEMA using RAFT chemistry yielded water-soluble nanoparticles. The second method involved the synthesis of core-cross-linked-shell particles (CCS) by the arm-first route. Linear arms of DMAEMA were synthesized using ACP-RAFT and subsequently used as macroinitiator for the ROP of ε-CL and BOD to form a degradable microgel that was water-soluble. Once again, molecular structure was analyzed by ¹H NMR, ¹³C NMR, and GPC and molecular size by TEM. Finally, GPC-MALLS was used to qualitatively investigate the different cross-link densities of the degradable core by the two different methodologies. Thus, we demonstrate two synthetic approaches for constructing water-soluble, degradable core-shell nanoparticles that exhibit varying degrees of cross-linking by combining RAFT and ROP.

Full text of DOI:10.1021/ma1027092

ARTICLE
pubs.acs.org/Macromolecules
Biodegradable Core-Shell Materials via RAFT and ROP:
Characterization and Comparison of Hyperbranched
and Microgel Particles
Yu Zheng,† William Turner,† Mengmeng Zong,† Derek J. Irvine,†,‡ Steven M. Howdle,† and
Kristofer J. Thurecht*,§
‡
†School of Chemistry and Department of Chemical and Environmental Engineering, University of Nottingham,
Nottingham, U.K. NG7 2RD
§Australian Institute for Bioengineering and Nanotechnology and Centre for Advanced Imaging, University of Queensland,
St. Lucia, Queensland 4072, Australia
S Supporting Information
b
ABSTRACT: Two methodologies for synthesizing novel,
degradable, core cross-linked copolymer particles were
investigated and the molecular properties of the resultant
polymers were compared. The first approach was to
synthesize hyperbranched poly(ε-caprolactone-co-N,N-di-
methylamino-2-ethyl methacrylate (PCL-co-PDMAEMA)
by combining metal-catalyzed ring-opening polymerization
(ROP) of ε-caprolactone (ε-CL) and reversible addition-
fragmentation chain transfer polymerization (RAFT) of N,
N-dimethylamino-2-ethyl methacrylate. First, the hyper-
branched core was prepared via ROP copolymerization of
ε-CL and branching agent 4,4-bioxepanyl-7,7-dione (BOD).
This polymerization was initiated using the hydroxyl moiety
of the bifunctional initiator 4-cyano-1-hydroxypent-4-yl
dithiobenzoate (ACP-RAFT) which resulted in reactive pendent RAFT groups located on the polymer chains. The hyperbranched
structure was confirmed by GPC-MALLS and NMR. Subsequent chain extension of this hyperbranched macromolecule with DMAEMA
using RAFT chemistry yielded water-soluble nanoparticles. The second method involved the synthesis of core-cross-linked-shell particles
(CCS) by the arm-first route. Linear arms of DMAEMA were synthesized using ACP-RAFT and subsequently used as macroinitiator for
the ROP of ε-CL and BOD to form a degradable microgel that was water-soluble. Once again, molecular structure was analyzed by 1H
NMR, 13C NMR, and GPC and molecular size by TEM. Finally, GPC-MALLS was used to qualitatively investigate the different cross-
link densities of the degradable core by the two different methodologies. Thus, we demonstrate two synthetic approaches for constructing
water-soluble, degradable core-shell nanoparticles that exhibit varying degrees of cross-linking by combining RAFT and ROP.
’ INTRODUCTION
Recently, 4,4-bioxepanyl-7,7-dione (BOD) was used as a
cross-linker to form a microgel or core cross-linked polymeric
structure.5,6 Both fully degradable (using ROP alone) and par-
tially degradable (using a combination of atom-transfer radical
polymerization ATRP and ROP) microgel particles were synthe-
sized using the arm-first methodology. These microparticles had
a highly dense and cross-linked, but potentially degradable, core
of polycaprolactone. This is of particular importance for the
design of new biomedical polymers, since discrete nanoparticles
with defined size and structure could be designed which also
exhibit good biocompatibility and degradability.7
Dendritic polymers have unique properties because of their
highly branched structures and large number of functional end
groups.1,2 Their unique, three-dimensional structure also makes
them attractive for new applications ranging from drug delivery to
nanobuilding blocks.2,3 Additionally, recent discoveries of new
controlled polymerization mechanisms have paved the way to
achievable synthetic procedures to deliver new macromolecular
architectures. Hyperbranched polymers are one class of dendritic
polymer that has a random branched structure. While the cost, and
most certainly the synthetic challenge, of developing hyper-
branched polymers is considerably lower than that of the much
morestructurally symmetrical dendrimer, they still exhibit manyof
the beneficial properties of a dendrimer. Perhaps the most
important of these is the presence of a high degree of functionality
bothat the surfaceofthe particle and within the branchedinterior.4
Received: November 29, 2010
Revised:
January 20, 2011
Published: February 10, 2011
r
2011 American Chemical Society
1347
dx.doi.org/10.1021/ma1027092 Macromolecules 2011, 44, 1347–1354
|

Macromolecules
ARTICLE
Scheme 1. Scheme of Degradable Core-Shell Polymer via RAFT and ROP Routea
a Routes A and B demonstrate the core-first method (for hyperbranched particles) and arm-first method (for CCS particles), respectively.
Polyester nanoparticles have become increasingly important
in the biomedical field due to the general ease of metabolization
of the degradation products that are produced.8,9 In addition,
readily available feedstocks and newly developed chemistries
have facilitated tuning of the polymer properties, for example,
hydrophilicity and postmodification with bioligands.7 Of equal
importance is the ability to have a handle on the degree of cross-
linking and hence degree of degradation and rate.
N-dimethylamino-2-ethyl methacrylate (DMAEMA, from Aldrich, 99%)
were passed through a columnof basic alumina to remove stabilizing agents
and then stored under a nitrogen atmosphere at -20 °C. Toluene
(Aldrich, reagent grade) and tetrahydrofuran (THF) were dried by
molecular sieves before use. 4-Cyano-1-hydroxypent-4-yl dithiobenzoate
(ACP-RAFT) was synthesized as previously described in the literature.11
Synthesis of 4,4-Bioxepanyl-7,7-dione (BOD). BOD was
synthesized as previously reported.10 Typically, a solution of urea hydro-
gen peroxide (CO(NH2)2 H2O2) (10.0 g, 106 mmol) in 50 mL of
3
In this paper, we present a new method for the synthesis of
core-shell, degradable hyperbranched polymers with varying
cross-link density. The polymer consists of a hyperbranched
PCL core and pH-sensitive shell. Our strategy combines the two
methodologies of ROP and RAFT; an hydroxyl-group-terminated
dithiobenzoate-based RAFT agent is used to initiate the ROP of
lactone in the presence of a suitable catalyst, i.e., Sn(oct2), while
acting efficiently as a controlling agent for methacrylate polymer-
ization. Here, we utilize a metal-free initiator and the RAFT agent
is seen as a biocompatible initiator if chosen judiciously. We
compare this structure to microgel particles synthesized (so-called
CCS particles) using RAFT by a modified methodology first
introduced by Qiao et al.10 The overall approach for synthesis of
hyperbranched (core-first) and microgel (arm-first) degradable
nanoparticles is outlined in Scheme 1. Dimethylaminoethyl
methacrylate (DMAEMA) was chosen as the shell of the particles
in both cases to impart solubility in aqueous solutions. The
molecular conformation of the particles is investigated by SEC-
MALLS which provides a qualitative comparison of cross-link
density between the polymers formed by each methodology.
formic acid (99%) was stirred at 23 °C for 90 min. 4,4-Bicyclohexanone
(5.0 g, 25.7 mmol) was then slowly added over 5-10 min and stirred for
a further 4 h. 200 mL of water was added to the mixture followed by
extraction with chloroform. The organic fractions were collected, washed
with a saturated aqueous sodium bicarbonate solution, and dried with
Na2SO4. The organic fraction was concentrated, and the solvent was removed
under reduced pressure to yield a white powder (3.50 g, 60% yield). 1HNMR
(300 MHz, CDCl3): 4.34 (R, R) 4.17 (S, R) (t, 2H, -CH2- OOC-), 2.73
(R, R) 2.60 (S, R) (t, 2H, -CH2COO-), 1.93-1.83 (m, 2H, -
CH2CH2OOC-), 1.70-1.60 (m, 2H,-CH2CH2COO-), 1.49 (q, 1H, -
CHCH2-).
Synthesis of ACP-RAFT-Functional Hyperbranched Poly-
(CL-co-BOD). A three-neck round-bottomed flask equipped with a
rubber septum, three-way tap, and condenser were flame-dried. BOD
(475 mg, 2.1 ꢀ 10-3 mol), ACP-RAFT (554 mg, 2.1 ꢀ 10-3 mol), and
ε-caprolactone (2.42 g, 2.1 ꢀ 10-2 mol) were added to 50 mL of dried
toluene. Stannous octanoate (Sn(Oct)2) (405 mg, 1 ꢀ 10-3 mol) was
added to the flask and then heated to 110 °C. The ratio [ACP-
RAFT]:[BOD]:[ε-caprolactone]:[Sn(Oct)2] was 1:1:10:0.5. After the
desired reaction time, a 10 mL sample was withdrawn, and the polymer
was then selectively precipitated in an excess volume of cold methanol,
filtrated, and dried under reduced pressure.
’ EXPERIMENTAL SECTION
Synthesis of Hyperbranched Core-Shell Poly(CL-co-
BOD)core-(DMAEMA)shell via RAFT. Poly(CL-co-BOD) (1.5 g,
1 ꢀ 10-4 mol) and ACP initiator (7.8 mg, 3.46 ꢀ 10-5 mol) were put in
a dry round-bottomed flask. DMAEMA (13.6 g, 8.65 ꢀ 10-2 mol) in dry
toluene (170 mL) was injected into the reaction vessel, and the system
Materials. ε-Caprolactone (ε-CL, from Aldrich, 99%) was dried over
calcium hydride(CaH2) for 48 h at room temperature and then distilled
under reduced pressure before use. Bicyclohexanone was purchased from
TCI Chemicals and used as received. Methyl methacrylate (MMA) and N,
1348
dx.doi.org/10.1021/ma1027092 |Macromolecules 2011, 44, 1347–1354

Macromolecules
ARTICLE
was freeze-pump-thawed three times to remove oxygen. Assuming
there are an average of 17.3 [ACP-RAFT] sites on each poly(CL-co-
BOD) core (see eq 2), the amount of ACP-RAFT sites in the reaction
mixture was 1.73 ꢀ 10-3 mol. Thus, the chemical ratio [ACP-RAFT
sites]:[DMAEMA]:[ACP initiator] was 1:50:0.2 with an overall DMAE-
MA concentration of 0.5 M. The solution was heated to 65 °C and
allowed to react for 12 h. The polymer was precipitated in excess cold
hexane and dried in a vacuum oven.
Degradation of Poly(PCL-co-BOD)core-(DMAEMA)shell Hy-
drolyzable Core. Dioxane (18 mL, 2.04 ꢀ 10-4 mol), hydrochloric
acid (1.5 mL, 30%) and poly(PCL-co-BOD)core-(DMAEMA)shell were
mixed in a flask; the solution was heated to 60 °C and stirred for 24 h.
After neutralization by NaOH and extraction, a fine pale yellow powder
was collected and analyzed by NMR and GPC.
Table 1. Polymerization Data for the Synthesis of Hyper-
branched Poly(CL-co-BOD)a
reaction
time/h
Mn/
Mw/
dn/dc
conversion/
%
entry
kDa
kDa
PDI (mL/g)
1
2
1
2
6.4
11.3
15.0
1.76
1.3
0.082
0.08
21
40
11.5
a Mn, Mw, and PDI are determined by gel permeation chromatography
equipped with MALLS detector. Yield of ε-caprolactone-co-BOD is
calculated gravimetrically.
Synthesis of Arm-First CCS DMAEMA-co-BOD. Linear
DMAEMA with a target molecular weight of 6 kDa (at 100% conversion)
was synthesized using ACP-RAFT as controlling agent at 65 °C. The
isolated and purified polymer (Mn = 4.7 kDa, PDI = 1.3, 79% conversion)
was used as the arms for the synthesis of CCS particles. 0.5 g of polymer
(1.06 ꢀ 10-4 mol) and 130 mg of BOD (6.28 ꢀ 10-4 mol) were
dissolved in dry toluene (20 mL) in a 100 mL round-bottomed flask,
followed by the addition of 25 mg of Sn(Oct)2 (6.28 ꢀ 10-5 mol). The
flask was then backfilled with argon and immersed in an oil bath at 110 °C
for 24 h. The solution was then precipitated into cold hexane with the
precipitate being collectedbyfiltration and dried overnight under vacuum.
The pure degradable CCS particle used in conformational analysis
was synthesized as previously described in the literature.12
Figure 1. MALLS traces showing molecular weight evolution of
hyperbranched poly(CL-co-BOD) with time. A clear shift to higher
molar mass is observed from 1 to 2 h.
Gel Permeation Chromatography (GPC). Hyperbranched
Polymer. Number-average molecular weight (Mn), weight-average mo-
lecular weight (Mw), and dispersity (Mw/Mn) were obtained by gel
permeation chromatography (PL-120, Polymer Laboratories) with an
RI detector. The columns (30 cm PLgel Mixed-C, 2 in series) were
eluted by THF and calibrated with polystyrene standards. All calibra-
tions and analyses were performed at 40 °C and a flow rate of 1 mL/min.
All of the products easily dissolved in THF and passed through a 0.2 μm
filter before injection with little or no backpressure observed—demon-
strating the absence of macrogelation.
CCS Polymer. GPC was performed using a Waters 333 system fitted
with an RI detector. Two Styragel HT3 columns were attached in series
and eluted with DMF at a flow rate of 1 mL/min at 40 °C. As with the
hyperbranched polymers, the polymer was dissolved in solvent to 5 mg/
mL and passed through a 0.2 μm filter before analysis.
and left to dry for 30 min. The TEM micrograph was measured using a
JEOL JEM-200 FXII electron microscope operating at 200 keV.
’ RESULTS AND DISCUSSION
Core-Shell Hyperbranched Poly(CL-co-BOD)core-DMAE-
MAshell. In this work, hyperbranched core-shell poly(CL-co-
BOD)core-DMAEMAshell polymers were synthesized via a two-
step process involving the synthesis of hyperbranched degradable
poly(CL-co-BOD) core via ring-opening polymerization, followed
by chain extension with DMAEMA via RAFT to give the shell.
In the first step, ring-opening copolymerization of ε-capro-
lactone and 4,4-bioxepanyl-7,7-dione (BOD) was performed in
the presence of a catalyst (stannous 2-ethylhexanoate) and an
initiator (ACP-RAFT) in toluene ([CL] = 0.42 M, 110 °C) to
produce an organic solvent-soluble hyperbranched poly(CL-co-
BOD) core with pendant RAFT groups. The reaction was follo-
wed over 2 h, and the molecular weight characteristics and
gravimetric yield are presented in Table 1. It was noted that
macrogelation occurred once conversion increased beyond 40%.
While the synthesis of hyperbranched polymers by this method
using free-radical chemistry has been reported and recently
reviewed,4 to our knowledge this is the first instance of hyper-
branched PCL being synthesized by this approach.
Table 1 shows that conversion reaches 40% after 2 h. As
suggested previously in reports on free-radical approaches to
hyperbranched polymers, the key to typically avoiding macrogela-
tion is to prevent the ratio of [initiator]:[cross-linker] from
exceeding 1.1,13-2321 Thus, the ratio of [ACP-RAFT]:[BOD]:
[ε-caprolactone] in this work is kept at 1:1:10 so as to provide a
hyperbranched poly(CL-co-BOD) core. The relatively low ratio of
monofunctional:difunctional monomer was used in order to obtain
highly branched molecules, and this typically necessitated termina-
tion of the polymerization at low conversion (<50%) in order to
prevent macrogelation. The GPC-MALLS traces (Figure 1) for the
synthesis of hyperbranched poly(CL-co-BOD) clearly show the
A Dawn 8-angle MALLS detector from Wyatt Technologies was used
for light scattering experiments in both solvents. The Astra software
package for Windows was used to process the data.
NMR Analysis of the Polymers. 1H was carried out on a 300
MHz Bruker spectrometer with MestRec processing software. The
chemical shifts were referenced to the residual solvent CHCl3. 13C
experiments were carried out in the solid state using a Bruker Avance III
300 spectrometer. Solid samples were spun to ∼5 kHz in a MAS probe
equipped with double-air bearings. The 13C NMR spectra were recorded
using a CP pulse sequence with contact time of 2 ms, a recycle delay of
3 s, and typically 2000 scans were sufficient to obtain very good signal-
to-noise. The spectra were calibrated to adamantane at 38.22 ppm.
Particle Size Determination. The size distribution of hyper-
branched and CCS PCL-co-DMAEMA was measured using dynamic light
scattering on a Zetasizer nano series (Malvern Instruments Ltd.). After
being filtered through a 0.2 μm filter, the samples were measured at a
temperature of 25 °C. The errors in the measurements of the molecular
size from DLS are within 5% of the mean value for 10 experiments over a
cumulative time of 1 min per experiment. Transmission electron micro-
scopy (TEM) was also used to investigate the size of the CCS particles.
PCL-co-DMAEMA CCS particles were dissolved in THF to a concentra-
tion of 50 μg/mL. 100 μL of solution was dropped on a holey carbon grid
1349
dx.doi.org/10.1021/ma1027092 |Macromolecules 2011, 44, 1347–1354

Macromolecules
ARTICLE
Figure 2. 1H NMR spectrum of hyperbranched poly(CL-co-BOD) measured in CDCl3 (entry 2, Table 1). The peaks due to ACP-RAFT initiator
(peaks 1 and 2) and ε-caprolactone units (or BOD units, peaks 3-6) are evident in the polymer spectrum.
molecular weight evolution of the hyperbranched polymer as a
function of time. As the reaction proceeds from 21% conversion to
40% conversion, the GPC peak shifts to shorter retention time
(higher molecular weight) and a multimodal peak is observed as is
typical for hyperbranched polymers and signifies an increase in
branching.
initiation of the RAFT hydroxyl groups, this is just higher than that
expected from the feed ratio of RAFT to ε-caprolactone and BOD
at 40% conversion; this was approximately 1 to 4. On the basis of
the feed ratio of initiator to monomers, the number of initiators
should be equal to the branching points—so at 40% conversion,
one BOD unit and one initiator per 4.2 ε-caprolactone units. The
assumption is also made that the RAFT agent is an efficient
initiator, and all chains are initiated by the hydroxyl group on the
RAFT agent.24-26 If these assumptions hold, then the number of
RAFT end-groups per polymer can be calculated based on the
molecular weight measured by an absolute GPC method (e.g.,
MALLS). Thus, at 40% conversion, there are on average 11.8 ACP-
RAFT end-groups in one hyperbranched poly(CL-co-BOD).
Using the core material that had a weight-average molecular
weight of 15 kDa (entry 2, Table 1; entry 1, Table 2), the polymer
was reacted with DMAEMA to chain extend the living RAFT end-
groups. This polymerization was conducted under typical RAFT
conditions in toluene at 65 °C, and 42% conversion of DMAEMA
monomer was achieved after 12 h (as determined gravimetrically).
Table 2 lists the polymerization data for this reaction (entry 2,
Table 2), and 1H NMR of the purified product was used to confirm
the structure (Figure S1, Supporting Information). Assuming that
there are ∼11.8 RAFT sites per hyperbranched molecule, the ratio
of RAFT sites, initiator, and monomer in the feed is kept at [ACP-
RAFT sites]:[DMAEMA]:[ACP initiator] = 1:50:0.2. After 12 h,
the poly(CL-co-BOD)core-DMAEMAshell is formed, and the Mn is
shown to have increased to 38 kDa as measured by GPC-MALLS
(entry 2, Table 2). In addition, the GPC chromatograms from the
RI detector (Figure 3) show the molecular weight evolution during
this chain extension. The GPC trace clearly shows the polymer
peaks shifted to higher molecular weight and became broader upon
chain extension with DMAEMA.
The molecular structure of hyperbranched poly(CL-co-BOD)
was confirmed by 1H NMR (Figure 2b; Table 1, entry 2). The
peaks due to the ACP-RAFT end-groups (peaks 1 and 2), ε-
caprolactone, and BOD units (peaks 3-6) in the polymer are
clearly discernible. In addition, the NMR spectrum shows that
the peaks due to residual BOD monomer are negligible. This
indicates that there are very few pendant, unreacted BOD rings
remaining in the hyperbranched poly(CL-co-BOD) core upon
purification. It should be noted that peaks due to polyBOD or
1
polyPCL are indistinguishable in the H spectrum as has been
previously published.12 The 1H NMR spectrum of BOD is shown
for comparison (Figure 2a).
The clear presence of the phenyl protons from the dithio-
benzoate of the RAFT (at 7.4-7.9 ppm) and the methylene
protons adjacent to the terminal hydroxyl groups (at 3.6 ppm)
and the fact that both rings of the BOD appear to have been
opened suggest that the composition of the hyperbranched
polymer can be calculated by following equation (eq 1):
RAFT end group
ε-CL þ BOD
ðintegrals of peak 1Þ=5
ꢀ
ꢁ
¼
¼ 0:19
ð1Þ
integrals of peak 1
integrals of peak 5 -
=2
2=5
The result shows that the ratio of RAFT end groups to ε-
caprolactone plus BOD is equal to 1:5.2. Assuming instantaneous
Since the hyperbranched poly(CL-co-BOD) core was formed
by random coupling of various numbers of primary chains, the
1350
dx.doi.org/10.1021/ma1027092 |Macromolecules 2011, 44, 1347–1354

Macromolecules
ARTICLE
Table 2. Chain Extension Polymerization Data for the Synthesis of Core-Shell Poly(CL-co-BOD)-(DMAEMA) and Subsequent
Hydrolysis
entry
reaction time/h
Mn/kDaa
Mw/kDaa
PDI
dn/dc
1. PCL-co-BODcore
0b
12c
11.5
38.3
3.1
15.0
68.6
3.9
1.3
1.8
1.2
0.08
2. PCL-co-BODcore-DMAEMAshell
3. hydrolyzed product
0.142
0.169
NA
a Mn and Mw are calculated by GPC-MALLS. b This entry is for the poly(CL-co-BOD) core, the same as entry 2, Table 1. c Yield of DMAEMA chain
extension determined gravimetrically (41%).
Table 3. Polymerization Data for the Synthesis of Core
Cross-Linked Core-Shell PolyMMA-co-BOD
reaction
time/h
Mn/
kDaa
Mw/
kDaa
entry
1. PDMAEMA
PDI yieldb
10
4
4.7
29.1
44.2
48.7
6.2
44.5
1.3
1.5
1.6
2.6
71
15
26
48
2. PDMAEMA-co-BODCCS
3. PDMAEMA-co-BODCCS
4. PDMAEMA-co-BODCCS
8
69.3
Figure 3. GPC trace using RI detector for chain extension from
poly(CL-co-BOD)core-DMAEMA. The GPC trace shows the evolution
of molecular weight following chain extension with DMAEMA and
subsequent broadening of the molecular weight distribution.
24
130.2
a Mn and Mw are calculated by GPC-MALLS using DMF as eluent at
40 °C. b Yield determined gravimetrically.
DMAEMA was determined to be 18.5 by integration of the
phenyl protons of the RAFT agent (∼7.6 ppm) and the methy-
lene protons of DMAEMA (2.6 ppm, Supporting Information
Figure S1). On the basis of the molar masses measured by GPC-
MALLS, the number of ACP-RAFT groups that were chain
extended was determined to be 9.2—this closely matches the
theoretical value described above.
The removal of the poly(CL-co-BOD) core from the core-
shell polymer after hydrolysis was also confirmed by NMR
analysis (Figure S1b, Supporting Information). In the spectrum
of the core-shell polymer, the peaks at 1.40, 1.65, and 2.30 ppm
(all assigned to the poly(CL-co-BOD) core27,28) are no longer
present upon hydrolysis, suggesting that the hydrolysis reaction
was successful.
Arm-First Core-Shell PCL-co-PDMAEMA. The second
method of forming degradable core-shell particles was by using
the arm first approach. This technique involved the synthesis of
microgel particles that were stabilized by presynthesized poly-
meric arms.17,29 Unlike the hyperbranched procedure, the de-
gradable cores of the CCS particles are synthesized using 100%
cross-linker (BOD); macrogelation of the system is minimized
by a combination of dilution effects and stabilization of the par-
ticles using linear polymer arms. This approach has been previously
discussed in the literature using ring-opening polymerization.6,10,12
PDMAEMA was chosen as a representative polymer for the arms of
the microgels, and low-molecular-weight polymers were synthesized
using ACP-RAFT. The oligomer obtained showed relatively narrow
PDI (Table 3, entry 1).
Utilization of PDMAEMA allows the formation of a CCS
particle with nondegradable arms and a fully degradable core and
should impart water solubility onto the particle; to our knowl-
edge, synthesis of such particles using RAFT has not been
reported. The microgel reaction was monitored by GPC
(Table 3, entries 2-4), and progression to higher molecular
weight was clearly observed with increasing reaction time. The
initial trace (Figure 5, DMAEMA arms) relates to an aliquot
removed from the reaction immediately after the injection of
BOD; this peak is not present (or low concentration) in the final
Figure 4. GPC trace of core-shell poly(CL-co-BOD)-DMAEMA
(Table 2, entry 2) before and after (Table 2, entry 3) hydrolysis. The
clean shift to lower molecular weight upon hydrolysis clearly demon-
strates that degradation of the core has occurred.
number of actual initiating sites per core is unknown although it
has been theoretically estimated to be 11.8 to allow the calcula-
tion of the molar ratio for the chain extension experiments. In
order to determine the average number of arms per molecule
(Narm) for hyperbranched polymers, it is necessary to determine
the length or Mn of a typical PDMAEMA arm. Thus, the PCL
core was degraded under acidic conditions, and the remaining
polymer was analyzed by GPC (GPC trace from RI detector
shown in Figure 4).
Following hydrolysis, the molecular weight of the remaining
polymer was determined by GPC-MALLS and found to be 3.1
kDa. Therefore, the average number of DMAEMA arms per
core-shell molecule can be calculated based on eq 2:
ðtotal MnÞ - ðcore MnÞ
Nav armsperpolymer ¼
ð2Þ
arms Mn
Once again, if we assume that the number of ACP-RAFT
groups present in each molecule is equal to the feed ratio as
described above, then 8.6 (73%) RAFT end-groups on the
poly(CL-co-BOD) have been chain extended with DMAEMA.
The remaining 27% of RAFT functionalities are inactive, pre-
sumably buried within the highly branched molecule and in-
accessible to chain extension. This result was also confirmed by
1H NMR whereby the degree of polymerization of the
1351
dx.doi.org/10.1021/ma1027092 |Macromolecules 2011, 44, 1347–1354

Macromolecules
ARTICLE
Figure 5. RI GPC trace in DMF of DMAEMA arms and CCS particles
as a function of reaction time (Table 1, entries 1-4).
CCS particle after 24 h reaction (Figure 5), indicating that
essentially all of the DMAEMA arms have been involved in the
stabilization of the CCS particles.
The progression toward higher molecular weights was mon-
itored after 4, 8, and 24 h. The final sample (removed 24 h after
the injection of BOD) displays a large and broad peak at short
1
retention time correlating to the CCS particles. H NMR was
Figure 6. 1H NMR spectrum of DMAEMA-BOD CCS in CDCl3. Note
that RAFT (peak a: phenyl groups in dithiobenzoate) and hydroxyl
(peak m: methylene adjacent to hydroxyl) end-groups are present in the
spectrum.
used to monitor the composition of the CCS particle (Figure 6),
and the structure was confirmed using solid-state 13C NMR
(Figure S2, Supporting Information). The peaks corresponding
1
to the DMAEMA and PCL constituents from H NMR are
assigned in Figure 6. Importantly, the aromatic peaks at 7.3-7.8
ppm belonging to the pendant RAFT dithiobenzoate groups
clearly show that the RAFT agent is present on the macromo-
lecules. Likewise, the methylene groups adjacent to hydroxyl
end-groups are also present at 3.65 ppm. Similarly, the methylene
groups immediately adjacent to the amino groups in DMAEMA
(2.6 ppm) are clearly well-separated from other peaks in the
spectrum and prove the presence of the PDMAEMA chains in
the copolymer.
The number of arms per CCS particle was estimated from the
GPC-MALLS data as outlined in the hyperbranched section
above. Assuming that all PDMAEMA-RAFT chains initiated the
ROP of BOD and knowing the feed composition, at 49%
conversion the number-average of arms per CCS particle was
calculated to be 9.5.
Size Analysis of PMMA-BOD CCS Particles. One advantage
of CCS and hyperbranched nanoparticles is that they are
intrinsically shape-persistent due to the high level of cross-linking
in the core. This is in contrast to micellar systems that require
solvent selectivity for a particular moiety to form a micelle above
the critical micelle concentration (cmc). This shape-persistent
character of these particles facilitates a wide range of size analyses.
TEM and DLS were used as two complementary methods of
characterizing the size of the CCS particles—DLS in solution
and TEM upon drying following deposition on a grid. Figure 7
shows the DLS and TEM results for entry 4 of Table 3 using THF
as solvent. Here, the particle size distribution maximum is at 8
nm, and the sizes cover the range from 5 to 15 nm. There also
exists a size distribution at around 50-60 nm, which we attribute
to particle agglomeration in the solvent. A TEM micrograph is
also shown in Figure 7 following deposition of the polymer from
THF directly onto a TEM grid. The image shows agglomerated
regions of sphere-like particles in the size range of 10-15 nm.
Unfortunately, because of concentration effects during the dry-
ing process, the particles tend to agglomerate on the grid during
evaporation of the solvent. Nonetheless, the TEM size range of
the particles is similar to the results calculated from DLS.
Particles within the size range of 5-10 nm fall within the
expected range for molecules having such a molecular weight
with a densely packed core.
Comparison of Molecular Structure of Hyperbranched
and CCS Degradable Particles. The cross-link density of the
degradable segment in nanoparticles for biomedical applications
can play a very important role in the degradation rate and
mechanism.30 Thus, it is important to distinguish, at least
qualitatively, any conformational differences between hyper-
branched poly(CL-co-BOD), linear PCL, and cross-linked core
star (CCS) PCL. Purely degradable hyperbranched and CCS
polymers were used in this comparison to minimize errors due to
refractive index differences in the respective copolymers and
hence yield a relationship directly related to cross-link density
(rather than molecular effects). Linear PCL was prepared using
ACP-RAFT as initiator and core-star particles (both CCS and
hyperbranched) using linear PCL arms initiated using ACP-
RAFT. Thus, the polymer end-groups were the same in all
molecules. Comparison of the plots of molecular weight against
elution volume between structurally different species can reveal
differences in the behavior of these molecular structures, since
SEC elution time is directly related to the hydrodynamic radius of
the polymer in solution (not necessarily the molecular size).
Figure 8 compares the log Mw (as calculated using an absolute
molecular weight method, i.e., MALLS) versus elution time for
the three species. The Mw of CCS polymer as synthesized by the
arm-first method is an order of magnitude higher than linear
polymer at the same elution volume. Furthermore, the molecular
weight at various elution times of hyperbranched poly(CL-co-
BOD) (entry 2, Table 1) lies between that measured for the
linear PCL and CCS sample. This suggests that the CCS BOD is
the most densely packed structure because it has the highest
molar mass at any particular hydrodynamic radius compared to
both the hyperbranched and linear polymer. This was expected
because the core of this polymer was synthesized using 100%
1352
dx.doi.org/10.1021/ma1027092 |Macromolecules 2011, 44, 1347–1354

Macromolecules
ARTICLE
Figure 7. DLS data showing size of PMMA-BOD CCS nanoparticles in THF. TEM micrograph is included as inset showing the existence of discrete
particles in the size range of 10-15 nm (scale bar is 50 nm).
the hyperbranched materials showed that the core underwent
facile degradation to a point at which only the water-soluble arms
remained. This suggests that they should be of interested in the
manufacture of controlled release drug delivery systems. In this
case, by varyingthe sizeand cross-linkdensity of the coreby choice
of the synthetic route and the BOD:ε-caprolactone ratio, the
degradation time could be tuned to suit the needs of a proposed
application. Therefore, these particles would show great potential
for use in drug delivery, and future work will involve investigation
into the degradation rate of the particles having a range of cross-
link densities—from loosely branched to microgels.
’ ASSOCIATED CONTENT
Supporting Information. 1H NMR spectrum of hyper-
Figure 8. Plots of the log Mw versus elution volume for linear PCL (a),
S
b
hyperbranched poly(CL-BOD) (b), and core cross-linked star poly-
BOD (c).
branched poly(CL-co-BOD)-(DMAEMA) before and after
hydrolysis; solid state 13C NMR spectrum of CCS BOD-b-
DMAEMA. This material is available free of charge via the
Internet at http://pubs.acs.org.
cross-linker and presumably forms a microgel structure. In this
latter case, the microgel core size, stability, and gelation was
influenced by the presence of the PDMAEMA arms which act in
a similar fashion to a traditional dispersant. The hyperbranched
species is less densely packed than the CCS, but more dense than
linear PCL at the same molecular weight. Again, this is expected
since the hyperbranched molecule contains a core that is a
mixture of both linear and branched monomers and hence will
be far less compact than the CCS microgel. A cartoon outlining
the various cross-link densities based on the MALLS data is
shown graphically in Figure 8 to aid interpretation.
’ AUTHOR INFORMATION
Corresponding Author
*E-mail k.thurecht@uq.edu.au; Tel þ44 (0)115 951 3486; Fax
þ44 (0)115 951 3058.
’ ACKNOWLEDGMENT
K.J.T. acknowledges the Australian Research Council
(DP0880032), The University of Queensland, and the Australian
Institute for Bioengineering and Nanotechnology (AIBN) for
funding. We acknowledge Dr. Lauren Butler and the Australian
National Fabrication Facility.
These data suggest that our methodology can be used to
synthesize polymers of varying cross-link density in the degrad-
able core, while being able to maintain the nanoparticle structure
and prevent macrogelation.
’ CONCLUSION
’ REFERENCES
(1) Costello, P. A.; Martin, I. K.; Slark, A. T.; Sherrington, D. C.;
Titterton, A. Polymer 2002, 43 (2), 245–254.
(2) Hult, A.; Johansson, M.; Malmstrom, E. Hyperbranched poly-
mers. In Branched Polymers II; Springer-Verlag: Berlin, 1999; Vol. 143,
pp 1-34.
(3) Hedrick, J. L.; Trollsas, M.; Hawker, C. J.; Atthoff, B.; Claesson,
H.; Heise, A.; Miller, R. D.; Mecerreyes, D.; Jerome, R.; Dubois, P.
Macromolecules 1998, 31 (25), 8691–8705.
(4) Rimmer, S.; England, R. M. Polym. Chem. 2010, 1 (10), 1533–
1544.
Two methodologies have been described for the synthesis of
biodegradable core-shell polymers. These were based on de-
gradable cores with varying cross-link densities. In both cases, the
shell of the particles was synthesized using RAFT polymerization
of DMAEMA—a water-soluble polymer. The molecular char-
acteristics of the polymers were characterized by GPC-MALLS
and NMR spectroscopy, and the morphological properties of the
materials were investigated by DLS and TEM. A comparison of
the hyperbranched methodology and the CCS methodology
using GPC-MALLS qualitatively demonstrated that the poly-
meric chains of the CCS particle were much more densely packed
than the hyperbranched molecule. Accelerated degradation on
(5) Qiao, G. G.; Connal, L. A.; Wiltshire, J. T. Porous polymeric
materials and polymer particles for preparation thereof. 2006-AU1645;
2007051252, 20061103, 2007.
1353
dx.doi.org/10.1021/ma1027092 |Macromolecules 2011, 44, 1347–1354

Macromolecules
ARTICLE
(6) Wiltshire, J. T.; Qiao, G. G. Macromolecules2008, 41 (3), 623–631.
(7) Pounder, R. J.; Dove, A. P. Polym. Chem. 2010, 1 (3), 260–271.
(8) Platel, R. H.; Hodgson, L. M.; Williams, C. K. Polym. Rev. 2008,
48 (1), 11–63.
(9) Place, E. S.; George, J. H.; Williams, C. K.; Stevens, M. M. Chem.
Soc. Rev. 2009, 38 (4), 1139–1151.
(10) Wiltshire, J. T.; Qiao, G. G. Macromolecules 2006, 39 (26),
9018–9027.
(11) Chiefari, J.; Chong, Y. K.; Ercole, F.; Krstina, J.; Jeffery, J.; Le,
T. P. T.; Mayadunne, R. T. A.; Meijs, G. F.; Moad, C. L.; Moad, G.;
Rizzardo, E.; Thang, S. H. Macromolecules 1998, 31 (16), 5559–5562.
(12) Wiltshire, J. T.; Qiao, G. G. Macromolecules 2006, 39 (13),
4282–4285.
(13) Konkolewicz, D.; Gray-Weale, A.; Perrier, S. Macromol. Theory
Simul. 2010, 19 (5), 219–227.
(14) Liu, B.; Kazlauciunas, A.; Guthrie, J. T.; Perrier, S. Polymer
2005, 46 (17), 6293–6299.
(15) Liu, B.; Kazlauciunas, A.; Guthrie, J. T.; Perrier, S. Macromole-
cules 2005, 38 (6), 2131–2136.
(16) Semsarilar, M.; Ladmiral, V.; Perrier, S. Macromolecules 2010,
43 (3), 1438–1443.
(17) Matyjaszewski, K. Polym. Mater. Sci. Eng. 2001, 84, 363–364.
(18) Tsarevsky, N. V.; Huang, J.; Matyjaszewski, K. J. Polym. Sci., Part
A: Polym. Chem. 2009, 47 (24), 6839–6851.
(19) Yan, D.; Mueller, A. H. E.; Matyjaszewski, K. Macromolecules
1997, 30 (23), 7024–7033.
(20) Bannister, I.; Billingham, N. C.; Armes, S. P. Soft Matter 2009,
5 (18), 3495–3504.
(21) Isaure, F.; Cormack, P. A. G.; Graham, S.; Sherrington, D. C.;
Armes, S. P.; Butun, V. Chem. Commun. 2004, 9, 1138–1139.
(22) Rosselgong, J.; Armes, S. P.; Barton, W.; Price, D. Macromole-
cules 2009, 42 (16), 5919–5924.
(23) Rosselgong, J.; Armes, S. P.; Barton, W. R. S.; Price, D.
Macromolecules 2010, 43 (5), 2145–2156.
(24) Chang, C.; Wei, H.; Quan, C.-Y.; Li, Y.-Y.; Liu, J.; Wang, Z.-C.;
Cheng, S.-X.; Zhang, X.-Z.; Zhuo, R.-X. J. Polym. Sci., Part A: Polym.
Chem. 2008, 46 (9), 3048–3057.
(25) Saeed, A. O.; Dey, S.; Howdle, S. M.; Thurecht, K. J.; Alexander,
C. J. Mater. Chem. 2009, 19 (26), 4529–4535.
(26) You, Y.; Hong, C.; Wang, W.; Lu, W.; Pan, C. Macromolecules
2004, 37 (26), 9761–9767.
(27) Zeng, F. Q.; Shen, Y. Q.; Zhu, S. P. Macromol. Rapid Commun.
2002, 23 (18), 1113–1117.
(28) Matyjaszewski, K.; Xia, J. H. Chem. Rev. 2001, 101 (9), 2921–
2990.
(29) Gao, H. F.; Matyjaszewski, K. Macromolecules 2008, 41 (12),
4250–4257.
(30) Amsden, B. G.; Tse, M. Y.; Turner, N. D.; Knight, D. K.; Pang,
S. C. Biomacromolecules 2006, 7 (1), 365–372.
1354
dx.doi.org/10.1021/ma1027092 |Macromolecules 2011, 44, 1347–1354