Simultaneous enzymatic ring opening polymerisation and RAFT-mediated polymerisation in supercritical CO2

Article abstract of DOI:10.1039/b611626d

This report presents the first simultaneous, metal-free synthesis of block copolymers through combination of enzymatic ring-opening polymerisation of ε-caprolactone with RAFT-mediated controlled radical polymerisation of styrene. The Royal Society of Chem

Full text of DOI:10.1039/b611626d

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Simultaneous enzymatic ring opening polymerisation and RAFT-
mediated polymerisation in supercritical CO2{
Kristofer J. Thurecht,a Andrew M. Gregory,a Silvia Villarroya,a Jiaxiang Zhou,a Andreas Heiseb and
Steven M. Howdle*a
Received (in Cambridge, UK) 14th August 2006, Accepted 11th September 2006
First published as an Advance Article on the web 26th September 2006
DOI: 10.1039/b611626d
Another issue surrounding the use of ATRP as a controlling
agent is the inhibiting effect that the catalyst can have on the
enzyme activity.7 This can be overcome by careful selection of the
ATRP catalyst, however it diminishes the versatility of the ATRP/
enzyme coupling.
This report presents the first simultaneous, metal-free synthesis
of block copolymers through combination of enzymatic ring-
opening polymerisation of e-caprolactone with RAFT-
mediated controlled radical polymerisation of styrene.
The combination of various living techniques for the controlled
polymerisation of monomers exhibiting different propagation
pathways has attracted particular interest over the previous
decade. This has been driven by the increasing demand for block
copolymers that can self-assemble into micellar structures in water
or other solvents. These polymers have particular importance for
the pharmaceutical industry as potential drug carriers or controlled
release vehicles.1 While such copolymers have been successfully
synthesised in the past, the methodology generally involves
multiple reaction and purification steps to achieve the desired
species.
In order to overcome some of these problems, we have designed
a metal-free synthetic strategy for block copolymers. This, com-
bined with the use of scCO2 as the reaction medium, removes the
necessity for purification following polymerisation. Additionally,
we have shown that RAFT mediated free-radical polymerisation
can occur simultaneously with enzymatic ROP leading to a high
degree of control over the constituents of the copolymer.
Our approach for the simultaneous preparation of block
copolymers is presented below (Scheme 1) and uses an hydroxy-
terminated trithiocarbonate.2 This species is similar to other
RAFT agents which have been shown to have low cytotoxicity
and are suitable for biomedical applications.8
Different approaches have recently been adopted to form
polyester based block copolymers. The ring-opening polymerisa-
tion (ROP) of lactides or lactones has traditionally been
accomplished using metal catalysts, followed by controlled free
radical polymerisation of a vinyl monomer. Stenzel and coworkers
synthesised a bifunctional RAFT agent, which initially acted as an
initiator for the ROP of DL-lactide and then allowed RAFT-
controlled free radical polymerisation of N-isopropyl acrylamide.2
However, this method involved purification and isolation of the
RAFT terminated polylactide before the free radical step.
More recently, others have shown that enzymatic ROP of
e-caprolactone can be combined with atom transfer radical
polymerisation (ATRP)3,4 or nitroxide mediated polymerisation
(NMP)5 to form copolymers. However, the polymerisation
requires two consecutive steps in order to achieve block copolymer.
Recently, we have shown that simultaneous enzymatic ROP and
ATRP can occur in supercritical CO2 (scCO2) with good control
over the resultant polymer product.6 This was attributed to the
plasticising effect of scCO2 on the polymer. Additionally, the use
of scCO2 removes the requirement for toxic organic solvents and
results in a clean, solvent-free product. Unfortunately, reactions
which utilise ATRP catalysts require purification of the product
following the polymerisation to remove the copper catalyst.
In a typical experiment (supplementary information), a high-
pressure reaction vessel was loaded with the monomers (e-capro-
lactone and styrene), the RAFT agent ((2-benzylsulfanyl
thiocarbonylsulfanyl)ethanol), the enzyme (Candida Antarctica
lipase B immobilised on an acrylic resin – Novozym-435) and the
free radical initiator (azobisisobutyronitrile (AIBN)). The reactions
were performed at 27.6 MPa (4000 psi) and 65 uC for 24 hours and
in all cases, the RAFT : AIBN molar ratio was kept at 2 : 1. We
have previously shown that for successful polymerisation utilising
ATRP and enzymatic ROP it is necessary to maintain a
homogeneous mixture in the scCO2.6 This was facilitated by the
aSchool of Chemistry, University of Nottingham, University Park,
Nottingham, UK. E-mail: steve.howdle@nottingham.ac.uk;
Fax: +44 (0)115 9513058; Tel: +44 (0)115 951 3486
bDepartment of Polymer Chemistry, Eindhoven University of
Technology, PO Box 513, 5600 MB, Eindhoven, The Netherlands
{ Electronic supplementary information (ESI) available: Polymerisation
procedure, NMR and oxalyl chloride treatment, MALDI-TOF results and
GPEC procedure. See DOI: 10.1039/b611626d
Scheme 1 Schematic showing the simultaneous enzymatic catalysed
ROP of e-caprolactone with RAFT-mediated free radical polymerisation
of styrene to form block copolymers with a RAFT linker group in scCO2.
This journal is ß The Royal Society of Chemistry 2006
Chem. Commun., 2006, 4383–4385 | 4383

View Article Online
Table 1 Block copolymer formation by simultaneous one-pot polymerisation of MMA and e-caprolactone
Conversionb
%
Block Copolymerc
Hydrolysedc
Feed ratioa
PCL : PSTY
mol%
PCLb
Crystallinityd
%
Entry
PCL
PSTY
Mw
PDI
Mw
PDI
—
1e
2
—
100
90
92
65
35
60
96
98
99
99
99
93
c
—
50
47
37
34
71
36000
17000
15500
11200
4500
1.6
2.1
1.7
1.7
1.5
1.8
—
2400
2300
2900
3200
3100
50
42
40
24
—
27
3 : 1
3 : 1
1 : 1
1 : 3
1 : 1
b
1.12
3f
4
1.09
1.14
1.25
1.15
5
6g
a
8500
Molar feed ratio. Calculated using 1H NMR. Molecular weight calculated by GPC against polystyrene standards with RI detector.
d
e
Degree of crystallinity calculated by DSC using enthalpy of melting. Polymerisation of e-caprolactone using RAFT agent as initiator in
f
g
absence of styrene (block copolymer molecular weight is the molecular weight of the RAFT-terminated PCL). Repeat of Entry 2. Three
times the RAFT and initiator concentration of Entry 4.
excellent plasticising ability of scCO2 in conjunction with the
cosolvent effect of e-caprolactone.7 In this case, where a RAFT
agent was used, we maintained very good control over the
polydispersity of the free-radical polymerisation product. This
suggests that e-caprolactone acts as a good cosolvent for
polystyrene (PSTY).
weight PCL can be obtained with high conversion. The relatively
low conversion of PSTY in the copolymers (Table 1, Entries 2–4)
is typical of precipitation polymerisation, whereby the hetero-
geneity of the system inhibits effective transport of monomer to the
propagating radical. However, we have shown that good control
over the ratio of the block lengths is obtainable in scCO2.
Moreover, the good reproducibility of this technique is highlighted
by the comparative results obtained in successive experiments
(Table 1, Entries 2–3).
The synthesis of block copolymers with reproducible block
lengths is an important consideration in copolymer synthesis. For
the case presented here, where the RAFT agent also acts as
initiator for the enzymatic ROP of e-caprolactone, the ratio of
block lengths can be controlled in a number of ways. The
polycaprolactone (PCL) block length can be increased by lowering
the number of OH initiators present – i.e. decreasing the RAFT
concentration. The other method for controlling the block lengths
is simply by varying monomer concentrations. Some typical
examples are presented where copolymers were prepared by
varying the monomer concentration in the feed (Entries 2–5,
Table 1). As a result of decreasing the e-caprolactone feed, smaller
PCL block lengths were observed in the copolymer. Indeed, while
the polydispersity of the copolymer is relatively high (due to the
less well controlled enzymatic ROP of e-caprolactone), the free
radical polymerisation is well controlled with polydispersities
typically less than 1.25 being observed for the PSTY block. It is
worth noting that the conversion of PSTY is affected by the
concentration of e-caprolactone; lower ratios of caprolactone to
styrene result in less styrene conversion (cf. Table 1, Entries 2, 4
and 5). Moreover, the highest polydispersity of the PSTY block is
observed when the amount of e-caprolactone in the feed is lowest.
This is typical of copolymerisation reactions in scCO2 and shows a
loss of control as the cosolvent effect of e-caprolactone is
diminished.7 At high e-caprolactone loadings the homogeneity of
the polymerisation mixture is maintained, facilitating good control.
In order to determine the effect of RAFT concentration on the
copolymer, a reaction was undertaken with three times the
concentration of RAFT agent used in the other reactions (Table 1,
Entry 6). As expected, the molecular weight of the PCL block
decreases due to the much higher concentration of OH-initiating
groups (cf. Entry 4). Interestingly, the conversion of STY increases,
presumably due to the higher concentration of initiator and hence
propagating chains. This example clearly shows the extent to
which variation in the reaction parameters can be utilised to
control the copolymer properties.
One disadvantage of our technique is that potentially both PCL
and PSTY homopolymers could be formed; water may initiate
PCL and RAFT-capped PSTY might not initiate PCL growth.
Thus, the presence of block copolymer was proven using a number
of different techniques. Gradient polymer elution chromatography
(GPEC) has recently been developed as a tool for distinguishing
between homopolymer blends and copolymer.9 The elution time of
PSTY and PCL homopolymers in GPEC depends on their
solubility in the gradient solvent and also on their affinity for the
stationary phase in the column. Thus, when methanol (a poor
solvent for PSTY and PCL) is used as the mobile phase with an
increasing gradient of tetrahydrofuran (see supporting information
for details), PCL will elute before PSTY due to the greater
solubility of PCL in the methanol–tetrahydrofuran mixture
(Fig. 1). It follows then, that the copolymer elutes in the period
between the two homopolymers. This is clearly seen for the GPEC
trace of the copolymer with 92 mol% PCL (Table 1, Entry 4)
which falls at longer retention time than PCL, but shorter than
PSTY (Fig. 1).
The enzymatic ROP of e-caprolactone in the absence of styrene
(Table 1, Entry 1) shows that the RAFT agent is compatible with
the enzyme and acts as a good initiator, such that high molecular
Fig. 1 GPEC traces of PSTY and PCL homopolymers and of the PCL-
b-PSTY copolymer. The copolymer elutes at a time between the two
homopolymers (gradient details in supporting information).
4384 | Chem. Commun., 2006, 4383–4385
This journal is ß The Royal Society of Chemistry 2006

View Article Online
affected, presumably due to phase separation of the two immiscible
polymers. The exotherms are measured on the second cooling scan
to eliminate the possiblity of kinetically induced blending of
homopolymers from the scCO2 processing.
An important factor for consideration in the enzymatic ROP of
e-caprolactone is the potential for initiation by adventitious water
in the system. This increasingly becomes an issue when bifunc-
tional initiators are used for copolymer formation without
intermediate purification steps. Water initiated PCL in the
RAFT mediated reaction was quantified using 1H NMR and
end-group functionalisation as previously described7 (supplemen-
tary information). Surprisingly, the amount of water initiated PCL
was extremely low, typically less than 1%. This is considerably less
than that achieved in analogous reactions involving a bifunctional
initiator for ATRP (typically y 10%).7 This suggests that either
the system used here was much drier than in previous experiments,
or that the bifunctional RAFT initiator has much better access to
the active site of the enzyme than those initiators used previously.
This may be due to beneficial interactions between the trithio-
carbonate moiety of the RAFT agent and the enzyme, such that
RAFT hydroxy groups are always present to act as initiator. In
any case, this high incidence of initiation greatly decreases the
amount of water initiated PCL and hence decreases the propensity
for homo-PCL formation.
Fig. 2 DSC traces showing the decrease in crystallisation temperature
for PCL-b-PSTY with increasing block lengths of PCL in the copolymer
(mole% PCL indicated on figure for each curve with corresponding Entry
from Table 1 in brackets). Note that no crystallisation peak is evident
when PCL contributes 35 mole% or less to the copolymer.
The peak is broadened by the large distribution of PCL block
lengths that are present. Additionally, a small peak is evident at
similar retention time to homo-PCL. This is due to a small amount
of linear and cyclic PCL that is present in the product. Thus,
GPEC data clearly demonstrate that block copolymer is formed.
Another common test for the presence of copolymer is the
hydrolytic degradation of the polyester from the copolymer.7 A
comparison of the GPC traces (Table 1, Entries 2–6) before and
after hydrolysis shows that in all cases, the product after hydrolysis
(PSTY block) had a lower molecular weight than the original
block copolymer (supporting information). This simple test shows
that block copolymer was predominantly formed rather than just a
blended mixture of homopolymers. NMR was used to demon-
strate the complete removal of PCL from the copolymer following
hydrolysis (not shown).
In this communication, we have presented the first simulta-
neous, metal-free synthesis of copolymers utilising enzymes and
RAFT-mediated polymerisation in scCO2. The low cytotoxicity of
the RAFT agent, coupled to the inherent bio-friendly nature of the
enzyme, makes this an important technique for the synthesis of
polymers for biomedical applications. The combination of these
techniques with scCO2 as a solvent makes this approach extremely
desirable from a green synthesis point of view.
This research has been supported by the Dutch Polymer
Institute through postdoctoral funding for KJT. We are also
grateful to Dr Mark F Wyatt at the EPSRC National Mass
Spectrometry Service Centre (Swansea, UK) for MALDI-TOF
analysis. We thank the Marie Curie Research Training Network
(BIOMADE, Contract No. MRTN-CT-2004-505147) for financial
support (JZ and SV). We would also like to express our thanks to
Mr Peter Fields and Mr Richard Wilson for their invaluable help
in developing and building the high-pressure autoclave. SMH is a
Royal Society Wolfson Research Merit Award Holder.
MALDI-TOF data also show that copolymer was indeed
formed (supplementary information). Additionally, there was
evidence for the formation of low levels of homopolymers of
PSTY and PCL. In the case of PCL, both linear and cyclic
homopolymer was observed. This is typical of enzymatic ROP
of e-caprolactone and such a product has been observed
previously.4,7
The effect on the crystallinity of the PCL upon changing the
ratio of the block lengths was also investigated. Crystallisation
exotherms were measured by differential scanning calorimetry
(DSC) for various copolymers (Fig. 2).
Notes and references
1 J. A. Hubbell, Science (Washington, D. C.), 2003, 300, 595.
2 M. Hales, C. Barner-Kowollik, T. P. Davis and M. H. Stenzel, Langmuir,
2004, 20, 10809.
The degree of crystallinity of the PCL block was calculated for
each copolymer and found to decrease with increasing proportions
of PSTY (Table 1). This can be explained by the inability of the
PCL chains to effectively pack into crystallites – i.e. the disruptive
effect of the PSTY block in the copolymer. Indeed, when the
PSTY block length becomes much larger than the PCL block
length (for example in Table 1, Entry 5), the ability of the PCL to
align and form crystallites is totally inhibited and a fully
amorphous copolymer is formed. By contrast, when RAFT
capped PCL and PSTY (50 : 50 wt%) were physically blended
prior to analysis, the degree of crystallinity of pure PCL was not
3 J. Peeters, A. R. A. Palmans, M. Veld, F. Scheijen, A. Heise and
E. W. Meijer, Biomacromolecules, 2004, 5, 1862.
4 M. de Geus, J. Peeters, M. Wolffs, T. Hermans, A. R. A. Palmans,
C. E. Koning and A. Heise, Macromolecules, 2005, 38, 4220.
5 B. A. C. Van As, P. Thomassen, B. Kalra, R. A. Gross, E. W. Meijer,
A. R. A. Palmans and A. Heise, Macromolecules, 2004, 37, 8973.
6 C. J. Duxbury, W. Wang, M. De Geus, A. Heise and S. M. Howdle,
J. Am. Chem. Soc., 2005, 127, 2384.
7 J. Zhou, S. Villarroya, W. Wang, M. F. Wyatt, C. J. Duxbury,
K. J. Thurecht and S. M. Howdle, Macromolecules, 2006, 39, 5352.
8 M. H. Stenzel, C. Barner-Kowollik, T. P. Davis and H. M. Dalton,
Macromol. Biosci., 2004, 4, 445.
9 V. Verhelst and P. Vandereecken, J. Chromatogr., A, 2000, 871, 269.
This journal is ß The Royal Society of Chemistry 2006
Chem. Commun., 2006, 4383–4385 | 4385