Macromolecular

Rapid Communications Communication

Controlled Ring-Opening Polymerization

of Substituted Episulfides for Side-Chain
Functional Polysulfide-Based Amphiphiles

Matthias Kuhlmann, Smriti Singh, Jürgen Groll*

We used initiation solutions of DBU and different thiols for the controlled ring-opening homo-
and copolymerization of ethoxy ethyl thio glycidyl ether (EETGE) and allyl thio glycidyl ether
(ATGE) to side-chain multifunctional polysulfides. Optimized preparation conditions allow
the syntheses of monomodal homopolysulfides and monomodal polysulfide-block-mPEG
copolymers. Furthermore, copolymers of EETGE and mPEG
are firstly synthesized, characterized, and finally deprotected
to yield intact poly(thio glycidol)-block-poly(ethylene glycol)
copolymers. These amphiphiles are suitable to form parti-
cles in aqueous solutions as confirmed by DLS and cryo-SEM
measurements.

1. Introduction hydrophilicity. The oxidative switch of the hydrophobic

Amphiphilic copolymers play an important role in mate-
rial science due to their self-assembly properties in
selective solvents and their potential application in life sci-
ences, for example in drug delivery.[1] Among others, bio-
compatible amphiphilic polyoxamers [e.g., poly(ethylene
oxide) (PEO) and poly(propylene oxide) (PPO) blockstruc-
tures] are in focus for biomedical applications. Napoli
et al.[1] substituted the hydrophobic PPO parts with even
more hydrophobic poly(propylen sulfide) (PPS) parts and
showed their self-assembly in aqueous solution. Introduc-
tion of oxidation-sensitive sulfur into the backbone gener-
ates unique properties of polysulfides with respect to their
M. Kuhlmann, Prof. J. Groll
Department and Chair for Functional Materials
in Medicine and Dentistry, University of Würzburg,
97070 Würzburg, Germany
E-mail: Juergen.groll@fmz.uni-wuerzburg.de
S. Singh
DWI e.V. and Institute of Technical and Macromolecular
Chemistry, RWTH Aachen University, Forckenbeckstr.
50, 52074 Aachen, Germany
PPS part to hydrophilic sulfoxides/sulfones resulting
in hydrophilic block copolymers makes these polymers
promising candidates for the controlled drug delivery in
inflammatory tissues.[2,3] Although PEO–PPO and PEO–PPS
copolymers are promising candidates for biomedical appli-
cations, there is still a lack of side-chain functionality for
further modification (e.g., labeling or the covalent attach-
ment of drugs with cleavable linkers). For linear polyox-
amers, this limit can be circumvented by polymerization of
protected polyglycidols resulting in block copolymers with
tunable removal of their protecting groups.[4] Keeping in
mind, the promising oxidative switch and side-chain func-
tionality, we focused on the sulfur analogue of protected
glycidols. Although there were attempts to polymerize
ethoxy ethyl thio glycidyl ether (EETGE) and allyl thio gly-
cidyl ether (ATGE), no detailed analysis with respect to dis-
persity and deprotection or block copolymers thereof are
reported in literature.[5,6] As the reported deprotection step
of poly(ethoxy ethyl thio glycidyl ether) (PEETGE) under
acidic conditions led to partial polymer degradation, no
further investigations were performed. Deprotection under
milder conditions was performed, but solubility problems
of the supposed poly(thio glycidol) (PTG) led to a lack of

Macromol. Rapid Commun. 2012, 33, 1482−1486

1482 © 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim wileyonlinelibrary.com DOI: 10.1002/marc.201200297

Controlled Ring-Opening Polymerization of Substituted Episulfides. . .
Scheme 1. Synthetic approach for the synthesis of thioglycidols and analogues.
analysis.[5] PTG was directly obtained by polymerization
of hydroxymethyl thiirane, but molecular weight analysis
was only possible after derivatization of hydroxy function-
alities and pure PTG also lacks of analysis. Polymerization
of ATGE was reported once with a water/ZnEt2 system in
benzene, but neither reports on dispersity nor any blockco-
polymers with EETGE have been reported. For polymer-
analogue functionalization of block copolymers, it might
be necessary to polymerize protected PTGs followed by
subsequent deprotection.
Here, we report the controlled ring-opening (block)
copolymerization of ATGE, EETGE, and PEG copolymers
thereof with the well-established thiol/DBU system in
DMF as a solvent. A modified experimental procedure
was used to reduce/avoid disulfide formation during
and after polymerization of episulfides. Mild acidic treat-
ment of PEETGE and PEETGE-block-PEG was used to yield
PTG homopolymers and PTG-block-poly(ethylene glycol)
amphiphiles without any observed chain scission. We
also show that these polymers do form nanoparticulate
aggregates in aqueous solution as measured by DLS and
cryo-SEM.
2. Results and Discussion
2.1. Polymerization
ATGE and EETGE were successfully homo- and copolymer-
ized under nitrogen atmosphere in dimethylformamide
(DMF; Scheme 1). As initiator an organic thiolate species,
obtained via the acid–base reaction of mercaptoethanol
(HOEtSH) with DBU in DMF was used. All polymerizations
were conducted at −20 °C − room temperature (RT) and
were finished within 1 h.1H NMR spectra (Figure S1, Sup-
porting Information) were used to determine the degree of
polymerization (DP) and the results were compared with
SEC measurements in THF (standard: polystyrene) to check
the accuracy of the eluting system on PEETGE and poly(allyl
thio glycidyl ether) systems.
Initiator signals at δ = 2.60 ppm and signals of the
aromatic ring system at δ = 7.45–7.20 ppm were used as
reference and compared with each other. However, both
references have disadvantages. The initiator signal is not
Macromol. Rapid Commun. 2012, 33, 1482−1486
www.MaterialsViews.com © 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Macromolecular
Rapid Communications
www.mrc-journal.de
always clearly separated from the polysulfide backbone
signals, whereas the aromatic ring system in the case of
block copolymers is not always present in all polymer
chains. In the case of oxidative coupling of two polymer
chains through disulfide formation before controlled end
capping by benzylchloride (BzCl), which was observed to
occur in significant amounts if not carefully prevented
through appropriate reaction conditions, only the non-
oxidized polymers can be determined using the aromatic
signals. The less oxidation occurred, the more accurate is
the DP determination via the aromatic system. Therefore,
the DPs as determined by NMR have always been care-
fully compared with the values determined using SEC. In
all experiments, the DPs determined by SEC in THF are
in good agreement with DPs determined by 1H NMR (see
Table 1).
Narrow molecular-weight distributions ranging from
— —
M w/M n = 1.07 −1.12 (see Table 1) were obtained in all
homopolymerization experiments in which polymeriza-
tion was quenched with a low-molecular compound as
BzCl or methylacrylate (MeAc). Literature reports similar
— —
M w/M n ratios for propylene sulfide polymerization ini-
tiated with thiolat species.[7] DPs ranging from 18−65
repeating units were under good control confirmed by 1H
NMR and SEC in THF.
In episulfide polymerization, the active thiolate spe-
cies are sensitive to oxidants and readily form disulfides
on oxidation. If the reactants already contain oxygen
from the beginning, the forming disulfides will result
in an overall decrease in molecular weight due to chain
transfer.[8] While these effects can be avoided by addition
of an in situ reduction reagent,[9] oxygen that is introduced
into the system during the polymerization and especially
during deactivation of the active species will mainly
cause dimerization of polymer chains. In order to avoid
oxygen contamination, all compounds used in this study
were thus degassed before polymerization and the active
species was trapped afterwards by quantitative reaction.
Unfortunately, bimodal distributions initially occurred
in a large number of experiments. Avoiding incorpora-
tion of disulfides via the initiator, mercaptoethanol was
recondensed before usage, but dimers were still forming
as was analyzed by SEC (e.g., Figure S2d, Supporting
1483
 Macromolecular
Rapid Communications M. Kuhlmann et al.

www.mrc-journal.de

Table 1 SEC results of the anionic ring-opening polymerization of EETGE and ATGE with different thiols and DBU as initiating system.
— a) — b) — /M—
Entry Initiator Monomer Terminating M c M n,SEC DPcc) DPNMRd) DPSECe) M w n
agent
1 HOEtSH ATGE BzCl 2343 2300 18 16 18 1.10
2 HOEtSH ATGE BzCl 8215 7600 63 60 58 1.11
3 HOEtSH EETGE BzCl 9783 6700 66 50 48 1.11
4 HOEtSH EETGE mPEG480Ac 3000 2300 17 13 12 1.16
5 HOEtSH EETGE MeAc 8831 8200 59 42 55 1.07
6 HOEtSH ATGE/EETGE BzCl 9038 9000 65 53 65 1.21
7 DTT EETGE BzCl 9528 10 400 2 × 31 n.d. 2 × 34 1.12
8 HOEtSH EETGE mPEG480Ac 2080 2800 11 12 15 1.08
9 HOEtSH “TGE” mPEG480Ac 1800 1700 15 n.d. 13 1.14
10 HOEtSH “TGE” mPEG480Ac 1800 1500 15 n.d. 11 1.07
a)Calculated molecular weight; b)Molecular weight determined by SEC; c)Calculated degree of polymerization; d)Degree of polymerization

determined by NMR; e)Degree of polymerization determined by SEC.

— /M— value of block

Information). Considering the high M w n the mixture. As all compounds have high boiling points
copolymer experiments (entry 6 in Table 1, Figure S3, Sup-
porting Information), it was supposed that the sequen-
tial addition of compounds via syringe-through-septum
is accompanied by addition of small amounts of oxygen.
Homopolymerization requires three syringe-through-
septum steps (thiol-to-base, monomer-to-initiation mix-
ture, termination reaction), and an additional step is nec-
essary for the block copolymerization. The assumption
that the experimental procedure is an important point in
disulfide formation was further confirmed as monomodal
homopolysulfides (see Figure S5, Supporting Informa-
tion) could also be synthesized using mercaptoethanol.
To further exclude disulfide introduction by the initiator,
we used dithiothreitol (DTT)/DBU as a initiating system.
DTT is known for the almost irreversible cyclization upon
oxidation due to its thermodynamically high stable six-
membered ring structure. Therefore, disulfides already
formed in the initiator would solely effect the DP but
do not participate in thiol-disulfide equilibrium during
the polymerization. Disulfide formation still occurred
resulting in polysulfide dimers with doubled molecular
weight (Figure S10, Supporting Information).
As monomodal narrow molecular weight distributions
(MWD) could be obtained with mercaptoethanol and
bimodal distributions were obtained with DTT as the ini-
tiating agent, we conclude that experimental accuracy to
avoid oxygen contamination during the polymerization
is one of the main reasons for disulfide formation in our
experiments. To further reduce the disulfide formation,
all further experiments were conducted in low-pressure
atmosphere to reduce the amount of dissolved gas in
(lowest at bp 153 °C), continuous degassing at reduced
temperature (−20 °C) could be performed. As reported in
the literature, episulfides polymerize fast at RT and even
at −20 °C polymerization is completed within minutes.[7]
Due to basic conditions and active thiolate species,
trapping the active species after polymerization is tre-
mendously important and was performed either by
nucleophilic substitution reactions on an excess of BzCl
or by Michael-addition to an excess of MeAc and mPEG-
acrylate (mPEG480Ac). Whereas quenching the thiolate
species with an aromatic group leads to a good reference
for the determination of the DP, the addition of the hydro-
phobic polysulfide chain to hydrophilic mPEG-acrylate
leads to amphiphilic polymers. Napoli et al.[10] showed the
lyotropic behavior of triblock copolymers obtained with a
PEG-macroinitiator polymerizing propylene sulfide and
quenching the reaction with PEG acrylate. As proof-of-
principle EETGE was polymerized with HOEtSH/DBU in
DMF and finally quenched with low-molecular MeAc as a
low-molecular Michael-acceptor (Table 1, entry 5). Narrow
molecular weight distribution and defined molecular
weight encouraged us using a mPEG-acrylate as a termi-
nating agent. Monomodal SEC trace (Figure S4, Supporting
Information) confirmed the efficient combination of con-
tinuous degassing procedure and polymeric acrylates as
the terminating agent for episulfide polymerization.
2.2. Hydrolysis
Polysulfide degradation during acidic PEETGE acetal hydrol-
ysis reported by Taton et al.[5] did not occur in our study

Macromol. Rapid Commun. 2012, 33, 1482−1486

1484 © 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.MaterialsViews.com
 Macromolecular
Controlled Ring-Opening Polymerization of Substituted Episulfides. . . Rapid Communications
www.mrc-journal.de

Figure 1. 1H NMR spectra of PEETGE before (above) and after (below) hydrolysis of acetal side chains resulting PTG. The area of polysulfide
backbone at δ = 3.05–2.7 ppm stays unaffected. Mercaptoethanol as the initiator was used as reference.

due to the use of milder reaction conditions. Using acetal 2.3. Particle Formation
hydrolysis of PEETGE in slightly acidified THF at RT, we were
As known for nanoparticles, the solely use of one analytical
able to obtain PTG and poly(thio glycidol)-block-(ethylene
method is not sufficient. Therefore, the supposed aggrega-
oxide) (PTG-block-mPEG) copolymers. After workup, the
tion of amphiphilic PTG-block-mPEG particles was studied
reported solubility problems were overcome by NMR anal-
by cryo-SEM and DLS. DLS measurements (Figure 3a) show
ysis in DMSO-d6. 1H NMR of PTG (Figure 1) proved the com-
plete removal of acetal functions and (with the initiator as
D=1,14
reference) the maintenance of the polysulfide backbone. D=1,07
D=1,08
1,0
As can be seen in Figure 1, parts of the backbone signals
(i.e., the CH-group) at δ ≈ 3.0 ppm are highfield shifted to
0,8
δ = 2.77 due to side-chain hydrolysis, whereas the overall
normalized ΔRI

integrated area confirms intact backbone.

SEC of PEETGE-block-mPEG copolymers in THF further 0,6

proved the cleavage-free PEETGE hydrolysis. The peak

signal of hydrolyzed copolymer (see Figure 2) is found 0,4
at M = 1700 g mol−1, with a theoretical M = 1800 g mol−1
(DPEETGE = 15, DPPEG = 11). PEETGE-block-mPEG can be found 0,2
at M = 3100 g mol−1 with a theoretical M = 2700 g mol−1.
In Figure 2, the acetal-protected copolymer (dotted line) is 0,0
compared to partially hydrolyzed copolymer (dashed line) 100 1000 10000
and PTG-block-mPEG (solid line). We suppose the slight molecular weight / g mol
-1

difference in the determined molecular weight behavior

Figure 2. SEC curves of PEETGE-block-mPEG copolymers (dotted
is due to different polymer−column interactions resulting line) with partial (dashed line, after 18 h of hydrolysis) and
from either acetal side chains or hydroxymethyl side fully hydrolyzed acetal groups (solid line, after more than 48 h
chains from the PTG moiety. hydrolysis).

Macromol. Rapid Commun. 2012, 33, 1482−1486

www.MaterialsViews.com © 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim 1485
 Macromolecular
Rapid Communications
www.mrc-journal.de
80
60
intensity / a.u.
40
20
0
1 10 100
diameter / nm
Figure 3. (a) DLS measurement of PTG-block-mPEG particles in
H2O and (b) cryo-SEM of particles formed upon dilution of PTG-
block-mPEG in D2O.
(by intensity distribution) the existence of particles in the
size range of 50−200 nm. As bigger particles give higher
intensities in DLS measurements, the volume distribu-
tion had to be used for analyzing the second species. For a
rough estimation, a mean diameter of 2.5 nm can be recal-
culated into diffusion coefficient (D = 1.7 × 10−10 m2 s−1)
and is in the range of single polymeric molecules [e.g., PEG
with Mn = 3400 g mol−1 (D = 9.9 × 10−11 m2 s−1)[1] indicating
still a large amount of single polymeric PTG-block-mPEG as
impurities.
Cryo-SEM pictures (Figure 3b) were analyzed and
showed particles with size ranges of 40−200 nm in dia-
meter. These pictures prove the broad distribution
obtained in DLS measurements for the bigger species.
Unfortunately, resolution restriction in cryo-SEM prevents
analysis of the small particle fraction. However, the max-
imum micelles diameter of our synthesized system should
be at approximately 20 nm and from both analytics, it can
1486 © 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
M. Kuhlmann et al.
be assumed that larger aggregates are formed instead of
micelles. A more detailed study on these aggregates and
on process control to facilitate micelle formation is cur-
intensity distribution
volume distribution
rently under investigation.
3. Conclusion
We used the established initiating system thiol/DBU
and successfully (co)polymerized EETGE and ATGE with
reduced appearance of disulfide formation using a con-
tinuous degassing procedure. Mild acetal hydrolyzes of
1000 10000
PEETGE homopolymers and mPEG-copolymers thereof
were successfully performed yielding PTG and sym-
metric poly(thio glycidol)-block-poly(ethylene glycol). No
polysulfide backbone degradation occurred during depro-
tection of PEETGE and PEETGE-block-mPEG. 1H NMR gave
hints of aggregate formation in aqueous solutions that
were confirmed by DLS and cryo-SEM measurements
indicating vesicle formation.
Supporting Information
Supporting Information is available from the Wiley Online Library
or from the author.
Received: April 26, 2012; Revised: June 10, 2012; Published online:
July 25, 2012 ; DOI: 10.1002/marc.201200297
Keywords: amphiphiles; functional polysulfides; self-assembly;
thioglycidol
[1] A. Napoli, N. Tirelli, G. Kilcher, J. A. Hubbell, Macromolecules
2001, 34, 8913.
[2] A. Napoli, M. Valentini, N. Tirelli, M. Muller, J. A. Hubbell,
Nat. Mater. 2004, 3, 183.
[3] A. Rehor, N. E. Botterhuis, J. A. Hubbell, N. A. J. M. Sommerdijk,
N. Tirelli, J. Mater. Chem. 2005, 15, 4006.
[4] M. Backes, L. Messager, A. Mourran, H. Keul, M. Möller,
Macromolecules 2010, 43, 3238.
[5] D. Taton, A. Le Borgne, M. Sepulchre, N. Spassky, Macromol.
Chem. Phys. 1994, 195, 139.
[6] J. Lal, J. Polym. Sci., B Polym. Lett. 1965, 3, 969.
[7] E. Nicol, C. Bonnans-Plaisance, G. Levesque, Macromolecules
1999, 32, 4485.
[8] G. Kilcher, L. Wang, N. Tirelli, J. Polym. Sci., Part A: Polym.
Chem. 2008, 46, 2233.
[9] L. Wang, G. Kilcher, N. Tirelli, Macromol. Chem. Phys. 2009,
210, 447.
[10] A. Napoli, N. Tirelli, E. Wehrli, J. A. Hubbell, Langmuir 2002,
18, 8324.
Macromol. Rapid Commun. 2012, 33, 1482−1486
www.MaterialsViews.com