Amino functional poly(ethylene glycol) copolymers via protected amino glycidol

Article abstract of DOI:10.1021/ma902245d

The synthesis of poly (ethylene glycol) (PEG) copolymers with multiple amino functionalities within the chain is described, relying on an epoxide comonomer bearing a protected amino group. N,Ndibenzyl amino glycidol (DBAG) and ethylene oxide (EO) were copolymerized via anionic polymerization, leading to well-defined polymers with varied comonomer content and low polydispersities (M_w/M_n in the range of 1.1 to 1.2). Subsequent hydrogenolysis with Pearlman's catalyst afforded polyethylene glycol-coamino glycerols (PEG-coPAG) with a precisely adjusted number of randomly incorporated amino groups in the range of 2-15%. For the first time, the kinetics of an. EO copolymerizations have has been directly monitored by ¹H NMR spectroscopy in real time. Monomer consumption and compositional drift in monomer feed have been studied for various reaction temperatures, revealing a slightly tapered yet random DBAG distribution in the copolymers. The random structure of the copolymers was confirmed by detailed ¹³C NMR characterization of EO- and DBAG-centered triad sequence distribution and DSC measurements.

Full text of DOI:10.1021/ma902245d

2
244 Macromolecules 2010, 43, 2244–2251
DOI: 10.1021/ma902245d
Amino Functional Poly(ethylene glycol) Copolymers via Protected
Amino Glycidol
Boris Obermeier, Frederik Wurm, and Holger Frey*
Institute of Organic Chemistry, Organic and Macromolecular Chemistry, Duesbergweg 10-14,
Johannes Gutenberg-Universit a€ t Mainz, D-55099 Mainz, Germany
Received October 9, 2009; Revised Manuscript Received January 22, 2010
ABSTRACT: The synthesis of poly(ethylene glycol) (PEG) copolymers with multiple amino functionalities
within the chain is described, relying on an epoxide comonomer bearing a protected amino group. N,N-
dibenzyl amino glycidol (DBAG) and ethylene oxide (EO) were copolymerized via anionic polymerization,
w n
leading to well-defined polymers with varied comonomer content and low polydispersities (M /M in the
range of 1.1 to 1.2). Subsequent hydrogenolysis with Pearlman’s catalyst afforded poly(ethylene glycol-co-
amino glycerol)s (PEG-co-PAG) with a precisely adjusted number of randomly incorporated amino groups in
the range of 2-15%. For the first time, the kinetics of an EO copolymerizations have has been directly
1
monitored by H NMR spectroscopy in real time. Monomer consumption and compositional drift in
monomer feed have been studied for various reaction temperatures, revealing a slightly tapered yet random
DBAG distribution in the copolymers. The random structure of the copolymers was confirmed by detailed
C NMR characterization of EO- and DBAG-centered triad sequence distribution and DSC measurements.
13
Introduction
synthesis of PEGs with different terminal functional groups
including carboxylic, amine, thiol, aldehyde, tosyl, epoxide, or
Poly(ethylene oxide) (PEO), often referred to as poly(ethylene
glycol) (PEG) for molecular weights below 20 000 g/mol, is the
established reference polymer in pharmaceutical and biomedical
applications today because it shows (i) high solubility in water
and most organic solvents, (ii) no immunogenicity, antigenicity,
and toxicity, and (iii) high flexibility and hydration of the main
8
,16,17
succinimidylsuccinate groups.
Various end-group-functio-
nalized PEGs are already commercially available, including the
1
8,19
amino-functional Jeffamines.
In general, the reported syn-
thetic strategies are either based on the modification of commer-
cial hydroxyl functional PEG diol or PEG monomethyl ether
(MPEG) or rely on the use of an initiator bearing a protected
functional group and an appropriate termination agent. The
latter strategy has been particularly valuable for the preparation
1-3
chain.
As a consequence, PEG is employed as a common
component by the pharmaceutical and cosmetic industry for a
wide range of applications and plays an important role in many
nonionic surfactants. On the basis of the pioneering works of
Davis and Abuchowski in 1977 on the covalent conjugation of
PEG to pharmaceutical agents, now known as “PEGylation”, the
possible applications of PEG have developed from a biocompa-
tible additive for healthcare applications to a powerful drug
modification tool. By increasing the molecular weight of drugs
and because of shielding from proteolytic enzymes, conjugation
with PEG permits us to overcome many deficiencies, particularly
of protein and peptide drugs. The consequences of PEGylation
are increased water solubility, enhanced plasma half-lifes, protec-
tion toward degrading enzymes, and reduced immunogenicity
2
0-26
of R,ω-heterobifunctional PEGs.
Recently, an elegant syn-
thetic access to these structures has been reported, relying on a
different approach, that is the N-heterocyclic carbene-induced
27
polymerization of EO.
To date, the major strategies to overcome the intrinsically low
loading capacity of PEG have been either dendronization of
4
-6
2
8,29
30-33
PEGs
or the synthesis of star- or dendrimer-like PEGs
that often requires demanding multistep synthetic procedures. In
addition, all methods described above for mono-, di-, or multi-
functionalization are limited to the chain end. The PEG back-
bone itself is left unaltered. The only synthetic approach to PEG
with several amino groups has been presented by Koyama et al.
In post polymerization reactions, random copolymers of EO and
allyl glycidyl ether were treated with 2-aminoethanethiol to yield
PEG derivatives with pendant primary amino groups linked to
7
,8
and antigenicity. Additionally, PEG monomethyl ether with a
molecular weight of 5000 g/mol is currently the most widely used
soluble polymeric supportfor catalysts (and reagents), combining
the homogeneous reaction kinetics of low-molecular-weight
compounds with the advantageous separation properties of
3
4
the PEG backbone via ether- and thioether bonds. Yoshikawa
et al. demonstrated this polymer to induce highly effective
9-12
heterogeneous catalysts.
35
compaction of long duplex DNA. With respect to multifunc-
tional PEG derivatives, linear copolymers of EO and glycidol to
date represent the most relevant strategy for the preparation of
PEG derivatives with functional groups along the backbone. In
this case, the functionality is restricted to hydroxyl groups. To
obtain linear copolymers of glycidol, the hydroxyl group must be
protected. A suitable protective group has to be stable under the
harsh basic conditions of anionic polymerization, has to show
little chain transfer reactions, and has to be cleavable in a facile
reaction to guarantee complete removal in subsequent polymer
A major drawback for many of these applications results from
the fact that PEG possesses no functional groups within the
chain. Commonly synthesized via anionic ring-opening polym-
erization of EO in the presence of alkali metal alkoxide initiators,
linear PEGs are limited to one or two hydroxyl groups at the
13-15
chain ends.
Because coupling reactions most often require
tailored functionalities, research efforts have been focused on the
*Corresponding author. E-mail: hfrey@uni-mainz.de.
pubs.acs.org/Macromolecules Published on Web 02/15/2010
r 2010 American Chemical Society

Article
Macromolecules, Vol. 43, No. 5, 2010 2245
modification reactions of the linear polymer. The ethoxy ethyl-
acetal protective group first applied to glycidol by Fitton et al.
over molecular sieve) were purchased from Aldrich. Deuterated
chloroform-d and DMSO-d were purchased from Deutero
1 6
3
6,37
fulfills these requirements.
polymerization of ethoxy ethyl glycidyl ether (EEGE) in 1994.
EEGE has been employed for the preparation of high-molecular-
Taton et al. reported the first
GmbH. Other solvents and reagents were purchased from Acros.
2-Methoxyethanol was cryotransferred from calcium hydride prior
to use. All other solvents and reagents were used as received.
General Procedures. N,N-Dibenzyl-3-chloro-2-hydroxypro-
pylamine (DBAG) (1a) Synthesis. Dibenzylamine (50.0 g, 253
mmol, 1 equiv) and 24.1 g of epichlorohydrin (26.1 mol, 1.03
equiv) were dissolved in 350 mL of methanol and stirred for 16 h.
Solvent and excess of epichlorohydrin were removed under
reduced pressure to obtain a pale-yellow oil (yield >90%).
38
3
9
weight random EO copolymers, copolymers with orthogonal
40
protective groups, and block copolymers with ethylene oxide,
4
1-43
propylene oxide (PO), and styrene.
As an interesting alter-
native to EEGE, the monomer isopropylidene glyceryl glycidyl
ether (IGG), providing a primary and a secondary hydroxyl
44
group after deprotection, was recently developed in our group.
The crude product is sufficiently pure to be directly converted.
H NMR (300 MHz, CDCl
Huang and coworkers described several polymer structures based
on multi-hydroxy functional PEGs such as amphiphilic PEO-
1
3
, δ): 7.42-7.32 (m, 10H, arom.),
3
.95-3.85 (m, 1H, CH), 3.80 (d, 2H, CH Ph), 3.53 (d, 2H,
4
5
2
graft-polystyrene copolymer brushes and amphiphilic macro-
cyclicgraftcopolymersbased on a multi-hydroxy functional PEG
CH
m, CHCH
N,N-Dibenzyl Amino Glycidol (1b) Synthesis. 0 (44.0 g, 152
mmol) was dissolved in 200 mL of tert-butanol. KOH (11.6 g,
76 mmol) dissolved in a minimal volume of water was added,
2
2
Ph), 3.50-3.46 (m, 2H, CH Cl), 3.26 (br s, OH), 2.74-2.55
(
2
N).
4
6
macrocycle and poly(ε-caprolactone) chains. Very recently, Li
and Chau reported a library of multifunctionalized PEGs prepared
by polymer modification reactions of functional PEGs with
1
47
hydroxyl groups via multiple synthetic steps. However, this
approach toward at least some of the multifunctional PEG
derivatives can be considered to be problematic because polymer
modification reactions often result in limited overall conversion.
Herein we describe a straightforward synthetic approach to
multi-amino functional PEGs using a novel epoxide monomer
with a protected amino group. The monomer N,N-dibenzyl
amino glycidol (DBAG) (1b) can be viewed as an amino analogue
of glycidol with the amino functionality protected by two benzyl
groups, guaranteeing high stability under oxyanionic polymeri-
zation conditions. The present study is focused on copolymers of
and the reaction mixture was stirred for 16 h. KCl was filtered
off, and the solvent removed under reduced pressure. The
-
2
product has been purified by distillation at 190 °C, 2 ꢀ 10
3
mbar, bp 150-160 °C and was obtained as a slightly yellow oil.
þ
1
FDMS: m/e 253 (100%, M ). H NMR (300 MHz, CDCl
7.42-7.32 (m, 10H, arom.), 3.84 (d, 2H, CH Ph), 3.60 (d, 2H,
CH Ph), 3.13-3.04 (m, 1H, CH), 2.85-2.67 (m, 2H, CH ring),
N). C NMR (75.5 MHz, CDCl , δ):
39.28, 128.73, 128.18, 126.19 (aromatic), 58.82 (CH Ph); 55.78
, δ):
3
2
2
2
1
3
2
1
.51-2.42 (m, CHCH
2
3
2
(
2 2
CH); 51.01 (CH ring); 44.90 (CHCH N).
Copolymerization of EO and DBAG. Exemplified for P(EG95
co-DBAG ): 352 mg of cesium hydroxide monohydrate (2.1
-
5
EO and DBAG with fractions of amino groups in the range of
1
mmol) were suspended in benzene in a dry Schlenk flask under
an argon atmosphere, and 160 mg of 2-methoxyethanol were
added. Stirring at 60 °C for 30 min and evacuation at 90 °C for
2
-15% to retain a mainly PEG-type structure. Online H NMR
spectroscopy has been used to monitor the copolymerization of
EO with the novel functional comonomer and proved to be a
powerful tool for the investigation of the epoxide copolymeriza-
tion kinetics. Monomer consumption and compositional drift in
monomer feed have been investigated in full detail for various
3
h afforded the cesium alkoxide. Dimethyl sulfoxide (50 mL)
and 50 mL of dioxane were added to the evacuated flask. After
dissolution, the flask was cooled to -40 °C, and 10 mL of EO
(
200 mmol) was cryotransferred from a graduated ampule to the
1
3
reaction temperatures. Employing C NMR spectroscopy, an
analysis of both the EO- and DBAG-centered triad sequence
distribution has been carried out. Thermal behavior has been
investigated by differential scanning calorimetry (DSC) both for
the benzyl-protected as well as for the final amino functional
PEG copolymers.
frozen solvents. DBAG (2.66 g, 10.5 mmol) was added via a
syringe. The mixture was rapidly warmed up to 70 °C, and
polymerization was performed in vacuo for 9 h. The polymer-
ization was terminated by the addition of methanol and acidic
ion-exchange resin. Filtration, removal of solvent, and precipi-
tation in cold diethyl ether resulted in the pure (co)polymer
(
yields >90%). For polymers with DBAG content>10%, the
polymer solution was dried in vacuo after filtration.
Deprotection. P(EG-co-DBAG) (500 mg) was dissolved in a
Experimental Section
1
Instrumentation. H NMR spectra (300 MHz) and C NMR
13
spectra (75.5 MHz) were recorded using a Bruker AC300.
Kinetic H NMR studies were conducted on a Bruker
solvent mixture of 20 mL of MeOH, 10 mL of H
THF. Pd(OH) /C (500 mg, 20%) was added, and the reaction
2
2
O, and 10 mL of
1
AMX400 apparatus. All spectra were referenced internally to
residual proton signals of the deuterated solvent. Field desorp-
tion mass spectra were measured using a Finnigan MAT 95. For
SEC measurements in DMF (containing 1 g/L of lithium
bromide as an additive), an Agilent 1100 Series was used as an
2
mixture was stirred at 8 bar H . We measured the progress of the
hydrogenation by drawing a sample every 12 h. The samples were
filtered, dried under high vacuum, and studied by H NMR with
1
respect to aromatic signals. The addition of acetic acid promoted
the benzyl-group cleavage. When the benzyl group cleavage
exceeded 95%, the catalyst was separated by filtration over Celite,
and the solvent was removed under reduced pressure to obtain the
PEG-co-PAG copolymers. Strong adhesion of the multi-amino
functional copolymers to the heterogeneous catalyst on activated
charcoal sometimes resulted in polymer yields as low as 30%.
Application of high hydrogen pressure is currently studied as a
means toreduce the amount of catalyst, whilemaintaining reason-
able reaction times and improved yields.
6
integrated instrument, including a PSS HEMA column (10 /
1
5
0 /10 g/mol), a UV (275 nm) detector, and a RI detector.
4
Calibration was carried out using poly(styrene) standards pro-
vided by Polymer Standards Service. DSC measurements were
carried out on a Perkin-Elmer 7 series thermal analysis system
and a Perkin-Elmer Thermal Analysis Controller TAC 7/DX
in the temperature range from -100 to 100 °C at heating rates of
-
1
1
0 K min under nitrogen.
3
Reagents. Ethylene oxide (99.5%) was purchased from Al-
9
Derivatization Reaction. P(EG90-co-AG ) (80 mg, 0.17 mmol-
drich. Cesium hydroxide monohydrate (99.95%), 2-methox-
yethanol (99.8%), epichlorohydrin (99%), potassium hydro-
xide, tert-butanol (99.5%), acetic anhydride (99þ%), potassium
hydroxide (85%), and acetic acid (99.7%) were purchased from
Acros Organics. Dibenzylamine, (97%), palladium hydroxide
on activated charcoal (puriss, moistened with water), dimethyl-
sulfoxide (puriss, over molecular sieve), and dioxane (puriss,
NH ), 173 mg of acetic anhydride (1.7 mmol), and 220 mg (1.7
2
mmol) of N,N-diisopropylethylamine were dissolved in 2 mL of
dry THF and stirred for 3 h. Dialysis (MWCO = 1000 g/mol)
1
against MeOH yielded the derivatized polymer in 86% yield. H
NMR (300 MHz, DMSO-d6, δ): 7.75 (br s, 1H, NH), 3.50
(backbone), 3.24 (s, 3H, H
CH N), 1.80 (s, H CON).
3
CO), 3.21-3.12, 3.11-2.98 (m, 2H,
2
3

2
246 Macromolecules, Vol. 43, No. 5, 2010
Obermeier et al.
1
1
H NMR Kinetics. In a conventional NMR tube, DMSO-d
6
composition, as determined by H NMR analysis. SEC
characterization showed monomodal molecular weight dis-
tributions with generally low PDIs in the range of 1.1 to 1.2.
Figure 1 shows representative SEC traces for P(EG -
solutions of the initiator and a mixture of DBAG and EO at a
concentration of ca. 150 g/L were separately frozen under an
argon atmosphere. High vacuum was applied, and the tube was
flame-sealed while the solutions were kept frozen. To reduce the
necessary time for locking and shimming of the polymerization
mixture, we measured a sample of the pure monomer mixture
in advance at the relevant temperature. Immediately after
melting and mixing, the first spectrum was recorded. Sample
spinning was turned off. Intervals between two measurements
were 30 s within the first 6 min and extended afterward.
90
DBAG ), using both a refractive index (RI) as well as a
9
UV detector. The overlap of the distribution modes obtained
with different detection methods evidences homogeneous
incorporation of DBAG units within the polymer chains
for all molecular weights of the molecular weight distri-
bution.
It is worth mentioning that a molar fraction of 16%
DBAG (run 4, Table 1) corresponds to a mass fraction of
3% DBAG because of the high molecular weight of DBAG
Results and Discussion
5
Monomer Synthesis. The synthesis of the comonomer
DBAG (1b) relies on a straightforward two-step procedure
presented in Scheme 1 and modified in comparison to a
compared with EO. An increase in DBAG content leads to
increasing hydrophobicity. In SEC analysis in DMF, this
effect corresponds to a decrease in the relative hydrodynamic
volume. Therefore, the molecular weight determined via
SEC versus theoretical mass also decreases (Table 1). Abso-
lute molecular weights were determined via H NMR spec-
troscopy. Because of a certain signal overlap, the error of this
method is estimated to be (10%.
4
8
previously reported synthesis. Epichlorohydrin is reacted
with dibenzylamine, and subsequent recyclization of the
epoxide ring leads to DBAG. Because this compound has
been used as a monomer for anionic ring-opening polymer-
ization in this study for the first time, it was important to
establish that the DBAG monomer can actually be purified
1
In addition to the copolymers listed in Table 1, PDBAG
homopolymers were also accessible, showing a monomodal,
narrow molecular weight distribution and polydispersities
M /M below 1.3. However, the degree of polymerization
-
2
by distillation at 190 °C and 2 ꢀ 10 mbar. It is obvious that
rigorous purification is essential for the use of DBAG in oxy-
anionic polymerization.
w
n
Synthesis of Random P(EG-co-DBAG) Copolymers. As
illustrated in Scheme 2, EO and DBAG were copolymerized
using the cesium salt of methoxy ethanol as an initiator. To
generate polymers that can be viewed as slightly modified
amino functional PEGs, we focused on molar fractions of
DBAG between 2 and 15% and theoretical degrees of
appears to be limited for the homopolymerization of DBAG.
We ascribe this to transfer reactions, an effect that is known
to represent a limiting parameter for achieving high molec-
ular weights in various epoxide polymerizations.
Deprotection of Amino Groups. Removal of the benzyl
protective groups, that is, liberation of the primary amino
groups, was achieved by catalytic hydrogenolysis. However,
already for low-molecular-weight compounds, the cleavage
of N-benzyl bonds is known to be often slow and not
polymerization (DP ) of 100 to preserve the characteristics
n
of PEG and provide multiple amino groups at the polyether
backbone at the same time. The DBAG content in the final
polymers could be precisely adjusted via the monomer feed
4
9
quantitative. The polymer characteristics of P(EG-co-
DBAG)s and the need for high conversions in this type of
polymer modification reaction impede the removal of the
protective groups even more. From the comparison of a
series of hydrogenation catalysts, Pearlman’s catalyst (20%
Scheme 1. Synthesis of the Protected Epoxide Comonomer
N,N-Dibenzyl Amino Glycidol (DBAG, 1b)
Pd(OH) on carbon) was found to be superior to Pd/C for
2
cleaving N-benzyl groups. Full deprotection of P(EG-co-
DBAG)s was achieved, resulting in the P(EG-co-AG) multi-
5
0,51
cleaving N-benzyl groups.
Hydrogenolysis was per-
formed for 1-8 days with mass equivalents of Pearlman’s
catalyst, 8 bar H in a solvent mixture of MeOH/H O/THF
2
2
2
/1/1 to guarantee solubility of the protected and depro-
tected polymers. Acidification with acetic acid was observed
to promote hydrogenolysis. Because of adhesion of the
multi-amino functional P(EG-co-AG)s on the carbon-sup-
ported catalyst, polymer yields were generally in the range of
3
0-50%. Current work in our laboratories is directed at
Scheme 2. Reaction Sequence for the Synthesis of Multi-Amino Functional PEG Derivatives Composed of Ethylene Oxide and Amino Glycidol
P(EG-co-AG)s
(

Article
Macromolecules, Vol. 43, No. 5, 2010 2247
improving the hydrogenolysis procedure and the respective
yields.
From H NMR characterization, a conversion of
The presence of free amino groups was also indicated by a
positive ninhydrin test. This test was negative for both the
protected polymer structure and the DBAG monomer. As
reported previously for the removal of benzyl groups from an
amino functional initiator for PEG, no chain cleavage or
other side reactions are expected to occur during hydroge-
nolysis, and the molecular weight distribution is expected to
1
9
5-100% for the hydrogenolysis reaction was obtained by
13
integration of the aromatic region. C NMR characteriza-
tion confirmed complete removal of the protective groups by
absence of resonances in the spectral region typical for
1
52
phenyl protons. Typical H NMR spectra for a copoly-
mer before and after deprotection are shown in Figure 2.
remain unchanged.
To demonstrate the accessibility of the primary amino
groups at the PEG chain for bioconjugation, P(EG -AG )
was derivatized with acetic acid anhydride in a model reac-
90
9
1
tion (Scheme 3). H NMR spectra confirmed complete
modification and attachment of 9 acetamide moieties. (See
the Supporting Information). Furthermore, the ninhydrin
test was negative after this modification, confirming the
absence of residual primary amino groups. SEC confirmed
Scheme 3. Derivatization Reaction of P(EG90-co-AG
9
) with Acetic
Acid Anhydride As Model Compound
Figure 1. Typical SEC traces (DMF) for P(EG90-co-DBAG
750 g/mol, M /M = 1.17) obtained with RI and UV detector,
respectively.
9 n
) (M =
5
w
n
1
Table 1. Characterization Data for Random Copolymers of Ethylene Oxide (EO) and N,N-Dibenzyl Amino Glycidol (DBAG) Obtained from H
NMR and SEC
a
run composition (th) % DBAG (th) % DBAG (NMR) composition (NMR)
a
a
b
b
M
n
(th) (g/mol)
M
n
(NMR) (g/mol)
M
n
(SEC) (g/mol)
w n
M /M
1
2
3
4
P(EG98-DBAG
P(EG95-DBAG
P(EG88-DBAG
2
5
9
)
)
)
2.00
5.00
1.80
4.27
P(EG110-DBAG
P(EG112-DBAG
2
)
)
4860
5440
5750
7540
5350
6200
6200
8300
5460
4950
4510
4710
1.15
1.14
1.17
1.18
5
10.00
15.00
9.09
16.13
P(EG90-DBAG
9
)
P(EG85-DBAG15
)
1
P(EG78-DBAG15
)
a
b
6 n
Determined from H NMR (300 MHz, DMSO-d ). M determined by SEC-RI in DMF.
1
Figure 2. H NMR spectra (300 MHz, DMSO-d
6
) of a random copolymer of ethylene oxide and N,N-dibenzyl amino glycidol before (black) and after
deprotection (green) with Pearlman’s catalyst, revealing complete removal of benzyl protective groups (blue ellipse).

2
248 Macromolecules, Vol. 43, No. 5, 2010
Obermeier et al.
the narrow molecular weight distribution after this transfor-
mation.
Online H NMR Copolymerization Kinetics. DBAG is an
manner represent a convenient technique for detailed inves-
tigation of monomer incorporation. Additionally, H NMR
1
1
spectroscopy itself is a standard technique in polymer ana-
lysis, and therefore, no additional instrumentation is re-
quired. In this experiment, for the first time, the growth of
the PEO backbone, manifested by a signal at 3.5 ppm, has
been monitored in real time, as illustrated for the copolym-
amine-containing epoxide monomer with two benzyl groups
in close proximity to the epoxide ring. No comparable
epoxide monomer structures have been mentioned in litera-
ture to date. Therefore, investigation of the reactivity of
DBAG with respect to copolymerization kinetics and a
detailed analysis of the evolution of the copolymer structure
in the course of the polymerization seemed to be crucial.
From a comparison with the copolymerization of EO and
PO, a higher reactivity of EO compared with DBAG had to
be expected. A drastic difference of reactivity would lead to
gradient or blocklike copolymers rather than the targeted
random copolymers.
erization of EO and DBAG at 50 °C in DMSO-d (Figure 3).
6
Monomer consumption and compositional drift in mono-
mer feed was followed by the decrease in the signal of the
methine proton of DBAG at 3.02 ppm and the signal of the
EO protons at 2.61 ppm, referenced internally to the integral
of the aromatic DBAG signals, which remain constant
throughout the polymerization. For the P(EG-co-DBAG)
copolymer synthesis in a flask, the addition of dioxane to
reduce the amount of high boiling DMSO did not have an
influence on polymer structure but permitted more conve-
nient purification of the copolymers.
To this end, we elaborated an experimental method,
permitting us to monitor the copolymerization of EO and
DBAG by online H NMR spectroscopy with minimum
1
1
intervals as short as 30 s between two measurements. In a
conventional NMR tube, DMSO-d solutions of the initiator
H NMR experiments were conducted at 25, 50, 60, and
70 °C with molar fractions of DBAG of 15%. The tempera-
ture dependence of the monomer conversion for 25, 50, and
60 °C is illustrated in Figures 4 and 5. Polymerization at
25 °C required 13 h for full monomer consumption. At 50 °C,
the polymerization rate was enhanced considerably. After
52 min, no monomer signals could be detected any more,
proving total consumption. At 60 °C, complete conversion
was already observed after 13 min. Interestingly, a short
inhibition interval was observed for 50 and 60 °C (Figure 5).
6
and a mixture of DBAG and EO were separately frozen
under an argon atmosphere. High vacuum was applied, and
the tube was sealed, while keeping the solutions frozen.
Immediately after melting and mixing, the first NMR spec-
trum was recorded. Compared with hitherto reported meth-
ods of monitoring polymerizations of the difficult-to-handle
gaseous EO, like gas chromatographic analysis of the recir-
5
3 1
culating monomers,
H NMR spectra recorded in this
1
Figure 3. Time-resolved 400 MHz H NMR spectrum and details with relevant DBAG (3.02 ppm), EO (2.61 ppm), and backbone signals for
6
copolymerization of ethylene oxide and N,N-dibenzyl amino glycidol (15%) at 50 °C monitored in DMSO-d for 52 min.

Article
Macromolecules, Vol. 43, No. 5, 2010 2249
Figure 4. Monomer conversion versus time plots for copolymerization
of ethylene oxide and N,N-dibenzyl amino glycidol (15%) at 25, 50,
and 60 °C in DMSO-d .
6
Figure 6. Percentage of initial monomer concentration of ethylene
oxide and N,N-dibenzylamino glycidol (ca. 15%) versus conversion
for copolymerization at 25, 50, and 60 °C measured in DMSO-d .
6
homopolymers and with various calculated spectra. The
signals were assigned as follows: (1) D-E -D, (2) D-E -E,
b
b
(
3) E-E -D, (4) E-E -E þ E-E -E, (5) E-E -D, and (6) D-E -
b
a
b
a
a
D þ D-E -E, where a and b, respectively, correspond to the
a
first and second CH groups of the central EO unit. Signals A
2
and B most probably stem from EO units at both termini.
The signal assignment is in good agreement with data for
EO-centered triads in random EO-PO copolymers. The
triad sequence distribution for DBAG can be evaluated via
the C signal of the tertiary carbon at 77 ppm (Figure 7, left):
(I) D-D-E, (II) E-D-D, and (III) E-D-E. D-D-D triads were
Figure 5. Monomer conversion versus time plots for copolymeriza-
tion of ethylene oxide and N,N-dibenzylamino glycidol (ca. 15%) at
13
5
6
0 and 60 °C in DMSO-d .
not observed. It is worth noting that DBAG units appear to
be incorporated almost exclusively in a head-to-tail orienta-
tion. Signals due to irregular addition mode (H-H or T-T)
were not observed. Because of partial signal overlap, quan-
tified analysis of inverse gated spectra was not performed.
However, the high intensities of E-E-E and E-D-E triads as
well as the missing D-D-D signal and low intensity for the D-
At 70 °C, no monomer signals were detected any more after
min. This very rapid polymerization is particularly remark-
7
able because stirring of the reaction mixture is impossible
inside the NMR tube and spinning was turned off. Effects of
monomer diffusion from and into the gas phase could not be
observed.
In contrast with the characterization of quenched samples
taken from the reaction mixture, strictly speaking, this NMR
method does not account for chain transfer. However, SEC
analysis for the NMR polymerizations demonstrated narrow
molecular weight distributions pointing to virtually no chain
transfer.
Figure 6 illustrates the evolution of monomer feed com-
position, revealing initially faster incorporation of EO
than DBAG into the growing polymer chain. Throughout
the polymerization, the molar ratio of DBAG units in the
polymer chain is slightly below the initial ratio of the
monomer feed and rises toward the final stages of the
copolymerization to the composition of the initial feed.
According to the NMR data obtained at different copo-
lymerization temperatures, the relative reactivities of EO
and DBAG were not measurably influenced by tempera-
ture. It is also important to emphasize that both monomers
are present until the end of the polymerization. There-
fore, by the copolymerization of EO and DBAG, random
copolymers are obtained, showing a slightly tapered struc-
ture. A strong gradient or even block formation is clearly
not observed.
1
E-D triads strongly support the findings of the kinetic H
1
3
NMR measurements. In summary, the C NMR analysis
clearly confirms random incorporation of DBAG along the
polymer chain.
Thermal Behavior. A crucial aspect in characterization of
the random copolymers based on DBAG or AG, respec-
tively, and EO was the investigation of the thermal behavior
with respect to monomer ratios (Table 2). PEG itself is a
highly crystalline polymer with a melting point of 66 °C.
With increasing amount of DBAG, the change in the materi-
als’ appearance is already obvious by going from a crystalline
powder to an amorphous material (cf. Supporting In-
formation). As expected, the incorporation of DBAG into
the polymer structure resulted in a strong decrease in the
melting point. Already, at a DBAG content of 2%, the
melting point decreased by 20 °C. This can be attributed to
the introduction of defects into the crystalline regions and,
consequently, destabilization of the ordered packing by the
sterically demanding DBAG units. After deprotection, the
multiple primary amino groups obtained after deprotection
permit hydrogen bonding interaction in the structure and
strongly reduced steric demand, which may contribute to
increased melting points and melting enthalpies. The sample
with the highest amount of DBAG (16.1%) is completely
amorphous, whereas the respective deprotected PAG-con-
taining polymer still shows a low degree of crystallization
(broad, flat endotherm in DSC). Interestingly, two samples
show recrystallization events during heating (sample 3 and
8), indicating strongly impeded crystallization. The trends
observed here are in good agreement with reports on other
1
3
C NMR Characterization of Copolymer Microstructure.
1
3
Figure 7 exemplifies the important regions of the C NMR
spectrum of P(EG -DBAG ) (run 3, Table 1) related to
signals of the EO (E)-centered triads and the DBAG (D)-
centered triads (complete C NMR available in the
Supporting Information). Peak assignment was performed
with reference to the literature data on random EO/PO
9
0
9
1
3
5
4-56
copolymers
and comparison with PEG and DBAG

2
250 Macromolecules, Vol. 43, No. 5, 2010
Obermeier et al.
1
3
Figure 7. C NMR (75.5 MHz, DMSO-d
6
) regions characteristic for DBAG CH backbone signals (left) and EO CH
2
backbone signals (right) for
P(EG90-DBAG ); assigned signals are explained in the text.
9
Table 2. Thermal Data Obtained from Differential Scanning Calorimetry (DSC) for Random Copolymers Based on EO and DBAG and AG,
Respectively
a
b
c
d
e
no.
composition (NMR)
% comonomer (NMR)
T
g
(°C)
T
rc (°C)
T
m
(°C)
ΔH (J/g)
1
2
3
4
5
6
7
8
P(EG110-DBAG
P(EG112-DBAG
2
)
)
1.80
4.27
9.09
16.13
1.80
4.27
9.09
-54
46
37
20
112
83
33
0
139
73
39
25
5
-40
-54
-43
f
P(EG90-DBAG
9
)
-28
P(EG78-DBAG15
)
P(EG78-AG
P(EG78-AG
P(EG78-AG
2
)
5
)
9
)
54
44
f
-38
-59
28
f
8
P(EG78-AG15
)
16.13
-40
). Glass-transition temperature, T
a
1
b
Percentage of comonomer incorporated, determined from H NMR (300 MHz, DMSO-d
6
g
, estimated error
c
d
e
(
5 °C. Recrystallization temperature, Trc, estimated error (2 °C. Melting temperature. T
m
: estimated error (2 °C. Melting enthalpy, determined via
f
integration of the melting signal, estimated error (2 J/g. Very broad transition.
5
6
57
random copolymers of EO and PO or EEGE. In sum-
mary, the gradual effects on melting points and melting
enthalpies observed by DSC characterization are in line
with a random distribution of amino groups along the
backbone, in agreement with the conclusions based on
NMR data.
The use of a monomer with a protected amino moiety broad-
ens the scope of PEG chemistry. Via the introduction of only a
few percent of the functional comonomer, the resulting materials
can be considered to be “multi-amino functional PEGs”, exhibit-
ing a significant increase in loading capacity for attachment of
other compounds. The high reactivity of primary amino groups
and the superior stability of the coupling products, combined
with the precisely defined polymeric structure of such amino-
functional PEG derivatives, offer intriguing potential for biome-
dical applications, for instance in multiple bioconjugation.
Conclusions
The synthesis of multi-amino functional PEG derivatives via
copolymerization of EO and the novel protected amino glycidol
monomer DBAG has been demonstrated. The fraction of amino
groups could be precisely adjusted between 2 and 15% by the
Acknowledgment. We thank Dr. Bernd Mathiasch for NMR
measurements and Colin Drabe for technical assistance. Heinz
Kohlshorn is acknowledged for valuable discussion. F.W. thanks
POLYMAT and the graduate school of excellence MAINZ for
funding. B.O. is grateful to the Fonds der chemischen Industrie
1
initial monomer feed composition. Online H NMR monitoring
of the EO copolymerization was performed and has been
introduced as a valuable method to follow the comonomer
incorporation in the course of the polymerization, revealing a
random distribution of DBAG units along the polymer chain and
rapid polymerization at elevated temperatures. The random
(FCI) the Graduate School of Excellence POLYMAT for fellow-
ships.
1
3
structure was confirmed by detailed C NMR analysis of EO-
and DBAG-centered triad sequences and is also reflected by the
thermal properties of the materials. Removal of the benzyl
protective groups was successfully conducted with Pearlman’s
Supporting Information Available: Additional data and
NMR spectra. This material is available free of charge via the
Internet at http://pubs.acs.org.
catalyst (20% Pd(OH) on carbon). This step ought to be further
2
improved by application of high H2 pressure, aiming at a
reduction of both the amount of catalyst and the reaction times.
In a model coupling reaction with acetic acid anhydride, the
primary amino functions have been quantitatively converted to
the respective amide.
References and Notes
(
(
1) Poly(ethylene glycol) Chemistry: Biotechnical and Biomedical Ap-
plications; Harris, J. M., Ed.; Plenum Press: New York, 1992.
2) Bailey, F. E. J.; Koleske, J. V. Alkylene Oxides and Their Polymers;
Marcel Dekker: New York, 1991.

Article
Macromolecules, Vol. 43, No. 5, 2010 2251
(
(
(
3) Bailey, F. E. J.; Koleske, J. V. Poly(ethylene oxide); Academic Press:
New York, 1976.
4) Abuchowski, A.; Vanes, T.; Palczuk, N. C.; Davis, F. F. J. Biol.
Chem. 1977, 252, 3578.
5) Abuchowski, A.; McCoy, J. R.; Palczuk, N. C.; Vanes, T.; Davis,
F. F. J. Biol. Chem. 1977, 252, 3582.
(31) Taton, D.; Saule, M.; Logan, J.; Duran, R.; Hou, S.; Chaikof,
E. L.; Gnanou, Y. J. Polym. Sci., Part A: Polym. Chem. 2003, 41,
1669.
(32) Berna, M.; Dalzoppo, D.; Pasut, G.; Manunta, M.; Izzo, L.; Jones,
A. T.; Duncan, R.; Veronese, F. M. Biomacromolecules 2006, 7,
146.
(
(
6) Davis, F. F. Adv. Drug Delivery Rev. 2002, 54, 457.
7) Harris, J. M.; Chess, R. B. Nat. Rev. Drug Discovery 2003, 2,
(33) Schiavon, O.; Pasut, G.; Moro, S.; Orsolini, P.; Guiotto, A.;
Veronese, F. M. Eur. J. Med. Chem. 2004, 39, 123.
(34) Koyama, Y.; Umehara, M.; Mizuno, A.; Itaba, M.; Yasukouchi,
T.; Natsume, K.; Suginaka, A.; Watanabe, K. Bioconjugate Chem.
1996, 7, 298.
214.
(
(
8) Pasut, G.; Veronese, F. M. Prog. Polym. Sci. 2007, 32, 933.
9) Dickerson, T. J.; Reed, N. N.; Janda, K. D. Chem. Rev. 2002, 102,
3
325–3344.
(35) Yoshikawa, K.; Yoshikawa, Y.; Koyama, Y.; Kanbe, T. J. Am.
Chem. Soc. 1997, 119, 6473.
(36) Hans, M.; Keul, H.; Moeller, M. Polymer 2009, 50, 1103.
(37) Fitton, A.; Hill, J.; Jane, D.; Miller, R. Synthesis 1987, 1140.
(38) Taton, D.; Le Borgne, A.; Sepulchre, M.; Spassky, N. Macromol.
Chem. Phys. 1994, 195, 139.
(39) Dimitrov, P.; Hasan, E.; Rangelov, S.; Trzebicka, B.; Dworak, A.;
Tsvetanov, C. B. Polymer 2002, 43, 7171.
(40) Erberich, M.; Keul, H.; Moeller, M. Macromolecules 2007, 40,
3070.
(41) Dworak, A.; Baran, G.; Trzebicka, B.; Walach, W. React. Funct.
(
10) van Heerbeek, R.; Kamer, P. C. J.; van Leeuwen, P. W. N. M.;
Reek, J. N. H. Chem. Rev. 2002, 102, 3717–3756.
11) McNamara, C. A.; Dixon, M. J.; Bradley, M. Chem. Rev. 2002, 102,
(
3275–3300.
(
(
12) Bergbreiter, D. E. Chem. Rev. 2002, 102, 3345–3384.
13) Hsieh, H. L.; Quirk, R. P. In Anionic Polymerization: Principles and
Practical Applications; Marcel Dekker: New York, 1996.
14) Penczek, S.; Cypryk, M.; Duda, A.; Kubisa, P.; Slomkowski, S.
Prog. Polym. Sci. 2007, 32, 247.
15) Penczek, S.; Duda, A.; Kubisa, P.; Slomkowski, S. Macromolecular
Engineering: Precise Synthesis, Material Properties, Applications;
Wiley-VCH: Weinheim, Germany, 2007; Vol. 1.
(
(
Polym. 1999, 42, 31.
(42) Halacheva, S.; Rangelov, S.; Tsvetanov, C. Macromolecules 2006,
39, 6845.
(43) Dworak, A.; Panchev, I.; Trzebicka, B.; Walach, W. Macromol.
(
(
16) Zalipsky, S. Bioconjugate Chem. 1995, 6, 150.
17) Herman, S.; Hooftman, G.; Schacht, E. J. Bioact. Compat. Polym.
1995, 10, 145.
Symp. 2000, 153, 233.
(
(
(
18) Jeffamines are distributed by Huntsman International LLC.
19) Polypure AS is a distributor of functional PEG derivatives.
20) Vadala, M. L.; Thompson, M. S.; Ashworth, M. A.; Lin, Y.;
Vadala, T. P.; Ragheb, R.; Riffle, J. S. Biomacromolecules 2008,
(44) Wurm, F.; Nieberle, J.; Frey, H. Macromolecules 2008, 41, 1909.
(45) Yu, Z.; Li, P.; Huang, J. J. Polym. Sci., Part A: Polym. Chem. 2006,
44, 4361.
(46) Pang, X.; Jing, R.; Huang, J. Polymer 2008, 49, 893.
(47) Li, Z.; Chau, Y. Bioconjugate Chem. 2009, 20, 780.
(48) Bakalarz-Jeziorna, A.; Helinski, J.; Krawiecka, B. J. Chem. Soc.,
Perkin Trans. 1 2001, 1086.
9, 1035.
(
(
(
21) Hiki, S.; Kataoka, K. Bioconjugate Chem. 2007, 18, 2191–2196.
22) Zeng, F.; Allen, C. Macromolecules 2006, 39, 6391.
23) Nakamura, T.; Nagasaki, Y.; Kataoka, K. Bioconjugate Chem.
(49) Greene, T. W.; Wuts, G. M. Protective Groups in Organic Synthesis,
3rd ed.; John Wiley & Sons: New York, 1999.
1998, 9, 300.
(
(
(
(
(
(
24) Nagasaki, Y.; Kutsuna, T.; Iijima, M.; Kato, M.; Kataoka, K.;
Kitano, S.; Kadoma, Y. Bioconjugate Chem. 1995, 6, 231.
25) Nagasaki, Y.; Iijima, M.; Kato, M.; Kataoka, K. Bioconjugate
Chem. 1995, 6, 702.
(50) Pearlman, W. M. Tetrahedron Lett. 1967, 8, 1663.
(51) Yoshida, K.; Nakajima, S.; Wakamatsu, T.; Ban, Y.; Shibasaki, M.
Heterocycles 1988, 27, 1167.
(52) Wurm, F.; R €a der, H. J.; Frey, H. J. Am. Chem. Soc. 2009, 131,
26) Cammas, S.; Nagasaki, Y.; Kataoka, K. Bioconjugate Chem. 1995,
7954.
6, 226.
(53) Stolarzewicz, A.; Becker, H.; Wagner, G. Acta Polym. 1980, 31,
743.
(54) Hamaide, T.; Goux, A.; Llauro, M.-F.; Spitz, R.; Guyot, A. Angew.
Macromol. Chem. 1996, 237, 55.
(55) Heatley, F.; Yu, G.; Booth, C.; Blease, T. G. Eur. Polym. J. 1991,
27, 573.
27) Raynaud, J.; Absalon, C.; Gnanou, Y.; Taton, D. J. Am. Chem.
Soc. 2009, 131, 3201.
28) Pasut, G.; Scaramuzza, S.; Schiavon, O.; Mendichi, R.; Veronese,
F. M. J. Bioact. Compat. Polym. 2005, 20, 213.
29) Choe, Y. H.; Conover, C. D.; Wu, D.; Royzen, M.; Gervacio, Y.;
Borowski, V.; Mehlig, M.; Greenwald, R. B. J. Controlled Release
(56) Rejsek, V.; Sauvanier, D.; Billouard, C.; Desbois, P.; Deffieux, A.;
Carlotti, S. Macromolecules 2007, 40, 6510–6514.
(57) Mangold, C.; Wurm, F.; Obermeier, B.; Frey, H. Macromol. Rapid
Commun. 2010, 31, 258.
2
002, 79, 55.
30) Hawker, C. J.; Chu, F.; Pomery, P. J.; Hill, D. J. T. Macromolecules
996, 29, 3831.
(
1