"Functional poly(ethylene glycol)": PEG-based random copolymers with 1,2-diol side chains and terminal amino functionality

Article abstract of DOI:10.1021/ma1015352

A series of poly(ethylene glycol-co-isopropylidene glyceryl glycidyl ether) (P(EO-co-IGG)) random copolymers with different fractions of 1,2-isopropylidene glyceryl glycidyl ether (IGG) units was synthesized. After acidic hydrolysis a new type of "functio

Full text of DOI:10.1021/ma1015352

Macromolecules 2010, 43, 8511–8518 8511
DOI: 10.1021/ma1015352
“Functional Poly(ethylene glycol)”: PEG-Based Random Copolymers
with 1,2-Diol Side Chains and Terminal Amino Functionality
Christine Mangold,† Frederik Wurm,‡ Boris Obermeier,† and Holger Frey*,†
†Institute of Organic Chemistry, Organic and Macromolecular Chemistry, Duesbergweg 10-14,
‡
€
ꢀ
Johannes Gutenberg-Universitat Mainz, D-55099 Mainz, Germany, and Institut des Materiaux,
ꢁ
ꢀ ꢀ
Laboratoire des Polymeres, Batiment MXD, Station 12, Ecole Polytechnique Federale de Lausanne,
CH-1015 Lausanne, Switzerland
Received July 9, 2010; Revised Manuscript Received September 15, 2010
ABSTRACT: A series of poly(ethylene glycol-co-isopropylidene glyceryl glycidyl ether) (P(EO-co-IGG))
random copolymers with different fractions of 1,2-isopropylidene glyceryl glycidyl ether (IGG) units was
synthesized. After acidic hydrolysis a new type of “functional PEGs”, namely poly(ethylene glycol-
co-glyceryl glycerol) (P(EO-co-GG)) was obtained. Using an initiator that releases a terminal amino moiety
after deprotection, functional end groups with orthogonal reactivity to the in-chain groups were obtained. All
polymers showed narrow molecular weight distributions (1.07-1.19), and control of the molecular weights
was achieved in the range 5000-30 000 g/mol. Random incorporation of both comonomers was verified by
monitoring the copolymerization kinetics via real-time 1H NMR spectroscopy during the polymerization and
by characterization of the triad sequence distribution, relying on 13C NMR analysis. Using the 1,2-diol
component of the side chains allows for attachment and facile acid-catalyzed release of molecules bearing
ketone/aldehyde functionalities. This renders the materials potentially useful as support for reagents, drugs or
catalysts. This was demonstrated using benzaldehyde as a model compound. DSC was carried out on all
samples, showing amorphous structures upon incorporation of IGG fractions exceeding 15%.
Introduction
monomer unitupon acidic hydrolysis orcatalytic hydrogenolysis.
At present the most prominent glycidyl ether is ethoxy ethyl
Poly(ethylene oxide) (PEO) is undoubtedly the most important
polymer for biomedical applications today. This is based on two
major key features: on the one hand PEO shows excellent sol-
ubility in aqueous media, and second PEO is biocompatible,
possessing low immunogenicity and antigenicity.1,2 Therefore,
PEO, or poly(ethylene glycol) (PEG), which mostly refers to PEO
with a molecular weight below 20 000 g/mol and its anionic or
cationic derivatives are used in numerous cosmetic3 and also
pharmaceutical products. Furthermore, molecular hybrids of
PEG and peptide or protein drugs play an important role to
enhance circulation times. The covalent linkage of monofunctional
PEG (mPEG) to a variety of different biomolecules is known as
“PEGylation” and represents a well established method today,4,5
which is already standard practice for different medications.
The introduction of functional groups other than hydroxyl
to obtain mono- or difunctional PEG derivatives can either be
realized by a suitable initiator system or via appropriate terminat-
ing agents in the anionic ring-opening polymerization of ethylene
oxide (EO). A detailed overview has been given by Riffle et al.6
A newly developed initiator, which allows the introduction of a
terminal amino moiety by catalytic hydrogenation subsequent to
the polymerization was recently introduced by our group.7
glycidyl ether (EEGE), which was first mentionedby Fitton et al.8
and has been employed by several groups9-11 to obtain block12,13
and recently also random copolymers with EO.14-16 Moller et al.
€
have reported an elegant synthesis of copolymers using various
glycidyl ethers for the orthogonal deprotection of the hydroxyl
functionality.17 For the introduction of other functional groups,
two different methods are possible: The first option is the deriv-
atization of hydroxyl-functionalized polymers in subsequent
polymer modification reactions.18 In this case the degree of
functionalization strongly depends on the yields and selectivity of
the respective modification reactions, often leading to inhomo-
geneous product mixtures. An alternative approach to func-
tional PEG copolymers relies on the copolymerization of
protected epoxide comonomers. In this context, we recently
reported the use of the protected amino analogue of glycidol to
introduce multiple amino groups via random copolymerization.19
While polymerization of the above-mentioned “classical”
glycidyl ethers results in linear poly(glycerol)s (linPG) after
deprotection, the use of 1,2-isopropylidene glyceryl glycidyl ether
(IGG),20 a monomer that was recently developed in our group,
permits the introduction of glycerol side chains with two adjacent
hydroxyl functions for each IGG unit incorporated after acidic
hydrolysis of the acetal protective groups. To date, the IGG
monomer has only been employed in the synthesis of block
copolymers or served as precursor for the synthesis of linear-
hyperbranched PEG-b-hbpoly(glycerol) (PG) copolymers.20,21
However, the presence of two vicinal hydroxyl groups also opens
a general pathway for attachment to the chain and facile, acid-
catalyzed release of any molecule with a ketone or aldehyde
functionality via formation of the cyclic acetal or ketal, respec-
tively. This leads to various interesting applications, such as
The preparation of well-defined PEGs with more than two
functional groups that are randomly distributed at the poly-
ether backbone requires a comonomer bearing a protecting
group, which must be stable under the strongly basic reaction
conditions of the oxyanionic polymerization and can be removed
quantitatively in postpolymerization reactions. Suitable comono-
mers are glycidyl ethers, which release one hydroxyl function per
*To whom correspondence should be addressed. E-mail: hfrey@
uni-mainz.de.
r 2010 American Chemical Society
Published on Web 09/29/2010
pubs.acs.org/Macromolecules

8512 Macromolecules, Vol. 43, No. 20, 2010
Mangold et al.
Figure 1. Reaction sequence for the synthesis of the random P(EO-co-IGG)-copolymers bearing one single terminal amino- and multiple 1,2-diol
side groups.
transport and controlled release of drugs or other biomolecules
and furthermore the use of these novel polymers as polymeric
support for organic synthesis.
added to hydrolyze the excess of phosphorus tribromide. The
organic and aqueous phases were separated, and the aqueous
phase was extracted with 2 Â 200 mL of diethyl ether. The
combined organic extracts were washed with 300 mL of H2O
and 300 mL of a saturated NaHCO3 solution, dried with MgSO4,
and concentrated in vacuo. A slightly yellow viscous liquid was
obtained, which was then distilled at 13 mbar. Bp: 113 °C. Yield:
35.5 g (83%) 1H NMR (300 MHz, CDCl3), δ (ppm) = 7.33 (d,
2H, aromatic), 6.88 (d, 2H, aromatic), 4.51 (s, 2H, BrCH2Ph),
3.80 (s, 3H, OCH3).
In the current work, we have studied the random copolymer-
ization of the recently introduced, protected comonomer IGG
with ethylene oxide by anionic ring-opening copolymerization.
Subsequent acidic hydrolysis results in PEG-copolymers with
1,2-diol side chains (Figure 1). Detailed characterization focusing
on polymerization kinetics, polymer microstructure and thermal
properties is presented. With this work we aim at the following
issues: (i) Can both comonomers be copolymerized in a random
manner, despite the steric bulk of the IGG comonomer and the
high reactivity of EO? (ii) How does copolymerization influence
the materials properties of the “functional PEGs” in compar-
ison to PEG? (iii) Can the resulting copolymers be employed
for the reversible attachment and acid-catalyzed liberation of a
model drug?
N,N-Di(p-methoxybenzyl)aminoethanol. A 15 g sample of the
freshly distilled p-methoxybenzyl bromide (75 mmol), 2.3 g
of aminoethanol (37 mmol), 13.8 g of K2CO3 (100 mmol), and
100 mL of DMF were refluxed for 24 h. After the reaction
mixture was allowed to cool to room temperature, the solution
was filtrated and 300 mL of diethyl ether was added. The organic
phase was then washed with water and a saturated NaHCO3
solution and dried with MgSO4. The organic phase was dried
in vacuo, and a highly viscous liquid was obtained. The crude
mixture was purified by column chromatography using acetic
acid ethyl ester and petrol ether as solvents. Yield: 9 g (80%
Experimental Section
1
Instrumentation. H NMR spectra (300 and 400 MHz) and
1
13C NMR spectra (75.5 MHz) were recorded using a Bruker
AC300 or a Bruker AMX400. All spectra were referenced
internally to residual proton signals of the deuterated solvent.
For SEC measurements in DMF (containing 0.25 g/L of lithium
bromide as an additive) an Agilent 1100 Series was used as
an integrated instrument, including a PSS HEMA column
(106/105/104 g/mol), a UV (275 nm) and a RI detector. Calibra-
tion was carried out using poly(ethylene oxide) standards
provided by Polymer Standards Service. DSC measurements
were performed using 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 80 °C at heating
theoretical) H NMR (300 MHz, CDCl3): δ (ppm)=7.33
(d, 4H, aromatic), 6.88 (d, 4H, aromatic), 3.80 (s, 6H, OCH3), 3.60
(s, 4H, NCH2Ph), 3.57 (t, 2H,CH2OH), 2.65 (t, 2H, NCH2).
General Procedure for the Copolymerization of EO and IGG.
N,N-Di(p-methoxybenzyl)-2-aminoethanol was dissolved in
benzene in a 250 mL Schlenk flask and 0.9 equiv of cesium
hydroxide monohydrate was added. The mixture was stirred at
60 °C under argon for 1 h and evacuated at 90 °C(10-2 mbar) for 2 h
toremove benzene and water, forming the corresponding cesium
alkoxide. Approximately 1 mL dry THF was then cryo-trans-
ferred into the Schlenk flask to dissolve the initiator salt. EO was
first cryo-transferred to a graduated ampule and then cryo-
transferred into the flask containing the initiator in THF (at around
-80 °C). Subsequently, the second comonomer (IGG) was
added via syringe and the mixture was heated to 60 °C and
stirred for 18-24 h. Precipitation in cold diethyl ether resulted in
the pure copolymers. For polymers with a high fraction of IGG, the
polymer solution was dried in vacuo. Yields: 95% to quantitative.
1H NMR (300 MHz, DMSO-d6): δ (ppm) = 7.24, 6.87 (d,
C6H4OMe), 4.16 (m, acetal-H), 3.98 (t, CHH-acetal), 3.74 (s,
C6H4OMe), 3.68-3.34 (polyether backbone), 1.3 (d, CH3 acetal).
r-N,N-Di(p-methoxybenzyl)-poly(ethylene oxide-co-glyceryl
glycerol). The acetal protecting groups were removed by the
addition of 1 N hydrochloric acid to a 20% solution of the
polymer in MeOH/THF 1:1 and about 500 mg of ion-exchange
resin (Dowex 50WX8). The reaction mixture was stirred at room
temperature for 12 h, filtrated (to remove the resin) and con-
centrated in vacuo. Yields: 80-90%. 1H NMR (300 MHz,
DMSO-d6): δ (ppm)=7.53, 7.00 (d, C6H4OMe), 4.01-3.80 (br,
OH), 3.74 (s, C6H4OMe), 3.67-3.09 (polyether backbone).
rates of 10 K min-1 under nitrogen. Matrix-assisted laser
3
desorption and ionization time-of-flight (MALDI-ToF) mea-
surements were performed on a Shimadzu Axima CFR MAL-
DI-TOF mass spectrometer using potassium trifluoroacetic acid
as a cationizing agent and dithranol (1,8,9-trishydroxyanthracene)
as a matrix.
Reagents. All solvents and reagents were purchased from
Acros Organics and used as received, unless otherwise stated.
Chloroform-d1 and DMSO-d6 were purchased from Deutero
GmbH. 1,2-Isopropylidene glyceryl glycidyl20 ether was pre-
pared according to a recently published procedure, dried over
CaH2, and freshly distilled before use.
p-Methoxybenzyl Bromide. A 25 g sample of p-methoxy-
benzyl alcohol (0.21 mol) and 400 mL of dry diethyl ether were
placed in 1-L three-neck round-bottom flask and 29 g of
phosphorus tribromide (0.11 mol) was added via dropping
funnel for 2 h. The reaction mixture was then allowed to stir
at room temperature for additional 12 h. Then, 300 g of ice was

Article
r-H2N-poly(ethylene oxide-co-glyceryl glycerol) or H2N-
Macromolecules, Vol. 43, No. 20, 2010 8513
experiments, which were carried out not only to verify the
kinetics of the reaction but also the reactivity of the two
comonomers, DMSO-d6 was used as a solvent. This is mainly
due to experimental and security reasons with regard to
the higher boiling point of DMSO in comparison to THF.
Depending on the IGG-content of the samples, the copoly-
mers obtained possessed different appearance, ranging
from amorphous solids (9% IGG) to highly viscous liquids
(53% IGG), in pronounced contrast to PEG, which is a
crystalline powder at room temperature. Detailed character-
ization of the thermal properties is given in a following
section subsequent to the structural characterization of the
materials.
poly(ethylene oxide-co-isopropylidene glyceryl glycidyl ether).
A 1 g sample of copolymer was dissolved in 50 mL of methanol
(or a methanol/THF mixture in the case of the IGG species) and
palladium on activated charcoal (10%) was added. The reaction
vessel was flushed with hydrogen (8 bar) and the reaction was
allowed to stir for 72 h at room temperature. The solution was
filtered, concentrated, and precipitated into cold diethyl ether.
Yields: quantitative. H2N-poly(ethylene oxide-co-glyceryl
glycerol) 1H NMR (300 MHz, DMSO-d6): δ (ppm)=4.54 (m,
OH, position varies dependent on concentration/temperature
etc.) 3.81-3.15 (m, polyether backbone). H2N-poly(ethylene
oxide-co-isopropylidene glyceryl glycidyl ether) 1H NMR (300
MHz, DMSO-d6): δ = 4.16 (m, acetal-H), 3.98 (t, CHH-acetal),
3.74 (s, C6H4OMe), 3.68-3.34 (polyether backbone), 1.3 (d,
CH3 acetal).
N,N-Di(p-methoxybenzyl)-poly(ethylene oxide-co-2-phenyl
1,3-dioxolan glycidyl ether). A 1 g sample of P(EO-co-GG) as
well as a 10-fold excess (per GG-unit) of benzaldehyde dimethyl
acetal was placed in a round-bottom flask. A catalytic amount
of p-toluenesulfonic acid was added, and the mixture was put
into an ultrasonic bath at room temperature and under argon
atmosphere for 3 h until a homogeneous mixture was obtained.
Piperidine was added, and all volatile compounds were removed
under reduced pressure. The crude product mixture was then
dialyzed in THF using benzoylated tubing with a molecular
weight cut off (MWCO) of 1000 g/mol. After 72 h THF was
removed in vacuo and the pure copolymer was obtained. Yield:
80%. 1H NMR (300 MHz, DMSO-d6): δ (ppm)=7.52-7.33 (m,
arom. side-chain), 7.22 (d, C6H4OMe), 6.84 (d, C6H4OMe), 5.74
(d, acetal-H ), 4.25-3.25 (m, polyether backbone).
Completion of the polymerization was monitored via 1H
NMR spectroscopy, relying on the absence of the oxirane-
signals corresponding to the two monomers after about
24 h stirring at 60 °C. The copolymer composition was also
studied by 1H NMR spectroscopy (Figure 2). Agreement of
the IGG fraction incorporated in the random copolymers
with the composition of the monomer feed is confirmed by
1H NMR spectra from the comparison of the polyether
backbone signals with the acetal protons at 4.16 ppm and
the methyl signals centered around 1.3 ppm. In addition,
from 1H NMR spectroscopy the number-average molecular
weight was calculated by integration of the aromatic reso-
nances of the initiator (6.87 ppm), the polyether backbone
(3.5 ppm), and the signals referring to IGG units (1.2 ppm). It
should be noted at this point that due to the high molecular
weights obtained the error of this method is on the order of
5-10%. Size exclusion chromatography (SEC) was carried
out for all copolymer samples, demonstrating narrow molec-
ular weight distributions (Mw/Mn =1.08-1.19). As can
be seen from Table 1, the molecular weights obtained from
the SEC measurements using PEG standards deviate from
the values calculated from the 1H NMR measurements. The
deviation is approximately half of the value for the 10% IGG
content samples and becomes gradually more pronounced
with increasing amount of IGG incorporated. This is due to
the peculiar nature of the molecular weight increase, since
the additional groups are located in the side chains, most
probably resulting in a stagnating hydrodynamic radius.
Removal of the cyclic protecting group and release of the
two adjacent hydroxyl-functionalities in the polymer-back-
bone was readily achieved under moderately acidic condi-
tions and stirring for 12 h at room temperature. Removal of
the protecting groups can be followed by NMR via dis-
appearance of the peaks due to the isopropylidene groups, for
instance the acetal proton at 4.16 ppm (1H NMR, Figure 2)
or the corresponding carbon atom and its signal at 99 ppm
(13C NMR). After release of the hydroxyl functions, the
apparent molecular weights obtained from SEC increase for
all P(EG-co-GG) samples, while a decrease would be ex-
pected due to the loss of the acetal group. Interestingly, this
increase is directly correlated to the IGG content and there-
fore to the number of hydroxyl functions formed. This can
most likely be attributed to the emerging diol functions
that interact with the SEC columns, thus leading to an
apparent increase in the hydrodynamic radii (compare with
Figure 7). The polydispersities remained low after deprotec-
tion and were in the range of Mw/Mn=1.09-1.19.
1H NMR Kinetics.19 In a conventional NMR tube, a mixture
of IGG and EO in DMSO-d6 was placed under an argon
atmosphere, cooled to -196 °C, and evacuated. The initiator
solution was added rapidly to guarantee that the first layer
stayed frozen and no reaction took place. Then the tube was
evacuated again while freezing the initiator-solution. The NMR
tube was flame-sealed by applying high vacuum. It is of crucial
importance to keep both solution-layers frozen until the NMR
measurements are started. Immediately after melting and
mixing of the two solutions, the first spectrum was recorded.
Intervals between two measurements were 5 s within the first
minute and extended afterward. A sample of the pure monomer
mixture was measured in advance to reduce the necessary time
for locking and shimming.
Results and Discussion
Synthesis. To copolymerize two monomers such as EO
and IGG with highly diverging boiling points a procedure
has to be chosen that levels these differences to a minimum.
All reactions were therefore carried out at 60 °C in vacuo, as
reported recently for another EO copolymerization.19 There
appears to be a general bias that the reactivity of ethylene
oxide is generally considerably higher than for all other
epoxide monomers, such as glycidyl ethers. It is an important
objective of this work to clarify this issue by probing the
possibility of random copolymerization in this system that
combines EO with a bulky epoxide as a comonomer.
As a novel initiator N,N-di(p-methoxybenzyl)aminoethanol
has been prepared. This initiator has been improved in com-
parison to the previously mentioned N,N-Dibenzylamino-
ethanol by introducing methoxy groups in the para-position
of the aromatic system, permitting more facile cleavage,
which is due to the inductive effect of the methoxy groups.
N,N-Di(p-methoxybenzyl)aminoethanol was deprotonated
with 0.9 equiv of cesium hydroxide monohydrate, and the
polymerization was carried out in a highly concentrated
solution of the two monomers and THF. In the NMR online
Similar to the deprotection of the acetals, the liberation of
the terminal amino moiety by catalytic hydrogenation can be
followed via 1H NMR, monitoring the disappearance of the
aromatic signals at 6.86 and 7.24 ppm. The reaction condi-
tions applied for the hydrogenolysis guarantee removal of
the benzylic protecting groups at the amine without removal
of the isopropylidene protecting groups. Figure 2 shows the

8514 Macromolecules, Vol. 43, No. 20, 2010
Mangold et al.
Figure 2. 1H NMR spectra of the copolymers synthesized. The deprotection steps can be interchanged.
Table 1. Overview of the Characterization Data for All Copolymer Samples Prepared
no.
composition (theor)
sum (NMR)a
IGG/GG-fractiona (%)
Mn (NMR)
Mn (SEC)b
PDI (SEC)b
1
2
3
4
MeOBn2NP(EO200-IGG20
)
)
MeOBn2NP(EO265-IGG26
MeOBn2NP(EO180-IGG32
)
)
9
14
26
30
53
9
14
26
30
53
9
26
30
9
26
53
16500
14000
5300
11800
30000
15700
12800
4800
10000
25200
15400
4500
7600
7000
3000
4700
4500
8000
8700
3700
5800
5700
8000
3900
5600
8000
3000
4400
1.11
1.15
1.08
1.11
1.13
1.15
1.17
1.08
1.17
1.18
1.15
1.14
1.30
1.15
1.11
1.13
MeOBn2NP(EO160-IGG40
MeOBn2NP(EO45-IGG15
)
MeOBn2NP(EO47-IGG17
)
)
MeOBn2NP(EO100-IGG50
)
MeOBn2NP(EO90-IGG40
MeOBn2NP(EO122-IGG136
5
MeOBn2NP(EO100-IGG100
)
)
1d
2d
3d
4d
5d
1dt
3dt
4dt
1t
3t
5t
MeOBn2NP(EO265-GG26
)
)
MeOBn2NP(EO265-GG26
MeOBn2NP(EO180-GG32
)
)
MeOBn2NP(EO180-GG32
MeOBn2NP(EO47-GG17
)
)
MeOBn2NP(EO47-GG17
)
)
MeOBn2NP(EO90-GG40
MeOBn2NP(EO90-GG40
MeOBn2NP(EO100-GG100)
MeOBn2NP(EO100-GG100
)
H2NP(EO265-GG26
)
H2NP(EO265-GG26
)
H2NP(EO47-GG17
)
)
H2NP(EO47-GG17
)
)
H2NP(EO90-GG40
H2NP(EO90-GG40
9700
H2NP(EO265-IGG26
)
H2NP(EO265-IGG26
)
16200
5000
27700
H2NP(EO45-IGG15
)
H2NP(EO47-IGG17
)
H2NP(EO122-IGG136
)
H2NP(EO122-IGG136
)
a Determined from 1H NMR (300 MHz, DMSO-d6. bMn determined by SEC-RI in DMF.
1
respective H NMR spectra of the four different types of
functional polyethers that can be synthesized after orthog-
onal reactions. Depending on the order of the deprotection
reactions that can be chosen freely due to their orthogonal
character, it is on the one hand possible to obtain P(EO-co-
GG)-copolymers with glyceryl side chains, where the termi-
nal amine still carries the benzyl protective groups, but on the
other hand also P(EO-co-IGG)-copolymers with a cleaved
amine in terminal position and isopropylidene-protected
glyceryl units along the backbone are accessible. Both poly-
mer types can finally be converted into the fully (terminal
and in-chain) deprotected polymers (Figure 2).
a copolymer with a variety of different possible comono-
mer combinations. The change in molecular weight after
acidic hydrolysis or hydrogenation can also be followed by
MALDI-ToF. The corresponding spectra are given in the
Supporting Information.
1H NMR Copolymerization Kinetics. To investigate the
copolymerization behavior of EO and IGG, we used an
experimental procedure that has recently been developed in
our group.19 First a comonomer solution was transferred
into a NMR tube, evacuated, cooled with liquid nitrogen,
and the initiator solution was added thereafter. The cold
NMR tubes were evacuated, flame-sealed, and the polymer-
ization was carried out in vacuo. Because of experimental
reasons the online-kinetic studies were carried out in DMSO-d6
instead of THF, which usually serves as the solvent in the
large-scale polymerizations. To examine the copolymerization
The different possible copolymer structures have also been
investigated by MALDI-ToF mass spectrometry. The
recorded spectra obtained are quite complex with rather
low resolution, but they clearly show the expected pattern for

Article
Macromolecules, Vol. 43, No. 20, 2010 8515
Figure 4. Monomer conversion versus time monitored by 1H NMR in
DMSO-d. resulting in polymers with the following final composition:
70 °C (green triangles) MeOBn2NP(EO61-IGG7) (7*), 50 °C (blue
squares) MeOBn2NP(EO31-IGG9), 25 °C (red, circles) MeOBn2NP-
(EO29-IGG9).
Figure 3. 1H NMR spectra for the copolymerization of IGG and EO at
50 °C in DMSO-d6 recorded after 0, 0.1, 2, 13, 21, 26, 87, 115, 190, and
264 min.
kinetics, three different temperatures, i.e., 25, 50, and 70 °C,
were applied. The molecular weight distributions of all
samples obtained in these online NMR experiments were
narrow with PDIs around 1.06. Incorporation of the two
comonomers and the growth of the polyether chain was
studied by following the decrease of the epoxide signals
located at 2.61 ppm for the four protons of the symmetric
EO monomer and at 3.09 ppm for the methine proton of
IGG, respectively. All signals are referenced to the aromatic
peaks corresponding to the initiator at 6.87 ppm, which were
set to 4 and should obviously remain constant in the course
of the reaction. The emerging backbone signal at 3.54 ppm
overlaps with different signals of the IGG monomer, which
complicates the calculation of the molecular weight. This is
unfortunate, since Mn determined in this independent manner
could otherwise serve as an affirmation for the molecular
weight calculated from the monomer decrease. Plotting the
monomer conversion versus molecular weight resulted in a linear
graph (see Supporting Information), confirming the living char-
acter of the polymerization, as it is expected for the anionic ROP.
Figure 3 displays a zoom-in (4.5-2 ppm) of different 1H NMR
spectra, showing the decreasing intensity of the monomer signals
and the growing backbone signal.
As expected, the rate of the polymerization is significantly
influenced by the reaction temperature. While the copoly-
merization at 50 °C takes approximately 3.5 h to completion
(5 h for the sample with higher molecular weight), approxi-
mately 1 week is required for the polymerization at 25 °C. At
70 °C, the reaction is completed already after 1 h, and even
the initial spectra taken after a few seconds already reveal
signals of the polymer backbone. Figure 4 shows the con-
version versus time plot for EO and IGG at different
temperatures, as derived from the NMR measurements.
The most relevant information clearly derived from the
NMR spectra is that there appears to be little difference in
the reactivity of the two monomers, despite the sterically
demanding structure of IGG in comparison to EO. The
conversion of both comonomers is virtually identical at all
stages of the reaction (Figures 4 and 5). Unexpectedly, in the
case of the two low temperature-samples (25 and 50 °C) the
decrease of the IGG signal appears to be even faster at the
beginning of the reaction than the decrease of the EO
resonances. Figure 5 shows exemplarily the conversion of
both comonomers plotted against the overall monomer-
consumption for one sample at 50 °C.
Figure 5. Momomer incorporation (percentage) versus the overall con-
version at 50 °C. Final copolymer composition: MeOBn2P(EO31-IGG9).
While the decrease of EO is directly related to the total
conversion, small fluctuations (to higher but also lower
values) are observed for the IGG content. At this point, it
should be noted that EO is the predominant component in the
copolymer studied (MeOBn2P(EO31-IGG9)) and that the EO
conversion contributes to a stronger extent to the overall
conversion than the conversion of IGG. Therefore, little
variation of the EO-incorporation is observed. In addition,
four proton signals of the two methylene groups of ethylene
oxide are used to calculate the decrease of EO in the course of
the polymerization, but only one methine signal is used as a
reference for the determination of the IGG conversion.
Obviously a larger error is expected in the latter case. Taking
the above-mentioned considerations into account, it can be
concluded that in all cases a linear EO and IGG conversion is
observed, that is the comonomers are incorporated equally,
and a random distribution is obtained in the polymer back-
bone. This unexpected behavior is valid independent of the
EO/IGG ratio and for all temperatures studied.
13C NMR Characterization. Random distribution of the
comonomer units within the PEG backbone is of crucial
importance with respect to potential applications of the
functional PEG copolymers. 13C NMR analysis permits

8516 Macromolecules, Vol. 43, No. 20, 2010
Mangold et al.
Figure 6. 13C NMR spectra of different P(EO-co-GG) copolymers with different GG fraction of (a) 9%, (b) 14%, and (c) 53%.
Table 2. Thermal Properties of Poly(ethylene glycol-co-isopropylidene glyceryl glycidyl ether) and Poly(ethylene glycol-co-glyceryl glycerol)
Random Copolymers (DSC measurements)
no.
formula
IGG/GG fraction (%)a
Tgb (°C)
Tmc (°C)
ΔHd (J/g)
Τrce (°C)
ΔHrcf (J/g)
1
2
3
4
MeOBn2NP(EO265-IGG26
)
)
9
14
26
30
53
9
10
23
9
14
26
30
53
-54
-51
-55
-50
-49
-58
-59
-51
-48
-42
-38
-32
-28
31
31
35
13
MeOBn2NP(EO180-IGG32
MeOBn2NP(EO47-IGG17
)
)
MeOBn2NP(EO90-IGG40
5
MeOBn2NP(EO122-IGG136
MeOBn2NP(EO85-IGG8)
MeOBn2NP(EO61-IGG7)
MeOBn2NP(EO31-IGG9)
)
6*
7*
8*
1d
2d
3d
4d
5d
13
14
39
49
-33
-38
34
33
MeOBn2NP(EO265-GG26
)
)
44
35
45
18
-27
-14
28
17
MeOBn2NP(EO180-GG32
MeOBn2NP(EO47-GG17
)
)
MeOBn2NP(EO90-GG40
MeOBn2NP(EO100-GG100
)
a Determined from 1H NMR (300 MHz, DMSO-d6). b Glass transition temperature, in °C. cMelting temperature Tm, in °C. d Melting enthalpy, in J/g,
determined by integration. e Recrystallization temperature Trc, in °C. f Recrystallization enthalpy, in J/g.
to investigate the monomer triad sequence distribution,
revealing details concerning the incorporation statistics of
IGG units. Besides the well-established signal assignment for
random copolymers based on propylene oxide and ethylene
oxide,22 only few other oxirane monomer combinations have
been investigated to date. However, EO has often been
assumed to possess considerably higher reactivity than other
epoxide monomers, such as glycidyl ethers. To confirm
random incorporation, the deprotected copolymer samples
were investigated, since fewer side group resonances overlap
with the triad signals in question than in the respective acetal-
protected copolymers. Figure 6 shows a section of the 13C
NMR spectra of samples 1, 2, and 5 with increasing comono-
mer fraction from spectrum a (front) to c. To keep abbrevia-
tions short for the triad sequences, the EO unit is referred to as
E and the GG (former IGG)-units are indicated as I.
Two of the side group signals (1 and 2 in Figure 6) resulting
from the glycerol unit are located at 63.2 and 72.9 ppm, the
other two (4 and 5) overlap with the triad sequence signals of
either the two methylene carbon atoms of EO or the one of
the GG unit. Several microstructure-related resonances
in the range 71.1-68.7 ppm occur, of which one signal at
69.9 ppm can clearly be assigned to the EEE triad. This signal
decreases with increasing GG content, while other triad
signals become more pronounced. In addition the signal at
68.9 ppm, which can clearly be assigned to the methylene
carbon in proximity of a GG unit (for example IEE) does not
only increase, but is split up into several signals with increas-
ing amount of GG units incorporated (increasing abundance
of IEI, IIE, and III triads). The same observations are made
for the methine carbon of the GG units. The sample with the
lowest GG-content shows only one single peak that is
assigned to the EIE triad. This resonance also splits up with
increasing GG-content, and the sample with 53% comono-
mer content shows three signals of different intensities,
which is most probably explained by the occurrence of the
III, EII, and IIE as well as the EIE triads. The two possible
end group signals of the primary or secondary carbons linked
to the terminal hydroxyl groups cannot be discerned in the
spectra due to the low concentration and high molecular
weight of the samples. Although unequivocal assignment
of all triad signals was not possible, the spectra permit to
exclude the formation of longer runs of one comonomer. All
observations confirm random incorporation of the IGG/GG
comonomer units in the PEG backbone and are in line with
the observations from the kinetic studies (vide supra).

Article
Macromolecules, Vol. 43, No. 20, 2010 8517
Table 3. Comparison of the Characterization Data of the Initial Polymer, the Deprotected Material and the Benzyl Derivatives
no.
formula
Mn(NMR)a
Mn(SEC)
PDI
Tg (°C)b
Tm (°C)c
ΔH (J/g)d
1
MeOBn2NP(EO265-IGG26
)
16500
15700
18100
5300
4800
6400
7600
9200
7400
3000
4000
3000
1.11
1.15
1.16
1.08
1.08
1.08
-54
-48
-49
-55
-38
-31
31
44
26
35
45
36
1d
1b
3
3d
3b
MeOBn2NP(EO265-GG26)
MeOBn2NP(EO265-PDG26
MeOBn2NP(EO47-IGG17
MeOBn2NP(EO47-GG17
)
)
)
MeOBn2NP(EO47-PDG17
)
a Determined from 1H NMR (300 MHz, DMSO-d6). b Glass transition temperature. c Melting temperature Tm. d Melting enthalpy in J/g determined
by integration of the melting peak.
Figure 7. Molecular weight distribution of the samples 3, 3d, and 3b from (a) SEC-measurements in DMF and (b) MALDI-ToF mass spectrometry
using trifluoroacetate as a cationizing agent and dithranol (1,8,9-trishydroxyanthracene) as a matrix.
Thermal Behavior. The thermal properties of the P(EG-co-
GG) copolymers are of central interest for future applica-
tions in the biomedical field as well as for use as polymer
supports for organic synthesis. The thermal characteristics
summarized in Table 2 have been studied using differential
scanning calorimetry (DSC). The glass transition tempera-
ture (Tg) of all samples is higher than for the PEG-homo-
polymer (-64 °C23 for MPEG-1000). The Tg is -54 °C for
sample 1 with an IGG content of 9% and increases slightly
with increasing IGG incorporation (sample 5 with 53% IGG
exhibits a Tg of -49 °C). For the samples with lower molec-
ular weights obtained from the 1H NMR kinetic experi-
ments, the Tgs were somewhat lower compared to those of
higher molecular weight, but similar IGG content, which
is consistent with expectation. The deprotected analogues
exhibit glass transitions that are approximately 5-10 °C
higher than those of the respective protected polymers,
which we ascribe to hydrogen bonding interaction of the
1,2-diol side chains.
Poly(ethylene oxide) is a crystalline polymer with a melt-
ing temperature of 66 °C.24 With increasing comonomer
content (IGG or GG), Tm is gradually shifted to lower
temperatures, until crystallization is completely inhibited,
if the incorporated comonomer fraction exceeds 15-20%.
For the P(EO-co-IGG) copolymer series crystallization can
be observed up to an IGG content of as much as 14%,
however, the melting temperature is significantly lowered in
comparison to PEG (31 vs 66 °C for the homopolymer). The
same observations are made for the deprotected samples,
but in this case the melting temperatures are higher and
additionally the melting enthalpies exceed those of the
protected samples. Thus, a higher fraction of crystalline
domains is present, which we ascribe to the lower steric
hindrance of the glyceryl units in comparison to the bulky
1,2-isopropylidene glyceryl units, which appear to strongly
impede ordering of the chains.
Derivatization. The use of MPEG (5000 g/mol) as soluble
polymeric support in organic synthesis is a widely established
method and is also conducted on an industrial scale.25-28 Its
success as a soluble support is based on its good solubility in
water and various other organic solvents, but on the other
hand good precipitability in diethyl ether and 2-propanol.29
However, PEG exhibits low loading capacity, since it pos-
sesses only two end groups. It has been suggested that this
drawback can be overcome by using hyperbranched poly-
(glycerol) (hbPG),30 which carries multiple hydroxyl groups
in either linear or terminal positions. Different reactions
have been carried out using PG as support for reagents or
catalysts and utilizing either the single hydroxyl-
functionalities31 or the vicinal position32 of the terminal
OH groups. To demonstrate the potential of the synthesized
copolymers as soluble polymeric support by the reaction
with ketones or aldehydes to the respective acetals, we
employed the dimethoxyacetal derivative of benzaldehyde
as a model compound. Removing the evolving methanol
in the course of the reaction favors the formation of the
polymeric acetal without cross-linking and leads to high
yields, which is evidenced by the absence of hydroxyl signals
in the 1H NMR. After dialysis of the crude product, quanti-
tatively acetalized polymers were obtained. Table 3 sum-
marizes the characterization data of the initial and the new
polymers with 2-phenyl-1,3-dioxolane (PDG) side groups.

8518 Macromolecules, Vol. 43, No. 20, 2010
Mangold et al.
1
The products were characterized by H NMR spectros-
school of excellence in the context of MAINZ for valuable
financial support. B.O. also acknowledges the Fonds der Che-
mischen Industrie for a scholarship. F.W. thanks the Alexander-von-
Humboldt Foundation for a fellowship. H.F. acknowledges the
SFB 625 of the DFG for valuable support.
copy. The newly emerging acetalic proton leads to signals at
5.78 and 5.69 ppm, and by integration of those and the
aromatic signals of the initiator as well as the new aromatic
side chains, quantitative derivatization can be confirmed.
Corresponding NMR spectra are shown in the Supporting
Information. The molecular weights obtained from SEC
measurements, which should increase by about 88 g/mol
per GG unit decreases slightly, which is consistent with the
previously discussed observations made for the IGG co-
polymers. Figure 7 shows the shift of the molecular weight
obtained from SEC-experiments and the corresponding
spectra obtained by MALDI-ToF, confirming the depen-
dence of the apparent molecular weight on the presence
of the hydroxyl groups. The MALDI-ToF spectra was
depicted, although the complexity of the situation does not
allow for distinct assignment of all peaks, but qualitative
information regarding the molecular weight can be gained.
The glass transition temperature increases significantly by
exchanging the isopropylidene side group with a benzylidene
side group. Melting enthalpy and melting points are slightly
lower in the case of the PDG-species.
Supporting Information Available: Additional characteriza-
tion data (Figures S1-S6). This material is available free of
charge via the Internet at http://pubs.acs.org.
References and Notes
(1) Harris, J. M. Poly(ethylene glycolChemistry: Biotechnical and
Biomedical Applications; Plenum Press: New York, 1992.
(2) Harris, J. M.; Zalipsky, S., Poly(ethylene glycol- Chemistry and
Biological Applications; American Chemical Society: Washington,
DC, 1997.
€
(3) Fruijtier-Polloth, C. Toxicology 2005, 214, 1–38.
(4) Abuchowski, A.; Vanes, T.; Palczuk, N. C.; Davis, F. F. J. Biol.
Chem. 1977, 252, 3578–3581.
(5) Davis, F. F. Adv. Drug Delivery Rev. 2002, 54, 457–458.
(6) Thompson, S.; Vadala, T. P.; Vadala, M. L.; Lin, Y.; Riffle, J. S.
Polymer 2008, 49, 345–373.
€
(7) Wurm, F.; Rader, H. J.; Frey, H. J. Am. Chem. Soc. 2009, 131,
7954–7955.
Conclusion. A novel class of poly(ethylene glycol)-based
random copolymers with predetermined amount of glycerol
side chain functionalities, low polydispersities as well as
adjustable molecular weights have been synthesized and
characterized in detail, focusing on macroscopic materials
properties and microstructure. The copolymerization kinetics
was investigated by 1H NMR spectroscopy using a recently
developed online NMR technique, and 13C NMR measure-
ments were employed to support random incorporation of
the two comonomers. Differential scanning calorimetry was
carried out for all samples confirming expectation for the
thermal properties of random copolymers. Incorporation of
the recently developed monomer 1,2-isopropylidene glyceryl
glycidyl ether (IGG) allows to obtain one glycerol per
comonomer unit upon acidic hydrolysis. Each glyceryl unit
offers two vicinal hydroxyl groups, which can serve as diol-
component in the reversible formation of a cyclic acetal/
ketal. In contrast to hbPG, which is used as polymeric support,
the number of vicinal hydroxyl groups is adjustable rather
independent of the molecular weight of the whole polymer.
This reaction can be used, as it has been shown exemplarily
with benzaldehyde, to attach and release molecules that bear
aldehyde or ketone functionality. The use of the new initiator
N,N-di(p-methoxybenzyl)aminoethanol leads to the facile
introduction of a terminal amino group, which can be
recovered by catalytic hydrogenation subsequent to the
polymerization. In addition the reaction conditions applied
allow to liberate the terminal amino moiety without removal
of the acetal protecting groups. This means the in-chain
functional groups and the terminal group are orthogonally
protected and can be addressed selectively, which is interest-
ing with respect to a variety of applications for bioconjuga-
tion, surface modification, and drug transport and release.
(8) Fitton, A.; Hill, J.; Jane, D.; Miller, R. Synthesis 1987, 1140–1142.
(9) Taton, D.; Le Borgne, A.; Sepulchre, M.; Spassky, N. Macromol.
Chem. Phys. 1994, 195, 139–148.
(10) Halacheva, S.; Rangelov, S.; Tsvetanov, C. Macromolecules 2006,
39, 6845–6852.
(11) Dworak, A.; Panchev, I.; Trzebicka, B.; Walach, W. Macromol.
Symp. 2000, 153, 233–242.
(12) Dworak, A.; Baran, G.; Trzebicka, B.; Walach, W. React. Funct.
Polym. 1999, 42, 31–36.
(13) Dimitrov, P.; Hasan, E.; Rangelov, S.; Trzebicka, B.; Dworak, A.;
Tsvetanov, C. B. Polymer 2002, 43, 7171–7178.
(14) Pang, X.; Jing, R.; Huang, J. Polymer 2008, 49, 893–900.
(15) Yu, Z.; Li, P.; Huang, J. J. Polym. Sci., Part A: Polym. Chem. 2006,
44, 4361–4371.
(16) Mangold, C.; Wurm, F.; Obermeier, B.; Frey, H. Macromol. Rapid
Commun. 2010, 31, 258–264.
(17) Erberich, M.; Keul, H.; Moller, M. Macromolecules 2007, 40, 3070–
€
3079.
(18) Li, Z.; Chau, Y. Bioconjugate Chem. 2009, 20, 780–789.
(19) Obermeier, B.; Wurm, F.; Frey, H. Macromolecules 2010, 43, 2244–
2251.
(20) Wurm, F.; Nieberle, J.; Frey, H. Macromolecules 2008, 41, 1909–
1911.
(21) Hofmann, A. M.; Wurm, F.; Huhn, E.; Nawroth, T.; Langguth, P.;
Frey, H. Biomacromolecules 2010, 11, 568–574.
(22) Hamaide, T.; Goux, A.; Llauro, M. F.; Spitz, R.; Guyot, A. Angew.
Makromol. Chem. 1996, 237, 55–57.
(23) Lestel, L.; Guegan, P.; Boileau, S.; Cheradame, H.; Laupretre, F.
Macromolecules 1992, 25, 6024–6028.
(24) Mandelkern, L. Chem. Rev. 1956, 56, 903–958.
(25) Dickerson, T. J.; Reed, N. N.; Janda, K. D. Chem. Rev. 2002, 102,
3325–3344.
(26) van Heerbeek, R.; Kamer, P. C. J.; van Leeuwen, P. W. N. M.;
Reek, J. N. H. Chem. Rev. 2002, 102, 3717–3756.
(27) McNamara, C. A.; Dixon, M. J.; Bradley, M. Chem. Rev. 2002, 102,
3275–3300.
€
(28) Bergbreiter, D. E. Chem. Rev. 2002, 102, 3345–3384.
(29) Lu, J.; Toy, P. H. Chem. Rev. 2009, 109, 815–838.
€
(30) Sunder, A.; Hanselmann, R.; Frey, H.; Mulhaupt, R. Macro-
Acknowledgment. The authors thank Dr. Mihail Mondeshki
for NMR measurementsand Alina Mohrfor technical assistance.
C.M., F.W., and B.O acknowledge the POLYMAT graduate
molecules 1999, 32, 4240–4246.
(31) Roller, S.; Siegers, C.; Haag, R. Tetrahedron 2004, 60, 8711–8720.
(32) Hebel, A.; Haag, R. J. Org. Chem. 2002, 67, 9452–9455.