Diblock copolymer formation via self-assembly of cyclodextrin and adamantyl end-functionalized polymers

Article abstract of DOI:10.1021/ma200048a

Two water-soluble polymers poly(2-methyl-2-oxazoline) and poly(N-isopropylacrylamide) with complexing moieties (β-CD and adamantane, respectively) located at the chain ends were prepared via controlled techniques. To verify the interaction of the β-CD- an

Full text of DOI:10.1021/ma200048a

ARTICLE
pubs.acs.org/Macromolecules
Diblock Copolymer Formation via Self-Assembly of Cyclodextrin
and Adamantyl End-Functionalized Polymers
Jan Stadermann,† Hartmut Komber,† Michael Erber,† Frank D€abritz,† Helmut Ritter,‡ and Brigitte Voit†,*
†Leibniz Institut fuer Polymerforschung Dresden e.V., Hohe Strasse 6, 01069 Dresden, Germany
‡Lehrstuhl f€ur Pr€aparative Polymerchemie, Institut f€ur Organische Chemie und Makromolekulare Chemie, Heinrich-Heine Universit€at,
Universit€atsstrasse 1, Geb. 26.33.00, D-40225 D€usseldorf, Germany
S Supporting Information
b
ABSTRACT: Two water-soluble polymers poly(2-methyl-2-
oxazoline) and poly(N-isopropylacrylamide) with complexing
moieties (β-CD and adamantane, respectively) located at the
chain ends were prepared via controlled techniques. To verify the
interaction of the β-CD- and adamantane-type polymer end
groups in aqueous solution, detailed complexation studies were
carried out by 1H NMR spectroscopy. It could be proved, that the
polymers undergo self-assembly to form the corresponding
supramolecular diblock structure. Furthermore, the double-
hydrophilic block assembly was observed to be switchable to a
hydrophilicꢀhydrophobic configuration by adjusting tempera-
ture leading to reversible aggregate formation.
’ INTRODUCTION
N-isopropylacrylamide (NIPAAm)14 and with N-vinylpyrrolidone,15
respectively, copolymers were obtained with up to 10 mol %
comonomer bearing covalently attached β-CD as backbone
functionality. Further, Nielsen et al. reported the synthesis of
β-CD-dextran polymers from alkyne-modified dextrans using the
click-approach.16 These polymers were studied with respect to
their binding properties to adamantane derivatives.14ꢀ16
In general, different types of noncovalent interactions can be
exploited for the construction of reversible supramolecular
architectures.17,18 Cordier et al. reported the design of reversible
networks, being rubber-like materials which show self-healing
behavior, where the assembly is formed via functional units able
to associate through multiple hydrogen bonds.19 The use of
strong hydrogen bonds is also described for the assembly of
linear polymer chains,20 diblock copolymers21 and other block
structures.22 Further, the design of diblock and triblock polymer
structures via directed metal coordination was demonstrated by
the group of Schubert.23ꢀ25
Cyclodextrins (CDs) are of particular interest in polymer
chemistry, e.g., for CD-mediated polymerizations.1 Further, CDs
have been used as building blocks for the construction of
polymeric networks. These gels are chemically cross-linked
materials where CD is covalently incorporated2,3 or physical
assemblies formed through noncovalent interactions of the CD
units.4 Various CD-containing structures have been designed and
studied, e.g., in view of pharmaceutical applications like drug
carrier systems.5ꢀ7
By using the complexation of CDs with nonpolar moieties of
appropriate size, the design of reversible polymer networks is
possible. Those can be broken by addition of competitive low-
molecular molecules or by increasing temperature.8 In particular,
the high binding forces in hostꢀguest inclusion complexes
between β-CD and adamantyl groups have often been exploited
in building up supramolecular assemblies.9ꢀ11 Resulting rever-
sible gel systems are considered to be interesting, e.g., as
injectable drug delivery depots.
Withrespecttoanincreasingdemandformoredefinedpolymer
structures, there is a high interest in a controlled design of CD-
mediated assembliesbytakingadvantageofthe highlyspecific click
chemistry. By using azide-functionalized CD, it is possible to
selectively equip alkyne-functionalized polymer materials or sur-
faces with CD-moieties.12,13 On the other hand, novel CD-
functionalized monomers are accessible via the “click-strategy”.
The group of Ritter reported the synthesis of monofunctionalized
β-CD-functionalized acrylate and methacrylate14 as well as N-
vinylpyrrolidone15 via the click reaction. By copolymerization with
However, reports which address the assembly of polymeric
structures through CD-adamantane complexation are so far
limited to reversible gel systems. In contrast, this work is
concerned with the building up of defined diblock copolymers
via complexation of β-CD and adamantane as polymer end group
functionalities. Herein, this concept is demonstrated with two
water-soluble polymers: poly(2-methyl-2-oxazoline) (PMeOxa)
Received: January 12, 2011
Revised:
March 24, 2011
Published: April 08, 2011
r
2011 American Chemical Society
3250
dx.doi.org/10.1021/ma200048a Macromolecules 2011, 44, 3250–3259
|

Macromolecules
ARTICLE
and poly(N-isopropyl-acrylamide) (PNIPAAm), both being of
high interest for use in physiological systems. We present a
versatile synthetic way to prepare defined end-functional poly-
mers able to undergo self-assembly through the formation
of strong inclusion complexes. Furthermore, studies on the
β-CD-adamantane end group interaction are demonstrated
yielding a well-defined diblock assembly.
The combination of PNIPAAm and PMeOxa in such a self-
assembled diblock can be described as a double-hydrophilic
system. However, due to the well-known thermoresponsive
character of PNIPAAm, this diblock structure is switchable to a
hydrophilicꢀhydrophobic configuration as a consequence of the
heat-induced PNIPAAm phase transition at a lower critical
solution temperature (LCST). The hydrophilic part is realized
by PMeOxa which does not show a thermoresponsive behavior
in this temperature range. Such a switching process strongly
influences the aggregation behavior, which is hence controllable
by adjusting temperature.26 Consequently, these materials are
highly interesting, if the change in aggregation can be related to a
controlled drug release.27,28 On the other hand, the supramole-
cular polymer coupling is controllable by means of temperature
owing to the enthalpic nature of the β-CD-adamantane inclusion
complex.29
Various block and graft copolymers composed of PNIPAAm
and PMeOxa have been prepared via conventional methods so
far.30,31 Here we present the synthesis of an adamantane-contain-
ing initiator for the cationic ring-opening polymerization and a β-
CD-containing initiator for the atom transfer radical polymeri-
zation (ATRP). These functional initiators are then employed to
initiate the polymerization of 2-methyl-2-oxazoline and N-iso-
propylacrylamide, respectively. NMR studies finally prove the
successful complex formation of the head-functionalized poly-
mers leading to noncovalent diblock assembly. Furthermore, the
influence of the thermoresponsive behavior of the PNIPAAm
block on the complex stability is studied.
controlled by the Bruker variable temperature accessory BVT-3000 and
was calibrated using the standard Wilmad ethylene glycol sample. In all
measurements the temperature was maintained constant within (0.2 K.
After each temperature step (ΔT = 4 K) the sample was allowed to
equilibrate for 15 min. The 2D ROESY spectrum was recorded with cw
spinlock using a loop of 690 (180°x ꢀ 180°ꢀx) pulses of 180 μs length of
the 180° pulse giving a total mixing time of 248 ms (phase sensitive using
TPPI, 24 scans, sweep width =4746 Hz, 2K data points in F2 and 256
experiments in F1). Further standard 2D NMR spectra were recorded to
1
verify the H and 13C signal assignments for initiators and polymers
using standard parameters.
DSC analysis were performed with a differential scanning calorimeter
Q1000 of TA Instruments (USA) in nitrogen atmosphere from ꢀ80 °C
to þ200 °C at a scan rate of 10 K/min. Tg values were determined using
the half-step method.
Mass spectra were acquired on a Biflex II MALDI-TOF MS system
(Bruker Daltonics, Germany). The measurements were carried out in
reflector mode and positive polarity using an ion acceleration voltage
of 20 kV.
Dynamic light scattering (DLS) measurements were performed using a
Zetasizer NanoZS Instrument (MALVERN Instruments, Worechestershire,
UK) equipped with a 4 mW HeꢀNe-laser (λ = 633 nm) and with
noninvasive backscattering (NIBS) detection at a scattering angle of 173°.
The solutions were prepared in ultrapure water and filtered using a 0.20 μm
disposable PTFE filter.
The cloud point temperature of the polymer solutions was measured
with a Specord M40 UV/vis spectrometer from Analytik Jena
(Germany). The spectrometer is equipped with a Peltier cell holder
for temperature control. The turbidity of the solutions was monitored as
a function of temperature at a specific wavelength λ of 400 nm and under
continuous stirring. Solutions were prepared in a concentration of 2 mg/mL
using distilled water.
Materials. The azide-functionalized β-cyclodextrin (β-CD) mono-
(6-azido-6-desoxy)-β-cyclodextrin was synthesized according to the
literature15 and monofunctionalization was proven (details on synthesis
and characterization see Supporting Information). 2-Methyl-2-oxazoline
(Aldrich, 98%) was purified by distillation over calcium hydride (CaH2).
2-(Isopropylamino)ethanol (Aldrich, 70%) was purified by distillation.
The following chemicals were purchased from Aldrich and usedas received
(purity level): 2-bromoisobutyryl bromide (98%), 3-(trimethylsilyl)-pro-
pargyl alcohol (99%), tetrabutylammonium fluoride solution (1.0 M in
THF), 1-adamantanemethanol (99%), p-bromomethylbenzoyl bromide
(96%), triethylamine (99.5%), N,N-diisopropylethyl amine (99%),
copper(I) iodide (99.5%), copper(I) bromide (99.999%), N-isopropyla-
crylamide (NIPAAm) (99%), and tris[2-(dimethylamino)ethyl]amine
(Me6TREN). All solvents were purchased from Aldrich and used without
further purification.
’ EXPERIMENTAL SECTION
Methods. GPC measurements were performed on a Polymer
Laboratories PL-GPC 50 Plus Integrated GPC System equipped with
a microvolume double piston pump of Agilent Technologies. N,N-
Dimethylacetamide (DMAc) containing 3 g/L LiCl was used as eluent at
a flow rate of 1.0 mL/min. A PolarGel-M column of Polymer Labora-
tories connected to a Mini DAWN light scattering (LS) detector of
Wyatt Technology was applied. All polymers were dissolved in DMAc
and filtered through a 0.2 μm PTFE filter. Data were processed with
ASTRA software.
Synthesis. 2-Bromoisobutyryl Propargylester (A). The alkyne A
was synthesized according to the given description (see Supporting
Information).
NMR spectra were recorded on a Bruker Avance III 500 NMR
spectrometer operating at 500.13 MHz for 1H and at 125.75 MHz for
13C using a 5 mm quad (1H, 13C, 19F, 31P) inverse probe, equipped with
a shielded z-gradient coil. The solvent signals were used as internal
standard: DMSO-d6: δ(1H) = 2.50 ppm, δ(13C) = 39.6 ppm and
CDCl3: δ(1H) = 7.26 ppm, δ(13C) = 77.0 ppm, respectively. Spectra
recorded from D2O solutions were referenced on external sodium
3-(trimethylsilyl)propionate-2,2,3,3-d4 (δ(1H) = 0 ppm). These spectra
were recorded applying a spectral width of 6000 Hz digitized into 64K
points and a recycle time of 10 s. The concentration of both components
in the complexation experiments was 5 ꢁ 10ꢀ3 M for the 1:1 complex.
However, due to titration of one component by the second one the
concentrations are different at different degree of complexation. For the
temperature-dependent experiments and for the 2D ROESY experi-
ments also 5 ꢁ 10ꢀ3 M solutions of the 1:1 complex were used. For
Initiator I-1. I-1 was synthesized according to the literature.32 Using
argon as protection gas, 1-adamantanemethanol (1 g, 6.02 mmol) was
dissolved in dichloromethane (30 mL) and dry Et3N (0.61 g, 6.01
mmol). After cooling to 0 °C, p-bromomethylbenzoyl bromide (2.39 g,
8.60 mmol) was added step by step. Afterward, the solution was stirred
under reflux for 1 h. The solvent and residual Et3N was removed under
vacuum and the crude product was purified by means of column
chromatography using silica gel with hexane/dichloromethane as eluent
(gradient: 10ꢀ30% dichloromethane) yielding I-1 as a colorless solid.
Yield: 1.77 g (81%). Mp: 75 °C.
1H NMR (CDCl3): δ (ppm) = 8.03 (d, 2H, Hh), 7.46 (d, 2H, Hi),
4.50 (s, 2H, Hl), 3.93 (s, 2H, He), 2.02 (m, 3H, Hb), 1.77 and 1.70 (AB
spin system, 2JHH = 12.1 Hz, 6H, Ha), 1.64 (d, 3JHH = 2.4 Hz, 6H, Hc).
13C NMR (CDCl3): δ(ppm) = 166.08 (Cf), 142.50 (Ck), 130.60 (Cg),
1
temperature-dependent H NMR measurements, the temperature was
3251
dx.doi.org/10.1021/ma200048a |Macromolecules 2011, 44, 3250–3259

Macromolecules
ARTICLE
130.05 (Ch), 129.01 (Ci), 74.37 (Ce), 39.45 (Cc), 37.00 (Ca), 33.56
(Cd), 32.25 (Cl), 28.09 (Cb).
Anal. Calcd for C19H23BrO2: C, 62.82; H, 6.38. Found: C, 62.87;
H, 6.51.
Initiator I-2. 2-Bromoisobutyryl propargyl ester (0.1 g, 0.49 mmol)
and copper(I) iodide (20 mg, 0.11 mmol) were added under nitrogen
atmosphere to a solution of mono(6-azido-6-desoxy)-β-cyclodextrin
(0.3 g, 0.26 mmol) (ratio alkyne/azide =1.9: 1) in 30 mL DMF,
Afterward, diisopropylethyl amine (0.13 g, 1.0 mmol) was added and
the solution was allowed to stir under nitrogen atmosphere for 14 h at
room temperature. For working up, the mixture was poured into 200 mL
of toluene. Next, about 100 mL of n-hexane were added, leading to
precipitation of the product. The yellowish solid was isolated by
filtration and washed several times with diethyl ether. Yield: 310 mg
(87%). Mp: 227 °C.
Figure 1. Synthetic strategy for noncovalent diblock formation via
β-CD-adamantane complexation.
1H NMR (DMSO-d6): δ (ppm) = 8.12 (s, 1H, H7), 5.9ꢀ5.6 (14H,
0
CHꢀOH), 5.23 (s, 2H, H9), 5.03 and 4.9ꢀ4.7 (7H, H1,1 ), 4.88 and 4.60
0
0
(m, 2H, H6 ), 4.55ꢀ4.2 (6H, CH2ꢀOH), 4.00 (t, 1H, H5 ), 3.8ꢀ3.5
1H NMR (DMSO-d6): δ (ppm) = 8.02 (H7), 7.5ꢀ6.9 (H16), 5.9ꢀ5.6
0
0
0
(23H, H3,5,6,3 ), 3.4ꢀ3.2 (H2,4,2 ,4 ; overlaps with DHO), 3.08 and 2.89
(m, 2H, H6 of the glucose unit bonded to C1 ),16 1.88 (s, 6H, H12). 13
C
0
0
(CHꢀOH), 5.05 (H9), 5.03 and 4.9ꢀ4.75 (H1,1 ), 4.85 and 4.62 (H6 ),
0
0
4.55ꢀ4.2 (CH2ꢀOH), 3.98 (H5 ), 3.95ꢀ3.75 (H17), 3.75ꢀ3.5
NMR (DMSO-d6): δ (ppm) = 170.44 (C10), 141.35 (C8), 125.67 (C7),
0
0
0
(H3,5,6,3 ), 3.5ꢀ3.2 (H2,4,2 ,4 ; overlaps with DHO), 3.13 and 2.94 H6
0
0
102.5 - 101.5 (C1), 101.24 (C1 ), 83.35 (C4 ), 82.0 ꢀ 81.0 (C4), 73.5 ꢀ
0
of the glucose unit bonded to C1 ), 2.2ꢀ1.8 (H14), 1.7ꢀ1.2 (H13), 1.05
0
0
0
71.5 (C2,3,5,2 ,3 ), 69.87 (C5 ), 60.5 ꢀ59.5 (C6), 58.96 (C6 of the glucose
(H18). 13C NMR: (DMSO-d6) δ (ppm) = 173.4 (C15), 141.9 (C8),
0
0
unit bonded to C1 ), 58.77 (C9), 56.94 (C11), 50.35 (C6 ), 30.16 (C12).
Ada-PMeOxa (P-1). The adamantyl-containing initiator (I-1) (0.10 g,
0.28 mmol) was dissolved in dry benzonitrile (4.3 mL) under argon
atmosphere. Next, dried 2-methyl-2-oxazoline (1.17 g, 13.76 mmol) was
added. The solution was heated to 100 °C and stirred for 1.5 h. For
determining the monomer conversion, a sample was taken and diluted
with deuterated chloroform for NMR measurement. Without cooling
the solution, 2-(isopropylamino)ethanol (0.28 g, 2.75 mmol) was added
all at once for chain termination. The solution was stirred for further 10
min at 100 °C. After cooling to room temperature, the solution was
diluted with dichloromethane and precipitated twice in diethyl ether
(400 mL). The product was separated and dried under vacuum at 40 °C
for 16 h. For subsequent dialysis, the solid was dissolved in water
(12 mL) and a small amount of acetone was added until the solution
became clear. After dialysis against water (800 mL) with cellulose
membranes of MWCO1000, freeze-drying and subsequent drying under
vacuum at 40 °C for 2 days, polymer P-1 was obtained as a yellow solid.
Yield: 957 mg (74%).
0
0
125.7 (C7), 102.5ꢀ102 (C1), 101.5 (C1 ), 83.4 (C4 ), 82ꢀ81 (C4),
0
0
0
73.5ꢀ71.5 (C2,3,5,2 ,3 ), 69.9 (C5 ), 61.5ꢀ59.0 (C6), 57.2 (C9), 50.3
0
(C6 ), 43ꢀ41 (C14), 41.5ꢀ41 (C17), 38ꢀ33 (C13), 24ꢀ22 (C18).
Self-Assembled Ada-PMeOxa/β-CD-PNIPAAm (sa P-1/P-2). In a
NMR tube, β-CD-PNIPAAm (P-2, 5.2 mg) was dissolved in D2O
(0.7 mL). Next, a D2O solution (4.8 mg/mL) of Ada-PMeOxa (P-1)
was added in portions until complete complexation was achieved as
confirmed by NMR spectroscopy. Because of weighing errors for small
amounts of substance and the inaccuracy of the Mn values, this is a
proper approach to prepare the 1:1 complex. Subsequently, the solvent
was removed under vacuum to yield sa P-1/P-2 as a white solid.
’ RESULTS AND DISCUSSION
Polymer Synthesis. Our approach toward the target PNI-
PAAm-PMeOxa diblock copolymers through self-assembly (sa)
includes the synthesis of initiators bearing the units for com-
plexation (β-cyclodextrin and adamantane, respectively) fol-
lowed by polymerization via controlled methods. Mixing the
obtained polymers in water is expected to give the respective
diblock structure via self-assembly (Figure 1).
Figure 2 displays the reaction sequence for the preparation
of the adamantane-functionalized poly(2-methyl-2-oxazoline)
(Ada-PMeOxa, P-1). Initiator I-1 was obtained via esterification
of 4-bromomethylbenzoyl bromide with 1-adamantanemetha-
nol. The cationic polymerization of 2-methyl-2-oxazoline with
I-1 was conducted at 100 °C for about 2 h and then stopped with
the capper molecule 2-(isopropylamino)ethanol.
1H NMR (CDCl3): δ (ppm) = 8.1 ꢀ 7.9 (Hh), 7.4ꢀ7.2 (Hi),
4.7ꢀ4.55 (Hl), 3.90 (He), 3.6ꢀ3.3 (Ho,s), 2.92 (Hp), 2.61 (Hr), 2.25 ꢀ
2.05 (Hn), 2.00 (Hb), 1.75 and 1.67 (AB spin system, Ha), 1.62 (Hc),
1.02 (Hq). 13C NMR (CDCl3): δ (ppm) = 172ꢀ170 (Cm), 166.1 (Cf),
130.5ꢀ129.5 (Cg,h), 126.3 (Ci), 74.5 (Ce), 53ꢀ48.5 (Cl), 48.5 ꢀ 43
(Co), 39.41 (Cc), 36.94 (Ca), 33.51 (Cd), 28.03 (Cb), 21.1 (Cn), 18.2
(Cq). Signals of Ck, Cr, and Cp could not be identified.
β-CD-PNIPAAm (P-2). In a 10 mL Schlenk flask, β-CD containing
initiator I-2 (0.1 g, 0.074 mmol), NIPAAm (0.282 g, 2.5 mmol) and
Me6TREN (0.017 mL, 0.074 mmol) were dissolved in a mixture of
1.0 mL of water (Millipore) and 1.9 mL of DMF. The solution was
degassed by means of four consecutive “freeze-pump-thaw cycles”. Next,
CuBr (7.3 mg, 0.0051 mmol) as catalyst was added to the mixture. After
one further “freezeꢀpumpꢀthaw cycle”, the mixture was allowed to
thaw in order to start the room temperature ATRP. After 1 h, the
polymerization was quenched by abrupt freezing with liquid nitrogen
followed by several drops of THF. The mixture was thawed and the
solvent was removed under reduced pressure. For purification, the crude
product was dissolved in 30 mL water which was slowly heated above the
lower critical solution temperature. The precipitated polymer was
separated and dried in vacuum at 50 °C to yield a colorless polymer.
Yield: 172 mg (45%).
1
Figure 3 shows the H NMR spectra of initiator I-1 and
polymer P-1 recorded in DMSO-d6. Both structures could be
clearly confirmed via signal assignments. The characteristic
signals of initiator I-1 are identified again in the polymer
spectrum. While the signal of the benzyl protons of I-1 (signal
l in Figure 3a) appears as a singlet, it changes to a multiplet in the
polymer spectrum (signal l in Figure 3b) because of hindered
rotation of the adjacent NꢀC bond due to amide mesomerism
resulting in a varying chemical environment at position l. The
characteristic adamantane signals (aꢀc) appear unchanged at
3252
dx.doi.org/10.1021/ma200048a |Macromolecules 2011, 44, 3250–3259

Macromolecules
ARTICLE
Figure 2. Reaction sequence for the synthesis of the adamantane-functionalized poly(2-methyl-2-oxazoline) (P-1).
Figure 3. 1H NMR spectra of initiator I-1 (a) and polymer P-1 (b) recorded in DMSO-d6.
1.5ꢀ2.0 ppm. P-1 was further studied by MALDI-TOF mass
spectrometry (see Supporting Information).
initiator (I-2) in hand allows for the synthesis of well-defined
CD-initiated polymers of all monomers polymerizable via ATRP.
In this work, initiator I-2 was employed to start the polymeriza-
tion of NIPAAm. The catalyst system CuBr/Me6TREN was
used, which allows for room temperature ATRP. In the 1H NMR
spectrum of P-2 (Figure 5b), the characteristic β-CD-signals of
the initiator I-2 are clearly identified.
The reaction sequence for the synthesis of the β-CD-functio-
nalized poly(N-isopropylacrylamide) (β-CD-PNIPAAm, P-2) is
shown in Figure 4. Starting with the monoazide-functionalized
β-cyclodextrin (CD-N3) (see ref 15 and Supporting Information
for synthesis and characterization), the ATRP-initiating moiety
is introduced via the click reaction with 2-bromoisobutyryl
propargylester (A) (ratio alkyne/azide =1.9:1). Next, the CD-
functionalized initiator (I-2) was employed for initiation of the
ATRP of NIPAAm to yield polymer P-2.
The GPC chromatograms of the polymers P-1 and P-2,
obtained with a light scattering detector, are shown in Figure 6.
The curves show narrow molar mass distributions, indicating a
high degree of control of the corresponding polymerizations. In
Table 1, synthesis data and characteristics of both polymers P-1
and P-2 are summarized.
The click reaction with the alkyne A resulted in a quantitative
functionalization of CD-N3, as confirmed by high yield of the
1
product and the fully assigned H NMR spectrum (efficient
For P-1, the molar mass calculated from the 1H NMR spectrum
(4200 g/mol), the one obtained by GPC using LS-detection (4600
g/mol) and the theoretical molar mass (Mn,th = 4300 g/mol) agree
triazole ring formation seen e.g. by appearance of signal 7 and
shift of signals 60 and 9, Figure 5a). Having the CD-functionalized
3253
dx.doi.org/10.1021/ma200048a |Macromolecules 2011, 44, 3250–3259

Macromolecules
ARTICLE
Figure 4. Reaction sequence for the synthesis of β-CD-PNIPAAm (P-2) via ATRP.
1
16
0
Figure 5. H NMR spectra of initiator I-2 (a) and polymer P-2 (b) recorded in DMSO-d6. Signals 6* are due to H6 of the glucose unit bonded to C1 .
molar mass of polymer P-2 was expected (4800 g/mol at a
conversion of 100%). The observed deviation is probably due to
a relatively low efficiency of initiator I-2. Additional ROESY
experiment performed with the initiator I-2 in a D2O/DMF-d7
mixture (1:1.9 v/v according to conditions of polymer synthesis)
indicated that the methyl groups of the bromoisobutyryl unit can
be included in the CD ring. Thus, partial and dynamic dimer
formation of I-2 is in principle possible which can explain the low
initiating efficiency. However, the narrow and symmetric molar
mass distribution together with the low PDI of polymer P-2 give
rise to assume a controlled ATRP process. Thus, two water-
soluble polymers, equipped with units for specific complexation
(adamantane and β-CD, respectively) have been prepared.
Complexation Studies. To verify the interaction of β-CD-
and adamantane-type polymer end groups in aqueous solution,
complexation studies were carried out by 1H NMR spectroscopy.
Generally, NMR spectroscopy is the most important method to
study the structure of hostꢀguest interactions. Mainly the
observations of chemical shift changes, the nuclear Overhauser
Figure 6. GPC chromatograms of P-1 (solid line) and P-2 (dashed
line) obtained with LS-detector.
very well. This result together with the low PDI of P-1 indicates a
controlled polymerization process.
For P-2, the value determined by GPC via LS-detection
(Mn,GPC = 9600 g/mol) and the one calculated from the 1H NMR
spectrum (Mn,NMR = 9300 g/mol) show a good agreement.
However, considering the ratio [monomer]/[initiator], a lower
3254
dx.doi.org/10.1021/ma200048a |Macromolecules 2011, 44, 3250–3259

Macromolecules
ARTICLE
Table 1. Synthesis data and characteristics of P-1 and P-2
polymer
[monomer]/[initiator]
yield [%]
conversion [%]
Mn,th [g/mol]
Mn,GPCe [g/mol]
Mn,NMR [g/mol]
PDIe
P-1
P-2
50
30
58a
45b
92c
-
4300d
4800g
4600i
9600j
4200f
9300h
1.10
1.10
a Yield = mpolymer/(mmonomer þ minitiator þ ninitiatorMcapper ꢀ ninitiatorMHBr) ꢁ 100, obtained after purification. b Yield = mpolymer/(mmonomer þ minitiator
)
ꢁ 100, obtained after purification. c Monomer conversion: determined via NMR. d Mn,th = conversion/100[monomer]/[initiator]Mmonomer þ Minitiator
þ Mcapper ꢀ MHBr. e Determined via GPC (Light scattering (LS)-detector, solvent: DMAc þ 3 g/L LiCl). f Calculated via NMR: integration of CH3
signal of the repeating units vs aromatic protons signal of the initiating moiety, estimated relative error: ( 5%. g Mn,th = [monomer]/[initiator]Mmonomer
þ Minitiator. (For this sample, Mn,th was calculated for 100% conversion; the actual monomer conversion could not be determined). h Calculated via
NMR: integration of CH3 and CH2 signal of the repeating units (0.7ꢀ1.8 ppm) vs. proton signals 1, 10, 6’ and 9 (4.55 ꢀ 5.2 ppm) of the CD-moiety,
estimated relative error: ( 5%. i dn/dc = ꢀ0.104 mL/g. j dn/dc = ꢀ0.105 mL/g.
Figure 7. Regions from 1H NMR spectra obtained for different ratios of (a) P-1 and successively added β-CD, (b) P-1 and successively added P-2, and(c) P-2
and successively added P-1 (D2O, 30 °C). The molar ratios of Ada and β-CD moiety are given. The final concentration of both moieties was 5 ꢁ 10ꢀ3 M.
effects (nOes) but also of the diffusion behavior are major tools
in structural studies of such complexes.33ꢀ35
mode was reported to be very stable for a structural similar
β-CD/sodium 4-(1-adamantanyl)benzoate complex.36
In a first step, the complexation of the adamantyl group of P-1
by β-CD was studied. Figure 7a depicts the signals of the benzoyl
spacer between adamantane moiety and PMeOxa chain for
different molar ratios of Ada-PMeOxa (P-1) and successively
added β-CD. Changes in chemical shift and signal shape which
were also observed for the signals of the adamantyl group itself
prove an increasing content of complexed adamantyl end groups
reaching almost complete complexation at 1:1 ratio in accor-
dance with the large equilibrium constants of such complexes.29
The significant narrowing of the signals of Hh and Hi in the
course of titration is probably due to the different surrounding of
the hydrophopic adamantane and benzoyl spacer moiety in D2O
and in the complex. The solvation should be poor in D2O maybe
resulting in ill-defined associates and interactions with the back-
bone whereas the well-defined surrounding in the β-CD cavity
results in narrower lines as also observed in the good solvent
DMSO-d6 (Figure 3). The structure of the formed inclusion
complex was studied by a ROESY spectrum (Figure 8). Intense
correlation peaks between the aromatic protons and H5 and H6
of β-CD indicate that β-CD deeply includes the whole adamantyl
end group together with the benzoyl spacer from the primary, i.e.,
the slightly narrower side of the β-CD ring. Accordingly, Ha and
Hb of the adamantyl unit show correlation peaks with H3 of
β-CD and Hc with both H3 and H5 of β-CD. Such an insertion
Parts b and c of Figure 7 represent regions from spectra
obtained by titration of P-1 with P-2 (Figure 7b) and of P-2 with
P-1 (Figure 7c), respectively. Whereas in the first case the
changes in signal position and shape of the benzoyl protons
indicate increasing content of complexed adamantane moieties,
the increasing amount of β-CD rings covering adamantane
0
moieties is indicated by changes in the H1,1 and H7 protons
region of the β-CD end group. Other signals which could also be
used to follow the complexation process, i.e., the adamantane
protons and H3 and H5 of β-CD, are hidden by the large polymer
signals. Finally, the same spectra were obtained at equimolar ratio
of guest and host end group (top spectra in Figures 7b and 7c).
Within the limits of the method, these findings prove formation
of the desired inclusion complex with 1:1 stoichiometry. Un-
fortunately, a ROESY spectrum of this polymer complex gives no
evaluable correlation peaks. Thus, the insertion mode could not
be revealed directly. However, the changes observed for the
benzoyl proton signals (Figure 7b) are similar to those observed
for the complexation with free β-CD (Figure 7a). This suggests
that also for polymerꢀpolymer complexation the adamantyl end
group is included in the β-CD cavity from the primary side. At
first glance, this seems to be unlikely because the triazole ring,
which connects the β-CD moiety with the PNIPAAm chain, is
also located on this side. However, its flexibility is considered to
3255
dx.doi.org/10.1021/ma200048a |Macromolecules 2011, 44, 3250–3259

Macromolecules
ARTICLE
Figure 8. 2D ROESY spectrum of a 1:1 molar mixture of P-1 and β-CD (both 5 ꢁ 10ꢀ3 M) in D2O at 30 °C (mixing time τm = 248 ms) and schematic
representation of the detected insertion mode in the hostꢀguest complex between the adamantane containing end group of P-1 and β-CD.
be sufficiently high to allow for insertion of the adamantane
moiety. Nevertheless, a direct prove of this insertion mode
cannot be presented and thus, also adamantane insertion from
the secondary side could occur or an equilibrium of both modes.
Harada et al. reported such multimodes for the β-CD complexa-
tion with the adamantyl moieties of poly(4-(1-adamantyl)phenyl
vinyl ether-co-sodium maleate).36
Furthermore, dynamic light scattering (DLS) was tried to
verify the complexation of P-1 and P-2. The interaction between
β-CD and adamantane is known to be particularly strong in
water. However, DLS measurements in water turned out to be
not conclusive due to additional strong intermolecular interac-
tions and hence the formation of unspecific aggregates of varying
particle sizes was observed. Therefore, further DLS measure-
ments were done in dimethylacetamide (DMAc) in which the
formation of unspecific aggregates is minimized and hence the
particle sizes should be determined properly. However, in DMAc
the binding constants for the Ada-CD complex are significantly
smaller, and thus, it was only possible to determine the average
size of the starting polymers P1 (4 nm) and P-2 (6 nm) but no
efficient block copolymer formation could be confirmed by this
methods (see Supporting Information). Similarly, GPC measure-
ments on the self-assembled block copolymer showed the same
limitations: strong aggregate formation in water as it was
expected for that type of double-hydrophilic block copolymer
and different complexation behavior in DMAc. The complex
solution and aggregation behavior of the sa P-1/P-2 block
copolymers thus has to be the subject of a forthcoming more
detailed study.
Thermoresponsive Behavior. With the self-assembled PNI-
PAAmꢀPMeOxa diblock copolymer in hand, the concept of
heat-induced hydrophilicꢀhydrophobic switching of the PNI-
PAAm block in aqueous solution was studied with respect to the
stability of the adamantane/β-CD inclusion complex. First, such
diblock copolymer was prepared by defined mixing of the
polymers in D2O under NMR observation (as described above)
until complete complexation of the adamantane moiety by the
β-CD unit was detected. Then the solvent was evaporated before
redissolving the diblock complex for further analysis. The lower
critical solution temperature (LCST) of the initial β-CD-PNI-
PAAm (P-2) and the yielded self-assembled complex (sa P-1/
P-2) was determined by turbidimetry measurements using
UVꢀvis spectrophotometry. Figure 9a shows the transmittance
vs temperature plots (cloud point curves) for P-2 and sa P-1/
P-2. Both polymers show distinct heat-induced phase transitions
with 50% transmittance points at about 37 and 41.5 °C, respec-
tively, with a significant broadening for sa P-1/P-2. Generally,
the LCST of PNIPAAm can be influenced by several factors, e.g.
Figure 9. Transmittance (a) and integrated NMR intensity of the H18
signal (b) as a function of temperature for an aqueous solution of P-2
and of the complex sa P-1/P-2, respectively (turbidimetry, 2 mg/mL
solution in H2O; 1H NMR, 5 ꢁ 10ꢀ3 M solution in D2O).
molecular weight, polydispersity and end groups, making the
interpretation of such small effects problematic.37ꢀ39 Compared
to the literature value of 32 °C reported for high-molecular-
weight linear PNIPAAm,38 the increased LCST of P-2 can be
attributed to both the rather low molecular weight (Mn,NMR
=
9300 g/mol) and to the hydrophilic end group. The effect of
complexation with P-1 toward block copolymer structure on the
LCST is moderate but obvious. Thus, extending the hydrophilic
part by complexation with the water-soluble PMeOxa further
increases the LCST by 4.5 K and leads to a less sharp transition.
The increase in LCST is consistent with the behavior observed by
Trellenkamp et al. with a different polymer system. In that study,
a linear correlation was found between the adjusted hydrophili-
city and the corresponding LCST of copolymers composed of
N-vinyl-2-pyrrolidone and 3-ethyl-1-vinyl-2-pyrrolidone.40
Several 1H NMR techniques provide a deeper insight into the
behavior of structural units during such phase transitions, i.e.,
allow for following the temperature-induced processes on mole-
cular level.41 Thus, we employed 1H NMR spectroscopy to study
the temperature effect on both the spectrum of β-CD-PNIPAAm
and of the block copolymer complex. The coilꢀglobule phase
transition at the LCST is accompanied by a virtual disappearance
of signals. This is due to the fact that at temperatures above the
LCST the mobility of most PNIPAAm units is reduced to such an
extent that the lines become too broad to be detected in solution-
state spectra. Thus, the integrated signal intensities can be used to
characterize the phase transition behavior quantitatively.
Figure 9b depicts the temperature dependence for the PNI-
PAAm methyl group signal obtained for P-2 and the complex sa
P-1/P-2. The intensity decrease with phase transition is less
steep than observed in the turbidity measurements but the
LCSTs correspond well: P-2 37 °C (turb.) vs 38.5 °C (NMR)
and sa P-1/P-2: 41.5 vs 43 °C. Thus, both methods show the
same trend in increase in LCST upon block copolymer formation
3256
dx.doi.org/10.1021/ma200048a |Macromolecules 2011, 44, 3250–3259

Macromolecules
ARTICLE
and the small differences in the absolute LCST values
determined by the two methods can be explained by the
different block copolymers concentration used for the mea-
surements. Again, the increase in LCST temperature due to
complex formation is significant and is attributed to a “soft-
ening” effect on the PNIPAAm globules by the extended
hydrophilic part in the self-assembled diblock and maybe also
due to partial incorporation of PMeOxa in the outer sphere of
the globules.
DSC measurements proved that both blocks are fully miscible
(Figure 10) with a single transition temperature (Tg) of 92 °C for
sa P-1/P-2 (P-1: 79 °C; P-2: 120 °C). Such a “softening” effect is
the most probable explanation for the slightly higher content of
still mobile component detected at the same temperature for sa
P-1/P-2 compared with P-2.
Figure 11 shows a series of 1H NMR spectra obtained for sa
P-1/P-2 at different temperatures. The previously discussed loss
of PNIPAAm signal intensity due to restricted mobility in the
collapsed globular structures is obvious. The question arises
whether the diblock copolymer complex still exists at higher
temperatures or a separation in PNIPAAm globules and free
moving PMeOxa coils occurs. If one compares the spectrum of sa
P-1/P-2 (top spectrum of Figure 11b) and the corresponding
spectrum of P-1 (Figure 11a), both recorded at 70 °C, the
differences in signal position and shape are obvious for the
aromatic and adamantyl protons of the Ada moiety. This suggests
that PMeOxa is not released from the complex to a great extent.
This is also in accordance with the almost unchanged signal
position for the signals of Hh and Hi over the whole temperature
region studied. The only change is in signal intensity and signal
shape. The first effect is attributed to partial incorporation of
connecting complexed PMeOxa block in the outer sphere of the
collapsed PNIPAAm globules thus also resulting in reduced
mobility and virtual signal loss. Only the still mobile chain
components of the diblock complexes were detected. Surpris-
ingly, the signal shapes at about the LCST become very similar to
those observed for β-CD/P-1 complexes (Figure 7a), which was
assigned by ROESY to the more stable insertion mode of Ada
moiety from the primary side. This suggests that below LCST the
complex in sa P-1/P-2 exists in different modes but at higher
temperatures the most stable one with insertion from the primary
side strongly dominates. The signal pattern of the PMeOxa
backbone signals change slightly, i.e. see signal group of Hn in
Figure 11b, but there is no loss of intensity. Furthermore, the
Figure 10. DSC curves (2nd heating cycle) of P-1 (cycles), P-2 (stars)
and sa P-1/P-2 (squares).
Figure 11. Regions from (a) the 1H NMR spectrum of P-1 (5 ꢁ 10ꢀ3 M in D2O) at 70 °C and (b) 1H NMR spectra measured at different temperatures
on a 5 ꢁ 10ꢀ3 M solution of sa P-1/P-2 (1:1 complex) in D2O.
3257
dx.doi.org/10.1021/ma200048a |Macromolecules 2011, 44, 3250–3259

Macromolecules
ARTICLE
backbone signals are identical for P-1 and sa P-1/P-2 at, e.g.,
70 °C (Figure 11a and top trace of Figure 11b) proving that this
effect is not related to complex formation but to conformational
changes of PMeOxa. Conclusively, at temperatures above LCST
a globular core of collapsed PNIPAAm chains is surrounded by
PMeOxa chains tethered by the β-CD/Ada complex formed
from the corresponding polymer end group functionalities
(scheme in Figure 11). This structure is similar to that of
PNIPAAm-graft-PAlkylOxa copolymers reported in previous
studies.31,42
analysis. Dr. Karin Sahre is thanked for MALDI studies. We
would further like to thank Dr. Maricica Munteanu for synthesis
of mono(6-azido-6-desoxy)-β-cyclodextrin.
’ REFERENCES
(1) Kollisch, H. S.; Barner-Kowollik, C.; Ritter, H. Chem. Commun.
2009, 1097–1099.
(2) Cesteros, L. C.; Gonzalez-Teresa, R.; Katime, I. Eur. Polym. J.
2009, 45, 674–679.
(3) Nozaki, T.; Maeda, Y.; Kitano, H. J. Polym. Sci., Polym. Chem.
1997, 35, 1535–1541.
’ CONCLUSIONS
(4) van de Manakker, F.; Vermonden, T.; van Nostrum, C. F.;
Hennink, W. E. Biomacromolecules 2009, 10, 3157–3175.
(5) Li, J.; Loh, X. J. Adv. Drug Delivery Rev. 2008, 60, 1000–1017.
(6) Li, J. S.; Xiao, H. N.; Li, J. H.; Zhong, Y. P. Int. J. Pharm. 2004,
278, 329–342.
In this contribution, we described the combination of two water-
soluble polymers to a defined thermoresponsive diblock copolymer
through the noncovalent interaction of β-CD and adamantane. For
that we developed a new synthetic strategy for the preparation of an
adamantane end-functionalized poly(2-methyl-2-oxazoline) and a
β-CD end-functionalized poly(N-isopropyl-acrylamide). The ap-
proach presented here includes the controlled polymerization of
NIPAAm and 2-methyl-2-oxazoline with initiators that are functio-
nalized with functional moieties able to form strong inclusion
complexes. In this way, well-defined polymers were obtained,
which undergo self-assembly to form the supramolecular diblock
structure poly(2-methyl-2-oxazoline-block-N-isopropylacrylamide).
The self-organized block formation could be verified undoubt-
edly by means of detailed NMR analysis.
Thus, we demonstrated for the first time the preparation of a
well-defined diblock copolymer formed through self-assembly of
cyclodextrin and adamantyl end-functionalized polymers. The
described diblock assembly represents a double-hydrophilic
system that is switchable to a hydrophilicꢀhydrophobic config-
uration by adjusting temperature. As detected by NMR spec-
troscopy, this thermal process leads to reversible polymer
aggregation where a globular core of collapsed PNIPAAm chains
is surrounded by PMeOxa chains, thus, representing a second
self-assembly process which is thermally switchable. Therefore,
this strategy leads to new reversibly thermoresponsive noncova-
lent diblock copolymers which can be fine-tuned in their
aggregation behavior with potential use as drug delivery system.
(7) Xin, J. Y.; Guo, Z. Z.; Chen, X. Y.; Jiang, W. F.; Li, J. S.; Li, M. L.
Int. J. Pharm. 2010, 386, 221–228.
(8) Li, L.; Guo, X. H.; Wang, J.; Liu, P.; Prud’homme, R. K.; May,
B. L.; Lincoln, S. F. Macromolecules 2008, 41, 8677–8681.
(9) Kretschmann, O.; Choi, S. W.; Miyauchi, M.; Tomatsu, I.;
Harada, A.; Ritter, H. Angew. Chem., Int. Ed. 2006, 45, 4361–4365.
(10) Kretschmann, O.; Steffens, C.; Ritter, H. Angew. Chem., Int. Ed.
2007, 46, 2708–2711.
(11) Charlot, A.; Auzely-Velty, R. Macromolecules 2007, 40,
9555–9563.
(12) Guerrouache, M.; Millot, M. C.; Carbonnier, B. Macromol.
Rapid Commun. 2009, 30, 109–113.
(13) Gonzalez-Campo, A.; Hsu, S. H.; Puig, L.; Huskens, J.;
Reinhoudt, D. N.; Velders, A. H. J. Am. Chem. Soc. 2010, 132,
11434–11436.
(14) Choi, S.; Munteanu, M.; Ritter, H. J. Polym. Res. 2009, 16,
389–394.
(15) Trellenkamp, T.; Ritter, H. Macromolecules 2010, 43, 5538–
5543.
(16) Nielsen, T. T.; Wintgens, V.; Amiel, C.; Wimmer, R.; Larsen,
K. L. Biomacromolecules 2010, 11, 1710–1715.
(17) Lehn, J. M. Polym. Int. 2002, 51, 825–839.
(18) De Greef, T. F. A.; Smulders, M. M. J.; Wolffs, M.; Schenning,
A.; Sijbesma, R. P.; Meijer, E. W. Chem. Rev. 2009, 109, 5687–5754.
(19) Cordier, P.; Tournilhac, F.; Soulie-Ziakovic, C.; Leibler, L.
Nature 2008, 451, 977–980.
(20) Folmer, B. J. B.; Sijbesma, R. P.; Versteegen, R. M.; van der Rijt,
J. A. J.; Meijer, E. W. Adv. Mater. 2000, 12, 874–878.
(21) Feldman, K. E.; Kade, M. J.; Meijer, E. W.; Hawker, C. J.;
Kramer, E. J. Macromolecules 2010, 43, 5121–5127.
(22) Tang, C. B.; Lennon, E. M.; Fredrickson, G. H.; Kramer, E. J.;
Hawker, C. J. Science 2008, 322, 429–432.
(23) Ott, C.; Kranenburg, J. M.; Guerrero-Sanchez, C.; Hoeppener,
S.; Wouters, D.; Schubert, U. S. Macromolecules 2009, 42, 2177–2183.
(24) Chiper, M.; Meier, M. A. R.; Wouters, D.; Hoeppener, S.;
Fustin, C. A.; Gohy, J. F.; Schubert, U. S. Macromolecules 2008,
41, 2771–2777.
(25) Chiper, M.; Fournier, D.; Hoogenboom, R.; Schubert, U. S.
Macromol. Rapid Commun. 2008, 29, 1640–1647.
(26) Ozyurek, Z.; Voit, B.; Krahl, F.; Arndt, K. F. e-Polym. 2010, 13.
(27) Chang, C.; Wei, H.; Quan, C. Y.; Li, Y. Y.; Liu, J.; Wang, Z. C.;
Cheng, S. X.; Zhang, X. Z.; Zhuo, R. X. J. Polym. Sci., Polym. Chem. 2008,
46, 3048–3057.
’ ASSOCIATED CONTENT
S
Supporting Information. Synthetic procedure of the
b
alkyne 2-bromoisobutyryl propargylester and the monoazide
β-CD, mass spectrometry results (EI-MS) of the adamantyl
functionalized initiator (I-1), experimental details and results
of the MALDI-TOF analysis of Ada-PMeOxa (P-1), proton
NMR results in DMSO of initiator I-1 and of polymer P-1 and
the DLS diagrams of P-1, P-2, and sa P-1/P-2 in DMAc are
provided. This material is available free of charge via the Internet
at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
*E-mail: voit@ipfdd.de.
(28) Wei, H.; Zhang, X. Z.; Cheng, C.; Cheng, S. X.; Zhuo, R. X.
Biomaterials 2007, 28, 99–107.
(29) Rekharsky, M. V.; Inoue, Y. Chem. Rev. 1998, 98, 1875–1917.
(30) David, G.; Alupei, V.; Simionescu, B. C.; Dincer, S.; Piskin, E.
Eur. Polym. J. 2003, 39, 1209–1213.
(31) Rueda, J.; Zschoche, S.; Komber, H.; Schmaljohann, D.; Voit, B.
Macromolecules 2005, 38, 7330–7336.
’ ACKNOWLEDGMENT
The authors thank Mr. Andreas Korwitz for assistance in
NMR measurements, Dr. Albena Lederer and Ms. Petra Treppe
for GPC measurements, and Ms. Liane H€aussler for DSC
3258
dx.doi.org/10.1021/ma200048a |Macromolecules 2011, 44, 3250–3259

Macromolecules
ARTICLE
(32) Gawronski, J. K.; Reddy, S. M.; Walborsky, H. M. J. Am. Chem.
Soc. 1987, 109, 6726–6730.
(33) Schneider, H. J.; Hacket, F.; Rudiger, V.; Ikeda, H. Chem. Rev.
1998, 98, 1755–1785.
(34) Wenz, G.; Han, B. H.; Muller, A. Chem. Rev. 2006, 106,
782–817.
(35) Cohen, Y.; Avram, L.; Frish, L. Angew. Chem., Int. Ed. 2005,
44, 520–554.
(36) Taura, D.; Taniguchi, Y.; Hashidzume, A.; Harada, A. Macro-
mol. Rapid Commun. 2009, 30, 1741–1744.
(37) Plummer, R.; Hill, D. J. T.; Whittaker, A. K. Macromolecules
2006, 39, 8379–8388.
(38) Schild, H. G. Prog. Polym. Sci. 1992, 17, 163–249.
(39) Xia, Y.; Burke, N. A. D.; Stover, H. D. H. Macromolecules 2006,
39, 2275–2283.
(40) Trellenkamp, T.; Ritter, H. Macromol. Rapid Commun. 2009,
30, 1736–1740.
(41) Spevacek, J. Curr. Opin. Colloid Interface Sci. 2009, 14, 184–191.
(42) Rueda, J. C.; Zschoche, S.; Komber, H.; Krahl, F.; Arndt, K. F.;
Voit, B. Macromol. Chem. Phys. 2010, 211, 706–716.
3259
dx.doi.org/10.1021/ma200048a |Macromolecules 2011, 44, 3250–3259