Thermoresponsive PS-b-PNIPAM-b-PS Micelles: Aggregation behavior, segmental dynamics, and thermal response

Article abstract of DOI:10.1021/ma902714p

We have studied the thermal behavior of amphiphilic, symmetric triblock copolymers having short, deuterated polystyrene (PS) end blocks and a large poly(N-isopropylacrylamide) (PNIPAM) middle block exhibiting a lower critical solution temperature (LCST) in aqueous solution. A wide range of concentrations (0.1-300 mg/mL) is investigated using a number of analytical methods such as fluorescence correlation spectroscopy (FCS), turbidimetry, dynamic light scattering (DLS), small-angle neutron scattering (SANS), and neutron spin-echo spectroscopy (NSE). The critical micelle concentration is determined using FCS to be 1 μM or less. The collapse of the micelles at the LCST is investigated using turbidimetry and DLS and shows a weak dependence on the degree of polymerization of the PNIPAM block, SANS with contrast matching allows us to reveal the core-shell structure of the micelles as well as their correlation as a function of temperature. The segmental dynamics of the PNIPAM shell are studied as a function of temperature and are found to be faster in the collapsed state than in the swollen state. The mode detected has a linear dispersion in q² and is found to be faster in the collapsed state as compared to the swollen state. We attribute this result to the averaging over mobile and immobilized segments.

Full text of DOI:10.1021/ma902714p

2
490 Macromolecules 2010, 43, 2490–2501
DOI: 10.1021/ma902714p
Thermoresponsive PS-b-PNIPAM-b-PS Micelles: Aggregation Behavior,
Segmental Dynamics, and Thermal Response
†
Joseph Adelsberger, Amit Kulkarni, Abhinav Jain, Weinan Wang,
†
†
†
‡
§
§
Achille M. Bivigou-Koumba, Peter Busch, Vitaliy Pipich, Olaf Holderer,
§
^
‡
†
Thomas Hellweg, Andr ꢀe Laschewsky, Peter M €u ller-Buschbaum, and
Christine M. Papadakis*
,†
†
Physikdepartment E13, Technische Universit a€ t M u€ nchen, James-Franck-Str. 1, 85747 Garching, Germany,
Institut f u€ r Chemie, Universit a€ t Potsdam, Karl-Liebknecht-Str. 24-25, 14476 Potsdam-Golm, Germany,
‡
§
Forschungszentrum J u€ lich GmbH, J u€ lich Centre for Neutron Science at FRM II, Lichtenbergstr. 1,
^
8
5747 Garching, Germany, and Physikalische Chemie I, Universit a€ t Bayreuth, Universit a€ tsstr. 30,
95447 Bayreuth, Germany
Received December 10, 2009; Revised Manuscript Received January 26, 2010
ABSTRACT: We have studied the thermal behavior of amphiphilic, symmetric triblock copolymers having
short, deuterated polystyrene (PS) end blocks and a large poly(N-isopropylacrylamide) (PNIPAM) middle
block exhibiting a lower critical solution temperature (LCST) in aqueous solution. A wide range of
concentrations (0.1-300 mg/mL) is investigated using a number of analytical methods such as fluorescence
correlation spectroscopy (FCS), turbidimetry, dynamic light scattering (DLS), small-angle neutron scatter-
ing (SANS), and neutron spin-echo spectroscopy (NSE). The critical micelle concentration is determined
using FCS to be 1 μM or less. The collapse of the micelles at the LCST is investigated using turbidimetry and
DLS and shows a weak dependence on the degree of polymerization of the PNIPAM block. SANS with
contrast matching allows us to reveal the core-shell structure of the micelles as well as their correlation as a
function of temperature. The segmental dynamics of the PNIPAM shell are studied as a function of
temperature and are found to be faster in the collapsed state than in the swollen state. The mode detected
2
has a linear dispersion in q and is found to be faster in the collapsed state as compared to the swollen state.
We attribute this result to the averaging over mobile and immobilized segments.
8
,27-36
1
. Introduction
Smart” polymeric hydrogels have received increasing atten-
two hydrophobic blocks and a PNIPAM block
are alter-
native routes to the formation of temperature-responsive micelles
or micellar gels, where the ability to self-organize in aqueous
solution is exploited. The effect of hydrophobic end groups on the
LCST is discussed in detail in ref 36. Few previous investigations
have focused on triblock copolymers with a PNIPAM middle
block and polystyrene (PS) end blocks;the system considered by
us in the present work. In all studies, the PS block was relatively
large. In very dilute aqueous solution, flowerlike micelles were
observed which collapse at the LCST of PNIPAM with the
magnitude of the size decrease depending on the volume fraction
of PNIPAM in the polymer. PS-b-PNIPAM-b-PS triblock
copolymers with high PNIPAM content and relatively large PS
blocks (not directly soluble in water) were reported to take up
large amounts of water upon immersion in water below the
LCST. Recently, the swelling of concentrated aqueous solu-
tions of PS-b-PNIPAM-b-PS triblock copolymers forming the
“
tion as they respond in a controlled and reversible way with a
volume change to a weak external stimulus, such as temperature,
light, electric or magnetic fields, or ionic strength. Especially
thermosensitive polymers are of great interest for medical and
drug delivery applications, bioseparation, and diagnostics
well as for porous membranes for molecular filtration where the
permeability can be controlled by a change of temperature across
1,2
3-5
as
6-8
the lower critical solution temperature (LCST). Polymers with
an LCST behavior, i.e., a temperature-dependent solubility in
water, are attractive candidates because they are swollen below
the LCST and collapsed, i.e., water-insoluble, above. A widely
used LCST polymer is poly(N-isopropylacrylamide) (PNIPAM).
In aqueous solution, its LCST at 32 ꢀC is attributed to alterations
3
1
8
9-11
in the hydrogen-bonding interactions of the amide group.
The LCST is almost independent of the PNIPAM molar mass
and of concentration (up to 20 wt %), which has been attributed
to sequential hydrogen bond formation between PNIPAM and
H O. So far, macroscopic gels,
shell particles with a cross-linked hydrophobic core and a cross-
3
7
spherical and the bicontinuous morphology was investigated.
In the former, the PS blocks form spheres which are connected by
PNIPAM blocks. During swelling, the PNIPAM blocks were
observed to be strongly stretched. Moreover, the PS spheres were
strongly deformed during the swelling process. However, in spite
of the already published works, at present, no comprehensive
data covering the structural and the dynamical aspects of such
thermoresponsive systems are available over a broad range of
concentrations.
12
13-15
16-19
microgels,
and core-
2
2
0-23
linked responsive shell
transitions have been characterized.
have been studied, and their collapse
The hydrophobic modification of both ends of PNIPAM by
24-26
alkyl end groups
or the use of block copolymers with one or
In this work, we present an investigation of aqueous micellar
solutions of ABA triblock copolymers consisting of a long
PNIPAM middle block and two short PS end blocks that can
*
Corresponding author: e-mail Christine.Papadakis@ph.tum.de;
Ph þ49 89 289 12 447; fax þ49 89 289 12 473.
pubs.acs.org/Macromolecules Published on Web 02/11/2010
r 2010 American Chemical Society

Article
Macromolecules, Vol. 43, No. 5, 2010 2491
be directly dissolved in water. A wide range of concentrations
Organics) was recrystallized from hexane and dried in vacuo.
AIBN (Wako) was recrystallized from methanol and dried in
vacuo.
(
0.1-300 mg/mL) is studied to characterize the crossover be-
tween weakly and strongly correlated micelles. The micelle
formation is investigated using fluorescence correlation spectros-
copy (FCS). The behavior of these micelles at the LCST of
PNIPAM is characterized using dynamic light scattering (DLS)
with respect to the collapse of the micellar shell. The changes of
the micellar structure and the correlation between micelles are
investigated using small-angle neutron scattering (SANS). For
the detailed and unambiguous SANS investigation, a strong
scattering contrast between the core block, PS, and the thermo-
responsive shell block is generated using fully deuterated PS
blocks (P(S-d )) which are contrast-matched by D O. Alterna-
2.2. Synthesis. 2.2.1. RAFT Agent Dibenzyltrithiocarbonate
DBTC). Triethylamine (12.0 mL, 0.0805 mol) was added
(
dropwise at ambient temperature with stirring to a mixture of
benzylmercaptan (9.5 mL, 0.0805 mol), CS (6 mL, 0.0805 mol),
2
and dry CHCl (120 mL). The solution became yellow/orange as
3
the addition proceeded with formation of the intermediate
triethylammonium butyltrithiocarbonate. After 30 min of stir-
ring, benzyl bromide (10 mL, 0.0805 mol) was added slowly,
causing the mixture to thicken with ongoing precipitation of
triethylammonium bromide. The reaction mixture was stirred
8
2
for 16 h and then diluted with CHCl
with three portions of 100 mL of water. The organic solution
was dried over MgSO , concentrated by evaporation, and
passed over a short column of basic Al . The solvent was
removed by rotary evaporation to give an orange oil. Yield:
1.62 g (92.46%). Elemental analysis (C15 , M = 290.47):
Calcd C15 . %C: 62.02; %H: 4.86; %S: 33.12. Found: %C
62.36; %H = 4.90; %S = 32.91. H NMR (300 MHz in
CDCl , δ in ppm): δ = 4.65 (s, 4H, CH -aryl), 7.38-7.28 (m,
3
(80 mL) prior to washing
tively, the PNIPAM block is contrast-matched by the appro-
priate D O/H O mixture, and selective information on the PS
2
2
4
core is obtained. The synthesis of these polymers follows the
2
O
3
previously reported synthesis of nondeuterated PS-b-PNIPAM-
34
b-PS analogues. Moreover, we have investigated the segmental
dynamics of the PNIPAM blocks using neutron spin-echo
spectroscopy on solutions and gels of the triblock copolymers.
Again, the P(S-d ) blocks are contrast-matched in D O, and the
2
H
14
S
3
r
14 3
H S
1
=
8
2
3
2
1
3
PNIPAM dynamics is investigated selectively. It is found to
change drastically at the LCST.
In our study, we address the case of very short hydro-
1
0H, aromatic H). C NMR (75 MHz in CDCl
3
, δ in ppm): δ =
4
1.48 (-S-CH -), 127.72 (CH aryl-para), 128.64 (CH aryl-
2
ortho), 129.20 (CH aryl-meta), 134.84 (C-aryl), 222.63
(-S-(CdS)-S-).
phobic blocks [the PS (or P(S-d )) blocks have degrees of
8
polymerization NPS = 10 or 11] and very long PNIPAM blocks
NPNIPAM = 280-390. This block length ratio is expected to
give better swelling ratios than the samples described in the
2.2.2. Polymerization of Styrene-d : Synthesis of Macro-
CTA1. Styrene-d (10.0 g, 0.089 mol) and DBTC (0.002 mol)
8
were purged with argon for 30 min and immersed in an oil bath
8
8,31,37
literature.
Moreover, our polymers are directly soluble in
at 110 ꢀC. After 19 h, the polymerization was quenched by
placing the flask into liquid nitrogen. The reaction mixture was
diluted with acetone and precipitated in methanol. The precipi-
tation was repeated twice, and the collected polymer was dried in
vacuo at 50 ꢀC for 16 h. Yield 4.0 g (40.0%) of yellow powder.
2.2.3. Polymer Chain Extension with NIPAM. In a typical
procedure, NIPAM (5.10 g, 0.045 mol), the selected macro-
water. Our previous investigations by turbidimetry and rheology
show that they only form gels at high polymer concentration
3
6
(
30 wt %). We attribute this behavior to the strong hydro-
phobicity of the PS blocks which results in lower bridging ratios
than for less hydrophobic blocks.
Recently, we reported first results on the aggregation behavior
of a P(S-d ) -b-PNIPAM390-b-P(S-d ) triblock copolymer,
n
RAFT agent (here macroCTA1; 0.2875 g, M = 2300 g/mol,
8
10
8 10
-
4
-5
1
(
.25 ꢀ 10 mol), and AIBN (0.0025 g, 1.56 ꢀ 10 mol) in THF
focusing on the structural changes at the collapse transition,
determined using turbidimetry, dynamic light scattering (DLS),
28 mL) were purged with argon for 15 min and immersed into
3
5
an oil bath preheated to 65 ꢀC. After 62 h, the reaction was
quenched by placing the flask in liquid nitrogen. The reaction
mixture was diluted with acetone and precipitated in diethyl
ether (300 mL). The polymer was collected by filtration, dis-
solved in acetone, and dialyzed against water for 3 days. The
polymer was recovered and freeze-dried. Yield: 4.08 g (80.0%)
of off-white powder.
and small-angle neutron scattering. In the present work, we
specify these preliminary results on several triblock copolymers
having different block lengths and having nondeuterated and
deuterated PS blocks. Moreover, we give the critical micelle
concentration found by FCS. The segmental dynamics of the
PNIPAM block is investigated by means of NSE. The results will
be compared to our previous investigations on other architec-
1
13
2
.3. Methods. 2.3.1. Spectroscopy. H and C NMR spectra
3
3,38
tures, such as PS-b-PNIPAM diblock copolymers.
were taken with an apparatus Bruker Avance 300 (300 MHz).
Mass spectra were recorded by a GC/MS-system Trace DSQII
(Thermo Scientific). IR spectra were taken from thin films cast
on KBr pellets with a FT-IR spectrometer Bruker IFS 66/s.
The paper is structured as follows. After the Experimental
Section, the synthesis and molecular characterization of P(S-d8)-
b-PNIPAM-b-P(S-d ) are described. The micelle formation in
8
dilute solution and the collapse transition of micellar solutions is
discussed next. Then, the mesoscopic solution and gel structures
are described. The segmental dynamics of the PNIPAM block
and its changes at the collapse are described. Eventually, the
results are summarized and compared to our previous results on a
PS-b-PNIPAM diblock copolymer.
2
.3.2. Size Exclusion Chromatography (SEC). SEC of the
polymers was run in N,N-dimethylacetamide containing 0.1%
LiBr as eluent at a column temperature of 45 ꢀC, with a setup
consisting of an Agilent 1200 isocratic pump, an Agilent 1200
refractive index detector, and two GRAM columns (10 μm, 8 ꢀ
300 mm, pore sizes 100 and 1000; PSS GmbH, Mainz,
Germany). The SEC setup was calibrated using low-polydis-
persity polystyrene standards (PSS GmbH, Mainz, Germany).
Assuming an ideal RAFT mechanism and that (i) the amount
of initiator derived polymer chains is negligible and that (ii) the
obtained yields correspond to monomer conversion, the theo-
2
. Experimental Section
.1. Materials for Synthesis. Carbon disulfide [75-15-0]
99.9%, Acros Organics), benzylmercaptan [100-53-8] (99%,
3
9
2
(
Sigma-Aldrich), benzyl bromide [100-39-0] (98%, Sigma-
Aldrich), dichloromethane [75-09-2] (Acros Organics), CHCl3
theor
retically expected number-average molar mass M
calculated in good approximation as
n
can be
[
37-297-8] (99%, Sigma-Aldrich), triethylamine [121-44-8]
(
(
(
99%, Acros Organics), and magnesium sulfate [7487-88-9]
97%, Acros Organics) were used as received. Tetrahydrofuran
Mtheor ¼ ðconversion ꢀ Mmonomer ꢀ ½MꢁÞ=½CTAꢁ þ MCTA ð1Þ
THF) was distilled over K-Na. Styrene-d [PS 07150] (99%,
8
Chemotrade) was distilled in vacuo in the presence of hydro-
quinone. N-Isopropylacrylamide [2210-25-5] (99%, Acros
where [M] is the initial concentration of monomer, [CTA] is the
concentration of RAFT agent, Mmonomer is the molar mass of

2
492 Macromolecules, Vol. 43, No. 5, 2010
Table 1. Characteristics of the Synthesized Polymers
Adelsberger et al.
theor a
M
n
b
b
c
entry
triblock copolymer
RAFT agent
yield (%)
M
n
(SEC)
PDI
n
M (UV)
1
2
macroCTA1
DBTC
macroCTA1
DBTC
macroCTA2
macroCTA2
40
80
40
86
89
2290
35000
2290
37000
27000
2300
1.19
1.47
1.21
1.47
1.31
d
P(S-d 10-b-PNIPAM390-b-P(S-d
8
)
8
)
10
19700
2300
17500
46000
e
3
macroCTA2
PS11-b-PNIPAM370-b-S11
PS11-b-PNIPAM280-b-S11
f
d
4
44000
34000
g
d
5
18000
a
8
Calculated according to eq 1, assuming 100% blocking efficiency and that the molar mass of the P(S-d ) or PS block corresponds to the one of the
b
macroRAFT agents used. In dimethylacetamide based on polystyrene standards. From the absorbance at 262 nm, assuming that the molar mass of
c
d
e
8
the P(S-d ) or PS block corresponds to the one of the macroRAFT agents used. Apparent values only. Polymerization conditions: 0.089 mol of
3 f g
-
styrene, 2.0 ꢀ 10 mol of RAFT agent, 110 ꢀC, 19 h. Molar ratio NIPAM/macro RAFT agent in reaction mixture = 350/1. Molar ratio NIPAM/
macro RAFT agent in reaction mixture = 250/1.
the monomer, and MCTA is the molar mass of the chain transfer
agent.
All data reduction was performed with the software qtiKWS
provided by JCNS.
The scattering length densities of the polymer blocks,
2
.3.3. Turbidimetry. Turbidity measurements were performed
on a temperature-controlled turbidimeter (model TP1, E. Tepper,
Germany). Solutions were prepared in demineralized H O at a
1
0
-2
10
-2
δ
were calculated based on the scattering length densities of the
P(S-d8) = 6.42 ꢀ 10 cm and δPNIPAM = 0.79 ꢀ 10 cm
,
2
concentration of 1 mg/mL. The transmittance of solutions was
set automatically to 90% at the beginning of each measurement.
The samples were heated from 20 to 60 ꢀC and cooled down
again with heating and cooling rates of 1.0 K/min, respectively.
Temperatures are precise within 0.5 K.
elements and the specific volumes, v = 0.952 mL/g for PS and
4
3
0.892 mL/g for PNIPAM in water. Contrast matching of the
core or the shell block was performed by using either pure D
2
O
1
0
-2
or a 20:80 v/v mixture of D
H
2
O (δD O = 6.38 ꢀ 10 cm ) and
2
1
0
-2
O (δH O = -0.56 ꢀ 10 cm ) as a solvent.
2
2
The SANS curves were analyzed using the software Scatter
.0. This software was able to fit the combination of the form
2.3.4. Fluorescence Correlation Spectroscopy. The setup and
the sample preparation have been described in ref 33. A Con-
4
4
2
factor of polydisperse spherical core-shell particles, P(q) (q is
the magnitude of the scattering vector), as well as a Percus-
Yevick structure factor, S(q), describing their liquidlike correla-
tion, to the curves. Above the LCST, additional forward
foCor 2 from Carl Zeiss Jena GmbH was used together with an
Ar laser (λ = 488 nm). In the solutions in demineralized H O,
the concentration of Rh6G was kept at 50 nM, whereas the one
þ
2
of the polymers (PS11-b-PNIPAM280-b-PS11 and PS11-b-PNI-
PAM370-b-PS ) was varied.
-
1
scattering at q < 0.15 nm was present. Prior to using Scatter
.0, the forward scattering was modeled by an algebraic decay
11
2
2
.3.5. Dynamic Light Scattering (DLS). Temperature-
resolved DLS experiments on filtered solutions of the polymers
at 0.1-0.2 mg/mL in demineralized H O were performed in
κ
I(q) ꢁ K/q . This contribution was subtracted from the SANS
curve, and the remainder was used with Scatter 2.0. Finally,
2
both contributions were summed up to give the fitted curve I(q)
=
polarized geometry using an ALV-5000/E correlator together
with a goniometer with an index matching vat filled with
toluene. The light source was a Nd:YAG laser operated at
λ = 532 nm. The scattered light was detected at a scattering
κ
I S(q)P(q) þ K/q . The incoherent background due to the
0
solvent and the polymer was treated as a fitting parameter and
was found to lie in a reasonable range.
2.3.7. Neutron Spin-Echo Spectroscopy (NSE). Experiments
3
3
angle θ = 90ꢀ using the setup described previously. At each
temperature, measurements of duration 4-10 min were per-
formed. After each temperature change, the waiting time was
4
5,46
were carried out at J-NSE at FRM-II.
The Fourier times
ranged from 0.1 to 60 ns. The measurement of one curve up to
Fourier times of 60 ns took ∼11 h. The sample temperature was
controlled by circulating thermostated water with a precision of
(0.1 K. Grafoil or graphite powder was used to measure the
resolution function of the spectrometers. The detector was a
multiwire area detector. The detector was placed at average
2
0 min. Additional θ-dependent measurements of the solution
of PS -b-PNIPAM370-b-PS11 at 0.1 mg/mL at 20 ꢀC showed
11
2
that the relaxation rate Γ ꢁ sin (θ/2); i.e., the process is diffusive.
The temperature-dependent average hydrodynamic radii of the
micelles or the clusters, R , were determined using the routine
h
-
1
40
q-values of 0.8, 1.1, 1.5, and 1.8 nm . Temperature scans were
REPES, as detailed in ref 33, which gives distribution func-
tions of hydrodynamic radii. At this, the Stokes-Einstein
equation, R = k T/6πηD, was used with D being the measured
-
1
carried out at q = 1.5 nm . Samples were prepared by dissolv-
ing P(S-d 10-b-PNIPAM390-b-P(S-d 10 in D O at concentra-
8
)
8
)
2
h
B
tions of 50, 200, and 300 mg/mL. The samples were mounted in
Hellma Quartz cells of size 30 mm ꢀ 30 mm with a sample
thickness of 4 mm for the lowest concentration (50 mg/mL) and
diffusion coefficient and η the viscosity of the solution.
.3.6. Small-Angle Neutron Scattering. SANS experiments
2
were carried out using the KWS-2 small-angle scattering instru-
ment of the JCNS outstation at the FRM II, Garching,
Germany. For the experiments, a wavelength λ = 0.7 nm (Δλ/
λ = 20%) and sample-to-detector distances (SDD) of 2 and 8 m
were chosen, resulting in q-ranges of 0.049-0.45 and 0.25-
2
mm for 200 and 300 mg/mL.
3
. Results
3.1. Synthesis and Molecular Characterization. The synth-
esis of the nondeuterated polymers (Table 1) by the RAFT
-
1
1
.7 nm . The detector design is based on a modified Anger
6
41
technology with a Li glass as a scintillator. The solutions of
P(S-d 10-b-PNIPAM390-b-P(S-d 10 having concentrations of
0 and 220 mg/mL (ref 42) were mounted in standard Hellma
39
method using the well-established bifunctional trithiocar-
8
)
8
)
bonate DBTC as chain transfer agent followed the procedure
described in ref 34. In this work, we describe the analogous
synthesis of the triblock copolymer P(S-d )-b-PNIPAM-
5
quartz cuvettes (light path of 1 or 2 mm). The exposure times
per image were between 5 and 20 min for different concentra-
tions and detector configurations. After each temperature
change, the waiting time was 15 min. The background of the
cuvette was subtracted from the sample scattering, taking the
transmissions into account. The intensities were corrected for
dark current and background using the scattering of boron
carbide. The scattering of poly(methyl methacrylate) was used
for measuring the detector sensitivity and for bringing the
intensities on an absolute scale. The 2D scattering images were
isotropic, and hence, the intensities were azimuthally averaged.
8
b-P(S-d ) with fully deuterated PS blocks (Scheme 1). In
8
the first step, styrene-d is thermally polymerized in the
8
presence of DBTC. The formed polystyrene-d is engaged
8
as bifunctional macroRAFT agent macroCTA1 in the sub-
sequent chain extension polymerization of NIPAM initiated
by AIBN to produce the symmetrical triblock copolymer
P(S-d )-b-PNIPAM-b-P(S-d ).
8 8
The bifunctional RAFT agent DBTC was used to prepare
the triblock copolymer. The initial thermal polymerization

Article
Macromolecules, Vol. 43, No. 5, 2010 2493
) via Successive RAFT Polymerizations of Styrene-d
Scheme 1. Synthesis of Triblock Copolymer P(S-d
8
)-b-PNIPAM-b-P(S-d
Using DBTC and NIPAM
8
8
,
of styrene-d in the presence of the DBTC (the ratio of
8
monomer to RAFT agent was set [M] /[RAFT] = 45)
0
0
was conducted to moderate conversions only, in order to
guarantee a high degree of end-group functionality of the
4
7,48
polymer formed,
and thus to produce a highly efficient
macro RAFT agent with a well-defined molar mass (see
Figure 1a).
While the FT-IR spectrum shows the typical features of
-
1
polystyrene-d , such as bands at 2270 cm (C-D aryl
8
-
1
stretching), 2195 cm (C-D chain asymmetrical stretch-
-
1
ing), 2102 cm
8
(C-D chain symmetrical stretching),
-
1
-1
40 cm (C-D aryl out-of-plane), and 821 cm (C-D
49 1
aryl out-of-plane), the H NMR spectrum of macroCTA1
exhibits mainly the signals of the initiating benzyl groups at
both chain ends (Figure 1b) due to the high degree of
deuteration of the styrene-d (99%). The analysis of macro-
8
CTA1 by SEC shows a monomodal, relatively narrow molar
mass distribution (Figure 1a), with a number-average molar
mass M of 2300 (corresponding to an average degree of
n
polymerization of 20.5) and a polydispersity index PDI of
1
.19. Assuming an ideal RAFT mechanism, the theoretically
theor
expected number-average molar mass Mn
290 using eq 1. Additionally, the average degree of poly-
is calculated as
2
merization can be calculated from the sulfur content of
the elemental analysis, assuming that all polymers contain
exactly the fragments of one molecule of DBTC, to yield a
number-average degree of polymerization of 22. The good
agreement between these values demonstrates the good
control over the molar mass of macroCTA1 as well as its
high content of RAFT active end groups, as needed for the
successful synthesis of the triblock copolymers.
Subsequently, the polystyrene-d was employed as bifunc-
8
tional macroRAFT agent for chain extension with NIPAM
in order to prepare amphiphilic triblock copolymers of the
BAB type. The successful chain extension of the macroCTA1
1
by NIPAM is qualitatively shown by H NMR. The spec-
trum exhibited in Figure 1c with strong signals at 4 ppm
(
6
characteristic for the CON-CHÆ methine proton) and at
.4-7.6 ppm (characteristic for the CONHÆ amide proton)
evidences the incorporation of NIPAM in large quantities.
The benzyl end groups which were visible in the spectrum of
macroCTA1 are camouflaged by the much larger signal of
the amide protons. The successful chain extension is further
corroborated by SEC. The polymer obtained shows a mono-
Figure 1. Molecular characterization of triblock copolymer P(S-d
b-PNIPAM390-b-P(S-d 10 and its precursor macroRAFT1. (a) SEC
elugrams of triblock copolymer (right curve) and of macroRAFT1
8 10
) -
8
)
modal molar mass distribution with a PDI value of 1.47 and
= 19 700
based on calibration with polystyrene (Figure 1a).
1
1
(left curve). (b) H NMR spectrum of macroRAFT1. (c) H NMR
app
an apparent number-average molar mass of Mn
spectrum of triblock copolymer (both spectra in acetone-d6).
It is difficult to determine the absolute mass of the triblock
copolymer precisely, as several problems are superposed.
This comprises the amphiphilic character of the block copo-
lymer with a strong tendency to aggregate in solution and the
unfavorable combination of a very large block, namely of
When assuming that the molar mass of the polystyrene-d
8
block of 2300 is preserved in the block copolymer, the C/N
ratio from elemental analysis enables in principle to calculate
the absolute molar mass. Still, due to the small size of the
polystyrene block, this analysis provides only an approxi-
mate value. The thus;formally;calculated value of 55 700 (
28 000 for the block copolymer implies a number-average
PNIPAM, that is hygroscopic with a very short one, namely
of polystyrene-d , that is virtually invisible in the H NMR.
1
8

2
494 Macromolecules, Vol. 43, No. 5, 2010
Adelsberger et al.
Figure 2. Results from FCS on aqueous solutions of PS11-b-PNI-
PAM280-b-PS11 (triangles) and PS11-b-PNIPAM370-b-PS11 (circles)
with Rh6G as tracers as a function of the copolymer concentration.
The concentration of Rh6G was kept constant at 50 nM. The tempera-
ture was 27 ꢀC. Open symbols: free Rh6G; closed symbols: micelles
carrying Rh6G. The dashed vertical line marks the resulting critical
Figure 3. Temperature-dependent light transmission of 1 mg/mL
aqueous solutions of PS11-b-PNIPAM280-b-PS11 (open triangles up),
PS11-b-PNIPAM370-b-PS11 (filled circles), and P(S-d
8
)
10-b-PNI-
micelle concentration at 0.9 μM. Results from DLS at 27 ꢀC on PS11
-
PAM390-b-P(S-d )10 (filled triangles down). The arrows denote heating
8
b-PNIPAM280-b-S11 (cross) and PS11-b-PNIPAM370-b-S11 (plus).
and cooling.
degree of polymerization of 500 ( 250 of the PNIPAM
block. A more reliable value can be derived from the
We conclude that the cmc of P(S-b-NIPAM-b-PS) is
located at 0.03-0.04 mg/mL for both polymers and that
the micellar hydrodynamic radius increases with PNIPAM
5
0
absorbance of the styrene moieties in the UV spectrum.
On the basis of the extinction coefficient of polystyrene at
3
1,33,52,54,55
molar mass, as expected.
similar to the one observed with the previously studied
PS -b-PNIPAM diblock copolymer in dilute aqueous
The behavior is very
-
1
-1
2
62 nm in CH Cl of 222.5 L mol cm , we calculate a
2 2
molar mass of 46 300 ( 13 000 for the triblock copolymer,
assuming again that the molar mass of the polystyrene-d8
block of 2300 is preserved in the block copolymer. This
means a number-average degree of polymerization of 390 of
the PNIPAM block.
4
8
159
3
3
solution.
3.3. Collapse Transition. Turbidity measurements were
carried out to determine the cloud point as a function of
NPNIPAM. To ensure that the majority of the polymers is
arranged in micelles, a concentration well above the cmc was
chosen (1 mg/mL) for all solutions. At low temperatures, the
solutions are clear and suddenly become turbid at 31-32 ꢀC.
This is measured in the light transmission curve which
decreases steeply in this temperature range (Figure 3). Thus,
the cloud point has been reached, and large clusters are
formed by the collapsed, hydrophobic micelles which scatter
light strongly. Above ∼33 ꢀC, all curves show a shoulder and
reach a plateau at 40-50% of the initial value at ∼40 ꢀC. This
shoulder may be due to the mechanism of cluster growth.
Upon cooling, hysteresis is observed, but nevertheless the
process is fully reversible. The resulting cloud point tempe-
ratures increase with NPNIPAM from 31.1 to 31.7 ꢀC. The
shoulder and the hysteresis have previously been observed
with a micellar solution of PS-b-PNIPAM diblock copoly-
3.2. Micelle Formation. The critical micelle concentration
cmc) of amphiphilic block copolymers is usually very low
(
and cannot be determined by scattering methods, but by
fluorescence correlation spectroscopy (FCS) which allows to
3
3,51-53
access sufficiently low polymer concentrations.
To
determine the cmc of PS-b-PNIPAM-b-PS as a function of
NPNIPAM below the LCST, poorly water-soluble fluorescent
dyes, such as Rh6G, at very low concentration (50 nM)
served as tracers in dilute solutions of PS -b-PNIPAM280-b-
1
1
PS11 and PS -b-PNIPAM -b-PS , which were directly
11
370
11
dissolved in H O. Below the cmc, the dye molecules stay
2
molecularly dissolved in H O, and only the diffusion of free
2
dye is expected. Above the cmc, in contrast, a fraction of the
dye molecules attach to the micelles, and an additional slow
diffusional decay is expected in the FCS correlation curve. At
low concentrations, the correlation curves can be fitted with
33
mers where bridging is absent. We thus ascribe this beha-
vior to the hydrophobic interactions between the collapsed
micelles.
a single decay leading to R = 0.8 nm, the value expected for
h
5
2
free Rh6G. The hydrodynamic radii deduced from the
decay times are shown as a function of polymer concentra-
tion in Figure 2. Above a polymer concentration of 0.9 μM, a
second, slower decay is present in the correlation curves with
hydrodynamic radii of 20.8 ( 0.7 nm for PS -b-PNI-
To characterize the structural changes in micellar solu-
tions at the LCST in more detail, we have carried out
temperature-resolved DLS experiments on PS -b-PNI-
11
PAM370-b-PS11 and P(S-d ) -b-PNIPAM390-b-P(S-d ) in
8 10
1
1
8 10
PAM -b-PS and 25.8 ( 0.9 nm for PS -b-PNIPAM
-
H O. To minimize the interaction between micelles, concen-
2
2
80
11
11
370
b-PS . Micelles are thus present in the solution, and a
trations of 0.1-0.2 mg/mL were chosen, which are only
slightly above the cmc. The hydrodynamic radii obtained
for PS -b-PNIPAM370-b-PS11 at 0.1 mg/mL show a shallow
1
1
fraction of the Rh6G molecules attaches to them. In another
amphiphilic block copolymer system, we have shown that
the cmc determined by using Rh6G is the same as when
using the identical, fluorescence-labeled block copolymer
1
1
decay from 27 nm at 25 ꢀC to 23 nm at 30 ꢀC (Figure 4). At
the cloud point determined from turbidimetry (31.4 ꢀC),
the hydrodynamic radius does not decrease, as expected
for single collapsing micelles, but increases within 1.6 K to
29 nm. We attribute this length scale to small clusters con-
sisting of several collapsed micelles. This behavior is very
different from what we observed with PS-b-PNIPAM di-
block copolymers, where single collapsed micelles could
5
3
as a tracer. We note that already below the polymer con-
centration of 0.9 μM micelles might be present, but too
few to solubilize Rh6G. The cmc is thus at 0.9 μM or lower.
The Rh value of the micelles determined using DLS at
concentrations of 0.1 mg/mL coincides with the values from
FCS, confirming the assignment of the slow FCS decay to
micelles.
3
3
clearly be detected above the LCST. We conclude that

Article
Macromolecules, Vol. 43, No. 5, 2010 2495
5
6-58
copolymers to bridge two micellar cores
results in more
pronounced cluster formation than in the diblock copolymer
where the clusters are only connected by hydrophobic inter-
actions. Only in very dilute solutions (< 10 mg/mL) of
-
3
PS-b-PNIPAM-b-PS triblock copolymers, the collapse of
3
1
single micelles could be observed.
.3. Collapse Transition in Concentrated Solutions. The
3
core-shell structure of the micelles, the collapse of the micellar
shell, and the correlation of the micelles as a function of
temperature were investigated by SANS. Two contrasts were
exploited to study the polymer P(S-d ) -b-PNIPAM390-
8
10
b-P(S-d ) : In D O, the P(S-d ) core was contrast-matched,
8
10
2
8
and we were able to highlight the thermoresponsive PNI-
PAM shell. Using a 20:80 v/v mixture of D O/H O, the
PNIPAM shell was matched and the P(S-d ) core was high-
2
2
8
h 2
Figure 4. Hydrodynamic radii, R of the micelles in H O from DLS.
lighted, however, at the expense of a higher scattering back-
ground. It must be kept in mind, though, that H/D exchange
can occur in the amide group of PNIPAM during the
Triangles up: PS11-b-PNIPAM370-b-PS11 at 0.1 mg/mL; circles:
P(S-d 10-b-PNIPAM390-b-P(S-d 10 at 0.2 mg/mL, both measured at
8
)
8
)
θ = 90ꢀ. The full lines are guides to the eye. The dashed line marks the
59
collapse transition temperature from turbidimetry (31.4 ꢀC, Figure 3).
equilibration time before the measurements (a few days).
At a relatively low concentration (50 mg/mL), we could
characterize the micelles in the absence of strong correlation.
Solutions with a higher concentration (220 mg/mL) enabled
us to investigate the correlations between the micelles as well
as their changes at the collapse transition.
already after dissolution in H O at room temperature the
2
micelles formed by PS -b-PNIPAM370-b-PS11 are bridged,
1
1
even at this low concentration. Above 35 ꢀC, additional large
objects (550-700 nm) are observed, which are presumably
very large clusters formed by collapsed micelles and the
existing small clusters.
The SANS curves at 50 mg/mL in D O/H O and in D O
2
2
2
are shown in Figure 5. Below the collapse temperature, the
curves in D O/H O are expected to show the scattering from
weakly interacting micellar P(S-d ) cores, since the PNIPAM
Similar characteristics are observed for the triblock co-
polymer P(S-d ) -b-PNIPAM390-b-P(S-d ) at 0.2 mg/mL
2
2
8
10
8 10
8
(Figure 4). With this polymer, the clusters formed above the
collapse temperature are larger than in PS -b-PNIPAM370-
shell is contrast-matched by the solvent (Figure 5c). The curves
measured at the same concentration in D O (Figure 5a) are
1
1
2
b-PS , presumably due to the slightly higher concentration.
1
In contrast, no large clusters are observed, even at high
temperatures.
expected to be entirely due to the scattering of the PNIPAM
blocks. They show qualitatively the same behavior: A weak
1
-1
correlation peak is observed at 0.16 nm , corresponding to
an approximate intermicellar distance of 40 nm. At higher
q-values, shoulders due to the micellar form factor are present.
As the collapse temperature is approached, the correlation
peak becomes less pronounced. Above the collapse tempera-
ture, the curve shapes change drastically (Figure 5b,d). The
curves right at the collapse temperature show a continuous
decay, indicating fluctuations on a wide range of length
scales. At higher temperatures, again a correlation peak is
The hydrodynamic radii of the swollen micelles formed
by the triblock copolymers below the collapse temperature,
e.g., 27.2 nm for PS -b-PNIPAM370-b-PS11 and 34.7 nm for
1
1
P(S-d ) -b-PNIPAM390-b-P(S-d ) at 25 ꢀC, are signifi-
8
10
8 10
cantly smaller than the value of the PS -b-PNIPAM
159
4
8
diblock copolymer studied by us previously (41 nm at
2
3
3
5 ꢀC). The geometric effect of the polymer architecture;
starlike vs flowerlike micelles;is not expected to have a signifi-
cant influence on the shell thickness, since the molar mass
of the PNIPAM block in the diblock copolymer is roughly
one-half of the ones of the triblock copolymers studied here.
We attribute the discrepancy to the difference in core size
resulting from the difference in NPS and, therefore, in
aggregation number, Nagg. Using the relations for the radius
-
1
observed at a much higher q-value (∼0.44 nm ) than below
the collapse temperature, which indicates that the inter-
micellar distance decreases by a factor of ∼2.8 upon collapse.
Moreover, the form factor is slightly altered. However, the
most prominent feature is the strongly increased forward
scattering which is due to the formation of large clusters of
micelles.
1
/2
of gyration of the PS block in the melt, R ꢁ N , together
g
3
with N ꢁ R , we estimate N of the triblock copolymers
to be a factor of 10 lower than the one of the diblock
copolymer, which is at the origin of the size difference.
The decrease of R of the swollen micelles with increasing
Below the collapse temperature, both sets of curves can be
modeled successfully with a form factor for spherical core-
shell micelles together with a Percus-Yevick structure fac-
tor. Fitting of a model for homogeneous spheres represent-
ing the P(S-d ) core to the curves measured in D O/H O was
agg
g
agg
h
temperature toward the collapse temperature has previously
been observed with diblock and triblock copolymers from PS
8
2
2
not successful, and it was necessary to include scatter-
ing from the shell. Reasons may be a slight bias in the
contrast matching or the H/D exchange between the amide
2
9,31,33
and PNIPAM.
(
Since the core size remains constant
see section on SANS below), the decrease must be due to a
relaxation of the PNIPAM shell blocks toward a more coiled
conformation.
group in PNIPAM and D O prior to the measurement. For
2
both sets of data, the core radius amounts to Rcore = 3.0 (
0.5 nm in the entire temperature range (Figure 6a). Using the
The clustering just above the collapse transition is readily
explained by the increased hydrophobicity of the formerly
hydrophilic PNIPAM block. Clusters of similar size (R =
3
mass density of bulk PS of 1.05 g/cm and the degree of
polymerization of both PS blocks (2300 g/mol), this size
corresponds to an aggregation number Nagg = 31 ( 15,
which is in the usual range. The micellar radius, Rmicelle, as
h
7
0-90 nm) have been observed with the above-mentioned
33
PS-b-PNIPAM diblock copolymer as well. However, in the
diblock copolymer, they were only observed above 40 ꢀC,
whereas in the triblock copolymers, they form right above
the collapse transition, and single, collapsed micelles cannot
be detected. We conclude that the ability of the triblock
determined in D O, is 17 ( 1 nm below the collapse
2
temperature and decreases slightly as the collapse tempera-
ture is approached, which points to an incipient collapse
already below the LCST. In the curves from D O/H O, the
2
2

2
496 Macromolecules, Vol. 43, No. 5, 2010
Adelsberger et al.
Figure 5. SANS curves of P(S-d
8
)
10-b-PNIPAM390-b-P(S-d
8
)
2 2 2
10 at 50 mg/mL in (a, b) D O and (c, d) in D O/H O. Experimental curves: filled squares,
2
3
0 ꢀC; open circles, 25 ꢀC; filled triangles up, 28 ꢀC; open triangles down, 29 ꢀC; filled diamonds, 30 ꢀC; open triangles left, 31 ꢀC; filled triangles right,
2 ꢀC; open squares, 33 ꢀC; filled circles, 35 ꢀC; open triangles up, 40 ꢀC. The lines are fits; see text. The curves at 35 and 40 ꢀC in (b) have lower overall
intensities which we attribute to sedimentation.
pretransitional changes of the shell structure. An initial concen-
tration gradient in the micellar shell with more dense packing
near the core may be equilibrated at higher temperature.
The presence of a structure factor indicates a certain
correlation already at this relatively low concentration,
which may be due to both the steric interaction and a certain
degree of bridging between the micelles. Such a correlation
was not observed with the PS-b-PNIPAM diblock copoly-
33
mers studied previously at a concentration of 30 mg/mL;
i.e., both the higher concentration and the bridging may be at
the origin of the correlation. Below the collapse temperature,
the hard-sphere radius, which corresponds to half the dis-
tance between micelles, is RHS = 16.3 ( 1.0 nm; i.e., within
the errors it is similar to Rmicelle. Thus, the micellar shells of
the correlated micelles touch each other, which is due to the
limited stretching of the bridging PNIPAM blocks.
The forward intensity, I , reflects the osmotic compressibility
0
6
0
of the micellar solution. In both solutions, I increases as the
0
collapse temperature is approached (Figure 6b); i.e., the com-
pressibility increases because of strong fluctuations of the
micellar distance. In PNIPAM homopolymers, the spinodal
temperature, T , has been determined using the mean-field
s
-
η 61
relation I ꢁ (T - T )
. In D O, where the PNIPAM shell
2
0
s
is seen, we find η = -0.39 ( 0.04 and T = 34.9 ( 0.6 ꢀC. T is
s
s
higher than the collapse temperature of 32.5 ( 0.5 ꢀC; i.e., the
collapse of the tethered blocks preempts the expected spinodal
decomposition. The exponent η differs strongly from the value
of -0.8 found for PNIPAM homopolymers which we attribute
to the tethering. The volume fraction of correlated micelles,
φ, as determined from the Percus-Yevick structure factor
Figure 6. SANS results from the fits in Figure 5 (50 mg/mL). Filled
symbols: in D O; open symbols: in D O/H O. The dashed line marks
2 2 2
the collapse transition. (a) Radii: squares, core radius; triangles up,
micellar radius; diamonds, hard-sphere radius. (b) Forward intensity
0
I . The lines below the LCST are fits; see text. (c) Volume fraction of
(
Figure 6c) decreases in both solutions, as the collapse transi-
tion is approached, which corroborates the notion of strong
fluctuations.
correlated micelles, φ. Full lines are guides to the eye.
apparent Rmicelle value increases from 6.7 nm at 20 ꢀC to
Above the collapse temperature, Rmicelle drops to 4.8 ( 0.6 nm
2
1
3.3 nm at 31 ꢀC. We attribute this unexpected increase to
in D O; i.e., the partially deuterated, collapsed PNIPAM

Article
Macromolecules, Vol. 43, No. 5, 2010 2497
42
8 8 2 2 2
Figure 7. SANS curves of P(S-d )10-b-PNIPAM390-b-P(S-d )10 at 220 mg/mL in (a, b) D O and (c, d) in D O/H O. Same symbols as in Figure 5.
shell is much thinner (∼1.5 nm) than below, whereas the core
radius is unaltered (Figure 6a). In contrast to telechelic
2
6
PNIPAM, the core-shell structure is preserved. RHS is
.1 ( 0.9 nm, thus slightly larger than Rmicelle; i.e., the
6
micelles form a close-packed structure in the collapsed state.
I0 and φ are kinetically determined above the collapse
temperature and are therefore not discussed further.
Heating the solutions above the collapse temperature
leads to a drastic increase in forward scattering together
with a shift of the correlation peak to a significantly higher
-1
q-value, 0.44 nm ; i.e., the distance between the micelles
decreases, and clusters are formed due to the collapse of the
shell. The forward scattering follows approximately a Porod
law; i.e., the clusters are compact objects.
Solutions of higher concentration (220 mg/mL) show the
same overall behavior (Figure 7) with higher scattering
intensities, as expected. The structure factor is more pro-
nounced than at lower concentration; i.e., a higher fraction
of micelles are correlated. A striking difference is that, in
D O solution, R
2
below the collapse temperature (11.7 (
micelle
1
.3 nm) is significantly lower than at 50 mg/mL (17 ( 1 nm)
which may be attributed to the higher degree of bridging
of the micelles at the higher concentration, hampering the
full swelling of PNIPAM (Figure 8a). Rmicelle increases
slightly toward the collapse transition. RHS stays constant
over the entire temperature range below the collapse tempe-
rature (12.9 ( 0.7 nm). Again, the value is smaller than at
5
0 mg/mL.
The temperature dependence of I in D O shows a power
0
2
law behavior as well with a slight deviation near the collapse
transition (Figure 8b). The value η = -0.76 ( 1.60 is close to
the value found for PNIPAM homopolymers (albeit the
Figure 8. SANS results from 220 mg/mL (Figure 7). Same symbols as
in Figure 6.
error is large), and T = 32.0 ( 0.05 ꢀC is close to the
s
collapse temperature. This better agreement with PNIPAM
homopolymer solutions is presumably due to the higher
concentration as compared to the value from 50 mg/mL
and the consequently more homogeneous PNIPAM matrix.
As expected, the volume fraction of correlated micelles
(Figure 8c) is higher than at 50 mg/mL and shows similar
temperature behavior.
Above the collapse temperature, all radii are very similar
to those at 50 mg/mL. In D O, R
decreases to 4.8 (
0.6 nm. The shell thickness thus decreases from 9.5 to 1.5 nm,
2
micelle

2
498 Macromolecules, Vol. 43, No. 5, 2010
Adelsberger et al.
a shrinkage of 84%, which is similar to the value obtained
with the lower concentration. RHS decreases to 7.2 ( 0.2 nm
(
inD O) or 8.0 (0.1 nm (in D O/H O), also similar to the values
2
2
2
at lower concentration. The shrinkage of the micellar distance
thus amounts to 44% in D O and to 37% in D O/H O.
2
2
2
From the above results, we conclude that a core-shell
model with a homogeneous shell fits the SANS curves
sufficiently well. We note here that the micelles (i.e., the
1
6-19
building blocks) are much smaller than the microgels
and the cross-linked core-shell particles
20-23
presented in the
literature. As a consequence, they are expected to react faster to
temperature changes than microgels and cross-linked core-
shell particles, which is of interest for applications. The PS-
b-PNIPAM-b-PS triblock copolymers discussed in the litera-
8,31,37
ture
have throughout higher degrees of polymerization of
the PS block (NPS = 35-200); i.e., the cores are significantly
larger than in our study, and the volume fractions of PNIPAM
are lower. Thus, our system is able to swell more strongly.
3
.4. Segmental Dynamics. The segmental dynamics of con-
centrated solutions of P(S-d ) -b-PNIPAM390-b-P(S-d )
8 10
8
10
triblock copolymers in D O have been investigated using
2
6
2
neutron spin-echo spectroscopy (NSE). Because the q-values
chosen are relatively large, the experiment is sensitive to the
local dynamics in the PNIPAM shell. The P(S-d ) cores do
8
not contribute to the intermediate scattering function, S(q,t),
because P(S-d ) is contrast-matched by the solvent, D O.
8
2
The signal obtained is thus entirely due to PNIPAM seg-
mental dynamics.
In the literature, two cases have been distinguished for the
segmental dynamics of the shell block: In loosely packed
shells of “hairy” micelles (i.e., long shell blocks), the single-
6
3
chain Zimm dynamics dominates. In contrast, in densely
packed shells of “crew-cut” micelles from diblock copoly-
mers (i.e., short shell blocks), the breathing mode was
observed to dominate, i.e., the cooperative motion of poly-
Figure 9. Representative intermediate scattering functions from NSE
on P(S-d
c) 300 mg/mL. Filled squares, 24.4 ꢀC; filled circles, 30.5 ꢀC; filled
trianglesup, 31.0 ꢀC; filledtriangles down, 32.2 ꢀC; opencircles, 33.9 ꢀC.
The solid lines are fits of eq 3 with Γ as a free fitting parameter.
8 8 2
)10-b-PNIPAM390-b-P(S-d )10 in D O: (a) 50, (b) 200, and
(
6
4,65
mer chains in the solvent.
This breathing mode was also
6
6
observed in micelles formed by triblock copolymers, where
the grafting density of the shell blocks at the core surface is
high. These two processes are distinguished by the q-beha-
vior of the relaxation rate Γ of the process: Γ ꢁ q and Γ ꢁ q
have been predicted for the breathing mode and the single-
chain dynamics, respectively. In addition to the segmental
dynamics, the diffusion of micelles has been found to contribute
to the intermediate structure factor at high Fourier times.
We have investigated three concentrations between 50 and
00 mg/mL. As evident from the SANS measurements, the
micelle
with Afast the amplitude of the fast decay due to the internal
2
3
2
2
dynamics and Γinternal = Dinternalq and Γ
= Dmicelleq
micelle
the relaxation rates of the fast and the slow decay, res-
pectively. Dinternal is the diffusion coefficient of the segmental
dynamics, and Dmicelle is the diffusion coefficient of the
micelles. t is the Fourier time. c denotes a factor accounting
for S(q,t)/S(q,0) values lower than unity. The coupled case is
modeled by
64,65
3
overlap between the PNIPAM shells becomes more and
more important as concentration is increased. The inter-
mediate scattering functions for the three concentrations are
shown in Figure 9. At temperatures up to 31 ꢀC, the curves
show two decays: a dominating, fast one decaying below 10 ns
and a slow one above. We attribute the fast decay to the
PNIPAM shell dynamics. The latter, slow decay has low
amplitude and is arrested above the collapse temperature.
We attribute it to the diffusion of micelles which is slowed
down above the collapse temperature due to the cluster
formation of the collapsed micelles.
Sðq, tÞ
¼
c½ð1 -AfastÞ
Sðq, 0Þ
þ Afast expð -ΓinternaltÞꢁ expð -ΓmicelletÞ
ð3Þ
Since we attribute the slow decay to micelle diffusion, we
measured Dmicelle at the lowest concentration (50 mg/mL) at
temperatures below the LCST using DLS. This value was
used afterward as a fixed input parameter in eqs 2 and 3,
which reduces the number of adjustable parameters. For
instance, we obtained from DLS at 24.4 ꢀC Dmicelle = 9.4 ꢀ
-
10
2
We have analyzed the curves in two ways: It was assumed that
i) the segmental dynamics and the micellar diffusion are com-
10
m /s, which corresponds to a micellar hydrodynamic
(
radius of 21.0 nm. This value is consistent with the micellar
radius of 17 nm determined using SANS (see above). As
expected, R is larger than the geometric radius because the
64
pletely decoupled from each other or (ii) that they are coupled.
The first case was described by the following equation:
h
swollen micelle drags along a shell of water when diffusing.
Using this value of Dmicelle as a fixed parameter in the fit of
eqs 2 and 3 (Figure 10), it is seen that the fitted curve lies in
both cases above the experimentally determined intermediate
scattering function for t = 20-60 ns. However, the fast decay
Sðq, tÞ
¼
c½Afast expð -ΓinternaltÞ
Sðq, 0Þ
þ ð1 -AfastÞ expð -ΓmicelletÞꢁ
ð2Þ

Article
Macromolecules, Vol. 43, No. 5, 2010 2499
Figure 11. Relaxation rates of the segmental dynamics, Γinternal, as a
2
function of q . The values were obtained using eq 2. Filled circles:
experimental data at 200 mg/mL at 24.4 ꢀC; full line: linear fit. Open
triangles: experimental data at 300 mg/mL at 28.6 ꢀC; dashed line:
linear fit.
Figure 10. Fits of eq 2 (a) and eq 3 (b) to the intermediate scattering
functions at 50 mg/mL measured at 24.3 ꢀC. Solid lines: fits with Dmicelle
as a free fitting parameter; dashed lines: fits with Dmicelle from DLS.
is very well recovered by both fits. This is also due to the low
amplitude of the slow decay.
For higher concentrations, the determination of Dmicelle is not
straightforward because the solutions are too viscous to be
filtered prior to the DLS experiments. Therefore, in an alter-
native approach, Dmicelle was used as a free fitting parameter at
50 mg/mL, and the effect on the fit of the fast decay was
examined. The resulting fits are shown in Figure 10 as well.
The fitting curves recover the experimental data at high Fourier
times very well; i.e., the decays are steeper with higher Dmicelle
values than the ones obtained from DLS. The resulting Dmicelle
values correspond to hydrodynamic micellar radii which are a
factor of ∼2 too small. The origin of the discrepancy may lie in a
different weighting of the DLS experiment (weight-average) and
the NSE experiment (number-average) over the distribution of
micelle sizes. The vastly different q values where DLS and NSE
experiments are carried out may also play a role. However, for
both equations and for both ways of treating Dmicelle, the fits are
equally good in the range of the fast decay, which is due to the
segmental dynamics. Therefore, we use Dmicelle as a fitting
parameter for all curves, even though it leads to an increase of
the uncertainties of the remaining fitting parameters.
Figure 12. Temperature dependence of diffusion coefficients obtained
from the fits of eq 2: stars, 50 mg/mL with Dmicelle from DLS; circles,
5
0 mg/mL with Dmicelle fitted; triangles up, 200 mg/mL with Dmicelle
fitted; triangles down, 200 mg/mL with Dmicelle fitted; filled symbols,
internal; open symbols, Dmicelle. The vertical dashed line indicates the
D
LCST from SANS. The horizontal full and dashed lines are the average
values of 50 and 200 mg/mL below the LCST.
scatter. Figure 12 shows that the values of Dinternal are slightly
lower when Dmicelle from DLS is used as a fixed fitting para-
meter; however, this effect is within the uncertainty. Below the
collapse temperature, the values of Dinternal are not tempera-
ture-dependent within the uncertainties. The average value of
Dinternal decreases with increasing polymer concentration, as
expected from the increase in local PNIPAM concentration due
to overlap of the micellar shells. All values are of the same order
of magnitude as the ones obtained for the shell dynamics of
The resulting relaxation rates Γ
are plotted as a
internal
2
64,65
function of q in Figure 11. For both measurements, the
proportionality is obeyed for the values obtained from the
decoupled model (eq 2), which indicates that the process
observed is due to the breathing mode of the PNIPAM shell.
Thus, the grafting density is sufficiently high, and the breath-
ing mode dominates over the single-chain dynamics. The
uncertainties of the relaxation rates from the coupled model
are too high to draw a significant conclusion, and the values
are therefore not shown.
chemically different diblock copolymer micelles in solution.
Above the collapse temperature determined from SANS,
the segmental diffusion coefficients are higher than below
and do not depend on polymer concentration any longer.
This seems counterintuitive since the swollen PNIPAM shell
collapses and forms a melt layer around the PS core, which
would imply slower dynamics. We attribute this unexpected
behavior to the following mechanism: The NSE experiment
measures a diffusion coefficient which is averaged over all
PNIPAM segments. We assume that a gradient of diffusion
coefficients exists for the segments close to the core and those
further away, which are surrounded by solvent. Below the
collapse temperature, the mobility of most segments is in
The average segmental diffusion coefficients Dinternal result-
ing from fitting eq 2 (decoupled case) are given for all three
concentrations as a function of temperature in Figure 12.
Fitting eq 3 leads to similar results, however, with much higher

2
500 Macromolecules, Vol. 43, No. 5, 2010
Adelsberger et al.
the time window of NSE, and we obtain an average value of
the segmental dynamics in the micellar shell. In contrast,
above the collapse temperature, most of the segments col-
lapse and form a relatively immobile PNIPAM melt with
only few PNIPAM blocks sticking out into the solution.
Some of these blocks may bridge two micelles. Only those
very mobile segments are monitored in the experiment,
resulting in an average diffusion coefficient higher than in
the swollen state. We conclude that NSE reveals the con-
centration dependence of the PNIPAM segmental dynamics
and its (significant) changes at the collapse transition.
(12) Okada, Y.; Tanaka, F. Macromolecules 2005, 38, 4465–4471.
(13) Hirokawa, Y.; Tanaka, T.; Matsuo, E. S. J. Chem. Phys. 1984, 81,
6
379–6380.
14) Shibayama, M.; Tanaka, T.; Han, C. C. J. Chem. Phys. 1992, 97,
829–6841.
15) Liao, G.; Xie, Y.; Ludwig, K. F., Jr.; Bansil, R.; Gallagher, P.
Phys. Rev. E 1999, 60, 4473–4481.
(
(
6
(16) Pelton, R. H.; Chibante, P. Colloids Surf. 1986, 20, 247–256.
(17) Hellweg, T.; Kratz, K.; Pouget, S.; Eimer, W. Colloids Surf., A
2
002, 202, 223–232.
18) Stieger, M.; Pedersen, J. S.; Lindner, P.; Richtering, W. Langmuir
004, 20, 7283–7292.
(
(
2
19) H o€ fl, S.; Zitzler, L.; Hellweg, T.; Herminghaus, S.; Mugele, F.
Polymer 2007, 48, 245–254.
4
. Conclusion
(20) Dingenouts, N.; Norhausen, C.; Ballauff, M. Macromolecules
998, 31, 8912–8917.
21) Hellweg, T.; Dewhurst, C. D.; Eimer, W.; Kratz, K. Langmuir
004, 20, 4330–4335.
1
Combining several analytical methods, namely turbidimetry,
(
fluorescence correlation spectroscopy, dynamic light scattering,
small-angle neutron scattering, and neutron spin-echo spectro-
scopy, we have investigated thermoresponsive PS-b-PNIPAM-
b-PS triblock copolymers in aqueous solution in a wide concen-
tration range, focusing on the micelle formation, the collapse
transition, and the resulting changes of the mesoscopic structure.
Moreover, the segmental dynamics was probed. Detailed investi-
gations were possible due to the use of fully deuterated PS blocks
and contrast matching in neutron scattering and spectroscopy.
The samples were directly soluble in water, facilitating the pre-
paration of the solutions.
The triblock copolymers investigated in this work differ from
the previously studied PS-b-PNIPAM diblock copolymers in two
respects which reflect the bridging between micelles: (i) The
micelles formed by triblock copolymers form clusters right after
the collapse temperature is crossed, and no single collapsed
micelles are observed. (ii) Correlation between the micelles is
already present below the collapse temperature.
As compared to the literature, our triblock copolymers have
relatively short hydrophobic end blocks. Thus, the investigated
system is in-between the previously studied telechelic PNIPAM
homopolymers and the PS-b-PNIPAM-b-PS triblock copoly-
mers with longer PS blocks. As a consequence, the PS cores of
the micelles are very small, and these micelles form small and pre-
sumably rapidly reacting building blocks of a thermoresponsive
network.
2
(22) Anderson, M.; Hietala, S.; Tenhu, H.; Maunu, S. L. Colloid Polym.
Sci. 2006, 284, 1255–1263.
(
(
(
23) Berndt, I.; Pedersen, J. S.; Richtering, W. Angew. Chem., Int. Ed.
006, 45, 1737–1741.
24) Kujawa, P.; Watanabe, H.; Tanaka, F.; Winnik, F. M. Eur. Phys. J. E
005, 17, 129–137.
2
2
25) Wang, W.; Troll, K.; Kaune, G.; Metwalli, E.; Ruderer, M.;
Skrabania, K.; Laschewsky, A.; Roth, S. V.; Papadakis, C. M.;
M u€ ller-Buschbaum, P. Macromolecules 2008, 41, 3209–3218.
26) Koga, T.; Tanaka, F.; Motokawa, R.; Koizumi, S.; Winnik, F. M.
Macromolecules 2008, 41, 9413–9422.
(
(
(
27) Nuopponen, M.; Ojala, J.; Tenhu, H. Polymer 2004, 45, 3643–
3
650.
28) Mertoglu, M.; Garnier, S.; Laschewsky, A.; Skrabania, K.;
Storsberg, J. Polymer 2005, 46, 7726–7740.
(29) Zhang, W.; Zhou, X.; Li, H.; Fang, Y.; Zhang, G. Macromolecules
2
005, 38, 909–914.
(
(
30) Tang, T.; Castelletto, V.; Parras, P.; Hamley, I. W.; King, S. M.;
Roy, D.; Perrier, S.; Hoogenboom, R.; Schubert, U. S. Macromol.
Chem. Phys. 2006, 207, 1718–1726.
31) Zhou, X.; Ye, X.; Zhang, G. J. Phys. Chem. B 2007, 111, 5111–5115.
(32) Kirkland, S. E.; Hensarling, R. M.; McConaughty, S. D.; Guo,
Y.; Jarrett, W. L.; McCormick, C. L. Biomacromolecules 2008, 9,
4
81–486.
(
(
(
(
33) Troll, K.; Kulkarni, A.; Wang, W.; Darko, C.; Bivigou Koumba,
A. M.; Laschewsky, A.; M u€ ller-Buschbaum, P.; Papadakis, C. M.
Colloid Polym. Sci. 2008, 286, 1079–1092.
34) Bivigou-Koumba, A. M.; Kristen, J.; Laschewsky, A.; M u€ ller-
Buschbaum, P.; Papadakis, C. M. Macromol. Chem. Phys. 2009,
2
10, 565–578.
35) Jain, A.; Kulkarni, A.; Bivigou Koumba, A. M.; Wang, W.; Busch,
P.; Laschewsky, A.; M u€ ller-Buschbaum, P.; Papadakis, C. M.
Macromol. Symp., in press.
36) Bivigou-Koumba, A. M.; G o€ rnitz, E.; Laschewsky, A.; M u€ ller-
Buschbaum, P.; Papadakis, C. M. Colloid Polym. Sci. 2010, 288,
DOI: 10.1007s00396-009-2179-9.
(37) Nyk €a nen, A.; Nuopponen, M.; Hiekkataipale, P.; Hirvonen, S.-P.;
Soininen, A.; Tenhu, H.; Ikkala, O.; Mezzenga, R.; Ruokolainen,
J. Macromolecules 2008, 41, 3243–3239.
38) Wang, W.; Metwalli, E.; Perlich, J.; Troll, K.; Papadakis, C. M.;
Cubitt, R.; M u€ ller-Buschbaum, P. Macromol. Rapid Commun.
Acknowledgment. We thank W. Doster and A.-K. Sommer
for assistance with the DLS experiments and fruitful discussions.
We kindly acknowledge given beam time by the J u€ lich Centre
for Neutron Science at FRM II. This work was supported by
the DFG priority program SPP1259 “Intelligente Hydrogele”
(
Pa771/4, Mu1487/8, La611/7, He2995/2-2). A.M.B.-K. grate-
fully acknowledges a personal grant from Deutscher Akade-
mischer Austauschdienst (DAAD).
(
References and Notes
2
009, 30, 114–119.
(
1) Das, M.; Zhong, H.; Kumacheva, E. Annu. Rev. Mater. Res. 2002,
(39) Moad, G.; Rizzardo, E.; Thang, S. H. Acc. Chem. Res. 2008, 41,
1133–1142.
36, 117–311.
(
(
(
2) Karg, M.; Hellweg, T. Curr. Opin. Colloid Interface Sci. 2009, 14, 438–450.
3) Hoffman, A. S. J. Controlled Release 1987, 6, 297–305.
4) Coughlan, D. C.; Quilty, F. P.; Corrigan, O. I. J. Controlled Release
(40) Jake ꢁs , J. Collect. Czech. Chem. Commun. 1995, 60, 1781–1791.
(41) Engels, R.; Clemens, U.; Kemmerling, G.; N o€ ldgen, H.; Schelten,
J. Nucl. Instrum. Methods Phys. Res., Sect. A 2009, 604, 147–149.
(42) These samples were erroneously labeled 170 mg/mL in ref 35.
(43) Heskins, M.; Guillet, J. E. J. Macromol. Sci., Part A 1968, 2, 1441–
1455.
2004, 98, 97–114.
(5) Schmaljohann, D. Adv. Drug Delivery Rev. 2006, 58, 1655–1670.
6) Feil, H.; Bae, Y. H.; Feijen, J.; Kim, S. W. J. Membr. Sci. 1991, 64,
(
2
83–294.
(44) F o€ rster, S.; Burger, C. Macromolecules 1998, 31, 879–891.
(45) Monkenbusch, M.; Sch €a tzler, R.; Richter, D. Nucl. Instrum.
Methods Phys. Res. 1997, A399, 301–323.
(46) Holderer, O.; Monkenbusch, M.; Sch €a tzler, R.; Kleines, H.;
Westerhausen, W.; Richter, D. Meas. Sci. Technol. 2008, 19, 034022.
(47) Souaille, M.; Fischer, H. Macromolecules 2002, 35, 248–261.
(48) Pallares, J.; Jaramillo-Soto, G.; Flores-Cata n~ a, C.; Vivaldo Lima,
E.; Lona, L. M. F.; Penlidis, A. J. Macromol. Sci., Part A: Pure
Appl. Chem. 2006, 43, 1293–1322.
(
(
7) Park, Y. S.; Ito, Y.; Imanishi, Y. Langmuir 1998, 14, 910–914.
8) Nyk €a nen, A.; Nuopponen, M.; Laukkanen, A.; Hirvonen, S.-P.;
Rytel €a , M.; Turunen, O.; Tenhu, H.; Mezzenga, R.; Ikkala, O.;
Ruokolainen, J. Macromolecules 2007, 40, 5827–5834.
(
9) Schild, H. G. Prog. Polym. Sci. 1992, 17, 163–249.
(
10) Lin, S.-Y.; Chen, K.-S.; Run-Chu, L. Polymer 1999, 40, 2619–
624.
11) Katsumoto, Y.; Tanaka, T.; Sato, H.; Ozaki, Y. J. Phys. Chem. A
002, 106, 3429–3435.
2
(
2
(49) Bhutto, A. A. J. Res. Sci. 2003, 14, 261–269.

Article
Macromolecules, Vol. 43, No. 5, 2010 2501
(
(
(
(
(
50) Larque, A. M.; Maire, C.; Belissent, H.; Galin, M.; Galin, J. C.
Actual. Chim. 199, 187–193.
51) Bonn ꢀe , T. B.; L u€ dtke, K.; Jordan, R.; St ꢁe p ꢀa nek, P.; Papadakis,
C. M. Colloid Polym. Sci. 2004, 282, 833–843.
52) Bonn ꢀe , T. B.; L u€ dtke, K.; Jordan, R.; Papadakis, C. M. Macromol.
Chem. Phys. 2007, 208, 1402–1408.
53) Bonn ꢀe , T. B.; L u€ dtke, K.; Jordan, R.; Papadakis, C. M. Colloid
Polym. Sci. 2007, 285, 491–497.
(58) Nojima, R.; Sato, T.; Qiu, X.; Winnik, F. M. Macromolecules 2008,
41, 292–294.
ꢁ
(59) Wang, W.; Metwalli, E.; Perlich, J.; Papadakis, C. M.; Cubitt, R.;
M u€ ller-Buschbaum, P. Macromolecules 2009, 42, 9041–9051.
(60) Grimson, M. J. J. Chem. Soc. 1983, 79, 817–832.
(61) Shibayama, M.; Tanaka, T.; Han, C. C. J. Chem. Phys. 1992, 97,
6829–6841.
(62) Ewen, B.; Richter, D. Adv. Polym. Sci. 1997, 134, 1–129.
(63) Kanaya, T.; Monkenbusch, M.; Watanabe, H.; Nagao, M.;
Richter, D. J. Chem. Phys. 2005, 122, 144905.
54) Proch ꢀa zka, K.; Limpoucho ꢀa , Z.; Webber, S. E. In Salamone, J. A.,
Ed.; Polymeric Materials Encyclopedia; CRC Press: Boca Raton, FL,
1
996; Vol. 1A,B; pp 764-772.
(64) Matsuoka, H.; Yamamoto, Y.; Nakano, M.; Endo, H.; Yamaoka,
H.; Zorn, R.; Monkenbusch, M.; Richter, D.; Seto, H.; Kawabata,
Y.; Nagao, M. Langmuir 2000, 16, 9177–9185.
(65) Castelletto, V.; Hamley, I. W.; Yang, Z.; Haeussler, W. J. Chem.
Phys. 2003, 119, 8158–8161.
(
(
(
55) Schuch, H.; Klingler, J.; Rossmanith, P.; Frechen, T.; Gerst, M.;
Feldthusen, J.; M u€ ller, A. H. E. Macromolecules 2000, 33, 1734–1740.
56) Semenov, A. N.; Joanny, J.-F.; Khokhlov, A. R. Macromolecules
1
995, 28, 1066–1075.
57) Kujawa, P.; Segui, F.; Shaban, S.; Diab, C.; Okada, Y.; Tanaka, F.;
Winnik, F. M. Macromolecules 2006, 39, 341–348.
(66) Yardimci, H.; Chung, B.; Harden, J. L.; Leheny, R. L. J. Chem.
Phys. 2005, 123, 244908.