Structure of Two-Compartment Hydrogels from Thermoresponsive

Article pubs.acs.org/Macromolecules

Structure of Two-Compartment Hydrogels from Thermoresponsive ABC Triblock Terpolymers Can Zhou,† Gilman E. S. Toombes,§ Matthew J. Wasbrough,§ Marc A. Hillmyer,† and Timothy P. Lodge*,†,‡ †

Department of Chemistry and ‡Department of Chemical Engineering & Materials Science, University of Minnesota, Minneapolis, Minnesota 55455-0431, United States § National Institute of Standards and Technology, Gaithersburg, Maryland 20899, United States S Supporting Information *

ABSTRACT: Aqueous dispersions of a poly(ethylene-alt-propylene)-bpoly(ethylene oxide)-b-poly(N-isopropylacrylamide) (PON) triblock terpolymer with block molecular weights of 3 000−25 000−10 000 and polymer concentrations ranging from 1 to 5 wt % were investigated at several temperatures from 25 to 55 °C using cryogenic scanning electron microscopy (cryo-SEM), cryogenic transmission electron microscopy (cryo-TEM), and small-angle neutron scattering (SANS). The cryo-SEM and cryo-TEM micrographs revealed that PON triblock terpolymer selfassembled into spherical micelles with PEP cores and PEO−PNIPAm coronae at room temperature and subsequently formed a twocompartment micellar network consisting of distinct spherical PEP and PNIPAm cores upon heating above the critical gelation temperature (42 °C). The formation of two discrete spherical PEP and PNIPAm hydrophobic domains was supported by detailed SANS analysis of the PON triblock samples in D2O, as the resulting intensity profiles can be successfully fitted using a scattering equation based on the two-compartment network structure. The two-compartment structure was further confirmed using contrast-matching SANS experiments on a PONd7 triblock sample with similar block molecular weights and a partially deuterated PNIPAm block. An important result of the SANS profiles in the gel state was the emergence of two distinct scattering peaks, which could be accounted for by considering spatial correlations between PEP and PNIPAm micellar cores. This study confirms the hypothesis that the formation of two-compartment networks in ABC terpolymer hydrogels results in better gelation properties, in comparison to other physically associated hydrogels, and can further guide the design and development of advanced hydrogel systems with enhanced performance.

■

INTRODUCTION The self-assembly of multiblock copolymers into multicompartment micelles or gels with distinguishable subdomains or compartments draws inspiration from biological systems such as eukaryotic cells, which possess distinct subunits that enable them to perform multiple functions simultaneously.1−3 For example, the discrete subdomains of these colloidal assemblies can facilitate the concurrent storage and delivery of two or more incompatible active agents such as drug molecules, gene therapy agents, or pesticides, in a prescribed manner.4 A variety of multicompartment micellar structures such as “hamburger” micelles, segmented worms, nanostructured bilayers and vesicles, and raspberry-like micelles have been successfully prepared in water or in water−organic solvent mixtures and well characterized by dynamic light scattering (DLS) and cryogenic transmission electron microscopy (cryo-TEM).5−13 Equally interesting are multicompartment physical gels, in which an ABC triblock terpolymer self-assembles into a network with distinct A and C micellar cross-links. In pioneering work, Weberskirch et al. studied the self-association behavior of poly(N-acylethylenimine) polymers end-capped with one fluorocarbon group and one hydrocarbon group and © XXXX American Chemical Society

showed the segregation of the end groups using NMR spectroscopy at relatively high polymer concentration (31 wt %).14 Similarly, Komenda et al. prepared multicompartment micellar hydrogels from poly(2-n-nonyl-2-oxazoline)-b-poly(2methyl-2-oxazoline)-b-poly(2-(1′H,1′H,2′H,2′H-perfluorohexyl)-2-oxazoline) (PNOx-b-PMOx-b-PFOx) and inferred the presence of spherical PNOx cores and ellipsoidal PFOx cores in the micellar network using small-angle neutron scattering (SANS) experiments.15 Yamaguchi et al. reported the structure of thermoplastic elastomer gels composed of polystyrene-bpolybutadiene-b-poly(methyl methacrylate) (PS-b-PB-bPMMA) in an aliphatic oil. They observed that the PS and PMMA end blocks are mixed into spherical cores at low polymer concentration and segregated into distinct cylindrical or lamellar microdomains at sufficiently large polymer concentration, using transmission electron microscopy (TEM) and small-angle X-ray scattering (SAXS) experiments.16 Other studies on ABC triblock hydrogels also suggest the Received: March 19, 2015 Revised: July 16, 2015

A

DOI: 10.1021/acs.macromol.5b00584 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

ization of ethylene oxide. The chain transfer agent S-1-dodecyl-S″(α,α′-dimethyl-α″-acetic acid) trithiocarbonate was then conjugated to the end of the PEP−PEO diblock, followed by RAFT polymerization with NIPIAm or d7-NIPAM (Polymer Source). The trithiocarbonate end groups were converted to ester end groups by sequential aminolysis and Michael addition with methyl acrylate. The product of each reaction step was confirmed by 1H NMR spectroscopy and characterized by size exclusion chromatography (SEC) (Figure S1). In the elugram, some amount of tailing was observed for both PON and PONd7 samples. The light scattering results confirmed that the tailing corresponded to the signals from PON triblocks rather than PO diblocks, suggesting that the apparent broadening is likely due to interaction between PNIPAm and the column, as is common for SEC of amide-containing polymers in THF without amine additives. The two samples investigated in this work are listed in Table 1 along with the molecular characteristics. Solution Preparation. All polymer solutions were prepared by the thin-film hydration method. Appropriate amounts of bulk polymer were dissolved in CH2Cl2, followed by evaporation of the solvent to yield a thin film on the walls of the vial. The thin film was then hydrated with a defined amount of water or heavy water, and the resulting mixture was stirred at room temperature for at least 1 month before further characterization. Water (H2O, Chromasolv grade) was used as solvent for cryogenic scanning electron microscopy (cryoSEM) and cryogenic transmission electron microscopy (cryo-TEM) experiments. For small-angle neutron scattering (SANS) experiments, heavy water (D2O, 99.9% D, Cambridge Isotopes) or H2O/D2O mixtures with compositions varying according to the desired contrast match were used as solvent. Cryogenic Scanning Electron Microscopy (Cryo-SEM). CryoSEM experiments were conducted on the 5 wt % PON sample following a reported protocol.20,27,28 A small amount of sample was sandwiched between two “freezing hats” (each about 100 μm deep). The sandwiched sample was left at 25 °C or heated at 50 °C for 10 min in the chamber of a Bal-Tec HPM 010 high-pressure freezing machine and then rapidly frozen at an operating pressure of 2100 bar. The frozen sample was then transferred into a liquid nitrogen bath and pried open with a scalpel to fracture the sample, exposing its interior. Vitrified water near the surface and a few micrometers deep into the sample was then partially sublimed in a Balzers MED 010 freeze-drying and sputtering device at −105 °C and ∼2 × 10−9 bar for about 5 min. The exposed surface was then coated with an 8 nm thick conducting Pt layer at −130 °C. The coated sample was transferred into a Hitachi S900 scanning electron microscope, maintained at about −175 °C, and examined at a low acceleration voltage of 2 keV to avoid excessive charging and radiation damage of the areas imaged. Cryogenic Transmission Electron Microscopy (Cryo-TEM). Cryo-TEM experiments were conducted on 1 and 2 wt % PON aqueous solutions. Vitrified samples were prepared using an FEI Vitrobot Mark III automated vitrification device. The aqueous solutions were stored at 25 °C or heated in an oil bath at 50 °C for 10 min and then rapidly applied to the lacey Formvar carbonsupported TEM grid placed within the climate chamber of the Vitrobot system, where the temperature was kept at 25 or 50 °C, respectively, and the relative humidity was kept at 100%. The excess solution was blotted with a piece of filter paper, resulting in the formation of thin films with thicknesses of ca. 100−300 nm in the mesh holes. The samples were quickly plunged into a reservoir of liquid ethane cooled by liquid nitrogen. The vitrified samples were then stored in liquid nitrogen until they were transferred to a cryogenic sample holder and examined with an FEI Technai Spirit BioTWIN transmission electron microscope with an Eagle CCD camera operated at an acceleration voltage of 120 kV at about −178 °C. The phase contrast was enhanced by acquiring images at a nominal underfocus of 10−15 μm. Small-Angle Neutron Scattering (SANS). SANS experiments were conducted on aqueous samples of PON and PONd7 over the concentration range of 1−5 wt %. Contrast matching was used to gain detailed information about the individual components. Contrast matching involves choosing a solvent environment with scattering

formation of segregated end block microdomains in the viscoelastic network, but no morphological characterization was reported.17−19 Recently, Taribagil et al. investigated the morphology of a compartmentalized hydrogel formed from a telechelic poly(ethylene oxide) (PEO) end-capped with mutually incompatible hydrophobic blocks (1,2-polybutadiene (PB) and poly(perfluoropropylene oxide) (PFPO)) and revealed a bicontinuous structure composed of PFPO disks distributed within a hydrophobic PB sheet covered by hydrophilic PEO brushes, by cryogenic scanning electron microscopy (cryo-SEM) and SANS experiments.20,21 We recently reported the gelation behavior of poly(ethylenealt-propylene)-b-poly(ethylene oxide)-b-poly(N-isopropylacrylamide) (PON) triblock terpolymers in water.22 These terpolymers form micelles in water at low temperatures, with hydrophobic PEP cores surrounded by hydrophilic PEO− PNIPAm coronae. These micelles associate to form a hydrogel upon heating above the lower critical solution temperature (LCST) of PNIPAm. The separation of micellization and gelation steps leads to formation of a two-compartment network with a very high fraction of bridging conformations for the PEO midblocks. Therefore, gelation could be achieved at a much lower concentration, with a much sharper sol−gel transition, as compared to poly(N-isopropylacrylamide)-bpoly(ethylene oxide)-b-poly(N-isopropylacrylamide) (NON) triblock copolymer hydrogels. The formation of discrete PEP and PNIPAm cores in the micellar network is therefore crucial for the preparation of compartmentalized hydrogels with improved gelation properties. However, the interpretation in terms of discrete PEP and PNIPAm micellar cross-links is based on rheology and cryogenic transmission electron microscopy; the former is at best an indirect probe of structure, and the latter cannot discriminate among PEP, PNIPAm, and mixed PEP/PNIPAM core micelles. In this article, therefore, we rely on small-angle neutron scattering with contrast matching to provide detailed morphological characterization of the aqueous solutions of PON triblock terpolymers, over the concentration range from 1 to 5 wt % and at varying temperatures. The results provide strong support for the hypothesis that PON terpolymers form two-compartment hydrogels with discrete spherical PEP and PNIPAm micelles in water upon heating above the LCST of PNIPAm.

■

EXPERIMENTAL SECTION

Materials. Two triblock terpolymers were prepared: one with a normal hydrogenous PNIPAm block and one with a partially deuterated PNIPAm block PONd7 (Scheme 1). Both were prepared

Scheme 1. Chemical Structure of the PONd7 Triblock Terpolymer

using a combination of anionic and reversible addition−fragmentation chain transfer (RAFT) polymerizations, followed by end-group modification.23 Briefly, 1,4-polyisoprene was prepared by anionic polymerization (91% 1,4) and terminated by addition of a single ethylene oxide unit to afford a terminal hydroxyl group. The resulting PI-OH was saturated to PEP-OH by catalytic hydrogenation. PEP-OH was used as a macroinitiator for the anionic ring-opening polymerB


Article

Macromolecules Table 1. Molecular Parameters of PON and PONd7 Triblock Terpolymers samplea

NPEPb

NPEOb

NPNIPAmb

f PEPc

f PEOc

f PNIPAmc

Đd

PON(3−25−10) PONd7(3−25−11)

45 45

565 565

89 95

0.11 0.10

0.63 0.62

0.26 0.28

1.09 1.11

The numbers in parentheses correspond to the molar masses of PEP, PEO, and PNIPAm (dPNIPAm), respectively, in kg mol−1 as determined by H NMR spectroscopy. bNumber-average degree of polymerization as determined by 1H NMR spectroscopy. cVolume fractions were estimated using the molecular weight and the RT bulk densities: ρ(PEP) = 0.856 g/cm3,24 ρ(PEO) = 1.12 g/cm3,25 and ρ(PNIPAm) = 1.07 g/cm3.26 dThe dispersity was measured by SEC equipped with both light scattering (LS) and refractive index detectors with THF as the mobile phase. a

1

length density (ρ*) equal to that of the component to be contrast matched. This renders that particular component invisible to neutrons and enables better resolution of information about other components. For aqueous systems, this is done by mixing water (ρ* = −5.6 × 109 cm−2) and heavy water (ρ* = 6.36 × 1010 cm−2). The scattering length density of the mixture is given by the sum of the scattering length densities of water and heavy water weighted by their volume fractions (= ∑ϕiρi*). The scattering length density of each polymer block was calculated with the NIST scattering length density calculator.29 Solutions of PON were prepared in D2O, as the scattering length densities of PEP (ρ* = −3.06 × 109 cm−2), PEO (ρ* = 6.34 × 109 cm−2), and PNIPAm (ρ* = 7.92 × 109 cm−2) are not very different, and therefore scattering intensity under contrast matched conditions is relatively low and cannot be clearly resolved from background. The scattering length density of deuterated PNIPAm was ρ* = 4.65 × 1010 cm−2, so the PONd7 samples were prepared under two contrast matching conditions, namely PEP contrast-matched (96 vol % H2O) and zero mean contrast (67.5 vol % H2O); see Figure S2 in Supporting Information. The zero mean contrast condition indicates that the scattering length density of the solvent equals that of the triblock as a whole, rather than any individual block. The scattering experiments were performed at the National Institute of Standards and Technology (NIST) on the NG-7 30 m instrument of the Cold Neutron Research Facility.30 A neutron wavelength of 7 or 8 Å was used in conjunction with detector distances of 1, 3, 13, and 15 m to cover scattering wave vectors from 0.001 to 0.3 Å−1. The samples were loaded into NIST sample cells with a path length of 1 mm and held at the desired discrete temperature for 5 min before taking measurements. Using Igor Pro macros, the resulting data were corrected for background, nonuniform detector efficiency, empty cell scattering, and sample transmission. The scattering intensities were then scaled to absolute values on the basis of direct beam flux measurements.31

■

Figure 1. Dynamic elastic modulus of aqueous solutions of PON(3− 25−10) at a frequency ω = 10 rad/s and heating rate of 1 K/min. Data reproduced from ref 22.

range of 50−70 nm (Figure 2a). This is consistent with the formation of spherical micelles with PEP cores and PEO− PNIPAm coronas at room temperature. Quantitative determination of the size of the PEP cores is complicated by the presence of a ca. 8 nm platinum coat and the difficulty in discriminating between the core and the corona. The cryo-SEM analysis at 50 °C reveals a network structure, with a significant number of voids (i.e., water-filled chambers) (Figure 2b). The typical dimension of these voids is 200−300 nm. The higher magnification image shows clusters of spheres with diameters in the range of 20−30 nm (Figure 2c). This dimension includes the ca. 8 nm thick platinum coat and contributions primarily from the hydrophobic cores (PEP or PNIPAm). Therefore, the radius of the hydrophobic cores can be estimated to be between 5 and 10 nm. However, all the spheres in Figure 1c appeared to be very similar. Thus, whether or not the micellar network contains two different hydrophobic domains (PEP and PNIPAm) cannot be verified. Overall, these cryo-SEM results are consistent with the conclusion that the PON hydrogels formed at elevated temperatures consist of networks of spherical micelles, with the presence of large voids.22 Cryo-TEM images of the 2 wt % PON(3−25−10) sample prepared at 25 and 50 °C are shown in Figure 3. In Figure 3a, the sample was vitrified after annealing at 25 °C, which is below the gel temperature. The formation of spherical PEP micellar cores (Rc ≈ 8 ± 2 nm) with liquid-like arrangement are clearly visible. The same solution gave the image shown in Figure 3b after annealing at 50 °C in the gel state. A 3−4-fold increase in the number of micellar cores is clearly evident, thereby suggesting the formation of a different micellar core

RESULTS AND DISCUSSION

The remarkably sharp gelation transitions for PON(3−25−10) solutions are illustrated in Figure 1. Upon heating at 1 K/min, the elastic modulus increases by 3−4 orders of magnitude over a ca. 4 deg window, even for polymer concentrations as low as 1%. We have previously proposed that this superior performance is attributable to a two-step gelation process, in which PEP core micelles are arrayed with liquid-like order at room temperature, with the PNIPAm blocks on the periphery of the coronas.22 Upon heating through the LCST of the PNIPAm chains, the latter assemble rapidly to form a second set of micelles; the fact that the PNIPAm chains are spatially “prearranged”, in roughly spherical shells around the PEP micelles, facilitates the temperature-induced self-assembly. Conversely, an equivalent symmetric NON triblock showed much broader gelation transitions and did not gel at all at 1− 2%. The main purpose of this paper, therefore, is to test this hypothesis with detailed structural studies and especially by SANS. Cryo-SEM images of the 5 wt % PON(3−25−10) sample prepared at 25 and 50 °C are shown in Figure 2. The sample at 25 °C suggests an array of spherical objects with radii in the C


Article

Macromolecules

Figure 2. Cryo-SEM images of the 5 wt % PON(3−25−10) sample at (a) 25 °C and (b, c) 50 °C.

Figure 3. Cryo-TEM images of the 2 wt % PON(3−25−10) sample at (a) 25 °C and (b) 50 °C.

(PNIPAm), which leads to the two-compartment network structure. The average micellar core radius (Rc ≈ 9 ± 2 nm) is comparable to that at 25 °C, in the micelle state, indicating that PNIPAm micelles have similar core radii to the PEP micelles. Note that under these conditions there is no exchange of PEP blocks among different micelles; i.e., the PEP cores are “frozen”. Based on the assumption that PNIPAm and PEP micelles have comparable core radii, the ratio of PNIPAm to PEP volume fractions suggests that there should be 2−3 times more PNIPAm cores, roughly consistent with the cryo-TEM observation; the SANS analysis to be discussed subsequently provides a more quantitative estimate. It is worth noting that reducing the polymer concentration to 1 wt % gave almost identical results to Figure 3, at both 25 and 50 °C (see Figure S3 and Figure 3 of ref 22). On the basis of these observations, we propose a morphology for the two-compartment micellar network as illustrated in Figure 4a, in which both PEP and PNIPAm blocks form spherical micelles bridged by PEO chains to give the twocompartment, polymer-rich gel phase, with water-filled voids forming the other polymer-lean phase. The presence of voids results from the low polymer concentrations (1−5 wt %) used. At such low polymer concentrations, there is insufficient material to permeate the entire sample volume at the preferred midblock extension, leading to large-scale heterogeneities. These heterogeneities give rise to local water-rich pockets, which appear as voids upon sublimation of the water in the cryo-SEM experiments; this picture is consistent with the turbidity that appears on heating.22 SANS Analysis. SANS measurements were performed to assess the viability of the proposed two-compartment morphology. The SANS intensity profiles of 2 wt % PON(3−25−10) in D2O measured at three different temperatures are shown in Figure 5. At room temperature (27 °C), the

Figure 4. (a) Proposed morphology of the two-compartment network for PON hydrogels. (b) Illustration of the different spatial correlations (d1 and d2) between micelles in PON hydrogels.

Figure 5. SANS intensity profiles obtained for the 2 wt % PON(3− 25−10) sample in D2O measured at three different temperatures. For clarity, the intensity data for higher temperatures have been shifted vertically: 46 °C (×10), 55 °C (×102). The open symbols represent experimental data, while the red curves represent model fits detailed in the text.

PON aqueous solution show scattering curves that are typical for block copolymer micelles; in particular, the intensity trace shows features that reflect the core−shell structure at higher q, but is flat at lower q, representative of a homogeneous solution. D


Article

Macromolecules

peak position for the 5 wt % solution at 27 °C in Figure S4 (∼0.009 Å−1) corresponds to a micellar center-to-center distance of ∼70 nm. This distance is approximately twice the hard-sphere radius (Rhs = 30−33 nm) derived from the model fits, supporting the validity of the fitting results. As the hardsphere radius remains nearly invariant with increasing polymer concentration, the hard-sphere volume fraction should be approximately proportional to polymer weight fraction. However, the hard-sphere volume fraction increases from 0.20 to 0.26 with increasing polymer concentration from 1 to 2 wt %, which is less than a factor of 2. The same trend holds as the concentration is further raised from 2 to 5 wt %. This is likely due in part to the limitation of hard-sphere (excluded volume) interactions for describing the structure factor in the higher concentration samples; note that the model fittings deteriorate slightly in the low-q region on increasing polymer concentration (Figure S4). The proposed model used to fit the intensity profiles of PON hydrogels at elevated temperatures consists of three individual contributions: form factors and structure factors for PEP and PNIPAm micellar cores, high-q scattering from PEO chains, and low-q scattering resulting from spatial heterogeneities. The total scattering intensity is therefore given by

Upon heating above the gelation temperature, for example, at 46 and 55 °C, in the gel state, the scattering curves change markedly. Two distinct scattering peaks appear at ∼0.01 and ∼0.02 Å−1, corresponding to two characteristic length scales in the spatial correlations among micelles; this feature alone suggests the presence of two different micelles rather than one type of mixed micelle (Figure 4b). There is also a strong upturn in intensity at lower q, consistent with the voids (water-rich heterogeneities) revealed by cryo-SEM. It should be noted that either decreasing the polymer concentration to 1 wt % or increasing to 5 wt % gives very similar intensity profiles at all three temperatures (Figure S4). The block copolymer micelle model used to fit the intensity profiles of PON micellar solutions at room temperature was developed by Pedersen and co-workers32−36 and is explained more fully in the Appendix. Briefly, the total coherent scattering intensity is described in terms of a form factor for a micelle with a spherical core and Gaussian corona chains and a monodisperse hard sphere structure factor (eq A1 in the Appendix). The micelle form factor consists of two selfcorrelation terms for the core and corona chains and two crosscorrelation terms between core and corona chains as well as between different corona chains, which are expressed in terms of the core radius (Rc), the radius of gyration (Rg) of the Gaussian corona chains, the width of the core−corona interface (σint), and the cubic B spline function parameters (a1, s) for the radial scattering length density distribution of the corona chains (eqs A2−A5 in the Appendix). The monodisperse hard-sphere structure factor depends on the hard-sphere radius (Rhs) and the hard-sphere volume fraction (ηhs) (eqs A7 and A8 in the Appendix). Dispersity in micelle size was accounted for by a Gaussian distribution of core radii with a distribution width of σR (eq A9 in the Appendix). Finally, instrumental smearing was accounted for by convolving the scattering intensity with the SANS instrumental resolution function.31 The model fitting was performed using IGOR Pro software. The core radius Rc was fixed at 8 nm, as determined by cryoTEM (see Figure 3). As the scattering length densities of PEO (ρ* = 6.34 × 109 cm−2) and PNIPAm (ρ* = 7.92 × 109 cm−2) are close, an average scattering length density of 6.81 × 109 cm−2 was used for the PEO−PNIPAm coronae. As shown in Figure 5 and Figure S4, the intensity profiles of PON micelles at 27 °C were successfully fit using this model, and all of the fitting parameters are summarized in Table 2. As there are

I(q) = n1 |A1(q)|2 S11(q) + 2 n1n2 A1(q)A 2 (q)S12(q) + n2 |A 2 (q)|2 S22(q) + a

1+ + b(ρgelphase − ρsolvent )2

a

Rc a (nm)

σint (nm)

Rg (nm)

a1

s (nm)

Rhs (nm)

ηhs

σR (nm)

1 2 5

8.0 8.0 8.0

2.5 2.5 1.7

9.5 9.5 9.1

0.27 0.34 0.29

41 44 39

30 33 31

0.20 0.26 0.37

0.1 0.2 0.2

exp(−q

2

q2

q 2R g 2

2 d int 2)

+ Iincoherent

(1)

The first three terms in eq 1 represent the scattering arising from the hydrophobic PEP (type 1) and PNIPAm (type 2) cores in the gel phase, where n1 and n2 are number densities, A1(q) and A2(q) are the form factor amplitudes, and S11(q), S12(q), and S22(q) are the interparticle structure factors describing the correlations between 1−1, 1−2, and 2−2 spheres. Because the PEP micelles should be almost completely dehydrated, the PEP core radius, R1,c, and number density, n1, are related by n1 =

fP fPEP 4πR1,c 3/3

(2)

where f P is the volume fraction of the PON triblock in water and f PEP is the volume fraction of PEP in the PON triblock. In contrast, even above the transition temperature, the PNIPAm cores may still contain some residual solvent, which can be characterized by η(H2O,PNIPAm), the fraction of solvent in the PNIPAm core. The PNIPAm core radius, R2,c, and number density, n2, are then related by

Table 2. Fitting Parameters for PON(3−25−10) Micelles in D2O at Room Temperature conc (wt %)

(ρPEO − ρsolvent )2

Value fixed on the basis of the cryo-TEM results.

n2 =

many parameters, the sensitivity of the fits was tested by monitoring the change of the results upon increasing or decreasing each parameter by 10% (Figure S5). We found that the most sensitive parameter is the PEP core radius (Rc), and 8 nm gives the best fitting results. This is consistent with the cryo-TEM experiments. The hard-sphere radius (Rhs) should be equal to half the minimum micellar center-to-center distance, which is the reciprocal of the scattering peak position q*. The

fP fPNIPAm + (1 − fP )η(H 2O, PNIPAm) 4πR 2,c 3/3

(3)

where f PNIPAm is the volume fraction of PNIPAm in the PON triblock. The form factor amplitude for the PEP and PNIPAm cores is given by (P(q) = |A(q)|2) ⎛ − q 2σ 2 ⎞ 4 i ,int ⎟⎟ Ai (q) = (ρi − ρPEO + solvent ) πR i ,c 3Φ(qR i ,c) exp⎜⎜ 3 2 ⎝ ⎠ (4) E


Article

Macromolecules where Φ(x) = 3[sin(x) − x cos(x)]/x3 is the hard-sphere form factor and σi,int represents the width of interface between core and corona for spheres of type “i”. The neutron scattering length density (SLD) of PEP is simply ρ1 = ρPEP. The SLD of PNIPAm is

function, where Rg is the radius of gyration of the PEO chains and a is a coefficient proportional to the number of chains. The fifth term accounts for the low-q scattering resulting from the spatial heterogeneities and is fitted empirically by the mass fractal scattering equation for a disordered network, where b is a coefficient proportional to the number of scattering objects, ρgel phase is the SLD of the gel phase, and dint is the characteristic size of the disordered scattering objects.41,42 Finally, given the difficulty in accurately estimating and subtracting the incoherent background scattering, a small background term (Iincoherent) is also included in eq 1. The resulting scattering intensity was then convolved with the SANS instrumental resolution function to account for instrumental smearing. In all, 12 parameters could be adjusted in the fits: the core radii of PEP and PNIPAm (Rc), the width of the core−corona interface for PEP and PNIPAm micelles (σint), the hard-sphere volume fraction of PEP and PNIPAm spheres (ηhs), interaction potential between PEP and PNIPAm spheres (τ), the fraction of H2O in the gel phase [η(H2O, gel phase)], the fraction of H2O in PNIPAm micelles [η(H2O, PNIPAm)], two prefactors (a, b) associated with the high-q scattering and low-q scattering terms, and the characteristic size of the disordered scattering objects (dint). The model fitting was performed using Matlab software. To constrain these multiparameter fits, Rc(PEP) and σint(PEP) were fixed at 8 and 2.4 nm, respectively, as determined by cryo-TEM and SANS analysis of PON micelles at room temperature. The range of Rc(PNIPAm) was constrained to fall between 7.5 and 10 nm, as the cryo-TEM observation suggests that PNIPAm micelles have similar or slightly larger core radii in comparison to PEP micelles. η(H2O, PNIPAm) has been fixed at 0 as this gives the best fitting results. As shown in Figure 5 and Figure S4, the intensity profiles of PON hydrogels at both 46 and 55 °C can be successfully fitted using this proposed scattering equation. The appearance of multiple peaks is a direct result of the nonrandom arrangement of PEP and PNIPAm micelles. The individual scattering contributions from each term in eq 1 for a representative PON sample are displayed in Figure S6. This illustrates that to a large degree the low-q scattering (and fitting) is independent of the scattering due to the micelles and that neither the high-q scattering nor the background is particularly significant. Thus, the parameters of most interest are those arising from the first three terms in eq 1. The fitting parameters are listed in Table 3, and the sensitivity is explored in Figure S7. The two prefactors (a, b) have no particular significance related to the purpose of this study and thus are not reported here. dint has a common value of ∼18 nm for all PON hydrogel samples, which suggests that the smallest disordered length scale in the sample is the size of the individual micelles. We observed a slight increase in the core radius of PNIPAm along with a small decrease in the core−corona interface thickness at all three concentrations when PON samples were heated from 46 °C, which is just a few degrees above the critical gelation temperature (Tgel, 42 °C),22 to higher temperatures, for example, 55 °C. PNIPAm becomes progressively more dehydrated upon heating across and above the LCST, and the relatively higher hydrophobicity at 55 °C leads to a larger aggregation number (larger core radius) with a sharper interface (smaller core−corona interface thickness) for PNIPAm micelles. This is consistent with variable-temperature 1 H NMR results of the same PON triblock solution, in which continuous dehydration of the PNIPAm block occurred on

ρ2 = fP fPNIPAm ρPNIPAm + (1 − fP )η(H 2O, PNIPAm)ρsolvent fP fPNIPAm + (1 − fP )η(H 2O, PNIPAm) (5)

Similarly, the SLD of the PEO/solvent regions within the gel phase is ρPEO + solvent = fP fPEO ρPEO + (1 − fP )[η(H 2O, gel phase) − η(H 2O, PNIPAm)]ρsolvent fP fPEO + (1 − fP )[η(H 2O, gel phase) − η(H 2O, PNIPAm)]

(6)

where η(H2O, gel phase) is the fraction of H2O in the gel phase. A crucial point is that in the picture of Figure 4 the PEP spheres are preferentially surrounded by PNIPAm spheres, and vice versa. Consequently, the two types of spheres are not packed randomly. Indeed, exploration of fitting models based on random arrays of two types of spheres never yielded two distinct scattering peaks, in contrast to the experiments. Therefore, a structure factor for binary spheres interacting through hard-sphere interactions is inadequate.37 To model the alternating tendency of the two spheres, a “sticky” hard-sphere structure factor (i.e., an interparticle structure factor for spheres with a narrow attractive well) was adopted.38−40 In this model, the short-range repulsion between spheres is accounted for by the hard-sphere radius (Rhs), and the increased probability for opposite sphere types to be nearest neighors is accounted for by a short-ranged adhesion energy between opposite sphere types. An attraction between opposite sphere types favors 1−2 contacts over 1−1 and 2−2 contacts. The interaction potential is given by ⎧ ⎫ Di + Dj ⎪∞ 0 < r < = Di , j ⎪ 2 ⎪ ⎪ ⎬ uij(r ) = ⎨ u D < r < D + Δ i,j i,j ⎪ 0 i,j ⎪ ⎪0 r > D + Δ ⎪ i,j i,j ⎩ ⎭

(7)

where Di,j is the hard-sphere diameter, Δi,j is the width of the square well, and −u0 is the depth of the well. The strength of the adhesion between hard spheres of types 1 and 2 is then described in terms of the “stickiness”, τ, as defined below: τ=

exp(u0 /kT ) 12Δ/(D + Δ)

(8)

The stickiness can be varied to adjust the interaction strength and is therefore used as a fitting parameter. For this binary sticky hard-sphere model, the interparticle structure factors, S11(q), S12(q), and S22(q) can be calculated using the formulas reported by Bergenholtz and co-workers,39 as functions of the volume fractions and number densities of the two sphere types within the gel phase, as well as τ, the stickiness between spheres of type 1 and type 2. The fourth term in eq 1 describes the high-q scattering from bridging PEO chains with the Lorentzian tail of the Debye F


Article

Macromolecules Table 3. Fitting Parameters for PON(3−25−10) Hydrogels in D2O at Elevated Temperatures 1 wt % 46 °C 55 °C Rc(PEP)a (nm) σint(PEP)a (nm) Rc(PNIPAm) (nm) σint(PNIPAm) (nm) η(H2O, gel phase)b (%) ηhs(PEP) (%) ηhs(PNIPAm) (%) micelle attraction (τ) dint (nm) n(PNIPAm)/n(PEP)c

8.0 2.4 8.6 3.3 63 10 26 13 19 1.9

8.0 2.4 9.6 2.7 31 3 17 19 19 1.3

2 wt %

5 wt %

46 °C 55 °C 46 °C 55 °C 8.0 2.4 8.5 3.2 100 15 30 11 17 2.0

8.0 2.4 9.5 2.7 40 4 28 12 19 1.4

8.0 2.4 8.9 3.1 100 20 43 10 18 1.7

8.0 2.4 9.6 2.0 88 5 33 13 14 1.3

Rc(PEP) and σint(PEP) were fixed at 8 and 2.4 nm, respectively. Fraction of H2O in the gel phase. cNumber of PNIPAm micelles relative to PEP micelles, as determined from eqs 2 and 3.

a b

heating at and above the Tgel (see Figure S8). In addition, the fraction of water in the gel phase drops significantly upon heating from 46 to 55 °C. This is likely due to the LCST phase behavior of PEO in water and the syneresis of PNIPAm above the LCST, consistent with a number of reports concerning PNIPAm-containing block copolymers.43,44 This trend is reversed on increasing the concentration at both 46 and 55 °C, as expected from the fact that the heterogeneity decreases with increasing polymer concentration.22 It is worth noting that no significant difference was observed for the size and interface thickness of PNIPAm spheres with varying concentration. The number of PNIPAm micelles relative to PEP micelles as determined from eqs 2 and 3 was estimated to be ∼2 at 46 °C and decreased to ∼1.5 at 55 °C, which is slightly smaller than the value (2−3) at 50 °C suggested by cryo-TEM analysis. Given the small sampling volume of cryo-TEM, and that multiple parameters were varied in the SANS fitting procedure, such a modest disagreement is not of great concern. Overall, the SANS analysis of PON(3−25−10) samples provides strong support for the two-compartment network structure proposed in Figure 4a. However, more detailed structural information for the individual blocks, especially PNIPAm, cannot be identified separately using the contract matching method because the scattering length densities of PEP, PEO, and PNIPAm are not that different. Therefore, we prepared the PONd7 triblock with a deuterated PNIPAm block and performed SANS experiments under two different contrast matching conditions, namely PEP contrast matched and zero mean contrast, to further elucidate the size, shape, and packing of the PNIPAm domains. When the PEP is contrast matched, the scattering is dominated by the deuterated PNIPAm block (Figure S2), and therefore the structural change of PNIPAm upon heating from the micelle to gel state should be easily resolvable. Representative SANS intensity profiles of 2 wt % PONd7(3− 25−11) samples under PEP matched conditions in both the micelle (25 °C) and gel states (55 °C) are shown in Figure 6a. The appearance of a single scattering peak of ∼0.016 Å−1 with a small shoulder on heating is consistent with the conversion of dPNIPAm from Gaussian chains in the corona to micellar cores and can be fitted using eq 1 with parameters listed in Table 4. It should be noted that the core radii of dPNIPAm micelles are ∼8 nm, irrespective of concentration (Figure S9), a reasonable value as expected from the cryo-TEM and SANS analysis of

Figure 6. SANS intensity profiles obtained for 2 wt % PONd7(3−25− 11) samples in (a) 96 vol % H2O (PEP contrast matched) and (b) 67.5 vol % H2O (zero mean contrast) measured at 25 and 55 °C. For clarity, the intensity data at 55 °C have been shifted vertically: 55 °C (×10). The open symbols represent experimental data, while the red curves represent model fits.

PON triblock samples. The various pairwise scattering contributions are also illustrated in Figure S10. The zero mean contrast signal is also informative, since the absence of scattering at low q would help establish that the PNIPAm really forms discrete micelles rather than 1D worms, 2D sheets, or 3D networks. The SANS intensity profiles of 2 wt % PONd7(3−25−11) samples under zero mean contrast conditions are shown in Figure 6b. The low-q upturn was not observed for PONd7(3−25−11) hydrogels, confirming the formation of finite sized PNIPAm micelles in the gel state. Again, we obtained reasonable agreement between experimental data and the model fitting results with PNIPAm core radii of ∼8 nm at all three polymer concentrations (1, 2, and 5 wt %, Figure S11 and Table 4). Furthermore, the results are broadly consistent with those in Tables 3 for the unlabeled triblocks. The combined SANS results and fits provide excellent support for the model proposed in Figure 4. Clearly, the structure of these samples is quite complicated to model in detail, and the fitting functions have multiple parameters, so we do not claim that this model is unique in describing the results, or that the parameter are all individually optimized. However, in G


Article

Macromolecules Table 4. Fitting Parameters for PONd7(3-25-11) Hydrogels at 55 °C PEP contrast matched a

Rc(PEP) (nm) σint(PEP)a (nm) Rc(dPNIPAm) (nm) σint(dPNIPAm) (nm) η(H2O, gel phase)b (%) ηhs(PEP) (%) ηhs(dPNIPAm) (%) micelle attraction (τ) dint (nm) n(dPNIPAm)/n(PEP)c

zero mean contrast

1 wt %

2 wt %

5 wt %

1 wt %

2 wt %

5 wt %

8.0 2.4 7.5 3.8 21 2 11 18 15 3.3

8.0 2.4 7.6 3.9 31 3 20 11 17 3.3

8.0 2.4 8.0 3.5 57 6 21 5 17 2.8

8.0 2.4 7.6 4.2 30 5 15 12 20 3.3

8.0 2.4 7.7 3.8 45 10 18 12 20 3.1

8.0 2.4 8.0 3.5 98 11 26 12 21 2.8

Rc(PEP) and σint(PEP) were fixed at 8 and 2.4 nm, respectively. bFraction of H2O in the gel phase. cNumber of PNIPAm micelles relative to PEP micelles, as determined from eqs 2 and 3. a

combination with the remarkably sharp rheological gelation transitions, and the cryo-TEM and cryo-SEM images, the SANS analysis reinforces the validity of the proposed mechanism.

Pmic(q) = Nagg 2βcore 2Acore 2 (q) + Naggβcorona 2Pchain(q) + 2Nagg 2βcoreβcorona Acore(q)Acorona (q) + Nagg(Nagg − 1)

■

× βcorona 2Acorona 2 (q)

SUMMARY In conclusion, we have investigated the morphology of PON micelles and hydrogels using a combination of microscopy and scattering experiments. The cryo-SEM and cryo-TEM analyses suggest that PON triblocks form spherical micelles with PEP cores and PEO−PNIPAm coronas at room temperature, and two-compartment micellar networks, in which both PEP and PNIPAm form spherical micelles, bridged by PEO chains with the presence of water-filled voids, upon heating above the critical gelation temperature. The formation of two discrete spherical PEP and PNIPAm cores in the micellar network is further confirmed by SANS experiments of PON triblocks with a normal PNIPAm and a deuterated PNIPAm block. This study confirms the hypothesis that ABC triblocks with two immiscible, hydrophobic end blocks can be beneficial for hydrogel formation, resulting from the formation of twocompartment networks with exclusively bridging conformations for the midblocks. Therefore, these results can help guide the design and development of new physically associated hydrogels with enhanced performance.

Here q is the scattering vector, Nagg is the aggregation number, and βcore and βcorona are total excess scattering lengths of core and corona blocks, respectively. They are defined as βcore = υcore(ρcore − ρsolvent) and βcorona = υcorona(ρcorona − ρsolvent), where υcore and υcorona are the volumes of core and corona chains, respectively. Also, ρcore, ρcorona, and ρsolvent are the scattering length densities of core block, corona block, and solvent, respectively. The first term is the self-correlation of the spherical core with radius Rc and a smoothly decaying scattering length density at the core−corona interface. Acore2(q) is given by Acore 2 (q) = Φ2(qR c) exp( −q2σint 2)

(A3)

where Φ(x) = 3[sin(x) − x cos(x)]/x3 is the hard-sphere form factor and σint takes into account the smoothly decaying density at the interface and represents the width of interface between core and corona. The radius of the spherical core and micelle aggregation number are directly linked by assuming the core contains no solvent. The second term is the self-correlation of the corona chains, which is approximated by a Debye function where the chains are considered as Gaussian chains with radius of gyration Rg. Pchain(q) is given by

■

APPENDIX. FITTING MODELS FOR SANS INTENSITY PROFILES OF PON MICELLAR SOLUTIONS The total coherent scattering intensity of PON micellar solutions at room temperature is expressed as a function of a micelle form factor and a monodisperse hard-sphere structure factor.32−36 I(q) =

(A2)

Pchain(q) =

2[exp(−x) − 1 + x] x2

x = q 2R g 2

;

(A4)

The last two terms are the cross term between core and corona and between corona chains, respectively. Both terms include the form factor of the corona chains, which is given as the normalized Fourier transform of the radial density distribution function of the corona chains (ρcorona(r))

∫ D(Rc)(Pmic(q) + (A mic(q)) [S(q) − 1]) dRc 2

(A1)

where Pmic(q) is the scattering form factor for a micelle consisting of a spherical core and Gaussian corona chains attached to the core surface, Amic(q) is the form factor amplitude of the radial scattering length distribution of the micelle, S(q) is the monodisperse hard-sphere structure factor, and D(Rc) is the Gaussian distribution for core radii. Four terms are considered for Pmic(q): the self-correlation of the core, the self-correlation of the corona chains, the cross term between the sphere and the corona chains, and the cross term between different corona chains.

Acorona (q) =

4π ∫ ρcorona (r )

sin(qr ) 2 r qr 2

4π ∫ ρcorona (r )r dr

dr

exp( −q2σint 2/2) (A5)

A linear combination of two partial cubic B spline functions was chosen for ρcorona(r), as explained elsewhere.34,45 The form factor amplitude of the radial scattering length distribution of the micelle (Amic(q)) is given by H


Article

Macromolecules A mic(q) = Nagg(βcoreAcore(q) + βcorona Acorona (q))

(5) Li, Z.; Kesselman, E.; Talmon, Y.; Hillmyer, M. A.; Lodge, T. P. Science 2004, 306, 98−101. (6) Li, Z.; Hillmyer, M. A.; Lodge, T. P. Nano Lett. 2006, 6, 1245− 1249. (7) Li, Z.; Hillmyer, M. A.; Lodge, T. P. Langmuir 2006, 22, 9409− 9417. (8) Pochan, D. J.; Chen, Z.; Cui, H.; Hales, K.; Qi, K.; Wooley, K. L. Science 2004, 306, 94−97. (9) Cui, H.; Chen, Z.; Zhong, S.; Wooley, K. L.; Pochan, D. J. Science 2007, 317, 647−650. (10) Kubowicz, S.; Baussard, J. F.; Lutz, J. F.; Thunemann, A. F.; von Berlepsch, H.; Laschewsky, A. Angew. Chem., Int. Ed. 2005, 44, 5262− 5265. (11) Von Berlepsch, H.; Bottcher, C.; Skrabania, K.; Laschewsky, A. Chem. Commun. 2009, 2290−2292. (12) Zhu, J.; Jiang, W. Macromolecules 2005, 38, 9315−9323. (13) Fang, B.; Walther, A.; Wolf, A.; Xu, Y.; Yuan, J.; Muller, A. H. E. Angew. Chem., Int. Ed. 2009, 48, 2877−2880. (14) Weberskirch, R.; Preuschen, J.; Spiess, H. W.; Nuyken, O. Macromol. Chem. Phys. 2000, 201, 995−1007. (15) Komenda, T.; Luedtke, K.; Jordan, R.; Ivanova, R.; Bonne, T. B.; Papadakis, C. M. Polym. Prepr. (Am. Chem. Soc., Div. Polym. Chem.) 2006, 47, 763−764. (16) Yamaguchi, D.; Cloitre, M.; Panine, P.; Leibler, L. Macromolecules 2005, 38, 7798−7806. (17) Li, C.; Buurma, N. J.; Haq, I.; Turner, C.; Armes, S. P.; Castelletto, V.; Hamley, I. W.; Lewis, A. L. Langmuir 2005, 21, 11026− 11033. (18) Nedelcheva, A. N.; Vladimirov, N. G.; Novakov, C. P.; Berlinova, I. V. J. Polym. Sci., Part A: Polym. Chem. 2004, 42, 5736− 5744. (19) Shen, W.; Zhang, K.; Kornfield, J. A.; Tirrell, D. A. Nat. Mater. 2006, 5, 153−158. (20) Taribagil, R. R.; Hillmyer, M. A.; Lodge, T. P. Macromolecules 2009, 42, 1796−1800. (21) Taribagil, R. R.; Hillmyer, M. A.; Lodge, T. P. Macromolecules 2010, 43, 5396−5404. (22) Zhou, C.; Hillmyer, M. A.; Lodge, T. P. J. Am. Chem. Soc. 2012, 134, 10365−10368. (23) Zhou, C.; Hillmyer, M. A.; Lodge, T. P. Macromolecules 2011, 44, 1635−1641. (24) Fetters, L. J.; Lohse, D. J.; Richter, D.; Witten, T. A.; Zirkel, A. Macromolecules 1994, 27, 4639−4647. (25) Smith, G. D.; Yoon, D. Y.; Jaffe, R. L.; Colby, R. H.; Krishnamoorti, R.; Fetters, L. J. Macromolecules 1996, 29, 3462−3469. (26) Shields, D. J.; Coover, H. W. J. Polym. Sci. 1959, 39, 532−533. (27) Jain, S.; Dyrdahl, M. H. E.; Gong, X.; Scriven, L. E.; Bates, F. S. Macromolecules 2008, 41, 3305−3316. (28) Lee, S.; Arunagirinathan, M. A.; Bates, F. S. Langmuir 2010, 26, 1707−1715. (29) NIST SLD calculator: http://www.ncnr.nist.gov/resources/ sldcalc.html. (30) Glinka, C. J.; Barker, J. G.; Hammouda, B.; Krueger, S.; Moyer, J. J.; Orts, W. J. J. Appl. Crystallogr. 1998, 31, 430−445. (31) Kline, S. R. J. Appl. Crystallogr. 2006, 39, 895−900. (32) Castelletto, V.; Hamley, I. W.; Pedersen, J. S. J. Chem. Phys. 2002, 117, 8124−8129. (33) Castelletto, V.; Hamley, I. W.; Pedersen, J. S. Langmuir 2004, 20, 2992−2994. (34) Pedersen, J. S.; Svaneborg, C.; Almdal, K.; Hamley, I. W.; Young, R. N. Macromolecules 2003, 36, 416−433. (35) Pedersen, J. S. J. Appl. Crystallogr. 1998, 31, 488−489. (36) Pedersen, J. S.; Hamley, I. W.; Ryu, C. Y.; Lodge, T. P. Macromolecules 2000, 33, 542−550. (37) Ashcroft, N. W.; Langreth, D. C. Phys. Rev. 1967, 156, 685−692. (38) Robertus, C.; Philipse, W. H.; Joosten, J. G. H.; Levine, Y. K. J. Chem. Phys. 1989, 90, 4482−4490. (39) Bergenholtz, J.; Romagnoli, A. A.; Wagner, N. J. Langmuir 1995, 11, 1559−1570.

(A6)

The structure factor (S(q)) is described by a hard-sphere interaction model that considers short-range repulsions between hard spheres of radius (Rhs). It is given by S(q) =

1 1 + 24ηG(2qR hs , η)/2qR hs

(A7)

where G is a trigonometric function of x = 2qRhs and the hardsphere volume fraction (η) G(x , η) = ax−2(sin x − x cos x) + βx−3[2x sin x + (2 − x 2) cos x − 2] + γx−5[−x 4 cos x + 4[2x 2 − 6] cos x + (x 3 − 6x) sin x + 6]]

(A8)

with α = (1 + 2η)2/(1 − η)4, β = −6η(1 + η/2)2/(1 − η)4, and γ = 1/2η(1 + 2η)2/(1 − η)4. The dispersity in micelle size was accounted for by a Gaussian distribution for core radii: D(R c) =

⎡ −(R − ⟨R ⟩)2 ⎤ 1 c c ⎥; exp⎢ 2 2πσR 2σR ⎦ ⎣

Rc > 0 (A9)

where ⟨Rc⟩ is the average radius and σR is the width of distribution truncated at Rc = 0. In all, eight parameters could be adjusted in the fits: the core radius (Rc), the radius of gyration of the corona chains (Rg), the width of the core−corona interface (σint), the width of the distribution for the core radius (σR), two terms (a1, s) in the cubic B spline function for the corona term, the hard-sphere radius (Rhs), and the hard-sphere volume fraction (ηhs).

■

ASSOCIATED CONTENT

* Supporting Information S

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.macromol.5b00584. SEC traces, contrast factor calculations, further cryoTEM images, further SANS traces, sensitivity analysis for SANS fitting parameters, and variable temperature 1H NMR spectra (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected] (T.P.L.). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported primarily by the National Science Foundation, through the University of Minnesota MRSEC (DMR-0819885 and DMR-1420013).

■

REFERENCES

(1) Moughton, A. O.; Hillmyer, M. A.; Lodge, T. P. Macromolecules 2012, 45, 2−19. (2) Laschewsky, A. Curr. Opin. Colloid Interface Sci. 2003, 8, 274− 281. (3) Lutz, J. F.; Laschewsky, A. Macromol. Chem. Phys. 2005, 206, 813−817. (4) Lodge, T. P.; Rasdal, A.; Li, Z.; Hillmyer, M. A. J. Am. Chem. Soc. 2005, 127, 17608−17609. I


Article

Macromolecules (40) Kline, S. R.; Kaler, E. W. J. Chem. Phys. 1996, 105, 3813−3822. (41) Teixeira, J. J. Appl. Crystallogr. 1988, 21, 781−785. (42) Beaucage, G. J. Appl. Crystallogr. 1995, 28, 717−728. (43) Convertine, A. J.; Lokitz, B. S.; Vasileva, Y.; Myrick, L. J.; Scales, C. W.; Lowe, A. B.; McCormick, C. L. Macromolecules 2006, 39, 1724−1730. (44) Magenau, A. J. D.; Martinez-Castro, N.; Savin, D. A.; Storey, R. F. Macromolecules 2009, 42, 8044−8051. (45) Castelletto, V.; Hamley, I. W.; Pedersen, J. S. J. Chem. Phys. 2002, 117, 8124−8129.

J


Structure of Two-Compartment Hydrogels from Thermoresponsive

Recommend Documents