Swelling properties of thermoresponsive/hydrophilic co-networks with functional crosslinked domain structures

Shohei Ida *, Hironobu Kitanaka , Tatsuya Ishikawa , Shokyoku Kanaoka and Yoshitsugu Hirokawa
Department of Materials Science, The University of Shiga Prefecture, Hikone 522-8533, Japan. E-mail: ida.s@mat.usp.ac.jp

Received 25th October 2017 , Accepted 2nd January 2018

First published on 3rd January 2018


For the fine control of the thermoresponsive properties of polymer hydrogels, we focused on the monomer and crosslinker sequence in a network. In this paper, we designed novel thermoresponsive/hydrophilic polymer co-networks with functional crosslinked domain (CD) structures such as a thermoresponsive network with hydrophilic CDs (DCDN gel) and a hydrophilic network with thermoresponsive CDs (NCDD gel). The gel synthesis was based on post-polymerization crosslinking of triblock prepolymers with reactive sites in the outer blocks. The obtained gels showed larger, sharper, and more rapid volume transitions in response to temperature change as compared to the corresponding gel with a random monomer/crosslinker sequence as well as a randomly crosslinked polymer co-network with the same composition. In addition, significant differences were observed in the transition temperature and shrinking kinetics between DCDN and NCDD gels with the same composition and a similar network structure, and the swelling behavior was also controlled by the composition of the triblock prepolymers. These results indicated that the domain structure in the network could effectively function and the monomer/crosslinker sequence exerts a strong effect on the swelling behavior of the gels.


Introduction

The thermoresponsive volume phase transition of polymer hydrogels, discovered by Tanaka et al.,1–3 has broadened the possibility for various applications such as actuators,4 microfluidics,5 drug delivery systems,6 biomedical devices,7 and catalysts,8,9 which would require precise control of the degree of volume change and/or properties in practical use. A key for achieving these advanced applications is the on-demand fine control of thermoresponsive properties including transition temperature, transition rate, and the amount of volume change. The main criteria for practical use have to be sharp and rapid transitions at a desired temperature with a targeted volume change. Such advanced control of swelling properties would be facilitated by the precise design of the network structure in a gel.

Efficient functionalization in a polymer gel is also essential for the construction of responsive advanced gel materials. A simple way to produce a gel with both stimuli-responsiveness and another characteristic function is by the use of more than two monomers besides a crosslinker in gel synthesis. In particular, a polymer co-network gel, which consists of two components but has homo-sequenced chains between crosslinking points, would exhibit properties derived from the two monomer sequences more distinctively. A typical example is amphiphilic co-networks (APCNs)10,11 consisting of two polymer chains with hydrophilic and hydrophobic characteristics. APCNs with independent hydrophilic and hydrophobic network chains show unique swelling properties both in water and in organic solvents.12–21 In a co-network structure, stimuli-responsive segments would have greater mobility, leading to a larger and faster volume change. For example, APCNs with thermoresponsive chains showed unique thermoresponsive behavior such as rapid response and mechanical toughening.22–33 It should be noted that such unique properties of APCNs result from the independent functions of two types of polymer chains in the network.

It would be preferable for a stimuli-responsive gel to undergo an isotropic volume change for a certain purpose. To this end, the precision design of segmented structures in the network and the control of the chain distribution including crosslinking points are crucial to maximize the dual effects from two polymer chains. However, the difficulty lies in preparing a co-network gel of two different types of polymers with random chain distribution since synthetic procedures usually involve random copolymerization of macromonomers/crosslinkers or post-polymerization crosslinking reaction of immiscible polymer chains.16–33 End-linking of block polymers via controlled polymerization is an interesting way to prepare a nearly homogeneous network and tune the sequence, but the distribution of crosslinking points along two polymers is hard to change.12–15 A new approach which allows the even distribution of two different components as well as crosslinking structures is desired; however, to the best of our knowledge the design of APCNs remains a matter of the combination of two polymer chains. The use of a triblock copolymer is a viable solution for realizing an even component distribution.

In a polymer gel, a crosslinking structure domain, an important component, can be a distinctive space with unique nature, compared to that of the corresponding linear polymers. A triblock copolymer with crosslinkable moieties in the outer segments is expected to be transformed into a gel with a network chain domain and a rather densely crosslinked domain (CD). A feature of this type of co-network is the possible compartmentalization of randomly distributed functional CDs and the continuous domain of highly mobile bridging chains. In this study, we examined the feasibility of synthesis of thermoresponsive APCNs with CDs, especially focusing on the spatial distribution of the crosslinking points in the APCN structure. This unique structure would give a clear and independent demonstration of the two characteristics of the APCN structure.

The designed APCNs with CD structures were synthesized by post-polymerization crosslinking of controlled ABA triblock prepolymers with reactive sites in the outer blocks as shown in Fig. 1. The crosslinking reaction takes place only at the outer blocks of a prepolymer, leading to the formation of functional CDs derived from the outer blocks and bridging polymer chains from the middle block. In this paper, we synthesized co-networks with functional CDs from triblock prepolymers consisting of thermoresponsive and hydrophilic blocks. Prepolymers were prepared by reversible addition–fragmentation chain transfer (RAFT) block polymerization34–36 of thermoresponsive N-isopropylacrylamide (NIPAAm) and hydrophilic N,N-dimethylacrylamide (DMAAm), and carried activated ester groups37,38 in the outer blocks as reactive sites for crosslinking. RAFT polymerization allows the facile control of the block sequence, the composition and the molecular weight of the prepolymers. We prepared the two kinds of prepolymers with inverse sequences as shown in Fig. 1 [we denote the P(DMAAm-b-NIPAAm-b-DMAAm) prepolymer as DND and the P(NIPAAm-b-DMAAm-b-NIPAAm) prepolymer as NDN]. DND yields a PNIPAAm network with hydrophilic DMAAm CDs (referred to as DCDN gel) and NDN produces a PDMAAm network with thermoresponsive NIPAAm CDs (referred to as NCDD gel). The thermoresponsive behavior of the obtained gels was evaluated to discuss the effects of the monomer/crosslinker sequence in the network on the swelling properties of co-networks. The effect of the composition was also examined for gels obtained from prepolymers with almost the same molecular weight.


image file: c7py01793f-f1.tif
Fig. 1 Schematic representation of APCNs with functional CDs prepared by post-polymerization crosslinking of controlled triblock prepolymers.

Experimental

Materials

DMAAm (Wako, 98%) was purified by distillation before use. NIPAAm (Wako, 98%) was purified by recrystallization from toluene/n-hexane. N-(Acryloyloxy)succinimide (NHSA)39 and a chain transfer agent for RAFT polymerization carrying two trithiocarbonate groups in the molecule (CTA-1)40 were prepared as described in the literature. Ethylenediamine (EDA; Wako, 98%), azobis(4-methoxy-2,4-dimethylvaleronitrile) (V-70; Wako, 95%), azobisisobutyronitrile (AIBN; Wako, 98%), N,N′-methylenebisacrylamide (BIS; Wako, 99%), 1,2,3,4-tetrahydronaphthalene (tetralin; Aldrich, 99%), 1,4-dioxane (Wako, 99%), and tetrahydrofuran (THF; Wako, 99.5%) were used as received.

Synthesis of triblock prepolymers

Triblock prepolymers were prepared by sequential RAFT block polymerization, followed by end-modification with a radical initiator. As a typical example, the synthesis of DND is given below. NIPAAm (3.40 g, 30.0 mmol), CTA-1 (122 mg, 0.300 mmol), AIBN (4.9 mg, 0.030 mmol), tetralin (1.5 mL), and 1,4-dioxane (13.5 mL) were added into a 50 mL round-bottomed flask equipped with a three-way stopcock, and then bubbled with nitrogen for 10 minutes. The flask was placed in an oil bath kept at 60 °C for 24 h. The reaction was terminated by cooling the reaction mixture to −60 °C. The monomer conversion was determined from the concentration of residual monomer measured by 1H NMR relative to the internal standard (tetralin). Then, the reaction mixture was poured into diethyl ether to obtain the purified PNIPAAm (2.97 g).

Then, the obtained PNIPAAm was employed as a macro-CTA for block polymerization. The PNIPAAm macro-CTA (DPn = 90 and Mn = 10[thin space (1/6-em)]600, which were determined by 1H NMR analysis; 2.11 g, 0.20 mmol), DMAAm (2.05 mL, 20.0 mmol), NHSA (338 mg, 2.00 mmol), AIBN (3.3 mg, 0.020 mmol), tetralin (1.0 mL), and 1,4-dioxane (7.0 mL) were added into a 50 mL round-bottomed flask equipped with a three-way stopcock, and bubbled with nitrogen for 10 minutes. The flask was placed in an oil bath kept at 60 °C for 48 h. The reaction was terminated by cooling the reaction mixture to −60 °C. Monomer conversion was determined from the concentration of residual monomer measured by 1H NMR. Then, the reaction mixture was poured into diethyl ether to obtain the purified triblock polymer, P[(DMAAm/NHSA)-b-NIPAAm-b-(DMAAm/NHSA)] (3.77 g). The monomer composition of the polymer was determined by 1H NMR analysis for NIPAAm and the calculation from monomer conversion for DMAAm and NHSA because of the overlapping of the signals in the 1H NMR spectrum.

Finally, the trithiocarbonate end-groups were eliminated by a radical coupling reaction with V-70. The triblock polymer (DPn, NIPAAm = 90, DPn, DMAAm = 90, DPn, NHSA = 10, Mn = 21[thin space (1/6-em)]200; 1.60 g, 0.075 mmol), V-70 (1.2 g, 3.0 mmol), and 1,4-dioxane (15.0 mL) were added into a 50 mL round-bottomed flask equipped with a three-way stopcock. The solution was stirred at 40 °C under nitrogen for 16 h. Then, the solution was concentrated by evaporation and poured into diethyl ether to obtain the purified DND (1.45 g).

Gel synthesis by post-polymerization crosslinking

As a typical example, the synthesis of DCDN gel is given below. DND (DPn, NIPAAm = 90, DPn, DMAAm = 90, DPn, NHSA = 10; Mn = 21[thin space (1/6-em)]200; 644 mg, 0.030 mmol) was dissolved in 1.5 mL of THF. 0.50 mL of a THF solution of EDA (containing 0.15 mmol of EDA; amino groups were set equimolar to NHSA units in the prepolymers) was added to this solution, and glass capillaries (internal diameter: 1300 μm; volume: 40 μL) were added to the reaction vessel. The mixture was kept at room temperature for 24 h to complete the crosslinking reaction. Then, the cylindrical gels were taken out from the capillaries and washed once with DMF for 24 h and several times with distilled water for 24 h by means of immersion to remove the residual molecules and replace the solvent in the gel.

Gel synthesis by copolymerization

NIPAAm (1.11 g, 9.80 mmol), DMAAm (1.00 mL, 9.80 mmol), and BIS (83.6 mg, 0.540 mmol) were all dissolved in methanol (5.0 mL). Then, the solution of AIBN in methanol (63.9 mg of AIBN in 2.0 mL of methanol) was added to the monomer solution. The obtained solution was transferred into a test tube containing glass capillaries (internal diameter: 1300 μm; volume: 40 μL) and then bubbled with nitrogen. The test tube was immersed in a water bath controlled at 55 °C and kept for 24 h for the completion of gelation. Under these conditions, almost all of the monomers were considered to be consumed and the composition of the gel was equal to the monomer feed ratio. Afterwards, the cylindrical gels were taken out of the capillaries and washed once with DMF for 24 h and several times with distilled water for 24 h by means of immersion to remove the residual molecules and replace the solvent in the gel.

Characterization

M n and Mw/Mn of polymers were determined by size-exclusion chromatography (SEC) in DMF containing 10 mM LiBr as an eluent at 40 °C using three polystyrene gel columns (PLgel 5 μm MIXED-C, PLgel 3 μm MIXED-E and Shodex KF-805L) that were connected to a Shimadzu LC-10AD precision pump and a Shimadzu RID-10A refractive index detector. The columns were calibrated against standard poly(methyl methacrylate) samples (Agilent). 1H NMR spectra were recorded on a JEOL JNM-LA400 spectrometer operating at 399.65 MHz. The swelling degree of gels was determined by measuring the diameter of the cylindrical gels. The gels were immersed in water at a predetermined temperature, and the equilibrium diameter at a given temperature, d, was measured using a digital microscope (MOTICAM2000, Shimadzu). The swelling degree was calculated as (d/d0)3; d0 is the internal diameter of the glass capillary (1300 μm), which could be regarded as the diameter of the as-prepared gel. The shrinking kinetics was evaluated using a temperature-jump experiment. The cylindrical gels were immersed in water at 5 °C until the gels reached the equilibrium swelling state. Then, the gels were quickly transferred into hot water at 70 °C, and the time dependence of the gel diameter was observed. A uniaxial tensile test was conducted with a Shimadzu EZ-SX using the rectangular specimens with the dimensions of ca. 1 × 10 × 20 mm. The cross-head speed was 5.0 mm min−1.

Results and discussion

Gel synthesis

Two types of triblock prepolymers with opposite sequences, DND and NDN, were synthesized by RAFT block polymerization using a bifunctional chain transfer agent (CTA) with two trithiocarbonate groups and subsequent end-modification by radical coupling reaction. Scheme 1 shows the synthetic route of DND as an example. First, RAFT polymerization of NIPAAm with a CTA (CTA-1) was conducted in 1,4-dioxane at 60 °C. In 24 h, monomer conversion reached 91%, and the obtained PNIPAAm had a narrow molecular weight distribution (Mw/Mn = 1.17) as shown in the SEC curve (Fig. 2a). The resulting polymer possessed ca. 90 monomeric units (DPn, NIPAAm = 90, Mn, NMR = 10[thin space (1/6-em)]600), as determined from the 1H NMR peak intensity ratio (Fig. S1 in the ESI). The observed DPn was very close to that calculated from the feed concentration and monomer conversion. These results indicated that a well-controlled synthesis of PNIPAAm was achieved.
image file: c7py01793f-s1.tif
Scheme 1 Synthesis of a triblock prepolymer, DND, by RAFT block polymerization and subsequent end-modification by radical coupling reaction with V-70.

image file: c7py01793f-f2.tif
Fig. 2 SEC curves of (a) DND and (b) NDN prepared by RAFT block polymerization followed by end-modification by radical coupling reaction. Reaction conditions: (1st polymerization) [Monomer]0 = 2000 mM, [CTA-1]0 = 20 mM, [AIBN]0 = 2.0 mM in 1,4-dioxane at 60 °C for 24 h; (2nd polymerization) [Monomer] = 2000 mM, [NHSA] = 200 mM, [macro-CTA] = 20 mM, [AIBN] = 2.0 mM in 1,4-dioxane at 60 °C for 48 h; (end modification) [polymer] = 5.0 mM, [V-70] = 200 mM in 1,4-dioxane at 40 °C for 16 h.

The resulting PNIPAAm was subsequently employed as a macro-CTA for RAFT random copolymerization of DMAAm and NHSA, a monomer with the activated ester group, in 1,4-dioxane at 60 °C. In 48 h, the conversion of DMAAm and NHSA reached 90% and 100%, respectively. After the second polymerization, the trithiocarbonate groups at the polymer-ends were eliminated by a radical coupling reaction with V-70 to avoid any side reaction during the crosslinking reaction with EDA since a trithiocarbonate group is susceptible to aminolysis.41 The SEC curve of the obtained triblock polymer (DND) clearly shifted toward the higher molecular weight region relative to that of macro-CTA (PNIPAAm) while keeping a narrow molecular weight distribution as shown in Fig. 2a. Based on 1H NMR analysis (Fig. S2 in the ESI) and calculations from the monomer conversion, the monomer composition in the DND polymer was determined to be 90 NIPAAm, 90 DMAAm, and 10 NHSA units in the chain (Table 1).

Table 1 Prepolymers for gel synthesis by post-polymerization crosslinking
Code DPn[thin space (1/6-em)]a M n[thin space (1/6-em)]a M w/Mn[thin space (1/6-em)]b
NIPAAm DMAAm NHSA
a Determined by 1H NMR analysis and calculations from conversion and feed concentration. b Determined by SEC analysis. c Synthesized by RAFT random copolymerization and subsequent end-modification. Polymerization conditions: [NIPAAm or DMAAm] = 2000 mM, [NHSA] = 100 mM, [CTA-1] = 10 mM, [AIBN] = 1.0 mM in 1,4-dioxane at 60 °C for 24 h. End-modification: [polymer] = 5.0 mM, [V-70] = 200 mM in 1,4-dioxane at 40 °C for 16 h. d Synthesized by RAFT block polymerization and subsequent end-modification. 1st polymerization: [NIPAAm] = 2800 mM, [CTA-1] = 20 mM, [AIBN] = 6.0 mM in 1,4-dioxane at 60 °C for 24 h. 2nd polymerization: [DMAAm] = 1200 mM, [NHSA] = 200 mM, [PNIPAAm macro-CTA] = 20 mM, [AIBN] = 6.0 mM in 1,4-dioxane at 60 °C for 24 h. End-modification: [polymer] = 5.0 mM, [V-70] = 200 mM in 1,4-dioxane at 40 °C for 16 h.
DND 90 90 10 21[thin space (1/6-em)]200 1.36
NDN 90 91 10 21[thin space (1/6-em)]400 1.31
N RP 183 10 22[thin space (1/6-em)]800 1.20
D RP 168 10 18[thin space (1/6-em)]700 1.27
DND(7[thin space (1/6-em)]:[thin space (1/6-em)]3) 131 55 10 22[thin space (1/6-em)]400 1.36


The prepolymer with the opposite sequence, NDN, was synthesized in a similar way by RAFT polymerization of NIPAAm and NHSA with PDMAAm macro-CTA, followed by end-modification with V-70. The reactions were also controlled, yielding a polymer with a relatively narrow molecular weight distribution (Mw/Mn = 1.31, Fig. 2b) and 90 NIPAAm, 91 DMAAm, and 10 NHSA units per chain (Table 1). Thus, DND and NDN with almost the same composition but the inverse block sequence were successfully synthesized.

These triblock prepolymers were employed for gel synthesis by post-polymerization crosslinking with EDA as a crosslinker. The total concentrations of NIPAAm and DMAAm units were set at 2800 mM for both prepolymers, and the feed ratio of amino groups in EDA was set equivalent to NHSA units in the prepolymers (ca. 150 mM for reactive sites: hence, the crosslinker concentration is about 75 mM). Injection of EDA into the prepolymer solution in THF quickly induced gelation for all prepolymers in a few minutes. The reaction mixtures were kept for 24 h for the completion of the reaction. After the reaction, the solvent of each gel was replaced by pure water by means of immersion for the swelling behavior analysis. Importantly, the gel obtained from DND, which contains crosslinkable NHSA units only in the outer hydrophilic PDMAAm blocks, has hydrophilic CDs (DCDN), whereas NDN affords the gel with thermoresponsive CDs (NCDD) (Fig. 1).

For comparison, we prepared two types of gels with almost the same composition but different monomer and crosslinker sequences with DCDN and NCDD gels. One type had PNIPAAm/PDMAAm co-network structures with randomly dispersed crosslinking points (NDRC in Fig. 3). This was synthesized by post-polymerization crosslinking of the P(NIPAAm/NHSA) random copolymer (NRP in Table 1) and the P(DMAAm/NHSA) random copolymer (DRP in Table 1), which were separately prepared by RAFT random copolymerization with CTA-1. The other type was a copolymerization gel (CPG in Fig. 3) with a random monomer/crosslinker sequence, which was prepared by free radical copolymerization of NIPAAm and DMAAm in the presence of a divinyl crosslinker, BIS. It should be noted that NDRC and CPG were synthesized under almost the same concentration conditions as those used in the synthesis of DCDN and NCDD ([NIPAAm unit] + [DMAAm unit] = 2800 mM, [crosslinker] = 75–77 mM).


image file: c7py01793f-f3.tif
Fig. 3 Schematic representation of NDRC (NIPAAm/DMAAm amphiphilic structure with random crosslinking) and CPG (copolymerization gel with a random monomer and crosslinker sequence).

Temperature dependence of the swelling degree of the gels with CDs

Fig. 4 shows the temperature dependence of the swelling degree of the obtained gels in pure water. Importantly, all the examined gels were prepared in the same monomer and crosslinker concentrations ([NIPAAm unit] + [DMAAm unit] = 2800 mM, [crosslinker] = 75–77 mM). Although the examined gels had almost the same monomer and crosslinker compositions, the swelling curves of the gels were totally different.
image file: c7py01793f-f4.tif
Fig. 4 Temperature dependence of the swelling degree of DCDN, NCDD, NDRC and CPG gels in pure water.

The gels with CD structures, DCDN and NCDD gels, apparently swelled more than the randomly crosslinked gels, NDRC and CPG gels. The swelling degree at low temperature (the swollen state) reflects well the crosslinking structure of the gels. The higher swelling degree of DCDN and NCDD gels indicated the apparently low crosslinking density of these gels. This fact also suggests the presence of the localized crosslinking points in these CD gels, which means an inhomogeneous distribution of crosslinking points in the network. In DCDN and NCDD gel synthesis, the crosslinkable units in the prepolymer were designed to be located only in the outer block, which was supposed to lead to the formation of relatively condensed crosslinking points, i.e. CDs. By contrast, a rather statistical distribution of the crosslinking points should be observed in the NDRC and CPG gels, prepared by crosslinking of the prepolymers with randomly distributed NHSA units or random copolymerization of the monomers and crosslinker. The non-designed and random crosslinking reaction system produced a gel with a decreased swelling degree because the apparent crosslinking density in the whole network became higher. The difference in the apparent crosslinking density was also confirmed by the value of the elastic modulus. The modulus of DCDN was smaller than that of CPG in the as-prepared state (DCDN: 53.5 kPa, CPG: 192 kPa, Fig. S3 in the ESI). This result indicated that the apparent crosslinking density of CD gels was lower than that of CPG because the modulus corresponds to the crosslinking density of the gel. Thus, the higher swelling degree and the lower modulus of DCDN and NCDD gels certainly indicated the presence of CDs in the network.

Fig. 4 also shows that gels with amphiphilic structures (DCDN, NCDD, and NDRC gels) showed large volume change at around 30–50 °C, while CPG gradually shrunk with the increase of temperature. This large volume change indicated the effective function of thermoresponsive PNIPAAm structures in the APCN. In particular, the gels with CD structures (DCDN and NCDD) gave sharper volume change in smaller temperature change compared to the randomly crosslinked gel (NDRC). In addition, this large volume change was reversible. When the shrunken gels were cooled, the gels swelled to the same extent as the first equilibrium swelling degree and shrank again upon re-heating. These results suggested that the PNIPAAm domains in DCDN (as bridging chains) and NCDD (as CDs) could function more independently of hydrophilic DMAAm domains than those of NDRC. We consider that this is due to the effect of the more segregated CD structure. DCDN and NCDD gels were prepared by the crosslinking of the outer blocks of the triblock prepolymers, and had a more concentrated domain structure of one part of the components compared with NDRC, in which PDMAAm and PNIPAAm were randomly crosslinked. It gave a relatively segregated structure and each polymer chain could function more independently and effectively. On the other hand, CPG possessed a random sequence of NIPAAm and DMAAm monomeric units. Randomly distributed DMAAm units made each network chain more hydrophilic than PNIPAAm homopolymer chains, preventing dehydration and hydrophobic interaction of NIPAAm units for shrinking. This deteriorated the sharpness of thermoresponsive volume change in CPG gels.

Surprisingly, the difference lies in the shrinking temperature among the gels having CD structures: NCDD shrunk at higher temperature than DCDN did. This indicates the effect of the sequence in the network on the swelling properties as these gels were prepared from triblock prepolymers with the same composition but inverse sequence, DND and NDN. The shrinking temperature was governed by the hydration state of the thermoresponsive chains in the network. NCDD had thermoresponsive CDs connected to many hydrophilic chains. Such a structure is likely to make the thermoresponsive CDs more hydrophilic. In addition, uncrosslinked hydrophilic bridging chains caused the higher swelling degree of the gel. On the other hand, in the DCDN gel, the thermoresponsive chains bridged the hydrophilic CDs, from which it can be regarded that each thermoresponsive chain has hydrophilic end groups. Therefore, these thermoresponsive chains in the DCDN gel were subject to a lower hydrophilic effect from the hydrophilic domains compared to those in the NCDD gel, since hydrophilic CDs can retain limited water due to the crosslinked structure, and are located only at the ends of thermoresponsive chains. Such differences in the hydrophilic environment of the thermoresponsive domains resulted in the higher shrinking temperature of NCDD.

Rapid shrinking of gels with CDs

Then, we evaluated the shrinking rate of the gels with CD structures and CPG by applying a temperature jump from 5 to 70 °C to the gels. This temperature range was wide enough for all the examined gels to undergo the transition from the swelling state to the shrunken state. Fig. 5 shows the changes in the gel appearance during shrinking after the temperature jump. The gels with CD structures (DCDN and NCDD) smoothly shrunk, maintaining their cylindrical shape (Fig. 5a and b). Such isotropic shrinking indicated that dehydrated water effectively went out of the interior of the gels. On the other hand, CPG deformed and bended upon temperature change (Fig. 5c), indicating that a shrinking layer (a so-called “skin layer”) was formed on the gel surface because of inefficient water discharge even with a high hydrophilic composition.
image file: c7py01793f-f5.tif
Fig. 5 Photographs of DCDN, NCDD, and CPG upon a temperature jump from 5 °C to 70 °C in pure water.

The time dependence of the swelling degree of the gels is shown in Fig. 6. Herein, the swelling degree was calculated as (d/dinit)3, where dinit is the equilibrium diameter at 5 °C of each gel. This normalized swelling degree was used to directly compare the shrinking behavior. It clearly showed that DCDN and NCDD shrunk much faster than CPG, which showed stepwise change of the swelling degree due to the deformation mentioned above. It should be noted that the swelling and shrinking behavior of the gel is significantly affected by the size of the gels, and the volume change of larger gels is slower than that of the smaller one.42 As shown in Fig. 4, DCDN and NCDD swelled much more than CPG did at 5 °C. Taking this into account, the shrinking of DCDN and NCDD was quite rapid relative to CPG.


image file: c7py01793f-f6.tif
Fig. 6 Shrinking behavior of DCDN, NCDD, and CPG upon a temperature jump from 5 °C to 70 °C in pure water. In this figure, the swelling degree was defined as (d/dinit)3, where dinit is the equilibrium diameter at 5 °C of each gel.

The rapid shrinking of DCDN and NCDD is likely attributed to the function of hydrophilic domains in the network derived from the hydrophilic blocks of the prepolymer. These gels possessed about 50% composition of hydrophilic DMAAm units in the network, and the hydrophilic domains are considered to form continuous domains. Such hydrophilic domains could behave as a water pathway during the shrinking, allowing for effective water draining. On the other hand, in CPG, almost no hydrophilic domain was formed because of the random distribution of hydrophilic monomers even with the same monomer composition. Thus, only the surface region of CPG shrunk before dehydrated water went out of the interior of the gel. Such a locally shrunken area at the surface suppressed further shrinking and led to deformation by the internal pressure of water.

Then, we performed a further analysis of the shrinking kinetics of DCDN and NCDD in order to understand the effect of the network sequence. For the swelling/shrinking of cylinder-shaped gels, the time dependence of diameter change can be described as

image file: c7py01793f-t1.tif
where dfin and τ, respectively, stand for the equilibrium diameter at the shrunken state and the relaxation time constant when t is larger than τ and no phase separation occurs.42–44 We applied this equation to the result of Fig. 6 and determined τ for DCDN and NCDD: τ = 5.9 × 102 s for DCDN and τ = 1.0 × 103 s for NCDD (Fig. S4 in the ESI). For cylindrical gels, the relaxation time (τ) is related to the cooperative diffusion coefficient (D):
image file: c7py01793f-t2.tif

From this relationship, we estimated D: D = 8.0 × 10−7 cm2 s−1 for DCDN and 5.2 × 10−7 cm2 s−1 for NCDD. These values indicated that DCDN shrunk faster than NCDD did. We suppose that this was due to the difference in mobility of thermoresponsive chains. Namely, thermoresponsive chains were crosslinked in NCDD, and hence the shrinking of thermoresponsive chains was restrained in NCDD, compared to non-crosslinked thermoresponsive chains in DCDN. Moreover, the shrinking behavior of DCDN and NCDD was quite different from that of CPG. The initial decrease of swelling degree was very large until the shrinking was in the exponential region (the linear region in Fig. S4) in the case of DCDN and NCDD, while CPG started the exponential decrease of swelling degree just after the plateau region. Such differences might be also derived from the domain structure with different mobility and functionality. That is, the aggregation of polymer chains in one domain and the aggregation of domains are considered to produce different modes of shrinking, and it may induce the unique shrinking behavior of domain gels. Thus, it was demonstrated that the difference in crosslinking structure greatly affected the shrinking kinetics.

Effect of the prepolymer composition on swelling behavior

Finally, we examined the effect of the prepolymer composition on the swelling behavior of the gels with hydrophilic CDs. The synthesis of a prepolymer was performed in a similar way to DND except for the monomer feed ratio. The obtained polymer had longer PNIPAAm blocks and shorter PDMAAm blocks but the total number of monomer units and the block sequence were the same as those of DND. We denoted this polymer as DND(7[thin space (1/6-em)]:[thin space (1/6-em)]3) because the composition of NIPAAm[thin space (1/6-em)]:[thin space (1/6-em)]DMAAm was about 7[thin space (1/6-em)]:[thin space (1/6-em)]3 (Table 1). This polymer was also employed for gel synthesis and the concentration conditions were similar to those of DCDN ([NIPAAm unit] + [DMAAm unit] = 2800 mM, [amino groups in EDA] = [NHSA unit] = 151 mM in THF at room temperature). Because DND and DND(7[thin space (1/6-em)]:[thin space (1/6-em)]3) had the same number of NHSA units per chain (Table 1), the obtained gel, DCDN(7[thin space (1/6-em)]:[thin space (1/6-em)]3), was expected to have the same concentration of crosslinking points in the as-prepared state. In addition, DCDN(7[thin space (1/6-em)]:[thin space (1/6-em)]3) was likely to have smaller hydrophilic CDs and longer thermoresponsive bridging chains relative to DCDN (Fig. 7).
image file: c7py01793f-f7.tif
Fig. 7 Temperature dependence of the swelling degree of DCDN(7[thin space (1/6-em)]:[thin space (1/6-em)]3) and DCDN in pure water.

D CD N(7[thin space (1/6-em)]:[thin space (1/6-em)]3) swelled much more than DCDN did at a low temperature in pure water (Fig. 7). Taking into account that both gels had almost the same concentration of crosslinking points in the as-prepared state, the large swelling degree of DCDN(7[thin space (1/6-em)]:[thin space (1/6-em)]3) is attributed to the longer middle blocks, leading to the larger distance between CDs. The results also indicated that the swelling degree of the gels can be controlled by the molecular weight of the middle block of the triblock prepolymer. Furthermore, DCDN(7[thin space (1/6-em)]:[thin space (1/6-em)]3) showed a larger and sharper volume change against temperature change than DCDN. This is due to the effect of the higher composition of NIPAAm units, and thermoresponsive PNIPAAm domains could function much more effectively than the gel with a lower NIPAAm composition (DCDN). The shrinking temperature of DCDN(7[thin space (1/6-em)]:[thin space (1/6-em)]3) shifted to the lower temperature region from that of DCDN. Thus, the swelling behavior of the gels having CD structures such as swelling degree and shrinking temperature can be easily controlled by the composition of the triblock prepolymers.

Conclusions

In this study, we successfully synthesized a novel type of thermoresponsive/hydrophilic amphiphilic co-network (APCN) with crosslinking domains (CDs), which exclusively contained crosslinking points, via RAFT polymerization and post-polymerization crosslinking reactions. The swelling behavior of the obtained gels with the designed structures differed from those of an APCN with randomly dispersed crosslinking points and a gel obtained from random copolymerization of two monomers and a crosslinker. The designed gels (DCDN and NCDD) showed a larger and sharper volume transition against temperature change in pure water in contrast to a totally random network (CPG) as well as randomly crosslinked APCN (NDRC) with the same monomer/crosslinker composition. In addition, the DCDN and NCDD gels were shown to shrink isotropically and much faster than CPG against the temperature jump. These results indicated that the CD structure effectively formed and functioned in the designed APCNs. Importantly, the swelling behavior was greatly affected by the network sequence: the DCDN gel showed a lower shrinking temperature and more rapid shrinking than the NCDD gel did. Moreover, the composition of the triblock prepolymer also exerted a powerful effect on the swelling behavior of the gels. Thus, the structure design of the prepolymer such as the sequence, molecular weight, and composition is of major significance for the control of swelling behavior. The actual distribution of CDs is unclear and perhaps the distribution or the size of CDs might be inhomogeneous; however, such a design of APCN with CD structures by using controlled triblock polymers certainly produces the unique swelling properties described here. We believe that a more precise design of network structures based on precision synthesis of prepolymers will contribute to the fine control of the swelling behavior of polymer gels, which is imperative in the pursuit of a more sophisticated design of thermoresponsive smart materials.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was partially supported by the Japan Society for the Promotion of Science through a Grant-in-Aid for Young Scientists (B) (No. 16K17962), for which the authors are grateful.

References

  1. T. Tanaka, Phys. Rev. Lett., 1978, 40, 820–823 CrossRef CAS.
  2. T. Tanaka, D. Fillmore, S.-T. Sun, I. Nishio, G. Swislow and A. Shah, Phys. Rev. Lett., 1980, 45, 1636–1639 CrossRef CAS.
  3. Y. Hirokawa and T. Tanaka, J. Chem. Phys., 1984, 81, 6379–6380 CrossRef.
  4. Y. Osada and J. Gong, Prog. Polym. Sci., 1993, 18, 187–226 CrossRef CAS.
  5. M. E. Harmon, M. Tang and C. W. Frank, Polymer, 2003, 44, 4547–4556 CrossRef CAS.
  6. A. Kikuchi and T. Okano, Adv. Drug Delivery Rev., 2002, 54, 53–77 CrossRef CAS PubMed.
  7. S. Chaterji, I. K. Kwon and K. Park, Prog. Polym. Sci., 2007, 32, 1083–1122 CrossRef CAS PubMed.
  8. S. Ikegami and H. Hamamoto, Chem. Rev., 2009, 109, 583–593 CrossRef CAS PubMed.
  9. D. Diaz Diaz, D. Kuhbeck and R. J. Koopmans, Chem. Soc. Rev., 2011, 40, 427–448 RSC.
  10. C. S. Patrickios and T. K. Georgiou, Curr. Opin. Colloid Interface Sci., 2003, 8, 76–85 CrossRef CAS.
  11. G. Erdodi and J. P. Kennedy, Prog. Polym. Sci., 2006, 31, 1–18 CrossRef CAS.
  12. M. R. Simmons, E. N. Yamasaki and C. S. Patrickios, Macromolecules, 2000, 33, 3176–3179 CrossRef CAS.
  13. A. I. Triftaridou, S. C. Hadjiyannakou, M. Vamvakaki and C. S. Patrickios, Macromolecules, 2002, 35, 2506–2513 CrossRef CAS.
  14. M. Rikkou-Kalourkoti, E. Loizou, L. Porcar, K. Matyjaszewski and C. S. Patrickios, Polym. Chem., 2012, 3, 105–116 RSC.
  15. E. N. Kitiri, C. S. Patrickios, C. Voutouri, T. Stylianopoulos, I. Hoffmann, R. Schweins and M. Gradzielski, Polym. Chem., 2017, 8, 245–259 RSC.
  16. G. Erdodi and J. P. Kennedy, J. Polym. Sci., Part A: Polym. Chem., 2005, 43, 4953–4964 CrossRef CAS.
  17. G. Erdodi and J. P. Kennedy, J. Polym. Sci., Part A: Polym. Chem., 2005, 43, 4965–4971 CrossRef CAS.
  18. C. Fodor, G. Kali and B. Iván, Macromolecules, 2011, 44, 4496–4502 CrossRef CAS.
  19. G. Kali, S. Vavra, K. László and B. Iván, Macromolecules, 2013, 46, 5337–5344 CrossRef CAS.
  20. G. Kali and B. Iván, Macromol. Chem. Phys., 2015, 216, 605–613 CrossRef CAS.
  21. C. Fodor, T. Stumphauser, R. Thomann, Y. Thomann and B. Ivan, Polym. Chem., 2016, 7, 5375–5385 RSC.
  22. H. Katono, A. Maruyama, K. Sanui, N. Ogata, T. Okano and Y. Sakurai, J. Controlled Release, 1991, 16, 215–227 CrossRef CAS.
  23. Y. Kaneko, S. Nakamura, K. Sakai, T. Aoyagi, A. Kikuchi, Y. Sakurai and T. Okano, Macromolecules, 1998, 31, 6099–6105 CrossRef CAS.
  24. J. Tobis, Y. Thomann and J. C. Tiller, Polymer, 2010, 51, 35–45 CrossRef CAS.
  25. J. Tobis, L. Boch, Y. Thomann and J. C. Tiller, J. Membr. Sci., 2011, 372, 219–227 CrossRef CAS.
  26. Q. Zheng and S. Zheng, J. Polym. Sci., Part A: Polym. Chem., 2012, 50, 1717–1727 CrossRef CAS.
  27. H. Cong, L. Li and S. Zheng, Polymer, 2013, 54, 1370–1380 CrossRef CAS.
  28. H. Kamata, U. Chung, M. Shibayama and T. Sakai, Soft Matter, 2012, 8, 6876–6879 RSC.
  29. H. Kamata, U.-I. Chung and T. Sakai, Macromolecules, 2013, 46, 4114–4119 CrossRef CAS.
  30. H. Kamata, Y. Akagi, Y. Kayasuga-Kariya, U.-I. Chung and T. Sakai, Science, 2014, 343, 873–875 CrossRef CAS PubMed.
  31. S. Nakagawa, X. Li, H. Kamata, T. Sakai, E. P. Gilbert and M. Shibayama, Macromolecules, 2017, 50, 3388–3395 CrossRef CAS.
  32. H. Guo, N. Sanson, D. Hourdet and A. Marcellan, Adv. Mater., 2016, 28, 5857–5864 CrossRef CAS PubMed.
  33. H. Guo, C. Mussault, A. Brûlet, A. Marcellan, D. Hourdet and N. Sanson, Macromolecules, 2016, 49, 4295–4306 CrossRef CAS.
  34. G. Moad, E. Rizzardo and S. H. Thang, Aust. J. Chem., 2005, 58, 379–410 CrossRef CAS.
  35. G. Moad, E. Rizzardo and S. H. Thang, Aust. J. Chem., 2009, 62, 1402–1472 CrossRef CAS.
  36. A. Gregory and M. H. Stenzel, Prog. Polym. Sci., 2012, 37, 38–105 CrossRef CAS.
  37. P. Theato, J. Polym. Sci., Part A: Polym. Chem., 2008, 46, 6677–6687 CrossRef CAS.
  38. A. Das and P. Theato, Chem. Rev., 2016, 116, 1434–1495 CrossRef CAS PubMed.
  39. A. Pollak, H. Blumenfeld, M. Wax, R. L. Baughn and G. M. Whitesides, J. Am. Chem. Soc., 1980, 102, 6324–6336 CrossRef CAS.
  40. A. M. Bivigou-Koumba, J. Kristen, A. Laschewsky, P. Müller-Buschbaum and C. M. Papadakis, Macromol. Chem. Phys., 2009, 210, 565–578 CrossRef CAS.
  41. H. Willcock and R. K. O'Reilly, Polym. Chem., 2010, 1, 149–157 RSC.
  42. T. Tanaka and D. J. Fillmore, J. Chem. Phys., 1979, 70, 1214–1218 CrossRef CAS.
  43. Y. Li and T. Tanaka, J. Chem. Phys., 1990, 92, 1365–1371 CrossRef CAS.
  44. H. Hirose and M. Shibayama, Macromolecules, 1998, 31, 5336–5342 CrossRef CAS.

Footnote

Electronic supplementary information (ESI) available: 1H NMR spectra of the prepolymers, stress–strain curves of the gels, and kinetic analysis of gel shrinking. See DOI: 10.1039/c7py01793f

This journal is © The Royal Society of Chemistry 2018