Structure–property correlation of hydrogels obtained via radical polymerization using the central cores of multiarm star polymers as crosslinkers

Abstract

To improve the mechanical properties of a hydrogel, the construction of uniform network structures and/or the incorporation of energy-dissipating structures is important. In this study, we focused on gel synthesis using multiarm star polymers with a microgel core, which is expected to establish the abovementioned structures in vinyl polymer hydrogels. A series of star poly(N-isopropylacrylamide)s (PNIPAAms) with different arm molecular weights and vinyl group contents in the core were synthesized via an arm-first method using reversible addition–fragmentation chain transfer (RAFT) polymerization. The obtained star polymers were employed as crosslinkers to prepare polyacrylamide (PAAm) hydrogels by free radical polymerization. The content of vinyl groups in the core was critical for producing a hydrogel, and significantly affected the mechanical properties of the produced gels, which is indicative of the high effectiveness of the star polymer core as a crosslinker. The molecular weight of the arm chains of the star polymers also played a pivotal role in controlling the mechanical properties of the produced gels: moderately long arm chains, which form hydrogen bonding, were shown to act as energy-dissipating units. An equally important feature is the nearly even dispersion of the star crosslinkers in the network structure, as confirmed by SAXS, which achieved an increase in toughness without impairing the elongation upon increasing the main chain monomer concentration in the gelation reaction.

---

Introduction

Precise structural design of hydrogels is essential for the development of new materials with more sophisticated functionality, approaching natural soft tissues for various applications, such as soft robotics and biomedical devices.^1–6 Reinforcement of gels by network design has also gained much attention, and various strategies have succeeded in producing hydrogels at the level of practical use.^7–9 In developing novel soft materials, a better understanding of the correlation between network structure and properties, as well as elaborate functionality design, is becoming crucial.

High-strength gel materials are generally synthesized in two ways: the construction of uniform network structures, and the incorporation of energy-dissipating structures. A typical method for producing a gel with a uniform network is a linking reaction of the terminal groups of well-defined precursor polymers.^10–17 Another effective way to construct a uniform network is crosslinking reaction with an inorganic material or a polymer particle at the nanoscale that not only acts as a multifunctional crosslinker but can also be homogeneously dispersed in water.^18–23 For the incorporation of energy dissipating structures, prominent cases at present are the combination of two different networks (so-called double networks or interpenetrating networks) and/or the utilization of non-covalent bonding, such as polymer entanglement, hydrogen bonding and hydrophobic interaction.^7,24–33 Without careful nanoscale design of interaction sites with appropriate strength, however, unintended non-covalent interactions would induce significant aggregation in water, which can result in a brittle material. Among various interacting moieties, dangling chains in an amphiphilic co-network structure^34–36 are particularly useful because the degree of interaction can be systematically controlled by varying the length and density of the dangling chains. For instance, the incorporation of stimuli-responsive dangling chains has been shown to be effective for the functionalization of hydrogels, such as improvement of the response speed of the entire gel network³⁷ and thermoresponsive mechanical toughening.^38,39 Moreover, gels with highly dense dangling chains, commonly known as “bottle-brush polymer networks”, have attracted much attention due to their unique mechanical and swelling properties.^40–44 Thus, even distribution of crosslinking points and the incorporation of energy-dissipating dangling chains into the network structure are essential for attaining highly functional gel materials with sufficient mechanical properties.

Based on this background, we focused on core-crosslinked multiarm star polymers as a key component in novel gel materials. Core-crosslinked star polymers consist of a crosslinked microgel core and a number of linear polymers connected to the core, exhibiting unique properties significantly different from those of the linear counterparts due to their highly dense structure.^45–47 For example, we have recently found that a multiarm star poly(N-isopropylacrylamide) (PNIPAAm) neither exhibits a distinct cloud point in water nor forms significant aggregates above the cloud point of linear PNIPAAm, which becomes insoluble in water at a high temperature, forming large aggregates.^48–50 This behavior is likely attributed to the high water dispersibility of the star PNIPAAm, which results from a highly dense structure and the repulsion between the nanospheres of shrunken star polymers at a high temperature. A core-crosslinked star polymer is generally synthesized by an arm-first method via living polymerization, in which living polymers are reacted with a small amount of a divinyl crosslinker.^51–53 This process often produces a core containing unreacted vinyl groups derived from the crosslinker. Due to its high dispersibility and unreacted vinyl groups, a core-crosslinked multiarm star polymer can serve as a homogeneously dispersed multifunctional crosslinker for gel synthesis. Terashima et al. demonstrated that a star polymer with poly(ethylene glycol) arm chains and a fluorinated microgel core can be used as a crosslinker in the synthesis of a gel for water purification.²³ We have also recently reported the synthesis of a polyacrylamide (PAAm) gel with densely packed dangling chains around the crosslinking points using star PNIPAAm as a crosslinker.⁵⁴ The produced gels exhibited higher compressive strength than a conventional gel obtained with a typical low-molecular-weight crosslinker, N,N′-methylenebisacrylamide (BIS). In addition, thermoresponsive mechanical toughening was observed in air in the absence of external water. Thus, a multiarm star polymer is a promising crosslinker for functionalizing hydrogels due to its high dispersibility and polymerization reactivity, leading to homogeneously distributed multifunctional crosslinking points. A variety of structural features, such as the monomer structures of the arm chains and the central core, as well as the content of vinyl groups in the core, can be controlled during the synthesis of the star polymers, which leads to the tuning of the gel properties. Furthermore, the arm chains of the star crosslinker integrate into the gel network as dangling chains, enhancing responsive properties due to their high mobility, also improving mechanical properties by forming non-covalent interactions that dissipate deformation energy. A more detailed study on structural diversification of star-crosslinked gels and a systematic analysis to understand the correlation between the network structure and the mechanical properties would lead to the development of novel advanced gel materials.

This study aims to elucidate the structure–property correlation of multiarm star-crosslinked hydrogels. We systematically synthesized a series of multiarm star PNIPAAms with various sizes (molecular weight of the arm chains) and vinyl group contents in the core via reversible addition–fragmentation chain transfer (RAFT) polymerization (Fig. 1). The obtained star PNIPAAms were employed for PAAm gel synthesis under different feed concentrations of the monomer and crosslinker, and the internal structures and mechanical properties of the product gels were evaluated by various measurements.

Fig. 1 Synthesis of multiarm PNIPAAm star-crosslinked hydrogels with various structures.

Experimental

Materials

NIPAAm (FUJIFILM Wako Chemical, >98.0%) was purified by recrystallization from toluene/hexane. Acrylamide (AAm; FUJIFILM Wako Chemical, for electrophoresis, >99.0%) was purified by recrystallization from acetone. Distilled water was obtained using an EYELA STILL ACE SA-2100A. The following compounds, reagents, and solvents were used as received: 4-cyano-4-[(dodecylsulfanylthiocarbonyl)sulfanyl]pentanoic acid (FUJIFILM Wako Chemical, 97%), 2,2′-azobisisobutyronitrile (AIBN; FUJIFILM Wako Chemical, 98%), BIS (FUJIFILM Wako Chemical, for electrophoresis, >99.0%), ammonium persulfate (APS; Fujifilm Wako Chemical, for electrophoresis, >99.0%), N,N,N′,N′-tetramethylethylenediamine (TMEDA; Fujifilm Wako Chemical, for electrophoresis, >99.0%), 1,2,3,4-tetrahydronaphthalene (tetralin; Sigma-Aldrich, 99%), 1,4-dioxane (Fujifilm Wako Chemical, for organic synthesis, >99.5%), methanol (Fujifilm Wako Chemical, for organic synthesis, >99.8%), N,N-dimethylformamide (DMF; FUJIFILM Wako Pure Chemical, 99.5%), diethyl ether (FUJIFILM Wako Pure Chemical, 99.5%), CDCl₃ (Cambridge Isotope Laboratories, 99.5%), and CD₃OD (Cambridge Isotope Laboratories, 99.8%).

Measurement and characterization

The number-average molecular weight (M_n), weight-average molecular weight (M_w), and polydispersity index (M_w/M_n) of the produced polymers were determined by size-exclusion chromatography (SEC; Shimadzu Prominence system consisting of an LC-20AD precision pump and an RID-20A refractive index detector) in DMF containing 10 mM LiBr as the eluent at 40 °C (flow rate: 1.0 mL min⁻¹) using three polystyrene gel columns (Shodex KF-805L). The columns were calibrated against standard poly(methyl methacrylate) samples (Agilent, M_n = 5.89 × 10²–1.68 × 10⁶).

¹H nuclear magnetic resonance (NMR) spectra were recorded on a JEOL JNM-ECS400 spectrometer operating at 399.90 MHz. The degree of polymerization (DP_n) and the M_{n, NMR} were calculated from the integral values of the peaks derived from the monomer units and the chain transfer agent (CTA) for RAFT polymerization.

The absolute M_w of the star polymers was determined using multi-angle laser light scattering (MALS) equipped with SEC on a Dawn HELEOS II instrument (Wyatt Technology, semiconductor laser, λ = 663 nm). The SEC was performed in DMF containing LiBr (10 mM) at 40 °C using three polystyrene gel columns (Shodex KF-805L) that were connected to a Jasco PU-2080 precision pump, a Jasco RI-1530 refractive index detector, and a Jasco UV-1570 UV/vis detector set at 270 nm. The average number of arm chains (N_arms) was calculated from the molecular weights of the star polymer and the linear precursor: N_arms = (weight fraction of the arm polymers according to the feed ratio and the conversion of BIS) × M_{w, star}/M_{w, arm}; M_{w, star} was determined by SEC-MALS, and M_{w, arm} was calculated from M_{n, NMR} and M_w/M_{n, SEC} of the linear precursor.

The overlap concentration of the star polymers in water was estimated from the intrinsic viscosity obtained by viscosity measurement using an A&D SV-1A viscometer.⁵⁵

Small-angle X-ray scattering (SAXS) experiments were conducted using synchrotron radiation at beamline BL-6A of the Photon Factory at the Institute of Materials Structure Science of the High Energy Accelerator Research Organization in Tsukuba, Japan. Two-dimensional scattering images were collected on a Dectris PILATUS 1 M detector. One-dimensional SAXS profiles were obtained by radial averaging of the two-dimensional images. The scattering angle was calibrated by using silver behenate with a periodical structure of 5.838 nm. The scattering vector was defined as q = (4π/λ)sin(θ/2), where θ and λ are the scattering angle and the wavelength of the incident X-rays, respectively.

The uniaxial tensile test at room temperature was conducted with a Shimadzu EZ-SX (load cell: 20 N) using rectangular specimens with dimensions of ca. 2 × 10 × 20 mm. The cross-head speed was 5.0 mm min⁻¹.

Dynamic viscoelasticity measurement was conducted with a TA Instruments Discovery HR-2 with roughened parallel-plate geometry using columnar specimens (diameter: 8 mm, height: 1 mm). The samples were prepared in a silicone mold and coated with paraffin oil. Temperature sweep measurement was conducted from 20–60 °C at a heating rate of 1.0 °C min⁻¹ (strain: 10%, frequency: 1 Hz), controlled by a Peltier plate.

Gel synthesis

Gel samples were synthesized according to our previous report.⁵⁴ In a typical method, a mixture of NIPAAm (11.3 g, 0.100 mol), 4-cyano-4-[(dodecylsulfanylthiocarbonyl)sulfanyl]-pentanoic acid as a CTA for RAFT polymerization (0.268 g, 0.664 mmol), AIBN (10.8 mg, 0.066 mmol), tetralin (as the internal standard for the calculations of monomer conversion by ¹H NMR analysis; 2.0 mL) and 1,4-dioxane (48.0 mL) was stirred under nitrogen at 60 °C for 24 h. The reaction was terminated by cooling the reaction mixture to −60 °C, after which the reaction mixture was poured into diethyl ether to obtain purified PNIPAAm (9.94 g; DP_n = 159 and M_n = 18 400, both of which were calculated by ¹H NMR analysis). Then, the obtained PNIPAAm (3.67 g, 0.199 mmol) as a macro-CTA, BIS (0.616 g, 4.00 mmol), AIBN (1.64 mg, 0.0100 mmol), tetralin (1.00 mL) and methanol (9.00 mL) were mixed and stirred under nitrogen at 60 °C for 24 h. The purified star PNIPAAm (3.14 g) was obtained by reprecipitation from diethyl ether. The obtained star polymer (0.500 g) and AAm (2.00 g, 28.0 mmol) were dissolved into distilled water (15.0 mL). After nitrogen bubbling for 10 min, 2.0 mL of aqueous solution of APS (containing 0.088 mmol of APS) and 0.50 mL of aqueous solution of TMEDA (containing 0.140 mmol of TMEDA) were added, and the reaction mixture was kept at 4 °C in a refrigerator for 24 h to reach the gelation state.

Results and discussion

Synthesis of star crosslinkers with various structures

Star PNIPAAm crosslinkers with various structures were synthesized by an arm-first method using RAFT polymerization as reported in our previous papers.^48,54 First, linear PNIPAAms with different molecular weights were synthesized by varying the feed concentration of NIPAAm to 4-cyano-4-[(dodecylsulfanylthiocarbonyl)sulfanyl]pentanoic acid as a CTA for RAFT polymerization (Fig. 1 and Table 1; the samples were denoted as LP_x, where LP and the subscript number, x, stand for “Linear Polymer” and the feed concentration ratio of NIPAAm and CTA, respectively). The molecular weight distributions of the obtained polymers were narrow in all cases (dashed line in Fig. 2). The DP_n of the polymers determined by ¹H NMR measurements ranged from 52 to 418, which were in good agreement with the feed concentration ratio (Table 1 and Fig. S1 in the ESI†). Thus, seven types of linear PNIPAAm with different molecular weights were successfully obtained.

Fig. 2 SEC curves of star PNIPAAms having arm polymers with various molecular weights and their precursor linear polymers obtained by RAFT polymerization (solid line: star polymers, and dashed line: linear polymers). Reaction conditions: see Table 1.

Table 1 Synthesis of linear PNIPAAms with different molecular weights^a

Code^b	[NIPAAm]/[CTA]	Time (h)	Conv. (%)	DPn, NMR^c	M n, NMR	M w/Mn, SEC^d
a Synthetic conditions: [NIPAAm] = 2000 mM, [CTA]/[AIBN] = 10 in 1,4-dioxane at 60 °C. b “LP” and the subscript number stand for “Linear Polymer” and the feed concentration ratio of NIPAAm and CTA ([NIPAAm]/[CTA]), respectively. c Calculated by ^1H NMR analysis. d Determined by SEC measurement.
LP50	50	24	88	52	6300	1.10
LP100	100	24	88	110	12 900	1.14
LP150	150	23	86	159	18 400	1.16
LP200	200	23	85	204	23 500	1.20
LP250	250	24	80	248	28 500	1.23
LP300	300	48	85	316	36 100	1.27
LP400	400	48	84	418	47 600	1.38

---

The produced linear PNIPAAms with different molecular weights were employed as macro-CTAs for the synthesis of star polymers, being treated with BIS as a crosslinker under the same conditions as previously reported⁵⁴ (Table 2; the samples were denoted as SP_x, where SP and the subscript number, x, stand for “Star Polymer” and the feed concentration ratio of NIPAAm and CTA in the synthesis of the linear precursors, respectively.). The concentration ratio of BIS to PNIPAAm macro-CTA was set at [BIS] : [macro-CTA] = 20 : 1 in this study to introduce as many unreacted vinyl groups in the star core as possible while preventing macroscopic gelation based on our preliminary study. In all cases, the main peaks in the SEC profiles shifted significantly towards the higher molecular weight region compared to that of the PNIPAAm macro-CTA (Fig. 2). The peak attributable to unreacted macro-CTA was very small in most cases, while the presence of unreacted macro-CTAs (ca. 30%) was observed in SP₂₅₀ and SP₃₀₀ due to low initiation and/or crosslinking efficiency. Notably, unreacted macro-CTAs were likely incorporated into a network as dangling chains during gel synthesis, and were considered to play a role similar to the arm chains of the star crosslinkers. Therefore, we employed these polymers in gel synthesis without further purification. The SEC profiles also indicated the formation of multiarm star polymers with relatively narrow molecular weight distribution, although slightly broader molecular weight distributions were observed when macro-CTAs with a high molecular weight were used. This broadening is probably due to radical coupling reactions between the star polymers. The ¹H NMR spectra of the produced polymers showed the absorption derived from unreacted vinyl groups of the crosslinker, and the integral value ratio of the peaks indicated that ca. 10% of the vinyl groups of the feed amount of BIS remained in the core (Fig. S2 in the ESI†). Furthermore, the absolute M_w of the star PNIPAAms was determined from the SEC-MALS measurement to be significantly larger than that obtained from SEC analysis with PMMA calibration, indicative of the formation of a multiarm structure. The average number of arm chains (N_arm) was calculated from these M_{w, MALS} values and the molecular weight of the PNIPAAm macro-CTAs (Table 2). The obtained N_arm values were around 30 in most cases within a fairly narrow range, irrespective of the molecular weight of the arm polymers.

Table 2 Synthesis of star PNIPAAms with various structures^a

Code^b	Macro-CTA	Time (h)	Conv.BIS^c (%)	Conv.Star^d (%)	M w, MALS	M w/Mn, SEC^f	R C=C (%)	N arm	c*^i (g L^−1)
a Synthetic conditions: [macro-CTA] = 10 (for SP250, SP300 and SP400) or 20 (for the others) mM, [BIS]/[macro-CTA] = 20, [AIBN] = 1.0 (for SP150 and SP300), 3.5 (for SP100_v6 and SP100_v1), 5.0 (for SP400 and SP100_v0.3) or 2.0 (for the others) mM in methanol at 60 °C. b “SP” and the subscript number stand for “Star Polymer” and the feed concentration ratio of NIPAAm and CTA in the synthesis of the linear precursors, respectively. The number after “v” stands for the ratio of the unreacted vinyl groups in the core. c Reaction conversion of BIS determined using ^1H NMR analysis of the reaction solutions. d Conversion of PNIPAAm macro-CTA into star polymers calculated from the peak area ratio in the SEC curves on the assumption that the refractive indexes of the star polymers and macro-CTAs were consistent. e Determined using SEC-MALS analysis. f Determined by SEC measurement. g The ratio of the unreacted vinyl groups in the core estimated by ^1H NMR analysis. h The average number of arm chains per star polymer molecule. i The overlap concentration in water estimated from the intrinsic viscosity obtained using viscosity measurement.
SP50	LP50	24	86	97	463 000	1.38	9.8	47	44
SP100	LP100	48	89	88	560 000	1.32	9.3	32	34
SP150	LP150	44	82	84	902 500	1.40	9.7	35	32
SP200	LP200	24	85	93	772 300	1.36	10	25	23
SP250	LP250	47	74	72	778 700	1.36	8.3	20	18
SP300	LP300	96	83	67	1 288 000	1.69	10	26	12
SP400	LP400	27	83	91	2 170 000	1.56	10	31	10
SP100_v6	LP100	19	90	87	660 000	1.32	6.1	37	—
SP100_v1	LP100	20	100	89	870 000	1.29	1.4	48	—
SP100_v0.3	LP100	24	97	89	770 000	1.36	0.3	42	—

---

The amount of residual vinyl groups in the core of the star PNIPAAms was roughly controlled by the feed concentration of AIBN and the reaction conversion of BIS during the synthesis of the star polymers. In the synthesis of star PNIPAAms using LP₁₀₀ as a macro-CTA, star polymers with similar molecular weights were obtained regardless of the feed concentration of AIBN (Fig. S3 in the ESI†). NMR measurements of the obtained star polymers showed that the amount of vinyl groups in the core decreased as the feed concentration of AIBN increased: for example, 10% (SP₁₀₀) and 0.3% (SP_{100_v0.3}) of the vinyl groups remained in the core at 2.0 mM and 5.0 mM of AIBN, respectively (Fig. S4 in the ESI†). The reaction with an AIBN feed concentration of 3.5 mM was performed twice to vary the reaction conversion, and star polymers with different amounts of unreacted vinyl groups, 6.1% and 1.4%, were obtained. Moreover, decreasing the amount of vinyl groups in the core resulted in a higher number of branches, corresponding to the progression of the crosslinking reaction (Table 2). Thus, star polymers with different amounts of vinyl groups in the core were successfully prepared depending on the reaction conditions.

Effect of vinyl content in the core on the gel properties

The effect of the amount of vinyl groups in the core of a star crosslinker on gelation behavior and gel properties was investigated. A series of star polymers with the same DP_n of the arm chains (DP_n = 110) but different amounts of vinyl groups in the core (9.3–0.3%) were employed as a crosslinker for radical random copolymerization with AAm in water. Based on our previous study,⁵⁴ the reactions were conducted at 10% of AAm and 2.5 wt% of star PNIPAAm crosslinker (Table 3). Transparent gels were obtained from the reaction with star PNIPAAms having vinyl groups in the core (entries 1–3), irrespective of the vinyl content, whereas no gelation occurred with a star polymer containing almost no vinyl groups, SP_{100_v0.3} (entry 4). These results indicated that the vinyl groups in the core of a star crosslinker are critical for inducing the gelation reaction, in other words, supporting that the core-crosslinked star polymer served as an effective crosslinker.

Table 3 Gel synthesis using star PNIPAAms with different amounts of vinyl groups in the core^a

Entry	Star crosslinker	R C=C (%)	Result^c	Gel code^d
a Synthetic conditions: AAm 10 wt%, star PNIPAAm 2.5 wt%, [APS] = 5.0 mM, [TMEDA] = 8.0 mM at 4 °C in distilled water. b The ratio of the unreacted vinyl groups in the core estimated by ^1H NMR analysis. c Determined by tilting the reaction vessels. d “SG” and the subscript number stand for “Star Gel” and the feed concentration ratio of NIPAAm and CTA in the synthesis of the linear precursors, respectively. The number after “v” stands for the ratio of the unreacted vinyl groups in the core.
1	SP100	9.3	Gelation	SG100
2	SP100_v6	6.1	Gelation	SG100_v6
3	SP100_v1	1.4	Gelation	SG100_v1
4	SP100_v0.3	0.3	No gelation	—

---

Sufficient crosslinking with the vinyl groups in the core is also supported by the mechanical properties of the products. The obtained gels from star crosslinkers containing more vinyl groups in the core exhibited higher Young's modulus and lower elongation in the uniaxial tensile test (Fig. 3 and Table S1 in the ESI†). This is probably because the increased number of vinyl groups in the core enabled more network chains to connect to a star polymer crosslinker molecule, which resulted in shorter network chains. Thus, the mechanical properties of the star-crosslinked gels can be altered by controlling the amount of residual vinyl groups in the core of the star crosslinker.

Fig. 3 Representative stress–strain curves obtained by uniaxial tensile tests of star-crosslinked gels prepared with star PNIPAAms with different amounts of vinyl groups in the core. Preparation conditions: see Table 3.

Effect of molecular weight of the arm chains on gel structure and properties

To examine the effect of the length of the free-end arm chains of the star crosslinker on the structure and properties, gels were synthesized by radical polymerization of AAm using a series of star PNIPAAm crosslinkers (SP₅₀–SP₄₀₀) with a similar amount of vinyl groups in the core but different arm molecular weights (Table 4). The feed concentration of the star crosslinker was set at 2.5 wt% for all the star polymers. Furthermore, gel synthesis was also conducted at feed concentrations of star crosslinkers proportional to the arm molecular weight, namely, at approximately equal molar concentrations of star polymers in a reaction mixture for a gel (e.g., 2.5 wt% for SP₁₀₀ and 5.0 wt% for SP₂₀₀). The reaction with SP₄₀₀ was performed only at 2.5 wt% of SP₄₀₀ due to its low solubility in water. Transparent gels were obtained under all the conditions examined (Fig. 4). The resulting gels were referred to as SG_{x_y}, where x and y stand for the feed concentration ratio of NIPAAm and CTA in the synthesis of the linear precursors and feed concentration (wt%) of the star polymer in the gel synthesis, respectively.

Fig. 4 Appearances of gels synthesized with a series of star PNIPAAms with different arm molecular weights. Preparation conditions: see Table 4.

Table 4 Gel synthesis using star PNIPAAms with different arm molecular weights^a

Gel code^b	Star crosslinker	Feed concentration of star crosslinker, c		c/c*
		(wt%)	(g L^−1)
a Synthetic conditions: AAm 10 wt%, [APS] = 5.0 mM, [TMEDA] = 8.0 mM at 4 °C in distilled water. b In the gel code “SGx_y”, SG, x and y stand for “Star Gel”, the feed concentration ratio of NIPAAm and CTA in the synthesis of the linear precursors, and the feed concentration (wt%) of star crosslinkers in the gel synthesis, respectively. c This sample is identical to SG100 in Table 3.
SG50_1.25	SP50	1.25	14.1	0.32
SG50_2.5	SP50	2.50	28.6	0.65
SG100_2.5	SP100	2.50	28.6	0.84
SG150_2.5	SP150	2.50	28.6	0.89
SG150_3.75	SP150	3.75	43.5	1.4
SG200_2.5	SP200	2.50	28.6	1.2
SG200_5.0	SP200	5.00	58.8	2.6
SG250_2.5	SP250	2.50	28.6	1.6
SG250_6.25	SP250	6.25	74.6	4.1
SG300_2.5	SP300	2.50	28.6	2.4
SG300_7.5	SP300	7.50	90.9	7.6
SG400_2.5	SP400	2.50	28.6	2.9

---

To evaluate the internal structure of the produced gels, SAXS measurements were performed with the gels obtained from the reaction with 2.5 wt% of star crosslinkers with differing arm lengths (Fig. 5a). The SAXS profiles of the obtained gels invariably exhibited intensity maxima, except for SG_{50_2.5}. The intensity maximum indicated the presence of internal particle structures with high electron density at a nearly regular interval, ascribed to the cores of the star crosslinkers. In particular, SG_{100_2.5} and SG_{200_2.5}, which exhibited distinct peaks, are likely to have better-ordered structures: a crosslinked structure with more evenly dispersed star polymers. As the star crosslinkers for the synthesis differed in arm molecular weight, but were of the same weight fraction, a larger number of star molecules were contained in the gels prepared from a star crosslinker with shorter chains. In SG_{50_2.5}, thus, too high molar concentration of the star polymers in the network likely caused increased intermolecular aggregation of the star molecules during the gel synthesis, as indicated by the absence of a clear intensity maximum in the SAXS profile.

Fig. 5 SAXS profiles of gels synthesized with a series of star PNIPAAms with different arm molecular weights: (a) under the same weight fraction (2.5 wt%) of star crosslinkers, and (b) under a similar molar concentration of star crosslinkers. Preparation conditions: see Table 4.

The average distances, d_star, between the star crosslinkers were estimated from the q value (q_max) of the maximum peak to be approximately 20 nm in all cases (SG_{100_2.5}–SG_{400_2.5}), regardless of the arm molecular weight (Table S2 in the ESI†). At 2.5 wt% (= 28.6 g L⁻¹) feed concentration, near or above the overlap concentration for SP₁₀₀–SP₄₀₀, the star polymers are forced to be in a shrunken or interpenetrating state. In particular, these shrinking and interpenetration effects were likely more pronounced in gels with longer arm chains, and the shrunken state increases the repulsive interaction between the star molecules. Conversely, star PNIPAAm potentially forms small aggregates from several molecules even below the response temperature,⁵⁰ and these small aggregates remain well dispersed in water without further significant aggregation, serving as crosslinking points. The balance between these repulsion and aggregation effects may result in similar d_star values, irrespective of the arm molecular weight, even under similar weight fraction of star crosslinkers. Furthermore, the d_star values of the gels (SG_{200_5.0} and SG_{300_7.5}) containing more star polymer molecules were larger than those of the gels with 2.5 wt% of the star crosslinkers (Fig. 5b and Table S2 in the ESI†). These results suggest that aggregate formation was more likely to occur to produce a less-ordered structure, which is responsible for an increase in d_star, when the concentration of a star polymer was increased above the overlap concentration. More disordered structure was observed in SG₃₀₀ and SG₄₀₀ with large arm molecular weights, which showed relatively high scattering intensity in the low q region and unclear maxima peaks. The formation of an inhomogeneous network structure is probably due to star–star aggregation. Thus, the star PNIPAAm with arms of moderate molecular weight served effectively as a crosslinker to produce a gel network with the star crosslinkers homogeneously dispersed.

The mechanical properties of the star-crosslinked gels with different arm molecular weights were evaluated by conducting uniaxial tensile tests. With the gels containing 2.5 wt% of the star polymers, the Young's moduli of the product gels declined with increasing arm molecular weight while the breaking elongations increased (Fig. 6a and b, and Table S3 in the ESI†). As the employed star polymers as crosslinkers had similar N_arm values, a star crosslinker was larger in molecular weight with longer arm chains, and hence the number of star molecules present in the gel network was smaller. The less amount of the star crosslinker also led to a decrease in the concentration of PAAm chains connected to the star polymers, which was responsible for a decrease in Young's modulus. SG_{400_2.5}, which exhibited low elongation, likely consisted of an incomplete network structure due to the significantly small number of star crosslinkers. Notably, the SG_{150_2.5} showed a slightly larger Young's modulus than the overall trend. The star polymers in SG_{150_2.5} were present at a concentration close to the overlap concentration, and therefore the moderate interaction of the arm chains between the star polymers likely served as effective physical crosslinking points.

Fig. 6 (a and c) Representative stress–strain curves and (b and d) the effect of the arm molecular weight on Young's modulus (red symbols) and breaking elongation (black symbols) of star-crosslinked gels synthesized with a series of star PNIPAAms with different arm molecular weights: (a and b) under the same weight fraction (2.5 wt%) of star crosslinkers, and (c and d) under a similar molar concentration of star crosslinkers. Preparation conditions: see Table 4.

A similar trend is also seen in the results of uniaxial tensile tests of the star-crosslinked gels synthesized at different concentrations. The product gels synthesized at similar molar concentrations of the star crosslinkers showed similar Young's moduli, except for SG₁₅₀, which had the highest Young's modulus (Fig. 6c and d, and Table S4 in the ESI†). In addition, SG₁₅₀ underwent the maximal elongation. The mechanical properties of the gels were likely influenced by the dispersion state of the star crosslinkers and the intermolecular interactions between the star crosslinkers. Star polymers with large arm molecular weight likely aggregated significantly, causing the formation of a heterogeneous network structure, as demonstrated by SAXS analysis, leading to poor mechanical properties. The disarray in the structure was probably due to the relatively higher concentration of the star polymers in the gel, well above the overlap concentration. Conversely, moderately long arm chains in SG_{150_3.75} not only suppressed the aggregation of star molecules but also facilitated appropriate intermolecular interaction based on arm chain interpenetration. This well-balanced interaction was key to dissipating the deformation energy, which augmented elongation as well as the Young's modulus.

The temperature dependence of the mechanical properties of the star-crosslinked gels was investigated using dynamic viscoelasticity measurement. The storage modulus (G′) of SG_{100_2.5} slightly changed when heated, whereas that of SG_{200_2.5} with a star polymer of longer arm chains increased significantly (Fig. 7a and b). The increase in G′ was more noticeable at a higher concentration of the star crosslinker (SG_{200_5.0}; Fig. 7c). This tendency in G′ resulted from thermoresponsive aggregation of the PNIPAAm chains at a high temperature, which became more apparent with longer arm chains and a higher concentration above the overlap concentration. Furthermore, the loss modulus (G′′) of SG_{200_5.0} decreased and then increased at around the response temperature. A similar phenomenon was observed with PNIPAAm gel under isochore conditions.⁵⁶ This characteristic behavior was probably due to the strong aggregation, which induced the immobilization of the chains to reduce the viscosity and finally to lead to phase-separation within the network. Thus, the star PNIPAAm-crosslinked hydrogels can alter their mechanical properties in response to the thermoresponsive behavior of the multiple free-end PNIPAAm arm chains. The mechanical properties were also variable depending on the arm molecular weight and star polymer concentration.

Fig. 7 Temperature dependence of G′ and G′′ of star-crosslinked hydrogels: (a) SG_{100_2.5}, (b) SG_{200_2.5}, and (c) SG_{200_5.0}. Heating rate: 1.0 °C min⁻¹.

Effect of monomer concentrations on gel structure and properties

The star-crosslinked hydrogels in this study are comprised of main chain polymers (PAAm) connected to the star polymer crosslinkers, which are homogeneously dispersed in a network. To examine the effect of main chain polymers on the structure and properties of the star-crosslinked hydrogels, gelation reaction was conducted at various AAm concentrations (the obtained gels are denoted as SG_Mz: z stands for the feed concentration of AAm; 20–40 wt%). Here, the star polymer with an arm of DP_n = 163 was used at a feed concentration of 2.5 wt% for gel synthesis (Fig. S5 in the ESI†). The SAXS profiles of SG_M20 and SG_M30 showed clear maxima, the intensity of which decreased as the concentration of AAm increased. Conversely, the SAXS profile of SG_M40 exhibited only a shoulder peak without any clear maximum in the same region (Fig. 8). Increasing monomer concentration led to a decrease in the maximal intensity, probably because the reduced amount of water declined the dispersibility of the star crosslinkers and the entanglement effect of PAAm chains on the structural inhomogeneity became more pronounced under high monomer concentration conditions. The q_max values of SG_M20 and SG_M30 were almost the same: the average distance between the star crosslinker was independent of the monomer concentration (d_star = 15 nm). The constant d_star values suggest that only the concentration, or number, of PAAm chains connected to the star crosslinkers increases even at a higher AAm concentration.

Fig. 8 SAXS profiles of star-crosslinked hydrogels with various AAm concentrations (SG_M20, SG_M30, and SG_M40). Preparation conditions: AAm 20–40 wt%, [APS] = 5.0 mM, [TMEDA] = 8.0 mM at 4 °C in distilled water.

The effect of monomer concentration on the mechanical properties of the star-crosslinked hydrogels was then evaluated by uniaxial tensile tests. The obtained gels were getting higher in Young's modulus and breaking stress with increasing AAm concentration in the reaction, whereas increased monomer concentration had little effect on the breaking strain (Fig. 9 and Table S5 in the ESI†). As the average distance between the star polymers remained almost constant regardless of the monomer concentration, the toughening of the gel is mostly attributed to a larger number of interconnecting PAAm chains between the star crosslinkers at a higher feed monomer concentration. The more PAAm chains would also lead to the increased entanglement of the PAAm network chains, which contributes to the increase in Young's modulus.

Fig. 9 Representative stress–strain curves of star-crosslinked hydrogels with various AAm concentrations (SG_M20, SG_M30, and SG_M40) obtained by uniaxial tensile tests.

Some of the star-crosslinked gels in this study showed elastomeric properties on a simple repeated tensile test by hand. Thus, cyclic tensile tests of the product gels were performed at varying strain levels (Fig. 10). An apparent hysteresis was observed between the loading and unloading curves for all the gels, and the hysteresis increased at higher strain levels and higher AAm concentrations (Fig. S6 in the ESI†). Furthermore, the stress–strain curves of the loading processes for each cycle overlapped in the low-strain region. These results indicate that the covalently-bonded network structure of the PAAm chains remained intact during the loading-unloading cycle at the strain levels examined in these experiments (<300%), and that the hysteresis was caused by the energy dissipation derived from the fracture of non-covalent interactions in the network. The presence of non-covalent interactions contributes to an increase in stress, and the energy dissipation prevents the stress concentration in the network, resulting in an improved elongation. Such an increase in stress and elongation led to the toughening of the materials. The presence of the hysteresis in addition to the decisive effect of the arm length of the star crosslinker on the mechanical properties (Fig. 6) indicates that the hydrogen bonds between the PNIPAAm arm chains contribute greatly to the energy dissipation. Furthermore, the interaction between the PNIPAAm arm chains and PAAm network chains may also be involved, as clearer hysteresis was observed at higher AAm concentrations.

Fig. 10 Cyclic tensile tests of star-crosslinked hydrogels at various AAm concentrations: (a) SG_M20, (b) SG_M30, and (c) SG_M40 (blue: 1st cycle with 50% strain, red: 2nd cycle with 100% strain, purple: 3rd cycle with 200% strain, and orange: 4th cycle with 300% strain; the holding time between each cycle was 60 seconds). The inset in panel (a) is an enlarged view of the stress–strain curves.

Conclusions

In this study, a variety of star-crosslinked PAAm hydrogels were synthesized using the core of multiarm star PNIPAAms as a crosslinker, and the effects of the structural factors of the building moieties on the whole structure and properties of the product gels were investigated. The content of the vinyl groups in the core, which can be altered by the reaction conditions during star polymer synthesis, was critical for producing a hydrogel, significantly affecting the mechanical properties of the product gels, which demonstrates that the core of the star polymer served as a crosslinker. The molecular weight of the arm chains of the star polymer also played a critical role in controlling the mechanical properties of the produced gels. Moderately long arm chains and an appropriate concentration of star molecules near the overlap concentration allowed suitable hydrogen bonding between the arm chains of the star polymers to serve as energy-dissipating moieties, resulting in high mechanical toughness. More importantly, SAXS analysis verified that star polymers homogenously dispersed in the network structure. Such a homogeneous structure led to an increase in toughness without diminishing elongation by increasing the main chain monomer concentration. Cyclic tensile tests revealed that the star-crosslinked hydrogel had a distinct hysteresis in the stress–strain curve, probably due to non-covalent interactions in the network, i.e. hydrogen bonding between the arm chains of the star crosslinkers. Thus, a multiarm star polymer with a microgel core was demonstrated to be an effective crosslinker, providing a hydrogel material with key features for good mechanical properties, such as evenly dispersed crosslinking points and an energy-dissipating structure. Moreover, a wide range of component monomers for the star polymers would permit structural diversity, which greatly expands the scope of soft material design.

Author contributions

Conceptualization: S. I. and S. K.; Formal analysis: S. I., S. S. and S. T.; Funding acquisition: S. I.; Investigation: S. S., S. T., H. T., M. O. and K. N.; Supervision: S. K.; Visualization: S. I.; Writing – original draft: S. I.; Writing – review & editing: S. K.

Data availability

The data supporting this article have been included as part of the ESI.†

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 Scientific Research (JP19K05602 and JP24K08529), by The Kyoto Technoscience Center, and by The Foundation of Institute for Chemical Fibers, Japan. The SAXS experiments at Photon Factory were performed under the approval of the Photon Factory Program Advisory Committee (Proposal No. 2023G567). The authors deeply thank Prof. Takaya Terashima and Mr Hiroyuki Kono (Kyoto University) for their assistance with SEC-MALS measurement.

References

1. H. Fan and J. P. Gong, Macromolecules, 2020, 53, 2769–2782.
2. X. Zhao, X. Chen, H. Yuk, S. Lin, X. Liu and G. Parada, Chem. Rev., 2021, 121, 4309–4372.
3. L. K. Rivera-Tarazona, Z. T. Campbell and T. H. Ware, Soft Matter, 2021, 17, 785–809.
4. E. Caló and V. V. Khutoryanskiy, Eur. Polym. J., 2015, 65, 252–267.
5. J. Chen, Q. Peng, X. Peng, L. Han, X. Wang, J. Wang and H. Zeng, ACS Appl. Polym. Mater., 2020, 2, 1092–1107.
6. D. Zhalmuratova and H.-J. Chung, ACS Appl. Polym. Mater., 2020, 2, 1073–1091.
7. X. Zhao, Soft Matter, 2014, 10, 672–687.
8. C. Creton, Macromolecules, 2017, 50, 8297–8316.
9. A. K. Means and M. A. Grunlan, ACS Macro Lett., 2019, 8, 705–713.
10. T. Sakai, T. Matsunaga, Y. Yamamoto, C. Ito, R. Yoshida, S. Suzuki, N. Sasaki, M. Shibayama and U.-I. Chung, Macromolecules, 2008, 41, 5379–5384.
11. T. Fujiyabu, Y. Yoshikawa, U.-I. Chung and T. Sakai, Sci. Technol. Adv. Mater., 2019, 20, 608–621.
12. S. Ida, R. Kimura, S. Tanimoto and Y. Hirokawa, Polym. J., 2017, 49, 237–243.
13. S. Ida, T. Hikida, A. Kawai, T. Matsuda, S. Suzuki, H. Imoto, K. Naka and S. Kanaoka, Polym. Chem., 2023, 14, 2771–2778.
14. H. W. Ooi, K. S. Jack, H. Peng and A. K. Whittaker, Polym. Chem., 2013, 4, 4788–4800.
15. X. Li, S. Nakagawa, Y. Tsuji, N. Watanabe and M. Shibayama, Sci. Adv., 2019, 5, eaax8647.
16. X. Huang, S. Nakagawa, X. Li, M. Shibayama and N. Yoshie, Angew. Chem., Int. Ed., 2020, 59, 9646–9652.
17. G. Gao, M. Hara, T. Seki and Y. Takeoka, Sci. Technol. Adv. Mater., 2024, 25, 2302795.
18. K. Haraguchi and T. Takehisa, Adv. Mater., 2002, 14, 1120–1124.
19. Q. Wang, J. L. Mynar, M. Yoshida, E. Lee, M. Lee, K. Okuro, K. Kinbara and T. Aida, Nature, 2010, 463, 339–343.
20. N. Morimoto, T. Ohki, K. Kurita and K. Akiyoshi, Macromol. Rapid Commun., 2008, 29, 672–676.
21. Z. Xu, J. Li, G. Gao, Z. Wang, Y. Cong, J. Chen, J. Yin, L. Nie and J. Fu, J. Polym. Sci., Part B:Polym. Phys., 2018, 56, 865–876.
22. T. Liu, F. Wang, Q. Wu, T. Chen and P. Sun, Mater. Horiz., 2021, 8, 3096–3104.
23. Y. Koda, T. Terashima, M. Takenaka and M. Sawamoto, ACS Macro Lett., 2015, 4, 377–380.
24. J. P. Gong, Y. Katsuyama, T. Kurokawa and Y. Osada, Adv. Mater., 2003, 15, 1155–1158.
25. J. P. Gong, Soft Matter, 2010, 6, 2583–2590.
26. J.-Y. Sun, X. Zhao, W. R. K. Illeperuma, O. Chaudhuri, K. H. Oh, D. J. Mooney, J. J. Vlassak and Z. Suo, Nature, 2012, 489, 133–136.
27. C. Norioka, Y. Inamoto, C. Hajime, A. Kawamura and T. Miyata, NPG Asia Mater., 2021, 13, 34.
28. J. Kim, G. Zhang, M. Shi and Z. Suo, Science, 2021, 374, 212–216.
29. C. Liu, N. Morimoto, L. Jiang, S. Kawahara, T. Noritomi, H. Yokoyama, K. Mayumi and K. Ito, Science, 2021, 372, 1078–1081.
30. Y. Zhang, C. Hu, X. Xiang, Y. Diao, B. Li, L. Shi and R. Ran, RSC Adv., 2017, 7, 12063–12073.
31. L. Yang, S. Li, Z. Zhao, J. Wang, H. Lv and X. Yang, Polym. Chem., 2023, 14, 2212–2219.
32. Y. Wu, W. Xing, J. Wen, Z. Wu, Y. Zhang, H. Zhang, H. Wu, H. Yao, H. Xue and J. Gao, Polymer, 2023, 267, 125661.
33. A. Varadarajan, R. M. Badani Prado, K. Elmore, S. Mishra and S. Kundu, Soft Matter, 2024, 20, 869–876.
34. C. S. Patrickios, Amphiphilic Polymer Co-networks: Synthesis, Properties, Modelling and Applications, RSC Publishing, Cambridge, UK, 2020.
35. C. S. Patrickios and T. K. Georgiou, Curr. Opin. Colloid Interface Sci., 2003, 8, 76–85.
36. G. Erdodi and J. P. Kennedy, Prog. Polym. Sci., 2006, 31, 1–18.
37. R. Yoshida, K. Uchida, Y. Kaneko, K. Sakai, A. Kikuchi, Y. Sakurai and T. Okano, Nature, 1995, 374, 240–242.
38. H. Guo, C. Mussault, A. Brûlet, A. Marcellan, D. Hourdet and N. Sanson, Macromolecules, 2016, 49, 4295–4306.
39. C. Mussault, H. Guo, N. Sanson, D. Hourdet and A. Marcellan, Soft Matter, 2019, 15, 8653–8666.
40. W. F. M. Daniel, J. Burdyńska, M. Vatankhah-Varnoosfaderani, K. Matyjaszewski, J. Paturej, M. Rubinstein, A. V. Dobrynin and S. S. Sheiko, Nat. Mater., 2016, 15, 183–189.
41. S. S. Sheiko and A. V. Dobrynin, Macromolecules, 2019, 52, 7531–7546.
42. F. Jia, J. M. Kubiak, M. Onoda, Y. Wang and R. J. Macfarlane, Adv. Sci., 2021, 8, 2100968.
43. M. Onoda, F. Jia, Y. Takeoka and R. J. Macfarlane, Soft Matter, 2022, 18, 3644–3648.
44. B. R. Clarke and G. N. Tew, Macromolecules, 2022, 55, 5131–5139.
45. A. Blencowe, J. F. Tan, T. K. Goh and G. G. Qiao, Polymer, 2009, 50, 5–32.
46. J. M. Ren, T. G. McKenzie, Q. Fu, E. H. H. Wong, J. Xu, Z. An, S. Shanmugam, T. P. Davis, C. Boyer and G. G. Qiao, Chem. Rev., 2016, 116, 6743–6836.
47. J. Hu, R. Qiao, M. R. Whittaker, J. F. Quinn and T. P. Davis, Aust. J. Chem., 2017, 70, 1161–1170.
48. S. Ida, Y. Toyama, S. Takeshima and S. Kanaoka, Polym. J., 2020, 52, 359–363.
49. K. K. Sharker, S. Takeshima, Y. Toyama, S. Ida, S. Kanaoka and S.-I. Yusa, Polymer, 2020, 203, 122735.
50. K. Terao, M. Abe, M. Nagase, S. Takeshima, S. Ida and S. Kanaoka, Macromolecules, 2023, 56, 5635–5641.
51. S. Bywater, Adv. Polym. Sci., 1979, 30, 89–116.
52. S. Kanaoka, M. Sawamoto and T. Higashimura, Macromolecules, 1991, 24, 2309–2313.
53. K.-Y. Baek, M. Kamigaito and M. Sawamoto, Macromolecules, 2001, 34, 215–221.
54. S. Ida, S. Toda, M. Oyama, H. Takeshita and S. Kanaoka, Macromol. Rapid Commun., 2021, 42, 2000558.
55. Y. Okaya, Y. Jochi, T. Seki, K. Satoh, M. Kamigaito, T. Hoshino, T. Nakatani, S. Fujinami, M. Takata and Y. Takeoka, Macromolecules, 2020, 53, 374–386.
56. M. Shibayama, M. Morimoto and S. Nomura, Macromolecules, 1994, 27, 5060–5066.

---

Footnote

† Electronic supplementary information (ESI) available: Additional ¹H NMR spectra and SEC curves of linear and star PNIPAAms; results of uniaxial tensile tests; and SAXS analysis. See DOI: https://doi.org/10.1039/d5py00014a

---