Article pubs.acs.org/Macromolecules
Elucidating Dynamics of Precoordinated Ionic Bridges as Sacrificial Bonds in Interpenetrating Network Hydrogels Jun Yang,* MingGuo Ma, XueMing Zhang, and Feng Xu Beijing Key Laboratory of Lignocellulosic Chemistry, Beijing Forestry University, Beijing, 100083, China S Supporting Information *
ABSTRACT: Interpenetrating polymer network (IPN) hydrogels were developed by engineering ionic bridges between carboxylated cellulose nanofibrils (CNFs) and amine-terminated poly(ethylene glycol) (PEG) into a covalent poly(acrylamide) architecture network, and the role of precoordinated CNF−PEG dynamic complexes in the IPN hydrogels viscoelastic dynamics was explored. The results shown that the ionic complexes significantly improved the IPN hydrogels energy dissipation and elastic recovery properties, leading to strain-rate dependent mechanics and notable enhancement in tensile toughness. The uniaxial deformation over a range of strain rates demonstrated that fracture energy peaked at 0.05 s−1 before decreased with further increasing strain rate, consistent with the crack propagation rate result. This straightforward sacrificial bonding strategy validates the hypothesis that elastomers with high toughness and excellent recovery can be achieved by incorporating precoordinated supramolecular associating motifs, which confers control over mechanical properties in a reversible, dynamic, and modular fashion.
1. INTRODUCTION Polymeric materials that contain reversible bonds and other dynamic interactions attract great interest due to their fascinating properties, such as self-recovery, adaptability, and malleability.1 These transient, noncovalent interactions (such as hydrogen bond, metal−ligand interactions, hydrophobic association, van der Waal forces, and aromatic stacking) are utilized as driving forces to build well-defined architecture for supramolecular materials, and a variety of supramolecular systems that respond to external stimuli have been developed.2−5 Among them one of prominent features of supramolecular polymers, especially compared to the most covalent polymers, is the thermal reversibility of the supramolecular networks.6 Recently, some research has highlighted the concept of interpenetrating polymer networks (IPNs), in which one component (semi-IPNs) or all of the components (IPNs) are cross-linked, to fabricate mechanically reinforced elastic materials.7 Consider the intimate contact in multicomponent of IPNs, each component could generate disparate characteristics, leading to a tunable platform toward exploration of interactions among the constrained components. Increasing experimental and theoretical reports on natural materials (e.g., collagen fibers in bone, protein fold in silk) have demonstrated that the “sacrificial bond” molecular mechanism is responsible for their unique mechanical properties.8 These sacrificial bonds can sustain stress at initial loading and then gradually rupture prior to covalent bonds in polymer matrix, during which a large amount of energy is dissipated and the materials integrity is survived thereby.9 Inspired by this © XXXX American Chemical Society
biomimetic strategy, various methodologies have been proposed to transfer the principle of sacrificial bonds into synthetic elastic materials with the aim of high toughness. Gong et al. produced robust self-healing polyampholyte hydrogels by sequential free-radical polymerization of opposite charged cationic and anionic monomers.10 Shull and Hawker et al. utilized well-defined ABA triblock copolyelectrolytes to synthesize phase-separation induced polyioncomplex hydrogels.11,12 Following mussel-inspired strategy, Hou7 and Menyo8a incorporated catechol−Fe3+ complexes and modified four-arm poly(ethylene glycol) as dynamic cross-linkers into a linear covalent polymer, respectively. Consider encoding desired mechanical properties and dynamics generally requires multiple types of cross-linking,9 these reports have demonstrated that the incorporation of dynamic supramolecular motifs into elastomers as sacrificial bonds enables efficient energy dissipation through reversible bond rupture, affording a unique combination of high toughness and the ability to recover after deformation.8,9 In our search for a suitable dynamic motif, the ionic-bridged assembly of carboxylated cellulose nanofibrils (CNFs) and tetra-arm poly(ethylene glycol) (PEG) with amine terminal groups attracted our attention due to its unique combination of thermoreversible properties.13 Heating the CNF−PEG assembly leads to a well-dissolved suspension, indicating that the Received: April 26, 2016 Revised: May 30, 2016
A
DOI: 10.1021/acs.macromol.6b00874 Macromolecules XXXX, XXX, XXX−XXX
Article
Macromolecules decrease in entropy drives the self-assembly of the gelator to a kinetically stable state. Whereas the substantial progress by applying ionic bonding as a potential candidate for renewable sacrificial bond strategy,10−12 most of them require intricate molecular design and tedious procedure step. To the best of our knowledge, no prior report has carefully investigated the dynamically mechanical properties of carboxylated CNF in covalent networks for controlling bulk dynamic properties. It is envisioned that the precoordinated PEG−CNF dynamic motif would provide a huge amount of untapped potential as the platform for dynamic materials design, affording us the unique opportunity to fulfill a library of materials with tunable dynamic properties. Specifically, we illustrate our concept by a facile one-pot sequential approach to design supramolecularly coordinated IPN hydrogels, derived from PEG and CNF bridged dynamic complexes in the second covalent polyacrylamide (PAAm) network, that is the skeleton of the IPNs built up with supramolecularly assembled architecture, rather than a composite of CNFs and PAAm. The results demonstrate that the two components act cooperatively. The precoordinated PEG−CNF dynamic complexes act as sacrificial bonds that preferentially rupture prior to the covalent bonds to dissipate energy and facilitate chain orientation, and the covalently crosslinked PAAm network maintains the original configuration after deformation.
Table 1. Composition of IPN Hydrogels sample CNF−PEG CNF−PEG CNF−PEG CNF−PEG CNF−PEG PAAm
2 5 7.5 10 15
VCNF (mL)
mPEG (mg)
AAm (g)
MBA (mg)
Irgacure 2959 (mg)
water (mL)
0.4 1 1.5 2 3 −
32 80 120 160 240 −
2 2 2 2 2 2
4.3 4.3 4.3 4.3 4.3 4.3
12.5 12.5 12.5 12.5 12.5 12.5
12 12 12 12 12 12
example, to prepare CNF−PEG 10 hydrogels, AAm (2 g), UV initiator Irgacure 2929 (12.5 mg, 0.2 mol % of AAm), N,N′-methylenebis(acrylamide) (MBA, 4.3 mg, 0.1 mol % of AAm), PEG (160 mg), and water (12 mL) were homogenized by stirring for 10 min. After bubbling N2 for 10 min, the CNF suspension (2 wt %, 2 mL) was added into the above aqueous solution and the mixture was ultrasonicated for 3 min by maintenance pH 8.7 (0.01 M NaOH/ HCl). The resulting suspension was injected into a tube, heated to 45 °C for 5 min. Then the suspension was gradually cooled down to room temperature to allow the CNF−PEG built the first network which was ionically bridged by amine-terminated tetra-arm PEG adsorption to carboxylated nanofibrils surface. Once the free-standing hydrogels were formed (∼15 min), the samples were subjected to UV irradiation (20 W, 365 nm wavenumber) for 2 h to achieve the second PAAm covalently cross-linked network. The attained IPN hydrogels were immersed in deionized water for 4 days (refreshed every 12 h) to remove unreacted monomers, and cut into required shapes for mechanical measurements. The covalently cross-linked PAAm counterparts were synthesized by UV photopolymerization in the same procedure. Mechanical Testing. The tensile properties of the achieved IPN hydrogels were measured using a Zwick Roell Testing Machine (Z005) equipped with a thermal chamber. The specimens, cut into rectangular shape with 12 mm in width, 8 mm in depth, and 50 mm in length, were uniaxially stretched at a constant speed of 0.05 s−1 with a gauge length of 30 mm. The nominal fracture stretchability (ε) was defined as the deformed length relative to the original length, and the nominal tensile strain (σ) was defined as tensile force divided by initial cross-sectional area (12 × 8 mm2) of the specimen. The Young’s modulus (E) was calculated by initial slope of the stress−strain curves (0−50% strain). The fracture energy (τ), a parameter to characterize the toughness of the sample, was determined by the area under the stress−strain curve. The strain rate-dependent tensile analysis was performed at strain rates ranging from stretching rate ranging from 8.33 × 10−4 to 0.167 s−1. For hysteresis test, the loading and unloading cycles was performed at a constant 0.05 s−1 to maintain a constant loading history. To measure residual strain after stretching, the specimens were deformed to a unified strain (500 mm/mm) and kept the deformation for a given time (1−15 min), then the stress was released and the length of specimen was recorded after 1 min resting time. The residual strain ratio was defined by (l − l0)/l0 × 100, where l0 and l were the length of sample before and after stretching, respectively. For recovery experiment, the specimens were initially stretched by a loading−unloading to achieve a maximum stretching (1200%), then the specimens were sealed in plastic bags and stored at 25 or 40 °C with different interval time (1−15 min) and tested again. For stress relaxation, the specimens were stretched to a strain of 400%, then the strain was held constant and the time-dependent relaxation of stress was recorded. For crack propagation calculation, the prenotched samples with 10 mm initial notch were stretched at a constant stretching rate at 25 °C until rapture. To examine the effect of salt ions on hydrogels fracture energy, the specimens were soaked in NaCl solution (0.05−1 M) for 4 h before mechanical testing, then the samples were dialyzed against DI water for 24 h to examine the reversible electrostatic screen process. The specimens were coated with a thin layer of mineral oil to prevent water evaporation before mechanical tests. To ensure accuracy, four specimens for each sample were tested and the average values were provided.
2. EXPERIMENTAL SECTION Extraction of Cellulose Nanofibrils. The cellulose pulp suspension (5 wt %) was mechanically stirred for 24 h. Subsequently, the suspension was homogenized using an Ultra-Turrax (IKA) at 12 000 rpm for 1 h. This initially refining step was performed to increase accessibility of cell wall for the following mechanical shearing treatment. The pulp slurry (3 wt %) was pumping into a high-pressure fluidizer with two different Z-shaped chamber pairs. First, the slurry was passed into a chamber pair with a diameter of 400 and 200 μm for 16 passes at 110 MPa, then passed through a chamber pair with a diameter of 200 and 75 μm for 12 passes at 170 MPa. The mixture was sonicated for 10 min (300 W) and converted to gel-like suspension with a solid content of 1.2 wt %. TEMPO-Mediated Oxidation of Cellulose Nanofibrils. Conversion of CNF to COOH−CNF, by oxidation of primary alcohol groups to carboxylic acid groups, was accomplished by the procedure reported in literature.14 About 5 g of CNF was suspend in deionized water (400 mL) containing TEMPO (2,2,6,6-tetramethylpiperidine-1oxyl, C9H18NO) 0.064 g, 0.4 mM) and NaBr (0.4 g, 4 mM) and homogenized by ultrasonic treatment at an ice−water bath for 10 min. Then certain amount of NaClO aqueous solution (10 wt %, 5 mM per gram of cellulose) was slowly added to the suspension to start oxidization at room temperature with mechanical stirring (300 rpm). The pH of the mixture was maintained at 10 by adding 0.5 M NaOH throughout the reaction. The reaction was quenched by adding ethanol (10 mL), and 0.5 M HCl was slowly dropped into the mixture to adjust pH to 7. Finally, the oxidized CNF was throughout dialyzed against water. The TEMPO-oxidized CNF has a carboxylic group content of 1.45 mmol/g, which was determined by conductivity titration. The average dimensions of CNFs were measured to be 20 ± 2 nm wide × 400 ± 30 nm long with TEM observation (Figure S1). IPN Hydrogels Preparation. 2-Hydroxy-4-(2-hydroxyethoxy)-2methylpropiophenone (Irgacure 2959) and tetra-arm amine terminated poly(ethylene glycol) (Mn 5000 Da) were purchased from Sigma-Aldrich and all other reagent were analytical grade and used as received without further purification. The code of IPN hydrogels was named as CNF−PEG x, where x was the percentage of CNF−PEG to acrylamide (AAm) with fixed CNF/PEG ratio and the both chemical cross-linker and UV initiator concentrations were fixed for all specimens based on an initial screening experiment (Table 1). For B
DOI: 10.1021/acs.macromol.6b00874 Macromolecules XXXX, XXX, XXX−XXX
Article
Macromolecules
Figure 1. (a) Schematic one-pot preparation of IPN hydrogels. (b) TEM images of CNF−PEG 10 hydrogels (inset of CNF hierarchical arrangement). (c) SANS profiles for the IPN hydrogels.
Figure 2. Dynamic rheological behaviors of IPN hydrogels. (a) Oscillatory frequency sweeps by 0.1% strain at 25 °C, (b) strain sweep by frequency of 6.28 rad/s at 25 °C, and (c) temperature sweep by 6.28 rad/s with 0.1% strain (G′ solid, G″ open, black for PAAm, red for CNF−PEG 10, and green for CNF−PEG 15). Small-Angle Neutron Scattering (SANS). The scattering measurements were performed at the National Institute of Standard and Technology Center for Neutron Research (NCNR). The SANS data were collected using neutrons with a wavelength of 6 Å and a detector distance of 30 m. The specimens were prepared by using D2O to enable adequate contrast between the hydrogen-rich gel and the deuterated solvent, and loaded into 1.5 mm thick titanium scattering cells with quartz windows. All the samples were incubated overnight at 25 °C prior to scattering test to ensure stable conditions. The date were exhibited by the radially averaged intensity (I) as a function of scattering vector (q), which was defined as (4π/λ)sin(θ/2), where λ and θ were the wavelength of the neutrons and the scattering angle, respectively. The corrected scattering intensity functions were normalized to the absolute intensity scale Rheological Measurements. Rheological measurement was performed by using TA AR2000 rheometer with a parallel-plate geometry (20 mm in diameter, 1 mm between gap distance). Frequency sweep was conducted at a constant strain of 0.1% by varying frequency 0.01 to 628 rad/s at 25 °C. For other tests, a constant frequency of 6.28 rad/s was applied. For temperature sweep, the experiment was conducted from 25 to 60 °C at a heating rate of 4 °C/min. The specimens were coated with a thin layer of mineral oil to prevent water evaporation. Morphological Observation. Scanning electron microscopy (SEM) was conducted on Hitachi S-3500 operating at 6 kV. The fractured samples were freeze-dried for 24 h and sputter-coated with gold before observation. For transmission electron microscopy (TEM), the samples were cryo-microtomed (Leica EM UC6) to obtain the thin section (∼100 nm) and stained with uranyl acetate (1.5
wt %) for 10 s and used for observation at an accelerating voltage of 100 kV (Hitachi H7600).
3. RESULTS AND DISCUSSION To prepare IPN hydrogels, the components of CNF, PEG, UV initiator, monomer (AAm), and cross-linker (MBA) were added into water at a pH of 8.7, followed by a simple heating/ cooling cycle and free-radical polymerization procedure (Figure 1a). The mixture was first heated above the CNF−PEG assembly sol−gel transition temperature to 45 °C with stirring, yielding a transparent suspension. The resulting suspension was then gradually cooled to room temperature, allowing CNF− PEG to form the first network via NH2−COOH coordination. Subsequently the AAm monomers in the mixture were photopolymerized to form the second covalently cross-linked network, yielding the IPN hydrogels with homogeneously distributed fibrillar networks. The formation of NH2−COOH ionic complex occurs at a pH value greater than or equal to the pKa (8.4) of the amine groups that used to prepare materials via this coordinating mechanism (Figure S2). It has been reported that gels became stronger when pH was above the pKa of the amine groups that conjugated to the carboxylic ligands.15 Thus, one can expect that supramolecular hydrogels stemmed from CNF−PEG complex exhibit dynamic mechanical properties across the range of evaluated pH values (>8.4). Transmission electron microscopy (TEM) images of the IPN hydrogels reveal a homogeneous fibrillar network with pores of C
DOI: 10.1021/acs.macromol.6b00874 Macromolecules XXXX, XXX, XXX−XXX
Article
Macromolecules
Figure 3. Mechanical properties of IPN hydrogels. (a) Typical stress−strain curves for IPN hydrogels with different CNF−PEG fractions (strain rate 0.05 s−1, 25 °C). (b) CNF−PEG 10 cyclic stress−strain curves at different temperatures (strain rate 0.05 s−1). (c) Influence of CNF−PEG fraction on damping properties (6.28 rad/s, 0.1% strain). (d) Influence of NaCl solution on mechanical properties of CNF−PEG 10 (strain rate 0.05 s−1, 25 °C).
the materials dynamics and became more sensitive to test frequency with increasing CNF−PEG fraction. The storage moduli (G′) increased with the increase of frequencies from 0.1 to 100 rad/s, which may relate to the relaxation of the coordination bonds and is in agreement with the tensile mechanics under different strain rates (discussion below). With increasing of strain, the rupture of CNF−PEG coordination bonds is expected to increase (Figure 2b). One can note that G′ of IPN hydrogels remained steady even at a strain of 22% while became more fluid than elastic with further increased strain, inferring softening of the network by the rupture of the sacrificial coordination bonds. Whereas for the PAAm counterparts, both G′ and G″ slightly decreased with increasing strain, which may be related to the covalent network defects (such as dangling chains and primary loops) at higher shear strain.6 Consider the tightly bundled fibrils bearing long multiarmed PEG chains with entanglement topology leads to the elastic energy of the hydrogel, the disruption of CNF percolated network may also contribute to the hydrogels mechanical stability to an external force (Figure S3), and the further discussion is required in future work. Additionally, the influence of temperature on strength of IPN hydrogels was explored by examining dynamic moduli under heating, and the results demonstrated that G′ decreased dramatically above 45 °C(Figure 2c), corroborating the gel−sol transition for CNF−PEG assembly at a higher temperature. The dynamic CNF−PEG complexes provide the capacity for energy dissipation within the coordinated network, as expressed by the moderate Arrhenius activation energy of ionic bonds (Figure S4). By measuring rheological properties of CNF−PEG motifs at different temperatures, one can note that the viscoelasticity is largely determined by the dynamic ionic cross-links, conveying the tendency of coordination bonds to
20−50 nm (Figure 1b), suggesting the excellent phase miscibility may contribute to the stress transfer. Indeed, this uniformly distributed fibrillar hierarchical structure could largely improve the mechanical properties of IPN hydrogels, since the interconnected fibrillar framework throughout the matrix enables a large contact area with polymer chains and leads to efficient stress transfer.16 Considering neutrons are sensitive to the nanometer scale length range, small-angle neutron scattering (SANS) experiments were performed on the D2O swollen IPN hydrogels to gain more refined picture with different CNF−PEG fractions (Figure 1c). The result indicates no pronounced changes in the curves with the amount of CNF−PEG, inferring isostructural of the materials. This lack of architecture differences between IPN hydrogels is crucial as any divergence in the viscoelasticity can be explicitly attributed to variation in the dynamics of the ionic coordination rather than gel cross-linked structure. It should be emphasized that contrast to previous polyelectrolyte hydrogels with self-aggregation induced microphase separation structures,10 the dynamic CNF−PEG coordination breaks in a reversible manner and no scattering peak was noted in SANS profiles, implying there was no pronounced electron density fluctuation at the nanoscale. This result further corroborates the homogeneous dispersion of fibrillar network within matrix, which contributes to the enhancement in mechanical performance. An oscillatory strain sweep measurement was initially conducted to determine the linear viscoelastic region for IPN hydrogels (
