Physical Gels Based on Charge-Driven Bridging of Nanoparticles by

Support. Get Help · For Advertisers · Institutional Sales; Live Chat. Partners. Atypon; CHORUS; COPE; COUNTER; CrossRef; CrossCheck Depositor; Orcid ...
1 downloads 0 Views 1MB Size
Download PDF
Article pubs.acs.org/Langmuir

Physical Gels Based on Charge-Driven Bridging of Nanoparticles by Triblock Copolymers Marc Lemmers,†,∥ Evan Spruijt,† Sabine Akerboom,† Ilja K. Voets,‡,⊥ Adriaan C. van Aelst,§ Martien A. Cohen Stuart,† and Jasper van der Gucht*,† †

Laboratory of Physical Chemistry and Colloid Science, Wageningen University, Dreijenplein 6, 6703 HB, Wageningen, The Netherlands ‡ Adolphe Merkle Institute, University of Fribourg, Route de l’ancienne papeterie, CP 209, CH-1723, Marly 1, Switzerland § Wageningen Electron Microscopy Center p/a Laboratory of Virology, Wageningen University, Droevendaalsesteeg 1, PB 16, 6700 AA Wageningen, The Netherlands ∥ Dutch Polymer Institute, John F. Kennedylaan 2, 5612 AB, Eindhoven, The Netherlands ⊥ Laboratory of Macromolecular and Organic Chemistry, Institute for Complex Molecular Systems, Eindhoven University of Technology, PO Box 513, 5600 MB Eindhoven, The Netherlands S Supporting Information *

ABSTRACT: We have prepared an aqueous physical gel consisting of negatively charged silica nanoparticles bridged by ABA triblock copolymers, in which the A blocks are positively charged and the B block is neutral and water-soluble. Irreversible aggregation of the silica nanoparticles was prevented by precoating them with a neutral hydrophilic polymer. Both the elastic plateau modulus and the relaxation time increase slowly as the gel ages, indicating an increase both in the number of active bridges and in the strength with which the end blocks are adsorbed. The rate of this aging process can be increased significantly by applying a small shear stress to the sample. Our results indicate that charge-driven bridging of nanoparticles by triblock copolymers is a promising strategy for thickening of aqueous particle containing materials, such as water-based coatings.

■

physical network, leading to “composite” or “hybrid” gels. The inorganic particles, often clay or silica, then act as additional physical cross-linking points in the network, resulting in higher values for the elastic modulus.13−18 Special cases are the “shake gels”, which show gelation upon shearing.11,19 There are, however, very few reports on physical gels in which inorganic nanoparticles are the only nodes in the transient network. Petit and co-workers have prepared hybrid gels of graft copolymers mixed with silica particles, in which the grafts physically adsorb onto the silica particle surface.20 Another example is the study by Xu and co-workers, where they describe a gel based on graphene sheets bridged by ssDNA.21 In both cases, the binding interactions are noncovalent, but they are so strong that they are essentially irreversible at room temperature. Weaker composite gels were studied by by Qi et al.,22,23 who reported particle clustering by diblock copolymers, and by Wang et al.,24 who described a water-based physical gel, presumably based on the electrostatic interaction between a positively charged

INTRODUCTION Polymers are commonly used as rheology modifiers because they can enhance the solvent viscosity at relatively low volume fractions by forming entanglements. These effects become more pronounced when “sticky” groups are incorporated in the polymer chains that physically cross-link the polymers. This leads to the formation of viscoelastic physical gels that can rearrange on time scales that are determined by the strength of the physical cross-links. It is well-known that the properties of polymeric materials can be modified by incorporating inorganic particles as fillers. This strategy is commonly used, for example, to enhance the mechanical properties of rubbers, melts, or chemically crosslinked gels.1−9 Incorporating large amounts of particles in a physical gel, while stabilizing these particles at the same time, is, however, a real challenge. Often, particle suspensions are unstable when mixed with polymers, caused by irreversible aggregation of the particles in a process known as flocculation.10−12 Learning how to avoid flocculation of particles in physical gels will be of great benefit to, for example, the coatings industry. There are several examples of water-based physical gels in which inorganic particles are introduced within an existing © 2012 American Chemical Society

Received: May 11, 2012 Revised: July 25, 2012 Published: July 26, 2012 12311

dx.doi.org/10.1021/la301917e | Langmuir 2012, 28, 12311−12318

Langmuir

Article

Figure 1. (a) Block copolymer PTMAEMA−PEO−PTMAEMA. (b) A schematic drawing of silica nanoparticles connected by triblock copolymers, thereby forming a transient network. The PEO middle block is drawn in green, the positively charged end blocks are drawn in red, and the physisorbed PVP is drawn in gray. The PVP-coated silica particles were subsequently added to a 10 mM salt triblock copolymer solution. The sample was shortly vortexed directly after addition of the silica particles. The samples were left to rest for 1 day before starting rheological measurements. DLS measurements were started directly after mixing, without any resting period. Rheometry. Rheological measurements were performed on an Anton Paar Physica MCR 301 stress-controlled rheometer. A cone and plate geometry was used with either a cone diameter of 50 or 25 mm. Both geometries have an angle of 1°. The effect of solvent evaporation was minimized by utilizing the solvent blocker system (Anton Paar), which ensures a completely sealed sample environment. Temperatures of both the plate and the hood were Peltier controlled at 20 °C. Sample loading was performed with caution; a limiting normal force of 0.5 N was set. Rheological measurements were started 2 h after sample loading to allow the sample to relax the loading stresses. A 5 mL gel sample was kept in a closed container, out of which small aliquots were taken to investigate the development of the complex composite gel in time. Oscillatory measurements were performed at 0.05% strain, which was checked to be in the linear viscoelastic regime for the entire time window studied (see Supporting Information). Creep tests were performed at a shear stress of 0.1 Pa. The reported viscosity values were obtained after smoothing the creep compliance data, fitting the curve locally to a power law. Dynamic Light Scattering. Dynamic light scattering experiments were performed on an ALV-125 goniometer, combined with a Cobolt Samba-300 DPSS laser (300 mW) operating at a wavelength of 532 nm, an ALV optical fiber with a diameter of 50 μm, an ALV/SO single photon detector, and an ALV5000/60X0 external correlator. Temperature was controlled at 20 °C using a Haake F8-C35 thermostatic bath. The angle of detection was 90°, corresponding to a scattering vector q = 3.1 × 107 m−1, except for the q-dependent measurement described in the main text, where we recorded the intensity correlation function at 20−60°, in steps of 10°. Because of the high scattering intensity, an optical density filter of 10% was used to ensure linearity of the detector. A total sample volume of 1 mL was measured at different moments in time. Intensity correlation functions were recorded for different time intervals, corrected for the specific measurement setup and shifted if necessary to obtain an autocorrelation function that starts to decay at unity for the smallest correlation times. To remove microbubbles, samples were centrifuged gently before DLS measurements. Measurements that were still influenced by small bubbles were identified by a strong increase in the scattered intensity and were not included in the analysis. From a dynamic light scattering experiment we obtain the intensity correlation function, g(2)(t), from which we compute the autocorrelation function g(1)(t):28

dendritic end group and the negatively charged surface of clay particles. Here, we present a new type of water-based composite physical gel, formed by easy to synthesize ABA triblock copolymers and silica nanoparticles. The ABA triblock copolymer consists of A blocks that are positively charged and a B block which is neutral and water-soluble. The precoated silica particles are negatively charged in water at neutral pH. Mixing these two oppositely charged components together leads to transient network formation, in which the silica particles are the nodes in the network, as is schematically depicted in Figure 1b. As we will show, the bonds between the triblock copolymers and the silica particles can rearrange, and the particles do not suffer from irreversible aggregation. We show that the resulting “complex composite” gels age slowly in time, a process that can be enhanced under shear. The interaction strength between the triblock copolymers and nanoparticles, and thus the rheological properties, are highly tunable by means of, e.g., composition, concentration, block lengths, and salt concentration.

■

EXPERIMENTAL SECTION

Triblock Copolymer. The triblock copolymer that we used, PTMAEMA−PEO−PTMAEMA, consists of a poly(ethylene oxide) middle block and two poly(trimethylaminoethyl methacrylate) end blocks (Figure 1a). The synthesis was described previously25,26 and, in brief, in the Supporting Information. On average, 68 monomers were attached per macroinitiator molecule, as determined by 1H NMR (see Supporting Information), corresponding to 34 positive charges per end block on average. Composite Gel Preparation. To prepare the coated silica nanoparticles, Ludox LS 30 (Sigma-Aldrich, 30 wt % in 10 mM salt solution in water) colloidal silica particles, with a particle radius of ≈7.5 nm and a surface charge of ≈−0.08 C/m2,27 were mixed with a 10 mM salt solution of poly(vinylpyrrolidone) K25 (PVP) (Fluka, Mw = 40 kg mol−1). The amount of PVP in solution was twice the amount that is expected to physically adsorb, based on reflectometry measurements (Γmax ≈ 0.6 mg/m2; see Supporting Information). The soaking of particles in the PVP solution was done at least 1 day before further use of the coated particles to saturate the silica surface with PVP. As solvent in all samples we used the composition of the Ludox solvent, which is 0.002% (w/w) of chloride as NaCl and 0.010% (w/w) of sulfate as Na2SO4. The pH of the solvent was adjusted to pH = 7.0 ± 0.2 to minimize the chance of hydrolysis of the ester bonds in the end blocks of the triblock copolymers. 12312

dx.doi.org/10.1021/la301917e | Langmuir 2012, 28, 12311−12318

Langmuir g(1)(t ) =

Article

g(2)(t ) − 1 A

lower triblock copolymer concentrations, the samples were turbid, indicating an inhomogeneous distribution of particles. The turbidity decreases slightly with increasing silica particle concentration. The effects of silica concentration and triblock copolymer concentration on the physical appearance of the gels are summarized in Figure 2.

(1)

where A is an experimental parameter close to unity. We assume that the contribution of multiple scattering events to the autocorrelation function is small, as we did not observe a diffusive cone around the laser beam and the measured decay times were found to scale with q−2. Small-Angle X-ray Scattering. Small-angle X-ray scattering (SAXS) experiments were carried out at the Adolphe Merkle Institute on a pinhole camera (S-Max 3000) from Rigaku Innovative Technologies equipped with a microfocus X-ray source operating at wavelength λ = 1.5405 Å (Cu Kα emission) and a Triton 2dimensional multiwire gas detector. The sample was loaded into a reusable 2 mm quartz capillary cell and subsequently placed into a sample holder thermostated to 20 °C by an external Julabo CF 30 circulator. The data are radially averaged and corrected for transmission, background scattering, and detector efficiency and converted into absolute scattering cross sections using water as a calibration standard. The SASfit program developed by Kohlbrecher and co-workers at the Paul Scherrer Institute (PSI) is used for data analysis.29

■

RESULTS AND DISCUSSION Sample Preparation. To avoid irreversible aggregation of the particles, we first coated the silica nanoparticles with a layer of poly(vinylpyrrolidone) (PVP), a neutral hydrophilic polymer. Coating of the particles was achieved by physical adsorption of PVP, which is known to be essentially irreversible on silica surfaces.30 The PVP coating also prevents physical adsorption of the poly(ethylene oxide) (PEO) middle block because the adsorption energy of PEO is much lower than that of PVP.30−32 The adsorbed amount of PVP on the silica surface is ≈0.5 mg m−2 (see Supporting Information). To physically bind the negatively charged silica nanoparticles in a transient network, we employ an ABA triblock copolymer consisting of a PEO middle block and two strongly charged cationic poly(trimethylaminoethyl methacrylate) (PTMAEMA) end blocks (see Figure 1a). Mixing this triblock copolymer with a solution of precoated silica nanoparticles can lead to transient network formation, as schematically illustrated in Figure 1b. The silica particle concentration and the triblock copolymer concentration are the two major factors that influence transient network formation. We have investigated the effect of varying these two parameters by making samples of different composition. To prepare a sample-spanning transient network, the concentration of potential nodes must be high enough to facilitate a percolating structure. We have found that the silica particle concentration should be above approximately 10% (w/w) to cause any noticeable increase in viscosity. As expected, the viscosity of the gels increased with increasing silica particle concentration, because a higher particle concentration gives more nodes in the transient network. The viscosity decreased with increasing triblock copolymer concentration. This decrease can be caused by an increase in the formation of loops and dangling ends of the triblock copolymer, thereby reducing the number of bridges between the particles. Another possibility is that the particles themselves become repulsive when we overcharge the particles with positive charges. From earlier work on polyelectrolyte complexes, it is wellknown that the driving force for charge-driven association is strongest at a 1:1 charge ratio.25,26,33−35 In this case we found that we needed to overdose the system with positive charge at least three times to obtain reasonably transparent gels.36 At

Figure 2. Diagram sketching the regions (particle and polymer concentrations) where the samples appear as liquids (+), turbid gels (■), or transparent gels (○). The calculated dashed line indicates charge stoichiometry between the silica particles and the triblock copolymers. The big circle indicates the sample used for further studies. The inset shows the observed direction of increased transparency and increased or decreased viscosity. The numbers in the phase diagram belong to the data points directly above the numbers and correspond to the photographs of the samples as shown in the inset.

We further tested the mechanical properties and time evolution of these gels using a 13% (w/w) silica nanoparticles and 3.4% (w/w) triblock copolymer gel, indicated by the big circle in Figure 2. When we mix the coated silica nanoparticles with an aqueous solution of the ABA triblock copolymer at this composition, we initially obtain a turbid gel, a sign of large-scale inhomogeneities in the sample. Within half an hour the gel becomes almost transparent, indicating that the sample becomes more homogeneous (see Supporting Information). Note that after 24 h the amount of air bubbles that appeared after vortexing has decreased, showing that this gel still has liquid-like properties. We assume that upon the physical binding of a TMAEMA group with an oppositely charged group on the silica surface, counterions from both components are released into the solution. The entropy gain related to this counterion release is an important part of the driving force in charge-driven association.37 As a result of this counterion release, the salt concentration in the gel rises, to a maximum of 0.1 M for this specific case. Gel Structure. To probe the structure of our gels, we performed small-angle X-ray scattering (SAXS) measurements on the sample. The results are given in Figure 3. The scattering curve can be fitted reasonably well with a form factor for polydisperse (Gaussian) spheres of radius R = 7.6 ± 1.2 nm, together with an effective hard-sphere structure factor.29 This gives an effective hard-sphere radius of 12 nm, meaning that the polymer layer around the particles is approximately 3−6 nm thick. This agrees reasonably well with the hydrodynamic radius of 13 nm of a polymer-coated particle found with dynamic light 12313

dx.doi.org/10.1021/la301917e | Langmuir 2012, 28, 12311−12318

Langmuir

Article

particles that are flocculated due to van der Waals or depletion attractions.38,39 In the present case, we have a network of particles bridged by the triblock copolymers. In polymeric gels, the value of the storage modulus is ∼1 kT per elastically active chain in the network.40 If every triblock copolymer would act as an elastically active bridge, G′ would reach several kPa. What we measure is orders of magnitude lower, meaning that the number of elastically active bridges is only a few percent of the maximum possible amount. This must mean that most polymers form loops or dangling ends. The slow increase in storage modulus that we see during the first 200 h (see Supporting Information) can then be attributed to an increase of the number of bridging chains in the network. We note that the distribution of particles does not change dramatically during this time. Cryo-SEM images taken from a 24 h old sample are not visually different from those of a 72 days old sample (see Supporting Information). Dynamics of the Gels. To study the relaxation dynamics of the complex composite gels, we use dynamic light scattering (DLS). The autocorrelation functions of a complex composite gel at 0.1 M salt, for different times after preparation, are shown in Figure 5. Two reference samples are also shown; 1% (w/w)

Figure 3. Small-angle X-ray scattering curve of a complex composite gel containing 13% (w/w) silica nanoparticles and 3.4% (w/w) triblock copolymer, 40 days after preparation. The thick line is a fit to the data obtained by the SASfit program, using a form factor for polydisperse spheres (R = 7.6 ± 1.2 nm) and an effective hard-sphere structure factor (with a hard-sphere radius of 11.9 nm and an effective volume fraction of 0.11).

scattering (see below). Cryo-SEM images of the same samples suggest that on larger length scales (around 100 nm) the gels are not completely homogeneous (see Supporting Information). In the SAXS data, we see no evidence for such large-scale density variations because the lowest q value that we can access in our SAXS measurements is 0.08 nm−1, corresponding to an upper length scale of ≈75 nm, roughly 5 particle diameters. Gel Elasticity. To study the viscoelastic properties of the physical gel, we performed rheological studies. The storage modulus G′ of the gel, measured at 1 rad s−1, was measured as a function of the age of the sample. For each measurement point, a new sample was taken from a developing gel and the modulus was measured 2 h after sample loading. We observed that G′ increases in the first 200 h, from 2 Pa to a final value of ≈8 Pa (see Supporting Information). Figure 4 (circles) shows the

Figure 5. Temporal evolution of the complex composite gel. Autocorrelation functions of reference samples containing 1% (w/w) PVP-coated silica particles (●) or 13% (w/w) PVP-coated silica particles (◆) and a complex composite gel sample with 0.1 M NaCl at 0.083 (○), 0.5 (△), 1 (□), 4 (×), 92 (◇), and 500 h (+) after preparation. Only 50% of the data points are shown. The intensity correlation functions at 0.083, 0.5, 1, and 4 h after preparation have been recorded for 5 min. The other intensity correlation functions were recorded for 4 h. Reference samples were recorded for 0.33 h.

PVP-coated silica particles and 13% (w/w) PVP-coated silica particles. The first shows a single-exponential decay, originating from the diffusive motion of the particles at this concentration. From a fit of the correlation function, we obtain a hydrodynamic radius of 13 nm, slightly larger than the effective hardsphere radius obtained by SAXS. At 13% (w/w) of PVP-coated silica particles, the autocorrelation function does not decay as a single-exponential function anymore because the particles interact with each other and hinder free diffusion. In the samples with triblock copolymer the decay of the autocorrelation function is shifted strongly to longer times, indicating a significant slowing down of the particles by the bridging polymers. We can distinguish three regimes in the autocorrelation function: first an initial exponential decay, then a second regime in which the autocorrelation function decays gradually, approximately as a power law, and finally a slower exponential

Figure 4. Frequency sweep of a complex composite gel, ∼1000 h (43 days) after preparation. Shown are two frequency sweeps, before (circles) and after (triangles) a creep test. Storage moduli are given by filled symbols, and loss moduli are given by the open symbols. The measurements shows that G′ > G″ over the measured frequency domain.

frequency-dependent viscoelastic moduli for an ∼1000 h (43 days) old complex composite gel. There is a weak frequency dependence and G′ > G″ over the measured frequency range, indicating that the material behaves solid-like on the investigated time scales. The final value of the storage modulus is still relatively low, around 10 Pa. This corresponds to roughly 0.02 kT per particle, much lower than values reported for colloidal gels formed by 12314

dx.doi.org/10.1021/la301917e | Langmuir 2012, 28, 12311−12318

Langmuir

Article

decay. We can use the following empirical function to fit the aurocorrelation functions:41,42 g(1)(t ) = A exp(−t /τf ) + B(1 + t /τi)−α /2 + C exp( −t /τl)

(2)

where τf is the relaxation time associated with the initial (fast) dynamics, τi is the time where the power-law regime begins, and τl is the longest relaxation time. From the fits we obtain estimates for the relaxation times of the gel at different aging times. The fits to the autocorrelation function are displayed as lines in Figure 5. The initial decay time τf in the complex composite gel is longer than for the reference samples. This indicates that the short-time motion of the particles is restricted by the presence of the triblock copolymers. It is independent of the time after preparation, however, as all autocorrelation curves overlap at short correlation times. Furthermore, τf is proportional to q−2 (see Supporting Information), indicating a diffusive process. We attribute this fast relaxation process to the local diffusive motion of the particles within the “cage” formed by the neighboring particles. Our observation that this diffusive motion does not become slower as the gel ages indicates that the particles remain stable and do not aggregate into larger clusters over time. At longer times, particle mobility is restricted because the particles are trapped in a network. Further relaxation must therefore imply network reorganization processes, which requires the breaking and reformation of bridges between the particles.41−45 For this, polyelectrolyte blocks must desorb from the particle surface. The gradual, power-law decay of the autocorrelation function means that there is a wide distribution of relaxation times associated with the reorganization of the network. The power-law exponent α decreases with the time after preparation, from close to 1 immediately after preparation, to ≈0.3 after 24 h (see Supporting Information). This means that the relaxation spectrum of the gel widens as the gel ages. The value of α in the fully developed gel is similar to values found by others for randomly cross-linked polymer networks.41−43 The final exponential decay is associated with the longest relaxation time of the network. The fact that the autocorrelation function decays to zero means that the gel is ergodic and the network is able to flow. In other words, the bonds between the triblock copolymers and the nanoparticles are reversible, and we are indeed dealing with a physical gel. While the initial decay is independent of time after preparation, the longest relaxation time increases strongly with time after preparation (see Figure 5). In Figure 6 we plot the development of the longest relaxation time, τl, as a function of time after preparation. From this figure we can see that the longest relaxation time keeps on increasing as the sample ages, initially (during the first 20 h) approximately linearly and more slowly in later stages. On the basis of the rheological and light scattering data, we propose the following mechanism to explain the temporal evolution of the gels. Initially, right after mixing triblock copolymer and particles, the sample consists of large clusters of bridged particles that are dispersed inhomogeneously throughout the sample, causing the turbidity. The bond between the triblock copolymer end blocks and the silica particles is initially relatively weak due to the intervening PVP layer. This enables the particles to redistribute in a more homogeneous way, thereby decreasing the turbidity and increasing the modulus.

Figure 6. Values for the longest relaxation time, τl, as a function of time after sample preparation, tprep. Shown are the values for the 0.1 M KCl gel (○) and the 0.2 M KCl gel (◆).

Gradually, the cationic blocks of the triblock copolymer adsorb more strongly on the particle surface, possibly by displacing the PVP from the surface. Hence, as the sample ages, the particles and polymers find continuously deeper local minima of the free energy, from which it becomes continuously harder to escape.46 This leads to an increase of the longest relaxation time of the network and to a broadening of the relaxation spectrum (as indicated by a decrease of the power-law exponent α). Effect of Salt. We assume that electrostatic interactions between the triblock copolymer and the silica particles are responsible for binding the particles in the transient network. As salt ions can screen these electrostatic interactions, we expect that addition of extra salt enhances the dynamics of the bonds between the oppositely charged components, as has been shown before for electrostatically assembled systems.35,47 To test this hypothesis, we prepared a complex composite gel with the same ratio of components, except that we doubled the salt concentration to 0.2 M. We noticed that the addition of extra salt decreased the time needed to reach the almost transparent appearance, indicating that electrostatic interactions are indeed important in our complex composite gels. We investigated this sample with DLS, and the results of these measurements are displayed in Figure 7.

Figure 7. Autocorrelation functions of reference samples containing 1% (w/w) PVP-coated silica particles (●) or 13% (w/w) PVP-coated silica particles (◆) and a 0.2 M complex composite gel sample at 0.083 (○), 0.5 (△), 4 (□), 15 (×), 64 (◇), and 218 h (+) after preparation. The intensity correlation functions at 0.083, 0.5, 4, and 15 h after preparation have been recorded for 5 min. The other intensity correlation functions were recorded for 4 h. Reference samples were recorded for 0.33 h. 12315

dx.doi.org/10.1021/la301917e | Langmuir 2012, 28, 12311−12318

Langmuir

Article

Figure 8. (a) Creep tests measured at a fixed shear stress of 0.1 Pa, but at different aging times. Shown is the creep compliance as a function of time. Symbols are explained in the graph. The graph displays only 10% of the data points. The inset shows the initial response of the 1000 h old sample on a log−log scale. We can see damped oscillations in the first 10 s of the creep test. (b) Time evolution of the apparent viscosity as a stress of 0.1 Pa is applied. The apparent viscosity keeps on increasing, approximately as η ∼ t0.5. The three curves correspond to samples of different age, from bottom to top: 500, 670, and 1000 h.

viscosity has increased during aging of the sample. This is in agreement with the observed increase of the relaxation time (Figure 6). After almost 3 h of creep, the creep curves do not yet show a clear linear increase of the creep compliance with time. We calculated the shear viscosity from the slope of the creep curve. The results for the three samples are shown in Figure 8b. The initial viscosity values are very noisy because of the creep ringing effect. Figure 8b shows that the viscosity of the samples increases with age, apparently without leveling off, even after 104 s. This continued viscosity increase points to shear-induced changes in the sample. This is confirmed by the increase of the storage and loss moduli measured by a frequency sweep 2 h after the creep experiment. The result of this frequency sweep is also plotted in Figure 4 and shows an approximate factor of 4 increase in moduli values as compared to the frequency sweep before the creep test. We conclude that the applied stress enhances the rate of aging of the gels. Such a stress-enhanced aging has been reported before for thixotropic gels.46,48 The applied stress probably helps the particles and polymer chains to escape from local free energy minima (which involves the desorption of polymer chains from the particles) and to find new, even deeper, minima. Drying of the Composite Gel. The gel that we have described is not completely transparent (see Supporting Infomation). For possible aplications in coatings this is not desirable. To test whether the gel maintains its turbidity when applied as a coating, we performed a simple drying test. Upon drying, the lump of gel shrinks, through the evaporation of water. This shrinking causes a flattening of the drop, which then becomes a thin layer of gel. Because of the decrease in volume, the components in the gel are concentrated, leading to a more or less homogeneous distribution of the particles in the gel. This in turn decreases the scattering of the drying gel, which becomes fully transparent after 2 h of drying (see Supporting Information).

The shape of the autocorrelation functions for the 0.2 M salt gel are qualitatively the same as for the 0.1 M salt gel. The longest relaxation time for this salt concentration is also plotted in Figure 6. We can see that, initially, it is lower in the 0.2 M salt sample than in the 0.1 M salt sample, which is as expected because salt is known to plasticize electrostatic assemblies.47 However, after ∼70 h, the longest relaxation time keeps on increasing and becomes even longer than for 0.1 M salt. At the same time, we observe that the intensity of the scattered light and the fluctuations therein increase over time at 0.2 M salt, while they are stable at 0.1 M salt (see Supporting Information). This suggests that the high salt concentration leads to irreversible flocculation of the particles. This can also be seen from the relaxation time of the fast process, associated with the short-time diffusive behavior of the particles. In contrast to what we observed for the 0.1 M salt sample, the fast relaxation time at 0.2 M salt is not constant, but increases as the sample ages (see Supporting Information). This increase can be attributed to the slow aggregation of the particles. This aggregation prevents us from melting the gel completely by adding salt. Stress-Induced Aging. We have seen that the complex composite gel slowly ages, as evidenced by an increase of the elastic modulus and the relaxation time. To measure the time evolution of the viscosity, we performed creep tests. We applied a constant shear stress of σ = 0.1 Pa to each sample and measured the creep compliance of each sample as a function of time, J(t). Figure 8a shows creep curves for three samples of approximately 500, 670, and 1000 h after preparation. The 1000 h sample is the same as studied in Figure 4. Qualitatively all three samples respond the same to the applied stress. There is an almost instantaneous response of the material, around 0.25 Pa−1 associated with the elasticity of the gel. The corresponding elastic modulus of ∼4 Pa agrees with the values of Figure 4. In the first 10 s after applying the shear stress, the measured deformation oscillates. Although present in all samples, for reasons of clarity only one response is shown in the inset of Figure 8a. This “creep ringing” phenomenon is seen also in creep tests of paste-like materials48,49 and is caused by the coupling between sample elasticity and instrument inertia.50 After the initial response, the gel continues to deform under the applied stress up to times of 104 s. The creep compliance decreases less rapidly for the older samples, indicating that the

■

CONCLUSIONS AND OUTLOOK In this paper we have presented an aqueous physical gel, existing of a transient network of precoated inorganic silica nanoparticles and ABA triblock copolymers, without causing irreversible particle aggregation. In this open network, the particle surfaces are frequently found 5−10 nm apart from each 12316

dx.doi.org/10.1021/la301917e | Langmuir 2012, 28, 12311−12318

Langmuir

Article

(6) Haraguchi, K.; Takehisa, T. Nanocomposite hydrogels: A unique organic-inorganic network structure with extraordinary mechanical, optical, and swelling/de-swelling properties. Adv. Mater. 2002, 14, 1120−1124. (7) Okay, O.; Oppermann, W. Polyacrylamide-clay nanocomposite hydrogels: rheological and light scattering characterization. Macromolecules 2007, 40, 3378−3387. (8) Liu, Y.; Liu, X. L.; Wu, Y. T.; Sun, B.; Zhu, M. F.; Takafuji, M.; Ihara, H. Peculiar nanocomposite hydrogel with controllable multiple thermosensitivity: double phase transition and ternary stable states. Chem. Commun. 2010, 46, 430−432. (9) Carlsson, L.; Rose, S.; Hourdet, D.; Marcellan, A. Nano-hybrid self-crosslinked PDMA/silica hydrogels. Soft Matter 2010, 6, 3619− 3631. (10) Iler, R. K. Relation of particle size of colloidal silica to amount of cationic polymer required for floculation and surface coverage. J. Colloid Interface Sci. 1971, 37, 364. (11) Cabane, B.; Wong, K.; Lindner, P.; Lafuma, F. Shear induced gelation of colloidal dispersions. J. Rheol. 1997, 41, 531−547. (12) Whitby, C. P.; Scales, P. J.; Grieser, F.; Healy, T. W.; Kirby, G.; Lewis, J. A.; Zukoski, C. F. PAA/PEO comb polymer effects on rheological properties and interparticle forces in aqueous silica suspensions. J. Colloid Interface Sci. 2003, 262, 274−281. (13) Agrawal, S. K.; Sanabria-DeLong, N.; Tew, G. N.; Bhatia, S. R. Nanoparticle-reinforced associative network hydrogels. Langmuir 2008, 24, 13148−13154. (14) Pek, Y. S.; Wan, A. C. A.; Shekaran, A.; Zhuo, L.; Ying, J. Y. A thixotropic nanocomposite gel for three-dimensional cell culture. Nat. Nanotechnol. 2008, 3, 671−675. (15) Puech, N.; Mora, S.; Porte, G.; Phou, T.; Grillo, I.; Oberdisse, J. Silica nanoparticles dispersed in a self-assembled viscoelastic matrix: structure, rheology, and comparison to reinforced elastomers. Braz. J. Phys. 2009, 39, 198−204. (16) Song, F.; Zhang, L. M.; Shi, J. F.; Li, N. N. Viscoelastic and fractal characteristics of a supramolecular hydrogel hybridized with clay nanoparticles. Colloids Surf., B 2010, 81, 486−491. (17) Loizou, E.; Porcar, L.; Schexnailder, P.; Schmidt, G.; Butler, P. Shear-induced nanometer and micrometer structural responses in nanocomposite hydrogels. Macromolecules 2010, 43, 1041−1049. (18) Puech, N.; Mora, S.; Phou, T.; Porte, G.; Jestin, J.; Oberdisse, J. Microemulsion nanocomposites: phase diagram, rheology and structure using a combined small angle neutron scattering and reverse Monte Carlo approach. Soft Matter 2010, 6, 5605−5614. (19) Zebrowski, J.; Prasad, V.; Zhang, W.; Walker, L. M.; Weitz, D. A. Shake-gels: shear-induced gelation of laponite-PEO mixtures. Colloids Surf., A 2003, 213, 189−197. (20) Petit, L.; Bouteiller, L.; Brulet, A.; Lafuma, F.; Hourdet, D. Responsive hybrid self-assemblies in aqueous media. Langmuir 2007, 23, 147−158. (21) Xu, Y. X.; Wu, Q. O.; Sun, Y. Q.; Bai, H.; Shi, G. Q. Threedimensional self-assembly of graphene oxide and dna into multifunctional hydrogels. ACS Nano 2010, 4, 7358−7362. (22) Qi, L.; Fresnais, J.; Berret, J. F.; Castaing, J. C.; Destremaut, F.; Salmon, J. B.; Cousin, F.; Chapel, J. P. Influence of the formulation process in electrostatic assembly of nanoparticles and macromolecules in aqueous solution: the interaction pathway. J. Phys. Chem. C 2010, 114, 16373−16381. (23) Qi, L.; Fresnais, J.; Berret, J. F.; Castaing, J. C.; Grillo, I.; Chapel, J. P. Influence of the formulation process in electrostatic assembly of nanoparticles and macromolecules in aqueous solution: the mixing pathway. J. Phys. Chem. C 2010, 114, 12870−12877. (24) Wang, Q.; Mynar, J. L.; Yoshida, M.; Lee, E.; Lee, M.; Okuro, K.; Kinbara, K.; Aida, T. High-water-content mouldable hydrogels by mixing clay and a dendritic molecular binder. Nature 2010, 463, 339− 343. (25) Lemmers, M.; Sprakel, J.; Voets, I. K.; van der Gucht, J.; Cohen Stuart, M. A. Multiresponsive reversible gels based on charge-driven assembly. Angew. Chem., Int. Ed. 2010, 49, 708−711.

other, which is the appropriate spacing for a triblock copolymer to bridge two particles. We have shown that electrostatic interactions are the driving force for the association between the positively charged end-blocks and the negatively charged silica nanoparticles. The complex composite gels age slowly in time and this aging can be enhanced by applying stress. The base materials that we use are cheap and the triblock copolymer can be easily produced on scales of hundreds of grams. In this paper, we have considered only one particular combination of particles and triblock copolymer and studied its evolution in time and the effect of salt thereon. However, it is anticipated that the properties of the gels can be tuned in a wide range by varying additional parameters. Figure 2 already shows qualitatively how the appearance and the viscosity of the gel change when the concentrations of the two components are changed. Other parameters that will influence the gel properties are the pH (by its effect on the surface charge of the silica particles), the length of the different blocks of the triblock copolymer, and the size of the nanoparticles. Exploring the effects of these parameters is beyond the scope of the present publication; it will be the subject of future work.

■

ASSOCIATED CONTENT

S Supporting Information *

Details of the synthesis, 1H NMR spectra of the synthesized triblock copolymer, reflectometry data, rheological amplitude sweep data and moduli versus time, cryo-SEM images, more detailed light scattering data, including fit parameters that are not mentioned in the main text, and images of the dried gel. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors thank Prof. M. A. J. Michels for interesting discussions and for bringing relevant literature under our attention. E.S. acknowledges financial support from The Netherlands Organisation of Scientific Research (NWO). I.K.V. gratefully acknowledges financial support by an European Marie Curie Fellowship (FP7-PEOPLE-IEF-2008, Contract 236723). This research is funded by the Dutch Polymer Institute (DPI), project #657.

■

REFERENCES

(1) Anderson, B. J.; Zukoski, C. F. Rheology and microstructure of entangled polymer nanocomposite melts. Macromolecules 2009, 42, 8370−8384. (2) Chang, C. W.; van Spreeuwel, A.; Zhang, C.; Varghese, S. PEG/ clay nanocomposite hydrogel: a mechanically robust tissue engineering scaffold. Soft Matter 2010, 6, 5157−5164. (3) Lin, W. C.; Fan, W.; Marcellan, A.; Hourdet, D.; Creton, C. Large strain and fracture properties of poly(dimethylacrylamide)/silica hybrid hydrogels. Macromolecules 2010, 43, 2554−2563. (4) Dach, B. I.; Rengifo, H. R.; Turro, N. J.; Koberstein, J. T. Crosslinked “matrix-free” nanocomposites from reactive polymer-silica hybrid nanoparticles. Macromolecules 2010, 43, 6549−6552. (5) Deng, F.; Ito, M.; Noguchi, T.; Wang, L. F.; Ueki, H.; Niihara, K.; Kim, Y. A.; Endo, M.; Zheng, Q. S. Elucidation of the reinforcing mechanism in carbon nanotube/rubber nanocomposites. ACS Nano 2011, 5, 3858−3866. 12317

dx.doi.org/10.1021/la301917e | Langmuir 2012, 28, 12311−12318

Langmuir

Article

(26) Lemmers, M.; Spruijt, E.; Beun, L.; Fokkink, R.; Leermakers, F.; Portale, G.; Cohen Stuart, M. A.; van der Gucht, J. The influence of charge ratio on transient networks of polyelectrolyte complex micelles. Soft Matter 2012, 8, 104−117. (27) Cardenas, J. Surface charge of silica determined using X-ray photoelectron spectroscopy. Colloids Surf., A 2005, 252, 213−219. (28) Lindner, P.; Zemb, T. Neutrons, X-rays and Light: Scattering Methods Applied to Soft Condensed Matter; Elsevier Science B.V.: Amsterdam, 2002. (29) Kohlbrecher, J.; Bressler, I. SASfit, 2010. (30) Cohen Stuart, M. A.; Fleer, G. J.; Scheutjens, J. M. H. M. Displacement of polymers 2. experiment - determination of segmental adsorption energy of poly(vinylpyrrolidone) on silica. J. Colloid Interface Sci. 1984, 97, 526−535. (31) Lafuma, F.; Wong, K.; Cabane, B. Bridging of colloidal particles through adsorbed polymers. J. Colloid Interface Sci. 1991, 143, 9−21. (32) Nelson, A.; Jack, K. S.; Cosgrove, T.; Kozak, D. NMR solvent relaxation in studies of multicomponent polymer adsorption. Langmuir 2002, 18, 2750−2755. (33) Cohen Stuart, M. A.; Besseling, N. A. M.; Fokkink, R. G. Formation of micelles with complex coacervate cores. Langmuir 1998, 14, 6846−6849. (34) van der Burgh, S.; de Keizer, A.; Cohen Stuart, M. A. Complex coacervation core micelles. colloidal stability and aggregation mechanism. Langmuir 2004, 20, 1073−1084. (35) Lemmers, M.; Voets, I. K.; Cohen Stuart, M. A.; van der Gucht, J. Transient network topology of interconnected polyelectrolyte complex micelles. Soft Matter 2011, 7, 1378−1389. (36) Cardenas, J. F. Surface charge of silica determined using X-ray photoelectron spectroscopy. Colloids Surf., A 2005, 252, 213−219. (37) van der Gucht, J.; Spruijt, E.; Lemmers, M.; Cohen Stuart, M. A. Polyelectrolyte complexes: bulk phases and colloidal systems. J. Colloid Interface Sci. 2011, 361, 407−422. (38) Grant, M. C.; Russel, W. B. Volume-fraction dependence of elastic moduli and transition-temperatures for colloidal silica-gels. Phys. Rev. E 1993, 47, 2606. (39) Trappe, V.; Weitz, D. A. Scaling of the viscoelasticity of weakly attractive particles. Phys. Rev. Lett. 2000, 85, 449. (40) Larson, R. The Structure and Rheology of Complex Fluids; Topics in Chemical Engineering; Oxford University Press, Inc.: New York, 1999. (41) Martin, J. E.; Wilcoxon, J.; Odinek, J. Decay of densityfluctuations in gels. Phys. Rev. A 1991, 43, 858−872. (42) Liu, R. G.; Gao, X.; Oppermann, W. Dynamic light scattering studies on random cross-linking of polystyrene in semi-dilute solution. Polymer 2006, 47, 8488−8494. (43) Adolf, D.; Martin, J. E. Ultraslow relaxations in networks evidence for remnant fractal structures. Macromolecules 1991, 24, 6721−6724. (44) Ren, S. Z.; Shi, W. F.; Zhang, W. B.; Sorensen, C. M. Anomalous diffusion in aqueous solutions of gelatin. Phys. Rev. A 1992, 45, 2416− 2422. (45) Blanco, M. C.; Leisner, D.; Vazquez, C.; Lopez-Quintela, M. A. Dynamic light scattering in transient reversible gels. Langmuir 2000, 16, 8585−8594. (46) Møller, P. C. F.; Fall, A.; Bonn, D. Origin of apparent viscosity in yield stress fluids below yielding. EPL 2009, 87. (47) Spruijt, E.; Sprakel, J.; Lemmers, M.; Cohen Stuart, M. A.; van der Gucht, J. Relaxation dynamics at different time scales in electrostatic complexes: time-salt superposition. Phys. Rev. Lett. 2010, 105, 208301. (48) Coussot, P.; Tabuteau, H.; Chateau, X.; Tocquer, L.; Ovarlez, G. Aging and solid or liquid behavior in pastes. J. Rheol. 2006, 50, 975− 994. (49) Ovarlez, G.; Barral, Q.; Coussot, P. Three-dimensional jamming and flows of soft glassy materials. Nat. Mater. 2010, 9, 115−119. (50) Ewoldt, R. H.; McKinley, G. H. Creep ringing in rheometry or how to deal with oft-discarded data in step stress tests! Rheol. Bull. 2007, 76. 12318

dx.doi.org/10.1021/la301917e | Langmuir 2012, 28, 12311−12318