Highly nanofilled polystyrene composite

Abstract

Highly filled systems, such as dental materials and tires, have some exceptional properties that make them very special for particular scientists and engineers. In this study, the thermal and dynamic properties of highly nanosilica-filled polystyrene were investigated. Thermal study predicts a phase in the filled system, named as adsorbed polymer, that has a different glass transition temperature (T _g) compared with the neat polymer. The adsorbed polymer seems to be responsible for special thermal properties of the highly filled system. The dynamic properties of the filled system are observed to have a similar trend as the thermal behavior at different particle sizes and concentrations, both increasing linearly with the increase of volume fraction of adsorbed polymer. However, at higher volume fractions or for smaller particles, this trend changes and the filler networking mechanism is considered to be the reason for this change. Effect of the filler network is studied through the Han plot and it is found that the contribution of the filler network to the dynamic behavior of the highly filled system increases by reducing the particle size and increasing the particle loading. Beside the particle size and concentration, the effect of filler surface physics on dynamic and thermal behavior of the highly filled system is investigated and it is found that surface modification of the particle surface with nonpolar groups tends to lower T_g and volume fraction for the adsorbed phase and lower strength of the filler network. In this work, the samples were prepared using the method of stabilizing suspension in polymer solution. For viscoelastic investigation, the dynamic rheometry in sweep mode was chosen, also for studying the thermal behavior, differential scanning calorimetric tests were performed. In addition, in order to study the structure of filler in low and highly filled samples, atomic force microscopic imaging was employed.

Keywords

Adsorbed phase microstructure surface modification highly filled filler network

Introduction

In recent years, a lot of attention has been paid to polymer nanocomposites due to their remarkable physical properties, especially their mechanical, thermal, and barrier properties.^1,2 The great diversity in viscoelastic properties of polymer nanocomposites is a consequence of nanoparticles’ contribution to molten polymer matrix.

The properties of composite systems depend on the properties of each component and the structure of the components in the composites. Since the composites structure varies with their dynamic history³ and thermodynamics,⁴ the structure of the system and consequently its properties could be controlled with different dynamic histories and changing the thermodynamics.

Some parameters like mixing time and method, and the time after mixing, determine the dynamic history systems. Capuano et al.⁵ prepared nanocomposites of poly(ethylene oxide) and fumed silica with melt-compounding and freeze-drying methods. They found systems that are prepared with various mixing methods have different properties and structures. Yin et al.⁶ showed same results. Wu et al. ⁷ studied the composites’ structure by tracing the time dependence of electrical resistivity and they found that the structure of the carbon black particle in polymeric matrix changes over time after mixing.

Bartholome et al.⁸ changed the surface physics of fumed silica and studied the structure of the resultant composites with transmission electron microscopic (TEM) micrographs. Those TEM micrographs show the morphology of the composites changes with particle surface physics. Moreover, composites containing particles with different types of surface modifier have different structures and properties.⁹

Trappe et al.¹⁰ described the jamming phase diagram for the suspension of two component systems. This diagram defines the border between the solid-like and liquid-like state of those systems. The variables of this diagram are volume fraction of the components, temperature or particle energy, and input energy to the systems. Based on this diagram for the polymer composites, increase in particle volume fraction and its surface energy would increase solidity of the systems and increase in temperature and applied stress on systems tend to refluidize the solid systems. The dependency of the properties of filled systems to the temperature, particle volume fraction and surface modification, and Payne phenomenon could be related to the phase diagram of the systems.

Polymer chains at the particle surface have different behavior compared with the free polymers. Nuclear magnetic resonance (NMR) study¹¹ predicts immobilized polymer layer around the particle surface and it is found that the adsorbed polymer transition on the particles surface has different glass transition temperature (T _g).^12,13 Blum and his coworkers did solid-state proton (¹H) NMR, Fourier transform infrared (FTIR), and differential scanning calorimetry (DSC) for the poly(methyl acrylate) (PMA) and polystyrene (PS) coated at the untreated and treated silica surface. The result showed that the mobility of the polymer chains at silica surface is lower compared to the free polymer and also the difference between the glass transition of the free and adsorbed polymer is higher for the polar polymers.¹⁴

In the filled systems, particles are in aggregate form and the adsorbed polymers cover them. The ratio of the single particle to the aggregate dimension could be considered as mixing quality or dispersion factor. With increase in particle volume fraction, aggregates become close enough to contact each other directly and create a three-dimensional (3-D) network of particles especially for polar particles in the media of nonpolar polymers.¹⁵ The possibility of particle network formation increases for the particles in the nanoscale.¹⁶ The number of particles in the nanoscale is high and particle dimension is in the order of gyration radius of polymer chains.¹⁷ The changes force the particles to trap in unstable direct network at highly filled system.

Kaully et al.^18,19 studied rheological properties of highly micro-filled systems with a capillary, dynamic, and rotational rheometer. The shear thickening is the unique behavior that is observed for highly loaded systems. They treated particle surfaces with fatty acids of various lengths. Changing the behavior to the shear thinning for highly filled systems with surface treatment is the remarkable observation of this study.²⁰

There is not enough information regarding nonlinear viscoelastic properties of filled materials with 3-D solid particles like silica, in comparison to polymers filled with treated clay nanoparticles, especially at low filler concentrations. However, the viscoelastic behavior of inorganic particles dispersed in polymer matrices from diluted solutions to molten polymer has been reported in some previous works.^21,22 The viscoelastic behavior of polymer nanocomposites could be attributed to interaction between three main effects. The first is the effect of polymer chains trapped between the filler surfaces, the second is due to the matrix molecular weight, and the last important factor is the effect of the polymer entanglement characteristics.²¹ One of the viscoelastic behaviors of filled polymers, usually referred as Payne effect, is the strain dependency of the viscoelastic properties of the filled rubbers or amorphous polymers, that is, the drop off of the storage modulus with rising deformation ratio.^23

–26 Recommended mechanisms by researchers for the Payne effect could be explained by either of the following: the Payne effect is a consequence of bonding and deboning between polymer chains and filler surfaces or filler structures breakdown at high strain amplitudes.^26,27 Since filler–matrix and filler–filler interactions both rely on filler characteristics (for instance surface treatment), the deboning theory and filler structure-based theories are difficult to differentiate experimentally.²⁶

In this study, dynamic rheometry and thermal behavior of highly nanosilica-filled polystyrene are investigated. To study the thermal behavior of polymer in composites, the DSC and rheometry are done. The data predict adsorbed polymer phase and it is found that the overall T _g value of composites increases with increase in particle concentration and decrease in particle size. DSC data of composites showed that the polymer at unmodified surface has higher T _g value compared with the polymer at modified surface. The filler network is considered responsible for the properties of the highly filled systems, and the effects of the particle size, concentration, and surface physics on filler network are studied through the cole–cole plot. Filler network strength increases with particle size reduction and increase in particle concentration. Modifying the particle surface with nonpolar group tends to decrease filler network strength. Atomic force microscopic (AFM) images are presented at different particle concentrations and surface physics to study the dispersion state.

Experimental

Materials

Industrial polystyrene grade 336 with density of 1.05 g cm⁻³ (ISO 1183) and a zero shear viscosity of 5000 Pa s (at T = 160°C) were purchased from En Chuan (Hsien, Taiwan). Nonporous nanosilica Aerosol 200, 90, OX50 (supplied by Degussa Chemical, Essen, Germany) and micro-silica (supplied by Ferroazna, Azna, Iran) were used as fillers for the sample preparation. Average primary hydrophilic particle sizes of nanosilica were 12, 20, and 40 nm. Also, micro-silica particles were hydrophilic, with an average primary particle size of 100 nm (0.1 lm). All silica particles have an individual density approximately 2.2 g cm⁻³. Nanosilica with 12 nm diameter was modified by vinyltriethoxysilane as described before.²⁸

Sample preparation

Hydrophobic silica particle was prepared using vinyltriethoxysilane (obtained from Merck, Germany) as a surface modifier. Aerosil was sonicated in toluene for 30 min and then vinyltriethoxysilane was added to the suspension, by twice its weight ratio, followed by magnetic stirring until the suspension was clear (about 3 h). Resulted suspension containing hydrophobic silica nanoparticles was washed with an excess of toluene, followed by filtration with the intention of removing the unreacted silane molecules. The purified product was afterwards dried at 165°C for less than 30 min.

In order to prepare composite samples, silica suspension in a polymer solution method as described earlier is used.²⁸ In this method, polystyrene was added to the silica particle suspension in toluene and sonicated for half an hour. Then, the polystyrene/silica suspension was stirred with a magnetic stirrer for 1 h. In order to breakdown the silica clusters and avoid silica particle sedimentation, during solvent evaporation, high-shear mechanical stirring (T25 Digital Ultra-Turrax, IKA, Germany) was used at 2000 r min⁻¹ and kept at room temperature for 6 h. The resulting solution was cast on a Teflon sheet followed by drying for 6 days and vacuum drying at 60°C for 1 day. Silica-filled polystyrene samples were molded into 1-mm thick and 25-mm diameter plates by hot pressing under 100 bars for 20 s at 200°C. The same procedure was used in the preparation of neat polystyrene sample and treated silica nanocomposites. Five types of polystyrene/silica composites were prepared. “N” refers to untreated nanoparticles and “T” refers to treated nanoparticles.

Characterization

For viscoelastic response measurement of polystyrene/silica composites, a strain-controlled rotary shear rheometer (Paar-Physica UDS 200, Graz, Austria) is used in plate–plate geometry. The temperature sweep rheometry was performed in a temperature range of 90–200°C at 1% deformation and frequency of 5 Hz to obtain T _g. The frequency sweep rheometry of neat polystyrene was performed in the frequency range of 0.01–500 Hz, at 1% deformation, and at the temperatures of 200°C. The strain sweep rheometry also has been used. Analysis of nonlinear behavior was performed in the deformation range of 0.1–60% at a frequency of 5 Hz and the temperature of 200°C.

For obtaining thermal T _g, in addition to dynamic T _g, DSC was conducted using Setaram DSC 141 (France). Samples were pressed in aluminum pans and were heated two times from the ambient temperature to 130°C with a heating rate of 5°C min⁻¹.

To study the microstructure and morphology of the nanocomposites, AFM (Veeco PC-Research, Plainview, New York, USA) was used. For low filled systems, the experiment is done in noncontact mode, whereas for highly filled systems, contact mode was used.

Results and discussions

Thermal study

Configuration of chains in the space changes at the solid or crystalline surfaces. With this change, the entropy of chains decreases and the change in configuration is related to the adsorption energy.²⁹ Chains that aren’t connected directly to the filler surface but are next to the directly adsorbed chains also have lower entropy compared with free chains, so these overall adsorbed chains form a shell near the particles. Polymer chains with different configuration have different T _g. Particle size, concentration, and surface properties are the parameters that can affect the overall T _g of the filled systems. Higher particle loadings and smaller particles provide more available surface for polymer chains adsorption.

In Figure 1(b), the effect of particle volume fraction on thermal T _g of nanocomposites is investigated. In the graph of DSC of the neat polystyrene, a peak appears, which is related to the T _g of polystyrene. Adding nanoparticles to the polystyrene changes the DSC graph of the composite. In this graph, a peak appears near the peak of the neat polystyrene and a peak at a higher temperature. The peak at the higher temperature is attributed to the existence of a phase with higher rigidity. This phase is related to the polymer chains that are affected by the surface of the particles or adsorbed chains. Therefore, this peak predicts T _g of the adsorbed chains. It is clear that with an increase in particle volume fraction the peak of the neat (free chains) and adsorbed chains are merged and appeared between the previous two peaks.

Figure 1.

The effect of the particle volume fraction and size on the T _g of the nanocomposites: (a) all data, (b) effect of particle volume fraction, and (c) effect of particle size. T _g: glass transition temperature.

In Figure 1(c), the effect of particle size on the thermal behavior of nanocomposites is shown. For a sample with 5% of 40 nm nanosilica, two peaks are appeared. In this sample, existence of the adsorbed chains is predicted. Smaller particle increases the contribution of adsorbed polymers, which is shown by merging the two peaks into a single peak that appears at a temperature between the T _g of the neat and adsorbed chains.

T _g for different nanosilica particles also can be obtained from viscoelastic data versus temperature, this kind of T _g could be called dynamic T _g versus the thermodynamic T _g obtained from the DSC test. Storage modulus of the PS-filled silica shows a wider glass transition region than the neat polymer. With increasing the silica content, storage modulus of PS-filled silica also increases. According to previous findings, T _g refers to the temperature at which the peak in the loss modulus curve occurs.^2,24 In Figure 2, the loss modulus of nanocomposites at different particle sizes and concentrations is shown. It is seen that addition of silica particles causes an increase in T _g. The increase of loss modulus at higher temperature and T _g shift could be related to hydrodynamic motion of filler particles,³⁰ interaction of filler with polymer,³¹ and interaction of fillers with each other or filler networks.^32,33 The hydrodynamic forces on particles are a function of shape and concentration of them. Filler–polymer interaction relates to the behavior of polymer chains at the filler surface.³⁴ Filler–filler interaction is related to agglomeration and structure of filler in a composite.³⁵

Figure 2.

Loss modulus of nanocomposites at different particle size and concentration.

There are some other studies that confirm these results, for example, Sternstein and Zhu²⁶ related the mechanical reinforcement of polymer matrix to these mechanisms. This trend is also reported by Goh et al.,³⁶ proving there is an optimum percent of nanoparticles to improve and reinforce the mechanical properties of nanocomposites.

In Figure 3, the thermal and dynamic T _g of the nanocomposites at various particle sizes and concentrations are presented. The experimental results show that reinforcement of nanocomposites and their T _g depend on the filler content for all sizes of nanosilica particles.

Figure 3.

T _g of PS/silica versus filler volume fraction for different sizes of nanosilica. Solid symbols represent DSC data and empty symbols represent the viscoelastic data. T _g: glass transition temperature; PS: polystyrene; DSC: differential scanning calorimetry.

Although the trend of dynamic T _g changes is similar to the ones shown in DSC results, the range of T _g values obtained by this method is around 100°C, even for neat polystyrene, whereas T _g values of DSC results are around 83–87°C for the same samples. This 15–20°C difference between dynamic and thermal T _gs could be attributed to the mechanism that each represents, that is, DSC defines the glass transition as a change in the heat capacity of the polymer matrix going from the glassy to rubbery state. In the DSC, the transition appears as a step transition and not as a peak that might be seen with a melting transition. Thermomechanical analysis (TMA) defines the glass transition in terms of the change in relaxation between chain segments, as the polymer goes from glassy to rubbery state. Effectively, each of these techniques measures a different physical attribute of the change from glassy to rubbery states. The DSC is measuring a heat flow effect, whereas the TMA is measuring a physical effect. Moreover, some polymers are more amenable to DSC or to TMA, because the transition is easier to observe using one technique over the other. All dynamic T _g points for complex samples were identified with the maxima of the tan δ absorption spectra. In fact, as mentioned before, there is usually a difference of 10–15°C or so between the values of T _g determined by the DSC and DMTA techniques.³⁷

T _g values of PS-filled silica versus filler volume fraction are plotted for both dynamic and thermal mode in Figure 3. The slope of T _g curve is a proper parameter for determining T _g shift velocity, which is due to nanoparticle size. This speed increases with decrease in particle size because of higher surface area of the small particles that are available for the polymer chains for adsorption. These results differ from Torkelson et al.³⁸ and the results of Bansal et al.³⁹ who reported both increase and decrease in T _g consider processing condition while pointing to these probable differences. The T _gs of nanocomposites estimated from the two tests of DSC and rheometry are different. In rheometry test, mechanical forces exist that could change the dynamics of polymer chains. Therefore, this difference could be related to the different mechanisms of relaxation in those tests, consequently the relaxation temperature of nanocomposites is found at higher temperature. Also the detection mechanism of T _g for those tests is different, which also results in different T _g detected from different methods.

It is clear from the data in Figure 3 that both types of T _g (thermal or dynamic) increase linearly with the volume fraction of particle for every size. The volume fraction of adsorbed polymer could be estimated from total surface of particles multiplied to the adsorbed layer thickness as follows:

φ_{a} = s υ ρ z,

where s is the surface area of filler particle that is available for the polymer chains for adsorption, υ is the volume fraction of filler in the composites, ρ is the density of particle structure, and z is the thickness of the adsorbed polymer layer. In this equation, volume fraction of the adsorbed polymer increases with particle volume fraction and overall dynamic or thermal T _g of the system increases with particle volume fraction. Thus, it is concluded that T _g of the system is a linear function of the adsorbed polymer volume fraction. In Figure 4, the dynamic T _gs of nanocomposites are plotted against the total surface of particles, that is, the volume fraction of adsorbed polymer.

Figure 4.

T _g of nanocomposites versus total surface of particles. T _g: glass transition temperature.

In order to estimate the total surface, surface areas of 40, 20, and 12 nm particles are considered as 37.5, 60, and 80, respectively.⁴⁰ From this figure, one can infer that the T _g of nanocomposites is only a function of amount of adsorbed polymer.

With an increase in filler volume fraction, available surface area for adsorption of polymer chain increases. Adsorbed polymer chains have different properties compared with the free chains,¹⁴ thus with an increase in the volume fraction of filler, composite rheological properties change.

Dynamic study

Storage modulus of the linear region versus filler volume fraction for different particle sizes is demonstrated in Figure 5. More filler content results in higher storage modulus that could be due to particle networks or filler–polymer interaction. Reinforcement of matrix increases by filler size reduction. Hence, in a constant amount of filler, silica with the lowest diameter will show the highest influence on polymer reinforcement. It can be understood from Figure 5 that by decreasing the filler diameter, effect of particles on storage modulus is more pronounced, that is, reinforcement of polymer will happen at lower filler content. By decreasing filler diameter at a constant amount of nano filler, volume fraction of adsorbed polymer chains on filler surface increases due to the increase of surface area.⁴¹ Hence, at a constant filler volume fraction, more increase in storage modulus is expected.

Figure 5.

Storage modulus versus filler volume fraction.

It is seen that the T _g of the studied nanocomposites increases linearly with particle volume fraction because of the increase in the volume fraction of adsorbed polymer. Rheological properties of the nanocomposites like storage modulus, as seen in Figure 5, at low volume fraction increase with this parameter because of hydrodynamic motion of particles and interaction of the polymer chains with particle surface. On the other hand, the number of particles in a unit volume increases with an increase in particle concentration, thus interparticle distance decreases. The reduction in the interparticle distance tends to increase the filler–filler interaction and tendency of the fillers to form a network. Highly filled composite properties are related to filler concentration because of relation of filler network properties to its concentration.

Effect of filler particle size on composite properties is similar to filler particle concentration. With the decrease in particle size, the number of particles in a given volume increases and the available surface of filler for adsorption of chains increases as well.

Based on the result of the DSC tests at those volume fractions, systems show two phases with different properties. Appearance of two peaks in the DSC graph (Figure 1) of low-filled samples confirms the existence of the adsorbed polymer around nanoparticles too and it is estimated from DSC data that the transition temperature for this type of chains is 8–10° higher than the free polymers. In our previous study,¹⁶ we predicted the structure for low-filled suspension from sedimentation of the nanoparticles in the media of polystyrene solution. The study proposed an adsorbed layer around particles. Berriot et al.¹¹ studied adsorbed layer with ¹H NMR and they found that this layer behaves as an immobilized glassy shell. The morphology of the low-filled system is be presented in Figure 6(a).

Figure 6.

Structure of particle in nanocomposite: (a) low filled and (b) highly filled.

At a certain volume fraction, called critical volume fraction, the functionality of those properties to the volume fraction is changed and increases rapidly. Thermal studies (both DCS and TMA) of those systems show single peak that is due to merging of adsorbed and free polymer peaks, and the location of this peak is between those two peaks. Thermal results of highly filled systems do not show a drastic change in the location and intensity of glass transition peak compared with the low-filled systems. Therefore, the structure and behavior of the particles are responsible for the rapid change in the rheological properties. Sedimentation tests for systems containing higher amount of nanoparticles predict network of particles with direct connection. A probable structure for the particles in the highly filled systems is presented in Figure 6(b).

Dense structure of particles, labeled as d in Figure 6, is an aggregate with d as its effective diameter. An aggregate consists of particles that adhere to each other strongly either by physical or chemical interactions, where polymer chains can’t diffuse into their structure. In this figure, white and gray zones represent adsorbed and free polymer, respectively. In Figure 7, AFM image of the low, percolation threshold and highly filled composites containing 12 and 20 nm particles is presented. This classification is selected from data in Figure 5. The percolation threshold is the concentration where the pattern of behavior is changed. Based on the data of this figure, the percolation threshold is 2.5, 5, and 15% for composites that contain 12, 20, and 40 nm nanoparticles, respectively. Discrete white zones in low-filled composites that represent nanoparticle aggregates confirm the presented structure for low-filled systems in Figure 6. AFM imaging of highly filled systems shows that the filler particles form 3-D structure and aggregates contact each other directly.

Figure 7.

AFM image of low filled systems: (a) low filled, (b) percolation threshold, and (c) highly filled. AFM: atomic force microscopic.

The volume of aggregate is partially filled with voids so its density is lower than the primary particle. Creation of aggregates is inevitable in the filled system and its dimension determines the state of dispersion and change with the method of mixing. Capuano et al.⁵ prepared a system with two different mixing methods and they found that the properties of a system are a function of the mixing quality and aggregate dimension. This dimension varies with the quality of mixing, which is responsible for this difference.³

Han plot (G′ versus G′′) is one of the methods for determining the proportion of solid-like behavior of composites and their rheological characterization.^42,43 For studying the effect of particle network on the rheological behavior of systems toward filler size and its concentration in the matrix G′ versus G′′ is plotted in Figure 8. Since the Han plot does not have the same trend for different samples, it could be deduced that rheological performance is affected by filler size, its concentration in the polymer, and the preparation process condition. In the previous work of Najjar,⁴⁴ none of the curves of their nanocomposites passed the line of G′=G′′, because of the process condition and surface treatment of the used nanoparticles.

Figure 8.

Storage modulus versus loss modulus.

In Han curve, increase of G′ against G′′ can be the result of existence of network-like structures in the system. For studying this pattern, the line of G′=G′′ is plotted in the Han curve and data on every sample with higher distance from this line at the left side of this curve have higher influence than filler network, which is shown in the lower diameter fillers, especially at high filler loadings. The result of this figure represents that the contribution of the particle network on the storage modulus increases with particle volume fraction and reduction in particle size.

Surface physics study

Filler network strength or filler–filler interaction is a function of filler and polymer surface tension⁴⁵ and also interparticle distance,⁴⁶ which itself is a function of particle size or surface area per weight and particle concentration. With increase in surface tension of filler, filler network strength increases. For constant filler concentration and same dispersion state, filler network strength changes with filler surface tension. Surface tension is the energy needed to separate fillers from each other. With decrease in surface tension of filler, energy for dispersion of filler decreases and surface area of filler that is available for free polymer increases, but the adhesion of polymer to filler surface or filler–matrix interaction decreases.^9,25,47 The effect of increase in surface area and decrease in filler surface adhesion with modification with nonpolar group has an inverse effect on viscoelastic properties of composites, thus a change in filler surface tension can increase or decrease the viscoelastic properties of composites.

Data of linear viscoelasticity (Figure 5) showed that the filler network has great effect on the rheological properties of highly filled systems. The elastic nature of the filler network is the interactions between neighbor particles. These interactions come from the polar and London forces. The elastic modulus of the filler network could be related to the amount of forces needed to separate two connected particles. Cohesive forces between two particles in contact are predicted as follows⁴⁸:

f_{c} = \frac{3}{2} π Γ R,

where f _c is cohesive force, R is radius of particles, and Γ is particle surface energy. Beside the cohesive force, the number of total connection between particles and the arrangement of particles in the system are the other parameters that determine the elastic properties of the filler network. Particle size and concentration determine the total number of connection in the system. The effect of the particle size and concentration on the properties of the low and highly filled systems is presented in Figure 5.

From equation (2), it can be inferred that the systems with particles having higher surface energy have higher modulus. For studying this parameter, particle surface is modified with nonpolar group,²⁵ and its surface energy is estimated from contact angle tests. Surface energies of modified and unmodified particles are 31.4 and 75.8 mJ m⁻ ², respectively.

Dependency of viscoelastic properties of the samples with modified and unmodified on the strain amplitude is studied. Storage modulus (G′) versus strain is shown in Figure 9. It can be seen that storage modulus decreases with surface modification of particle. This reduction is related to the lower adsorbed polymer fraction and lower filler network strength because of the lower filler–polymer and filler–filler interactions, respectively.

Figure 9.

Storage modulus of modified and unmodified 12 nm silica.

Filler agglomeration and network formation are responsible for higher reinforcement and nonlinear behavior of materials at high strain levels. As it is obvious in Figure 9, the fall in storage modulus happens at lower strains where high filler concentration is added. In all of the nanocomposites, by increasing the filler volume fraction, which results in higher filler surface area, interfacial layer volume enhances. As mentioned in our previous work,²⁴ increase in the interfacial layer volume ratio results in more low-mobility polymer segments that increase T _g and modulus. Arevalillo et al.⁴⁹ also proved that in complex systems such as nanocomposites, double constraints are suffered by the polymer chain. The first constraint is the blocking effect due to the nanostructure (ordered polymer domains or nanoparticles) on the whole chain motion and the second one is the hindering effect of entanglements on local motions.

The network of particles has viscoelastic behavior.^10,50 In this network, particles are in chain-like structure. Connected particle chains have elastic behavior because of the enthalpic and entropic nature. Entropic nature of a chain comes from its configuration. In this case, particles are connected to each other with reversible bond.⁵¹ Cohesive forces like London and hydrogen bond are responsible for the bonding between particles. Equation (2) is an estimation of the amount of these forces. Therefore, reversible bond with enthalpic nature between particles results in elastic behavior for particle chains. Viscose behavior of a system is related to the dissipating energy during movement. Friction of particles with suspending media and rupture of the bond between particles are mechanisms that determine and control viscose behavior of the viscoelastic system. Reduction in the storage modulus with surface modification and strain amplitude is an evidence for existence of filler network. In addition, it shows that the contribution of enthalpic on the nature of elasticity in highly filled systems is higher than the entropic effects that is originated from filler network.

Loss modulus relates to the energy dissipated per cycle. Dissipated mechanism in a composite system outlines the irreversible deformation and breakdown of the 3-D structure and relative movement of components. With the increase in amplitude, the 3-D structure of fillers is distorted, since this distortion needs energy. The effect of a filler network on the viscoelastic properties of a composite is a function of direct interaction of particles on each other, number of particles surrounding a particle, and interparticle distance. Therefore, particle size, concentration, and surface physics are responsible parameters for loss modulus alteration. In Figure 10, the effect of surface modification at different particle concentrations for composites containing 12 nm particles is shown.

Figure 10.

Effect of particle concentration and surface modification on loss modulus.

According to Figure 10, a noticeable peak in the loss modulus for highly filled system is observed. This peak is related to the filler network strength.⁴⁶ From this figure, it is clear that the surface modification reduced loss modulus and intensity of appeared peak. The surface area under the loss modulus curve is estimated, and it is found that this parameter is decreased with surface modification. Thus, it can be concluded that surface modification reduced the strength of filler network.

Blum and Krisanangkura¹⁴ have used the FTIR spectra of the neat and covered silica particles with PMMA and found that the hydrogen bonding between the particle surface and polymer chains is the driving force for the polymer adsorption on the particle surface. Also, they proved with NMR that the adsorbed polymer chains have lower mobility compared with the free chains. At last, they concluded from the DSC tests that the modification of the particle surfaces with nonpolar groups decreases the particle–polymer interaction because of the lower hydrogen bonding. For studying the mobility of the polymer chains at surfaces with different physics, DSC test results for modified and unmodified particle-filled nanocomposites are performed and shown in Figure 11.

Figure 11.

DSC test results for composites containing modified and unmodified nanoparticles. DSC: differential scanning calorimetric.

The results of DSC test for systems with different particle surface properties show that transition peaks shift to the lower temperature. This means polymer at surface with lower polarity and surface tension has higher mobility, that is, lower polar surface induces lower effect on the polymer chains.

As mentioned above, peak in the lower temperature is related to the free polymer and the other peak is the adsorbed peak. Intensity of the adsorbed peak is decreased with modification, thus properties and volume fraction of the adsorbed polymer also decrease with modification.

In Figure 12, the Han plot is presented for studying the effect of modification on the filler network. Higher storage modulus, compared to the loss modulus, represents higher solidity. It is clear from the presented data in this figure that the modification of particle surface with nonpolar group reduces the strength of the formed particle network in the highly filled system. It is shown that the properties of the solid filler network are functions of the particle size and concentration. Thermal behavior and storage modulus of the modified particle-filled system are studied, and it is found that the storage modulus and the effect of filler on the behavior of the polymer at solid interface are decreased with modification. From the result of the Han plot and the properties of the highly filled system, the change in those properties with modification could be related to the effect of the modification on the particle network strength.

Figure 12.

Han plot for studying the effect of surface modification.

In this study, the effect of particle size and concentration on the thermal and rheological properties of highly filled systems are examined. It is found that the overall T _g of these systems increases with particle size and concentration because of the increase in volume fraction of adsorbed polymer on the particle surface. At low filler volume fraction, rheological properties of filled systems have the same trend as the T _g of the system and vary with the filler–polymer interaction. For all ranges of the particle size, at higher concentration of particles, this trend is changed and beside the particle–polymer interaction, another mechanism is considered that controls rheological properties of highly filled systems. This mechanism is attributed to the filler networking, and the effect of the particle size and concentration of this mechanism are studied with Han plots. It is seen that the contribution of the filler network on the properties of highly filled systems increases with particle size reduction and increases with particle volume fraction. The change in the particle size and concentration varied the number of particle–particle connection in the network of particles. Strength of the particle network is a function of particle connection numbers, particle arrangements in space, and particle–particle interactions. To study the effect of particle–particle interaction on the filler network, two particles with different surface properties are used, and it is found that particle–particle interaction changes with particle surface tension. Lower particle surface tension tends to lower the particle–particle interaction and filler network strength, and consequently lower storage modulus.

Conclusion

The highly filled systems like dental materials and tires are important to engineers in the view of the processing and properties. Rheology as a powerful tool is used to estimate processing condition and properties of these systems. In this study, thermal behavior and rheology of the highly nano-filled systems are investigated. The thermal study predicts the existence of an adsorbed polymer phase where this phase is responsible for increase in the overall T _g of the filled system. The effect of the particle size and concentration on thermal behavior of nanocomposites are studied, and it is found that T _g of filled polymers increases with increase of particle loading and decrease of particle size. This is attributed to the increase in available surface area of particle for polymer chains to adsorb. At low particle volume fraction, rheological properties and thermal behavior have the same trend. At higher particle volume fraction or lower particle size, rheological properties have different trend compared with the thermal behavior. These differences are related to the filler network. The effect of particle diameter and loading on the contribution of filler network on the properties of the filled systems is presented through the Han plots, and it is found that the created filler network of smaller particles or networks of highly filled systems is stronger. The theory predicts that the strength of the filler network is a function of the physics of the particle surface. For studying this parameter, two particles with different surface tension are chosen and their resultant composites rheological properties and thermal behavior are investigated. Rheological properties are reduced with surface modification with nonpolar groups and thermal study shows the amount of adsorbed polymer, and the effect of surface on the properties of adsorbed polymer is decreased with this change. Based on the Han plot, it is concluded that the network of the modified particles is looser compared with the other ones.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Chen

Yang

. Synthesis of epoxy–montmorillonite nanocomposite. J Appl Polym Sci 2002; 86: 414–421.

Mortezaei

Famili

MHN

Kokabi

. Influence of the particle size on the viscoelastic glass transition of silica-filled polystyrene. J Appl Polym Sci 2010; 115: 969–975.

Bumm

White

Isayev

. Breakup of silica agglomerates in corotating twin-screw extruder: modeling and experiment. J Elastomers Plast 2014; 46: 527–552.

Yeganeh

Goharpey

Foudazi

. Anomalous phase separation behavior in dynamically asymmetric LCST polymer blends. RSC Adv 2014; 4: 12809–12825.

Capuano

Filippone

Romeo

. Universal features of the melt elasticity of interacting polymer nanocomposites. Langmuir 2012; 28: 5458–5463.

Yin

C-L

Liu

Z-Y

Yang

. Effect of processing method on morphological and rheological properties of PC/CaCO₃ nanocomposites. Polym Plast Technol Eng 2009; 48: 788–793.

Asai

Sumita

. Carbon black as a self-diagnosing probe to trace polymer dynamics in highly filled compositions. Macromolecules 2002; 35: 1708–1713.

Bartholome

Beyou

Bourgeat-Lami

. Viscoelastic properties and morphological characterization of silica/polystyrene nanocomposites synthesized by nitroxide-mediated polymerization. Polymer 2005; 46: 9965–9973.

Robertson

Lin

Rackaitis

. Influence of particle size and polymer-filler coupling on viscoelastic glass transition of particle-reinforced polymers. Macromolecules 2008; 41: 2727–2731.

10.

Trappe

Prasad

Cipelletti

. Jamming phase diagram for attractive particles. Nature 2001; 411: 772–775.

11.

Berriot

Lequeux

Monnerie

. Filler–elastomer interaction in model filled rubbers, a ¹H NMR study. J Non-Cryst Solids 2002; 307–310: 6.

12.

Berriot

Montes

Lequeux

. Evidence for the shift of the glass transition near the particles in silica-filled elastomers. Macromolecules 2002; 35: 7.

13.

Montes

Lequeux

Berriot

. Influence of the glass transition temperature gradient on the nonlinear viscoelastic behavior in reinforced elastomers. Macromolecules 2003; 36: 8107–8118.

14.

Blum

Krisanangkura

. Comparison of differential scanning calorimetry, FTIR, and NMR to measurements of adsorbed polymers. Thermochimica Acta 2009; 492: 55–60.

15.

Filippone

Romeo

Acierno

. Viscoelasticity and structure of polystyrene/fumed silica nanocomposites: filler network and hydrodynamic contributions. Langmuir 2009; 26: 2714–2720.

16.

Kourki

Famili

MHN

. Particle sedimentation: effect of polymer concentration on particle-particle interaction. Powder Technol 2012; 221: 137–143.

17.

Leblanc

. Rubber-filler interactions and rheological properties in filled compounds. Progr Polym Sci 2002; 27: 61.

18.

Kaully

Siegmann

Shacham

. Rheology of highly filled natural CaCo₃ composites. I. effects of solid loading and particle size distribution on capillary rheometry. Polym Compos 2007; 28: 512–523.

19.

Kaully

Siegmann

Shacham

. Rheology of highly filled natural CaCO₃ composites. II. Effects of solid loading and particle size distribution on rotational rheometry. Polym Compos 2007; 28: 524–533.

20.

Kaully

Siegmann

Shacham

. Rheology of highly filled natural CaCO₃ composites: IV. Effect of surface treatment. Polym Adv Technol 2007; 18: 696–704.

21.

Cassagnau

. Payne effect and shear elasticity of silica-filled polymers in concentrated solutions and in molten state. Polymer 2003; 44: 2455–2462.

22.

Cassagnau

. Melt-rheology of organoclay and fumed silica nanocomposites. Polymer 2008; 49: 2183–2196.

23.

Karrabi

Mohammadian-Gezaz

. The effects of carbon black-based interactions on the linear and non-linear viscoelasticity of uncured and cured SBR compounds. Iran Polym J 2011; 20: 15–27.

24.

Mortezaei

Famili

MHN

Kalaee

. Effect of immobilized interfacial layer on the maximum filler loading of polystyrene/silica nanocomposites. J Reinf Plast Compos 2011; 30: 593–599.

25.

Mortezaei

Famili

MHN

Kokabi

. The role of interfacial interactions on the glass-transition and viscoelastic properties of silica/polystyrene nanocomposite. Compos Sci Technol 2011; 71: 1039–1145.

26.

Sternstein

Zhu

A-J

. Reinforcement mechanism of nanofilled polymer melts as elucidated by nonlinear viscoelastic behavior. Macromolecules 2002; 35: 7262–7273.

27.

Cassagnau

Melis

. Non-linear viscoelastic behaviour and modulus recovery in silica filled polymers. Polymer 2003; 44: 6607–6615.

28.

Mortezaei

Farzi

Kalaee

. Evaluation of interfacial layer properties in the polystyrene/silica nanocomposite. J Appl Polym Sci 2011; 119: 2039–2047.

29.

Sarvestani

Jabbari

. Modeling and experimental investigation of rheological properties of injectable poly(lactide ethylene oxide fumarate)/hydroxyapatite nanocomposites. Biomacromolecules 2006; 7: 1573–1580.

30.

Havet

Isayev

. A thermodynamic approach to the rheology of highly interactive filler-polymer mixtures. Part II. Comparison with polystyrene/nanosilica mixtures. Rheol Acta 2003; 42: 9.

31.

Sarvestani

Jabbari

. The role of filler-matrix interaction on viscoelastic response of biomimetic nanocomposite hydrogels. J Nanomater 2008; 2008: 1–9.

32.

Heinrich

Klüppel

. Recent advances in the theory of filler networking in elastomers. Adv Polym Sci 2002; 106: 44.

33.

Fröhlich

Niedermeier

Luginsland

H-D

. The effect of filler–filler and filler–elastomer interaction on rubber reinforcement. Compos A Appl Sci Manuf 2005; 36: 449–460.

34.

Léger

Raphaël

Hervet

. Surface-anchored polymer chains: their role in adhesion and friction. Adv Polym Sci 1999; 138: 41.

35.

Santos

PHS

Campanella

Carignano

. Brownian dynamics study of gel-forming colloidal particles. J Phys Chem B 2010; 114: 13052–13058.

36.

Goh

. Dynamic mechanical behavior of in situ functionalized multi-walled carbon nanotube/phenoxy resin composite. Chem Phys Lett 2003; 373: 277–283.

37.

Huh

P-H

Choi

M-G

. Effect of salt concentration on the glass transition temperature and ionic conductivity of poly (ethylene glycol)-polyurethane/LiClO₄ complexes. Macromol Res 2004; 12: 422–426.

38.

Roth

McNerny

Jager

. Eliminating the enhanced mobility at the free surface of polystyrene: fluorescence studies of the glass transition temperature in thin bilayer films of immiscible polymers. Macromolecules 2007; 40: 2568–2574.

39.

Bansal

Yang

. Quantitative equivalence between polymer nanocomposites and thin polymer films. Nat Mater 2005; 4: 693–698.

40.

Kourki

Mortezaei

Famili

MHN

. Prediction of the viscoelastic response of filler network in highly nanofilled polymer composites. J Compos Mater. 2015; 49(30): 3799–3807.

41.

Pandis

Logakis

Kyritsis

. Glass transition and polymer dynamics in silver/poly(methyl methacrylate) nanocomposites. Eur Polym J 2011; 47: 1514–1525.

42.

Ezzati

Ghasemi

Karrabi

. Correlation between the rheological behaviours and morphologies of PP/EPDM blends in various dynamic vulcanization systems. Iran Polym J 2008; 17: 265.

43.

Kim

Park

Kim

. Multiwall-carbon-nanotube-reinforced poly(ethylene terephthalate) nanocomposites by melt compounding. J Appl Polym Sci 2007; 103: 1450–1457.

44.

Disfani

Jafari

. Assessment of intertube interactions in different functionalized multiwalled carbon nanotubes incorporated in a phenoxy resin. Polym Eng Sci 2013; 53: 168–175.

45.

Mahaling

Kumar

Rath

. Effects of rubber–filler interaction on the developments of physical, mechanical, and interfacial properties of vamac–silica nano-composites. J Elastomers Plast 2007; 39: 253–268.

46.

Kourki

Mortezaei

Famili

MHN

. Filler networking in the highly nanofilled systems. J Thermoplast Compos Mater. 2016; 29(8): 1047–1063.

47.

Lee

(2000) Adhesion at polymer-solid interfaces: sticker and receptor group effects. ProQuest Dissertations and Thesis. Chemical Engineering, University of Delaware, USA.

48.

Moreno-Atanasio

. Energy dissipation in agglomerates during normal impact. Powder Technol 2012; 223: 12–18.

49.

Arevalillo

Muñoz

Calafell

. Thermo-mechanical properties linked to rheological features in high molecular weight SEBS copolymers: effect of styrene content and ethylene/butylene proportion. Polym Test 2012; 31: 849–854.

50.

Yziquel

Carreau

Tanguy

. Non-linear viscoelastic behavior of fumed silica suspensions. Rheologica Acta 1999; 38: 14–25.

51.

Zweistra

HJA

Besseling

NAM

. Mean chain length of adsorbed supramolecular polymers. Phys Rev Lett 2006; 96: 078301.