SciELO - Scientific Electronic Library Online

vol.23 issue1Advanced ceramics: evaluation of the mechanism of stock removal and ground surface quality author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Services on Demand




Related links


Journal of the Brazilian Society of Mechanical Sciences

Print version ISSN 0100-7386

J. Braz. Soc. Mech. Sci. vol.23 no.1 Rio de Janeiro  2001 

Effect of Particle Morphology on the Mechanical and Thermo-Mechanical Behavior of Polymer Composites


R. L. Oréfice,
Federal University of Minas Gerais.
LEPCom-Laboratory of Polymers and Composites Engineering
Department of Metallurgical and Materials Engineering
Rua Espírito Santo 35, 2° andar
30160-030 Belo Horizonte, MG. Brazil
L. L. Hench,
Imperial College-Department of Materials Science
London. UK
A. B. Brennan
University of Florida
Department of Materials Science and Engineering
32611 Gainesville, FL. USA



Fiber reinforced polymer composites have been used in many applications, such as in automobile, aerospace and naval industries, due basically to their high strength-to-weight and modulus-to-weight, among other properties. Even though particles are usually not able to lead to the level of reinforcement of fibers, particle reinforced polymer composites have been proposed for many new applications due to their low cost, easy fabrication and isotropic properties. In this work, polymer composites were prepared by incorporating glass particles of different morphologies on poly(aryl sulfones) matrices. Particles with aspect ratios equal to 1, 2.5 and 10 were used. The prepared composites were characterized using electron microscopy and thermal analysis. Mechanical properties of the composites were evaluated using a four-point bending test. The thermo-mechanical behavior of the obtained composites was also investigated. The results showed that the morphology of the particles alter significantly the mechanical properties of composites. Particles with larger values of aspect ratio led to large elastic modulus but low levels of strain at failure. This result was explained by modeling the thermo-mechanical behavior of the composites using a viscoelastic model. Parameters of the model, obtained from a Cole-Cole type of plot, demonstrated that interactions at the polymer-reinforcing agent interface were higher for composites with large aspect ratio particles. Higher levels of interactions at interfaces can lead to higher degrees of stress transfer and, consequently, to composites with large elastic modulus, as experimentally observed.
: Composites, particulate, viscoelastic model




Although many definitions of "composites" are available, they can be described as materials comprised of two or more constituents with very distinct compositions, structures and properties separated by an interface. The aim in producing composites is to combine different materials in a single device with properties that cannot be obtained from the individual components. Therefore composites for optical, structural, electrical, opto-electronic, chemical and other applications are easily found in modern devices and systems. During the past 30 years, there has been a substantial development of composites for structural applications. The main support of this tendency is the possibility of producing composites with high mechanical properties and low density that can replace traditional materials such as steel and wood. The combination of high performance polymers with high modulus-high strength ceramic or polymer fibers allowed the production of composites with a group of properties per weight superior than those of steel, aluminum and others (Mallick, 1993).

Another important feature of composites is that they can also be easily designed to fit in a specific application due to the capability of having their properties tailored by changing one of a series of variables. Some of these variables are the type, concentration, size and shape of the constituents. Among these variables, the shape of the reinforcing agent is a component of recognized importance in the design of composites. The influence of the aspect ratio of short fibers on the properties of composites is well documented and dictates the overall stress transferability phenomenon. Free fiber ends do not contribute to stress transfer. Stress is built progressively from the fiber ends up to a certain length (fiber critical size) where the stress transferred to the fibers reaches the characteristic maximum value for a specific system (Mallick, 1993).

In terms of particulate composites, model composites often use spherical particles that are more easily fitted in mathematical models. In this case, tensile stresses lead to a stress field at the poles of the spheres that ultimately can cause debonding (Mallick, 1993). However, in most of the engineering cases, fillers are usually brittle and produced by milling processes. The resulting particle shapes can be very complex and sometimes difficult to characterize. Glass particles, for example, when produced by milling can have very angular, needle-like shapes with aspect ratios other than 1 (Rothon, 1995). Thus, in order to design properly composites reinforced by non-spherical agents, it is important to evaluate the relationship between the morphology of the phases and the properties of the systems, including interfacial properties.

The aim of this work is to evaluate the effect of particle shape on the properties and thermo-mechanical behavior of composites. The hypothesis that interfacial properties can also be affected by the shape of the constituents will also be tested.


Experimental Procedure

The effect of the shape of reinforcing agents on the properties and behavior of polymer composites was investigated by preparing composites with the following reinforcing agents:

1. Soda-lime-phosphate silicate glass particles (125-106 µm), prepared by ball milling a batch of melted glass (properties of the bulk glass: density = 2.65 g/cm3, elastic modulus = 40 GPa, bending strength = 60 MPa).

2. Soda-lime-phosphate silicate glass spheres* (125-106 µm), prepared by passing particles obtained similarly as above, through a heated (1000°C) column (Wilson at al., 1996).

3. Calcium metaphosphate short fibers** with average aspect ratio equal to 10 and size between 50 to 100 µm (Kasuga et al., 1995) Those short fibers were obtained by dissolving one of the two phases of a calcium phosphate glass-ceramic. The reported properties of the original multiphase bulk glass-ceramic were: 400-600 MPa bending strength and 60 GPa elastic modulus. The elastic modulus of the final material (fiber) after dissolution was predicted to be close to the modulus of the glass particles (i.e. 40 GPa) by using microhardness tests. The surface chemistry of the fibers is equivalent to the surface chemistry of the glass particles used, since both have a surface layer rich on calcium and phosphate ions.

Ball milled glass particles had their aspect ratio measured by optical microscopy. The results showed that particles with 125-106 µm in size have an average aspect ratio of 2.5. A representative histogram of the distribution of aspect ratios of glass particles is shown on the top of Figure 1.



Structural density of the calcium metaphosphate fiber was measured by helium pycnometry and a value close to 3.0 g/cm3 was obtained.

Poly(aryl sulfones) (PAS) composites were prepared by dissolving the polymer in a suitable solvent and adding the reinforcing agents to the solution. The solvent was then removed, leading to the production of homogeneous polymer-reinforcing agents blends. The blends were then hot-pressed at 220°C. The measured densities of individual components were used to prepare composites with 20% volume fraction of reinforcing agent.

The dynamic thermo-mechanical behavior of the composites was analyzed by DMS (Dynamic Mechanical Spectroscopy). A Seiko DMS110 interfaced with a Seiko SDM5600H Rheostation was used to study the thermo-mechanical behavior of samples in flexural mode. Frequencies used were 0.1, 0.5, 1, 5 and 10 Hz from -140°C to 250°C at a heating rate of 0.75°C/min in nitrogen atmosphere.

Micrographs were obtained with a Scanning Electron Microscopy (SEM) JEOL 6400.

Poly(aryl sulfones) composites were cut and polished (SiC sand paper # 600). Mechanical properties of samples were measured using four point bending test apparatus in an Instron machine. The ASTM D790M-92 protocol was used during the test.


Results and Discussion

Mechanical four-point bending test was performed on composites and results are depicted in Figure 1. In this figure, values for elastic modulus, strength and strain at failure are shown for all the composites studied. The highest modulus was obtained for the short fiber reinforced composite, while the composite reinforced with spherical particles showed the lowest modulus. Again, large aspect ratios, i.e. large surface areas per volume, can lead to higher levels of stress transferability since this phenomenon is governed by shear mechanisms between matrix and fiber/filler at the interface. Moreover, polymer entrapped in angular shaped particles leads to higher levels of stress transfer.

Since strength of the different types of composite was almost the same, the total energy to failure (an indication of toughness) will be directly influenced by the strain at failure (total energy to failure = area under the stress-strain curve). The composite with spherical glass particles showed the highest value of strain at failure, while composites with reinforcing agents with progressively larger aspect ratios led to lower values of strain at failure. Thus, composites with larger aspect ratio reinforcing agents would have the lowest toughness, mainly because angular particles have the ability to induce crack formation. Moreover, interfacial debonding (formation of voids at the interface) can consume energy during crack propagation, leading to higher levels of toughness. This result demonstrates the influence of the interfacial strength on the properties of the system and also indicates that the interfacial properties vary according to the morphology of the reinforcing phase.

Figure 2 shows SEM micrographs of the fractured surface of PAS composites. The homogeneity of fiber/particle distribution can be demonstrated, confirming the achievement of good dispersion by the process used to manufacture the composites. The fractured surface of composites, in Figure 2, yields information related to the interfacial properties. The strength of the interface between components of a composite can be divided into three types (Sinien et al. 1992) by using fractographic analysis: (a) weak (where few reinforcing agents are noted on the fractured surface), in this case the crack propagates around the particles; (b) strong (a thin layer of polymer covers the fiber/particle surface), where the crack "climbs over" the fibers/particles; and (c) intermediary between weak and strong (where few fibers/particles were removed during the fracture process and some polymer can be observed on the surface of fibers/particles). The micrograph in Figure 2-a shows the fractured surface of a PAS-short fiber composite, where a large number of fibers can be seen, indicating that fibers were not totally extracted from the matrix during the fracture process. In Figure 2-b, a micrograph of the fractured surface of composites with angular particles is revealed. The morphology of the fractured surface shows the presence of a large number of particles on the surface. Moreover, the exposed particles are also partially covered by polymer. According to the fractographic analysis just presented, the interface between poly(aryl sulfones) and both short fibers and angular particles can be described as having intermediate values of strength. In this case, interfacial strength is high enough to avoid complete fiber/particle elimination during fracture, but is still low to restrict crack propagation through the interface region.



On the other hand, the fractured surface (Figure 2-c) of the composite having glass spheres as reinforcing agents shows an extensive presence of voids due to the elimination of spherical particles during fracture. In this case, the interfacial strength should be low (according to the fractographic analysis), since cracks are allowed to easily propagate through the composite interface, leading to debonding and particle elimination from the fracture surface.

Based upon the fratographic analysis, it can be suggested that interfacial properties in polymer composites are affected by the shape of the reinforcing agent. This result agrees with the discussion regarding the mechanical properties of composites, since high levels of debonding can reduce the energy of cracks leading to higher levels of toughness (as observed). This information was also further studied by using DMS (Dynamic Mechanical Spectroscopy).

Dynamic mechanical behavior of materials is often studied by inputting a sinusoidal strain and measuring the returned stress. For a pure elastic body, stress and strain will be in phase, while for a pure viscous body, a 90° out-of-phase measurement will be recorded. A viscoelastic behavior, displayed by polymers, is the one that combines both elastic and viscous behaviors and the stress response for an input sinusoidal strain will be out-of-phase by a value between 0 and 90°. Viscoelastic behavior is usually expressed by complex notation. For the elastic modulus,

where, E* is the complex elastic modulus, E’ is the storage modulus that represents the elastic behavior and E" is the loss modulus that designates the viscous behavior. The loss tangent (tan d) is defined by

and can be indirectly related to the energy spent during a phase transition, internal friction and damping. Therefore, transitions such as the glass transition and even smaller segmental transitions can be detected by studying the loss tangent in a temperature spectrum.

In terms of composites, incorporation of fibers or fillers leads to changes on the thermo-mechanical spectrum that can be used to qualitatively or even quantitatively give information about the degree of adhesion between the constituents.

Figure 3 shows graphically the results of the loss tangent of different composites evaluated as a function of temperature (frequency = 1 Hz). It can be observed that the introduction of reinforcing agents with larger aspect ratio led to a much more pronounced drop in the loss tangent at Tg (Tg for poly(aryl sulfones) is around 190°C). Reinforcing agents with rough surfaces and high aspect ratio confine the polymer chains into spaces with restricted mobility. Moreover, the surface area per volume gets larger for higher aspect ratios. The presence of a large surface area reduces the mobility of the polymer at the surface of reinforcing agents, creating a layer of polymer with high viscosity. Since chain mobility is reduced at the interface, the amount of energy lost in processes related to the glass-transition is lower and the loss factor is then also lower for samples with high interfacial areas. The shift in Tg, seen in Figure 3, to higher temperatures is also a result of the immobilization of polymer chains at the interface.



In Table I, a parameter called "B" is also reported. This parameter intends to account for the usual non-ideal behavior of the loss tangent (tan d) at Tg, when reinforcing agents are introduced into polymeric matrices (Dong, S. and Gauvin, 1993). The drop of loss tangent at Tg due to the introduction of the inorganic particles/fibers ideally would follow the rule of mixtures. "B", in this case, would be equal to 1 in equation (3).

where: subscript c is related to composite, g to the glass and m to the matrix. Vg is the volume fraction of glass and tan d is the loss tangent. B is a parameter that quantifies the difference between the expected result from the rule of mixtures and the obtained one (B is larger for stronger interfacial interactions). As shown in Table I, the drop in the loss tangent of composites is larger than that predicted by the rule of mixtures. Moreover, the values of B calculated for composites with different particle shape, but same (20%) volume fraction, increase for high aspect ratio agents. Therefore, this model shows again that the extension of the interfacial area of composites, in this study, affects the viscoelastic behavior of the system. Consequently, interactions between matrix and filler that scale with the interfacial area, are indeed occurring.



In order to get more information about the effect of the interface on the relaxation processes of the matrix, a physical model that relates the dynamic (complex) elastic modulus to the unrelaxed, relaxed modulus, relaxation times and uses parameters from the Cole-Cole plot to fit the experimental data to the model, was used.

A Cole-Cole type of plot was first developed for studying dielectric phenomena but is also applied to viscoelastic phenomena, since this one has also an intrinsic time-energy relationship. Relaxation processes during viscoelastic transitions are well emphasized by Cole-Cole plots. During transitions, relaxation occurs whenever polymer chains rearrange themselves to adopt, for example, lower energy conformations or to fill free volume. These relaxation processes are seldom represented by only a single relaxation time and a distribution of relaxation times might be required in order to express the whole process occurring in a common polymer. The complexity of the relaxation phenomenon can be attributed to factors such as differences in chemistry and structure of segmental units within a polymer repeat unit, different environments for the chains, inhomogeneities due to inherent processing deficiencies (molecular weight distribution, free volume distribution, etc.) and thermo-mechanical history. When reinforcing agents are introduced into a polymeric matrix, the mobility of the chains are constrained by both the physical presence of the fillers or fibers and any possible chemical interaction between the phases. Moreover, the relaxation processes of polymeric chains near rigid surfaces are often modified by changes in mobility as well as conformational structure of the chains. In this situation, relaxation processes are enthalpically and entropically modified, since different levels of energy are required for transitions to occur and new conformational modes are imposed to the chains. The magnitude of the changes in the relaxation phenomena due to the presence of the interface is definitely related to the degree of interaction between phases of the composite. When lack of wetting and therefore adhesion between the phases is present, the relaxation processes will be less affected.

Although as commented before, a distribution of relaxation times is a more realistic image of the relaxation processes occurring in a polymer, a simple model that employs two relaxation times can be used to physically model the behavior of polymers during a viscoelastic transition (Bergeret et al., 1992; Bergeret et al., 1996; Cavaille et al., 1986; Harris et al., 1993; Ibarra et al.,1995); An equation such as the one below can usually give a good correlation between experimental data and model.

where: E*= complex modulus; Er = relaxed modulus; Eu = instantaneous modulus; t1 and t2 = relaxation times; w = frequency; h = long time parameter; k = short time parameter.

Values for the relaxed and unrelaxed modulus can be obtained from a Cole-Cole plot by extrapolating the data to near zero loss (E" = 0). Parameters h and k are used to fit the model to the data. They are usually obtained from an experimental Cole-Cole plot, where the storage modulus is plotted against the loss modulus. Both parameters can be calculated from angles in which the curves reach the E’ axis of the Cole-Cole plot: h = 2qr/p and k = 2qu/p, where h is the long time parameter and k is the short time parameter. The angles qr and qu are defined in the insert on the top of Figure 4. Both h and k parameters can be related to inhomogeneities within the material that would disturb the relaxation processes. Thus, lower values for h and k can represent a less Gaussian distribution of relaxation times with more pronounced tails. In terms of a composite, the interface is a type of inhomogeneity that can lower the values of k and h, by interfering in the relaxation phenomena of chains nearby.



The dynamic mechanical data (DMS) at 1 Hz of poly(aryl sulfones) composites containing 20% volume fraction of reinforcing agents with different particle shapes is shown in Figure 4 in a Cole-Cole type of plot (E’ vs. E"). The values of k and h and unrelaxed and relaxed modulus for the composites and pure polymer are displayed in Table I. A more pronounced reduction in the values of h can be seen for composites with high aspect ratio agents when compared with the pure polymer. This fact indicates that the presence of a larger interfacial area affects more the relaxation processes, by imposing constraints to the mobility of the chains. The parameter k is the short term parameter and it is less sensitive to changes due to the interfacial area. However, it also shows a decrease in its magnitude as the interfacial area increases, as a consequence of new constraints for chain mobility. Since relaxation processes of chains were modified by the presence of interfaces, it is clear that chains were interacting either physically or chemically (or both) with the surface of fillers. Basically, polymer chains chemically or physically associated with surfaces are forced to adapt to fewer numbers of conformation modes. Moreover, these confined chains have also the enthalpy of their transitions altered. The overall result is that poly(aryl sulfones) chains are interacting with the surface of glass particles. These interactions scale with the interfacial area and can be responsible for the observed trend in transference of stress performed by the interface.



In this work, polymer composites were prepared by reinforcing poly(aryl sulfones) matrices with agents of different morphologies. The influence of the shape of the phases on the interfacial properties as well as the overall properties of the composites was then investigated.

The results obtained from flexural mechanical tests showed that the elastic modulus of composites is enhanced when particles with higher aspect ratios are used. This result confirms that the interface is actually allowing the stress to be transferred from the matrix to the reinforcing agent. The strain at failure of the composites was substantially reduced by using reinforcing agents with higher aspect ratio, indicating that angular entities can act as crack initiator.

Fractographic analysis applied on SEM micrographs showed that interfacial strength on composites with spherical particles was lower than that for composites reinforced with high aspect ratio agents. The interfacial properties of the composites was also studied by applying a viscoelastic model to the collected dynamic mechanical data. This model showed that reinforcing agents with higher aspect ratio affect more pronouncedly the viscoelastic behavior of the systems, by imposing constraints to the mobility of the polymer chains. Thus, it was demonstrated that interfacial properties are influenced by the morphology of the phases.



The authors gratefully acknowledge the financial assistance of CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico) of the Brazilian Government.



Mallick, P.K., 1993 "Fiber-Reinforced Composites". Marcel Dekker, New York.        [ Links ]

Rothon, R., 1995 "Particulate Filled Polymer Composites". Ed. Longman Scientific and Technical, London.        [ Links ]

Wilson, J.; Proctor, D.C.; Elsberg, L. and Brownlee, R.E., 1996 "Injectable Bioglass for soft tissue augmentation: (Part 1). Preliminary". Transactions of the Annual Meeting of the Society of Biomaterials, 671.        [ Links ]

Kasuga, T.; Tsuji, O.; Abe, Y.; Niimi, A. and Ueda, M., 1995 "Preparation of calcium metaphosphate ceramics containing b-PO3)2 fibers for biomedical use". In: Bioceramics 8, Ed. Pergamon, 415-20.        [ Links ]

Sinien, L.; Lin, Y.; Xiaoguang, Z. and Zongneng, Q., 1992 "Microdamage and interfacial adhesion in glass bead-filled high-density polyethylene". J. Mat. Sci, 27, 4633-8.        [ Links ]

Dong, S. and Gauvin, R., 1993. "Application of dynamic mechanical analysis for the study of the interfacial region in carbon fiber/epoxy composite materials". Polym. Comp., 14, 414-20.        [ Links ]

Bergeret, A.; Agbossou, A.; Alberole, N.; Cassagnau, P. and Sarraf, T., 1992. "Micromechanical properties of random copolymers of styrene with methacrylic acid experimental and theoretical approaches". Eur. Polym. J., 28, 1201-18.        [ Links ]

Bergeret, A. and Alberola, N., 1996. "A study of the interphase in styrene-methacrylic acid polymer/glass bead composites". Polymer, 37, 2759-2765.        [ Links ]

Cavaille, S.; Etienne, J.; Perez, L. and Johari, J.P., 1986. "A comparison of the micromechanical properties of various poly(2,6-dimethyl-1,4-phenylene oxide)-polystyrene (PPO-PS) polymer blends". Polymer, 27, 549.        [ Links ]

Harris, O.G.; Bradell, D.P.; Almond, C. and Verbist, J., 1993. "Study of carbon fibre surface treatments by dynamic mechanical analysis". J. Mat. Sci., 28, 3353-66.        [ Links ]

Ibarra, L.; Macias, A. and Palma, E., 1995. "Viscoelastic properties of short carbon-fiber thermoplastic (sbs) elastomer composites", J. Appl. Polym. Sci., 57, 831-842.        [ Links ]



Manuscript received: September 1999. Technical Editor: Átila P. S. Freire.



*Provided by L. Elsberg, University of Florida
**Provided by Dr. Y. Abe, Nagoya Institute of Technology, Japan

Creative Commons License All the contents of this journal, except where otherwise noted, is licensed under a Creative Commons Attribution License