Processing of nanocomposites PLA/graphite using a novel elongational mixing device.
With the advent of graphene in the last decade , research on polymer/graphite nanocomposites has been gaining relevance since graphite is a natural source of this revolutionary material. Graphene is a honeycomb shaped atomic thick monolayer of [sp.sup.2] hybridized carbon and it is the primary constituent of graphite and carbon black particles. Literature has remarkably emphasized its properties , Back to graphite, this is one of the most common forms of carbon and, as silicates, its layered structure makes it a natural filler candidate for the elaboration of polymer nanocomposites. It has the susceptibility to undergo exfoliation under proper conditions, that is, to render pristine tactoids into graphite nanoplatelets (GNP, several layers stacks) and/or graphene. Actually, there is a large volume of work in the literature devoted to describe different ways to sustainably obtain GNP and graphene from graphite, in their basic and modified versions (named, in general, chemically modified graphene, CMG). Derived from such research, a series of graphitic materials is to be considered as precursor or structural fillers for the elaboration of polymer nanocomposites, i.e. expanded graphite (EG), graphite intercalation compounds (GIC), graphite oxide (GO), graphene oxide (G-O), functionalized graphene sheets (FGS), chemically reduced GO (R-GO), thermally reduced GO (TrGO), polymer modified GO (P/GO), and others [2, 3].
Up to now, there are three main methods to prepare graphene or GNP-based polymer nanocomposites: solvent blending, in situ polymerization, and melt compounding. The first and second methods have been, so far, the most successful to produce compounds with filler inclusion in the range of monolayers to several layers thick. In this regard, Stankovich et al.  and Brinson and coworkers  have reported remarkable results at working on polystyrene (PS) and polymethylmethacrylate (PMMA) solution-made nanocomposites, respectively. In the former case, the authors found an electrical percolation threshold as low as about 0.1% vol/vol of graphene. In the latter, it was found a glass transition temperature ([T.sub.g]) around 25[degrees]C higher than the neat polymer. Nevertheless, despite these promising results, important limitations arises concerning chemical or solvent approaches: the use of monomers and/or solvents, laborious procedures, small amounts of final material, and limited capability to expand to a larger scale. In view of this, melt compounding continues to be a strong alternative to produce polymer/ graphite nanocomposites in a more reliable manner, even though results obtained by this method have been so far not that remarkable. However, the great interest in melt compounding lies in that it presents very significant advantages, for instance, it may involve continuous, semi-continuous, and batch processes that yield high production of material; the use of solvents is normally avoided and operation is simple and economic. Furthermore, large and intensive research on clay-based nanocomposites vial melt blending provides an important and fundamental reference since the morphology of this type of filler i.e., montmorillonite possess layered microstructure similar to graphite.
The main challenge in melt processing about the inclusion of nanostructured particles is to attain good dispersion and distribution at the nanoscale level, searching for the maximum improvement in properties. In this regard, literature on polymer/graphite nanocomposites has reported the employment of diverse graphitic materials and strategies to disperse them. For example, Steurer et al.  used TrGO as Filler into different engineering matrixes by solution masterbatchs in a first step and subsequent melt blending using a twin screw microcompounder. Independently of polymer polarity, large exfoliated nanocomposites were obtained. Wilkie and coworkers  found that virgin and EG composites did not undergo exfoliation and presented a slight improvement of the mechanical properties when blended with acrylonitrile butadiene styrene (ABS) and high impact polystyrene (HIPS) using an internal mixer. Also, Murariu et al.  utilized an internal mixer for melt blending of polylactic acid (PLA) and EG; they obtained a composite with heterogeneous graphite morphology consisting of discrete domains of exfoliated graphite and large particles. Katbab et al.  compared dispersion of natural graphite, G, GIC, and EG in polypropylene (PP)/ethylene-propylene-diene rubber (EPDM) blends, employing concentrated masterbatch of functionalized PP, followed by dilution with rubber phase; both steps done by internal mixer. GIC gave the best results when it came to exfoliation. Macosko and coworker  compared dispersion of graphite and FGS in polycarbonate using a twin-screw microcompounder, finding that FGS nanocomposites form a filler network at very low concentrations that imparts a strong elastic character to the melt of the nanocomposite. Nevertheless, a result that seems unexpected is that tensile modulus was fairly comparable between the two types of composites at similar range of concentration. On the other hand, Zhao et al.  worked on polyphenylene sulfone (PPS)/EG systems by means of an internal mixer. The authors argue that low viscosity of PPS allows for the molecular diffusion into the pores of EG, giving place to intercalated systems and, thus, matrix reinforcement.
From the literature above, and more elsewhere, it has been strongly demonstrated the key role of particle pre-treatment before mixing and processing to succeed in obtaining GNP/graphene nanocomposites. However, surprisingly nothing or little is said about the specific influence of the flow geometry in the final dispersion of the compounds. As it is known, high shear contributions to dispersive mixing are common in conventional mixing devices, as those just mentioned in the literature. This is because sustained shear flow is easy to build up using relatively simple geometries compared to elongational flow field, also present at a different extent and whose contribution depends on the type of mixer. Nonetheless, on the other hand, it has to be pointed out that shear stresses necessary for filler dispersion often involve ranges of high shear rates that considerably reduce the polymer viscosity in the shear thinning zone, which demands high energy inputs to reach reasonable dispersion efficiency. Also, this efficiency is affected by the rotational component of the shear flow that offers no contribution to filler dispersive mixing. In this context, it has been stated that a way to highly improve dispersive mixing is to enhance the contributions of elongational flow fields in the overall mixing process [12, 13]. Accordingly, greater hydrodynamic stresses are generated in elongational flow and stresses are transferred more efficiently to the agglomerates because rotational motion is absent in pure elongational flow (irrotational flow); as a consequence, the energy consumption is expected to be less in elongational flow relative to shear flow , This is why, in the present work, a new mixer design with predominantly elongational flow contribution (RMX) has been used as a different approach to obtain polymer/graphite nanocomposites.
Elongational Flow Mixer and Reactor
In the last two decades, new designs have been developed to create strong elongational flow to improve dispersive mixing. Developments have been principally applied on the existing continuous systems like single and twin-screw extruders. Among these improvements are the tapered slots in screw flights, optimized kneading configuration for twin-screw extruders, and static elements like screens. In the operation of these mixers, a critical feature is the multiple passage of the material through the points of elongational stresses, which compensates for the relative low velocities needed to develop the elongational flow field. Also, the idea of promoting elongational flow for enhancing mixing efficiency has given place to the development of devices specifically designed to study elongational flow. In particular, Meller et al.  studied the deformation and break up of dispersed droplets in molten polymer blends of different viscosities. They used a capillary rheometer equipped with dies having different entry profiles and showed that the mixing efficiency in the converging flow zone was dependent on both the shape of the convergence and the flow rate. This kind of works has enabled some authors to revisit and to bring back old mixers designs like those by Hausman  and Westover  concerning the concept of flow between two chambers. Mackley et al.  adapted this geometry to design their so-called "multipass rheometer" in which the influence of the number of passes through the central die on the rheological properties can be studied. Recently, Son et al.  have shown that the concept of multipass rheometer can be adapted to design a batch mixer in which an unlimited number of convergent/divergent flows can be applied to the material to be mixed.
The RMX is based on similar principles described in the literature just mentioned. In this mixer, the material is induced to pass throughout a central die from opposites cylindrical chambers in a back and forth fashion. An unlimited number of convergent/divergent flows can be produced at different speed, allowing the generation of elongational flow in an efficient manner. A mixing sequence is defined simply by the piston velocity (v) and the number of cycles (N)\ the pressure in one of the chambers is continuously measured during the mixing sequence by a pressure transducer. Figure 1  shows a general view of the mixer.
The selected polymer matrix for this work was PLA, given the great interest at present on biodegradable polymers and those derived from renewable sources. The diversity of PLA properties makes it attracting to many applications, like packaging, films, textiles, biomedical products, tableware, etc. being some of them typically covered by well-known petroleum-based thermoplastics like polyethylene terephthalate (PET). PLA belongs to a family of biodegradable polyesters which are characterized for their potentially hydrolysable ester bonds . It is synthesized through ring opening polymerization of lactide, an intermediary form of lactic acid stereoisomers, which can be obtained by fermentation of carbohydrate rich substances coming from agricultural by-products. In general, properties of PLA depends on the final stereoisomer structure: poly(L-lactide) (PLLA), poly(D-lactide) (PDLA), and poly(DL-lactide)(PDLLA). For instance, polymers with stereo regular chain microstructure present a medium degree of crystallinity unlike amorphous PDLLA with racemic mixture of d- and L-lactide. As a consequence, some physical properties varies in the same way but, it can be said that some of them like elastic modulus, impact resistance, dimensional stability, and gas barrier remain low in comparison with equivalent thermoplastics for similar applications. Because of that, a lot of interest is still being devoted to improve this property by the inclusion of nanofillers like silicates and different types of carbon particles as described by some recent results: Carreau and coworkers  obtained PLA/ clay nanocomposites with improved mechanical and barrier properties, avoiding PLA degradation through the addition of chain extenders. Tait et al.  compared the effect of carbon nanofibers and GNP on PLA using extrusion and injection molding. A more significant increase in elastic modulus was attributed to GNP and flexural strength to carbon nanofibers. Also, injection molding was the most efficient method of improvement. Murariu et al.  increased the elastic modulus of PLA in a significant way using EG.
On the other hand, even though degradability is the most important feature for PLA as eco-friendly material, it is precisely this characteristic that, along with low thermo-mechanical properties, limits its use up to now. As a matter of fact, kinetics of PLA degradation, promoted by hydrolysis of the ester groups and chain scission, is strongly catalyzed at temperatures needed for mixing and processing operations. Cordari et al.  found that the value of the kinetics constants for the hydrolysis of ester groups presented at least a threefold increase for every 20[degrees]C of temperature increment. That is why in the present project, the assessment of RMX performance is accompanied by key considerations on the degradative process of PLA during the compounding process. This is possible to do in a much more controlled manner compared to conventional systems since RMX defines very specific operational parameters, as speed and number of cycles, from which residence times are very simple to calculate and to correlate with viscosity and molecular weight changes in PLA.
Amorphous PLA grade 4042D from NatureWorks[R] was used as polymer matrix with average molecular weight, [M.sub.w], 157,000 g/mol, [M.sub.n], 71,000 g/mol, and polydispersity 2.21. The mol% of D-isomer units in PLA is 4.3  and density, [rho] = 1.24 g/[cm.sup.3] (technical data sheet). As filler, it was used EG Ecophit[R] from SGL Group, The Carbon Company, Germany; carbon content [greater than or equal to] 95%, [D.sub.50] ~5-7 [micro]m, [[rho].sub.powder] ~0.115-0.135 g/[cm.sup.3], [[rho].sub.real] = 2.25 g/[cm.sup.3], moisture content [less than or equal to] 2 (technical data sheet).
Rheological characterization was carried out in both Bohlin Inst capillary rheometer HR2000 and Anton Paar MCR301 rheometer. In the first one, shear and elongational viscosities of pristine PLA in the range of the selected RMX speeds were obtained, as well as the global shear power law index, n. In order to estimate elongational properties from capillary measurements, Cogswell approach  is commonly used:
[[eta].sub.e] = [9/32] x [[(n + 1).sup.2]/[eta]] x [([DELTA][P.sub.e]/[??]).sup.2] (1)
[[sigma].sub.e] = [3/8] x (n + 1) x [DELTA][P.sub.e] (2)
[??] = [[sigma].sub.e]/[[eta].sub.e] (3)
where [[eta].sub.e] is the elongational viscosity, [eta] is the shear viscosity, [DELTA] [P.sub.e] is the entrance pressure drop, [??] is the shear rate and [[sigma].sub.e] and [??] are the elongational stress and elongational strain rate, respectively.
In the Anton Paar rheometer, by using oscillation mode, the complex viscosity, [[eta].sup.*], of pristine, neat processed PLA, and PLA/EG samples as a function of frequency (0.1-600 rad/sec) was measured. Characterization by X-ray diffraction (XRD) was done in a Siemens 5000 diffractometer (Cu-[K.sub.[alpha]] radiation) [lambda] = 1.54 [Angstrom], generator voltage of 35 kV, and current 25 mA. A 2 theta sweep from 10[degrees] to 30[degrees] was applied. Morphology characterization was realized by transmission electron microscopy (TEM) in an electronic microscope Hitachi 7500. Also, for the assessment of molecular weight and polydispersity of pristine and neat processed PLA, gel permeation chromatography (GPC) measurements were carried out in an Agilent Technologies[R] Chromatograph 1200 employing a refraction index detector (RDI) and chloroform as a solvent. Dynamical mechanical properties were evaluated in a DMA Nietzsch 202D in a three point bending geometry.
Previous to blending, PLA and EG were dried under vacuum at 80[degrees]C for 12 h. Before feeding of the RMX with the raw materials, the powder was incorporated into the polymer by means of an internal mixer Haake Rheomix 600 Thermo Electron[R] at 200[degrees]C and 20 rpm for 5 min. Afterwards, the integrated material was cut into pieces to be introduced into the RMX. This integration step was considered to be necessary since the RMX, at the moment, does not count on an effective powder feeding system. However, it has to be pointed out that the selected conditions for this incorporation stage were kept mild in order to avoid a significant influence in the final morphology of RMX composites. Discussion on morphology goes back to this point.
The mixing operation was carried out in the RMX which presents the next main features: controlled pistons speed (v) range of 3-180 mm/sec which corresponds, using cylinders of 3.2 cm diameter, to volumetric flow (Q) values in the range of 2.4-125 [cm.sup.3]/sec. A basic scheme of its operation is presented in Fig. 2 . Samples were prepared at 200[degrees]C, the concentration of EG was kept constant at 3% wt/wt and different combinations of pistons speed, v (mm/sec) and number of cycles, N, were employed according to Table 1. In general, these conditions were selected to cover in a first approach a relatively low range of speeds and number of cycles with respect to the RMX capacity, trying to prevent large PLA degradation. Also, since the RMX configuration is analogous to capillary rheometer geometry, corresponding shear rates at the different speeds were calculated applying the Rabinowitsch equation to non-Newtonian flow,
[??] = [4Q/[pi][R.sup.3]] x [3n + 1/4n] (4)
where [??] is the true or corrected shear rate, Q is the volumetric flow, R is the radius of the die, and n is the power low index, that, at the processing temperature is 0.285. The mixing element used in the RMX was a round die of [phi] = 4 mm and L/D= 7. Also, a following up of the pressure, P, during the mixing sequence was considered to be fundamental as to describe the flow behavior within the RMX in real time. Data acquisition of P values during the RMX mixing sequence was done by Data-Xport[R] software. For the sake of morphology comparison and evolution as a function of the different mixing conditions, a reference sample was prepared at the lowest piston speed allowed by the mixer, 3 mm/sec without mixing sequence. Finally, a mixing efficiency comparison between the RMX and an internal mixer Haake Rheomix 600 was carried out by preparing samples at the same specific mechanical energy input (SMEI), calculated for each mixer as follows:
[W.sub.INT] = [omega]Tt/m (5)
[W.sub.RMX] = [DELTA]PxN/[rho] (6)
where N is the number of cycles, [rho] is the polymer density, m, mass of polymer, t, time, [omega], the angular speed, and T, torque.
RESULTS AND DISCUSSION
Effect of RMX Mixing Conditions on Flow Properties and PLA Integrity
A basic understanding of neat PLA flow behavior under certain mixing conditions in the RMX was considered to be fundamental in order to address the role of the degradative phenomenon of PLA.
Rheology and Molecular Weight Assessment. Figure 3 presents the shear and elongational viscosity behavior of neat PLA as a function of shear rate as obtained by capillary rheometry. It can be observed that shear viscosity, [eta], decreases in a more significant way than the elongational viscosity, [[eta].sub.e], thus making the ratio [[eta].sub.e]/[eta] higher as shear rate increases. This fact is important since the elongational to shear ratio remains also high (Fig. 4) toward shear rates ranges estimated to be reached in the RMX. Intensive or dispersive mixing of components with significant cohesive strength is said to be dependent on the magnitude of shear and elongational stresses [12, 13, 25], contributing the last to a greater dispersion efficiency. In conventional mixers, an important constraint is to reach significant rates of elongation for the built-up of critical elongational stresses, whereas in the RMX, due to the easiness with which elongational flow is produced, this limitation is thought to be far surpassed.
As it was described above, the RMX architecture resembles that of a capillary rheometer; thus, capillary measurements were useful to correlate the piston speed, v, one of the main parameters of the RMX, to corresponding shear and elongational strain rates, [??] and [??], respectively, during mixing. This was possible by approximating convergence ratios, [[empty set].sub.c]/[[empty set].sub.d] (diameter of the reservoir to that of the die), for both, the capillary rheometer and the RMX. So that, we can observe in Fig. 5 that, on superimposing data of two different convergence ratios (15/2 and 15/1) used in the capillary rheometer, corresponding data series seem to describe an almost linear relationship between [??] and [??] at least in the vicinity of this ratio values. According to that, once a true [??] is estimated in the RMX by means of the Rabinowitsch equation and, keeping a similar convergence ratio to that of the capillary rheometer (32/4), Fig. 5 leads to the estimation of the correspondent [??] value in the mixer. The whole values are shown in Table 2. Thus, for the selected RMX conditions it is possible to have an estimation of the balance between shear and elongational strain permitting to envisage better correlations mixing/ properties for the compounds.
On the other hand, it was also necessary to address the effect of the number of cycles, N, in the PLA melt flow at specific shear and elongational strain rates. In order to do this, a following up of the pressure, P, for every passage was done in the RMX for 10/40 and 40/10 (v/N) samples, since P is a response of the system directly related to flow. A transducer fixed in one of the chambers senses the overall pressure. Figures 6 and 7 show the total pressure cycle as a function of the piston displacement, forward (high P) and backward (low P). Even though the general trajectory of P is clearly dependent on the piston speed, v, from both figures it can be suggested a three stage pattern as a function of the piston forward displacement: an initial pressure increase because of the induction of the material to get into the die; a stable zone derived from the well-developed flow inside the die and, a pressure shot possibly due to over compacting of the last portion of material. The subsequent decrease and constant low value of P corresponds to the filling stage of the chamber or retracting movement of the piston. However, it is evident a great difference of pressure behavior especially in the steady flow regimes between samples. On this respect, we can say that at high piston speeds, sample 40/10 (8220 [sec.sup.-1]), there is not time enough to develop steady flow in contrast to sample 10/40 (2070 [sec.sup.-1]), where pressure stabilizes at a large displacement range in the first cycles. A large difference in shear viscosity at the corresponding shear rates (Fig. 3) may also contribute to this behavior.
A second significant finding from Figs. 6 and 7 is the decrease in P
at each cycle as mixing sequence evolves. Total P losses of about 50% and 30% for mixing conditions at 10/40 and 40/10, respectively, were estimated at the end of the sequence. Based on these results and, considering the principle of multiple passage of the RMX as well as the sensitivity of PLA to degradation, the effect of viscous heating as a result of the number of cycles was addressed. This phenomenon is thought to be especially significant at high strain rates and/or cycles number . Accordingly, the estimation of the average adiabatic temperature rise (AT) was done using the next relationship from Ref. 17]:
[DELTA]T = [DELTA]P/[[rho].sub.PLA][Cp.sub.PLA] (7)
where [DELTA]P is the pressure difference between chambers, [[rho].sub.PLA] and [Cp.sub.PLA] are the density and the specific heat of PLA at the reference temperature, respectively. It is worthy to mention that P values from Figs. 6 and 7 were approached to [DELTA]P in Eq. 7, given a very low P in the opposite chamber. Also, P values were taken from the assumed "steady state" region of the curves. Finally, an accumulated [DELTA]T of about 65[degrees]C for the sample 10/40 and 35[degrees]C for the sample 40/10 were estimated at the end of the mixing sequence. High shear contributions, according to [DELTA]T [varies] [eta][[??].sup.2]  accounts for large [DELTA]T increments at relatively high piston speed (40 mm/sec), whereas a high number of passages (40) produces even a much higher viscous heating in spite of employing a relatively low shear rate. Moreover, viscous heating seems to be greater in the first cycles, what is especially noteworthy in the sample at 10/40. For example, after the first 10 cycles, differences in P are much smaller, almost disappearing when approaching the end of the mixing sequence. Also, if we consider a very large decrease in viscosity, at those conditions subsequent viscous heating may be limited in turn by [DELTA]T [varies] [eta][[??].sup.2]. Another observation that calls the attention is the minimum in the trajectory of P as N increases; it can be argued that additional effects of viscous heating and PLA degradation in flow stability.
Figure 8 depicts the balance of PLA in terms of complex viscosity, [[eta].sup.*], and correspondent weight molecular weight average, [M.sub.w], after being subjected to different RMX conditions. First, the large drop in both, [[eta].sup.*] and [M.sub.w] for all samples, indicates that [DELTA]P decrements observed during the mixing sequence in the RMX was a direct result of PLA degradation. Secondly, somewhat surprisingly, v (related to shear rate) seems to have a little bit more influence in PLA degradation than N (related to residence time) given a lower [M.sub.w] for the sample 40/10 with respect to the sample 10/40; although, at the end, degradation phenomenon is a combination of both. On the other hand, effects of mild and hard mixing conditions are clearly identified since low v and N tend to preserve [M.sub.W] (sample 20/10) and, on the contrary, high v an N promotes large [M.sub.W] detriments (sample 40/20).
Flow Properties. Figure 9 shows the corresponding [[eta].sup.*] behavior of EG filled PLA. Because of the presence of EG, results are somewhat not aligned to those of neat processed PLA. As an illustrative example of this, the sample at highest v (40/10) presents the highest [[eta].sup.*] even tough neat PLA at these conditions reported the lowest [M.sub.w]. A possible explanation could lie on the ease of diffusion and/or interlocking of shorter polymer molecules, as a result of chain scission, into the high porosity of EG particles, enhancing hydrodynamic effect on viscosity. On the other hand, samples at increasing N (10/20, 10/30, and 10/40) exhibit lower [[eta].sup.*] than those of high v, also in contrast to the
observations of the precedent section. In this regard we could state that, because of the increase in viscosity of filled samples, degradation by viscous heating during mixing at long times becomes predominant. Samples 10/10 and 20/10, in turn, processed at mild mixing conditions, are expected to preserve in a greater extent the structural integrity of the polymer matrix and, hence, show higher [[eta].sup.*].
Morphology. XRD. Figure 10 presents XRD diffractograms of PLA/EG samples, focusing only on the small 2 theta range of the characteristic peak of graphite (2 theta ~26.38[degrees]) since it is expected to be modified as a function of mixing conditions. For the sake of a valid comparison, a reference sample was processed at the lowest RMX specification speed, 3 mm/sec, and 1 cycle to try to preserve the EG structure. On one hand, when comparing, the 2 theta value of the graphite diffraction peak for all samples remains practically the same (26.40-26.52) which states the lack of intercalated structures. On the other hand, we can observe very important differences in peak intensities as a function of mixing conditions. In this respect, diffraction peak intensity is related to the amount of graphitic layers that scatter, it means, to the mass fraction of crystalline phase . Thus, differences observed may obey to the development of different dispersion patterns, even partial exfoliation, as a function of mixing , where the reference sample is expected to present a much more agglomerated or preserved filler microstructure, which coincides with the highest peak. Also, although a tendency about the influence of v or N in the signal intensity is not clear in the processed samples, it is noteworthy that those at an extreme (40/10) and mild (10/10) mixing condition presents the highest and lowest peak, respectively. It can be said that differences in molecular weight of the PLA matrix as a function of mixing (Fig. 8) may strongly account for the development of a different filler dispersion pattern for each sample, situation to be clarified in the subsequent sections.
TEM. Morphology was tried to be complemented by TEM. Figures 11 and 12 show different samples processed by the RMX and compared to the reference. The image of the reference sample shows very large EG particles, suggesting that the previous incorporation step carried out in the Haake mixer had not a striking effect or contribution to the mixing process in terms of morphology. An improvement of filler distribution and dispersion is clearly observed in the RMX processed samples, which evidence also a lack of filler networking at this concentration. In general, most of the samples presented a broad distribution of particle sizes, ranging roughly from 100 nm to 1 pm thick. However, in many cases, the identification of particles much thinner than 100 nm could be evident, giving support to X-ray results about partial exfoliation. Even though direct comparison between samples at different mixing conditions did not report great visual differences, samples at higher v and/or N seem to exhibit, in general, lower particle size but, again, it is not possible to establish a tendency in this sense. It is important to mention that special difficulties arose to carry out a more accurate morphology assessment, since the high hardness of the samples, made it that thin EG edges have been slightly oriented in the sense of the cut. More work has to be done on this.
Mixing Efficiency. With the aim of having a qualitative reference to ponder the RMX performance, for the particular system studied, two PLA/EG blends were prepared by means of an internal mixer and their morphology were compared to RMX samples at the same estimated specific mixing energy inputs (SMEI) as described in the experimental section. Figure 13 presents micrographs from the two mixing devices at the lowest (62 J/g) and highest (246 J/g) energy inputs used in the RMX. In both cases, direct comparison evidences a less homogeneous EG particles distribution and significantly higher particle size for the internal mixer blends.
Physical Properties. In order to obtain a basic look at the physical behavior of PLA/EG compounds, Table 3 is presented. A general remarkable increase of elastic modulus, from 20 up to 38%, was found. It must be recalled that a fixed concentration of only 3% wt/wt of EG has been used; therefore, only the influence of mixing on properties is addressed. In a closer look to the effects, the particular results and comparison between different RMX conditions indicate several scenarios that would have to do with the issue of molecular weight role in polymer reinforcement. As it has been stated above, [M.sub.w] changes at every commanded mixing condition. About this matter, it is known that, under proper compatibility between components, high molecular weight and viscosity of the polymeric matrix leads to the generation of high shear and elongational stress needed to effectively disperse and/or exfoliate the filler particles [25, 29], thus improving mechanical properties through the high increase of interfacial area. But, on the other hand, low molecular weight promotes better chain diffusion into particle galleries (interlocking) what is said to also promote reinforcing and exfoliation [30, 31]. In the present case, this could be better explained starting out from samples at extreme v/N conditions. For example, sample at the mildest condition, 10/10, expected to have the highest molecular weight, presents the second largest E' increment, around 31%. On the other hand, sample at the highest speed, 40/10, with the lowest [M.sub.w] of the series surprisingly presents the largest E' increment, 38%. Evidently, reinforcement mechanisms in terms of molecular weight role are different, suggesting for the first one a predominant shear/elongational stress contribution to dispersion and reinforcement and, for the second one, a more diffusional, molecular interlocking phenomenon. Although, samples just described seem to adhere in a greater extent to one mechanism or another, it is difficult to discuss in more detail on mechanisms for samples at relatively medium conditions, having among them a similar E' increment of about 20%. In this point, the work by Bousmina  gives an important insight; it points outs to a balance between mechanical stresses and diffusion process that requires rather low medium viscosity as to favor the best level of exfoliation in, for example, clay-based nanocomposites. In the present system, a range of viscosities coming up from the mixing process, and resultant degradation, favors different balances between the two mechanisms, diffusion and stress dependent dispersion and, thus, reinforcement. In spite of the aim to ponder the importance of flow geometry in dispersion, it is not possible to assess so far the punctual role of elongational flow contribution since the drastic effect of viscous heating in PLA produces significant changes in the rheology of the system in the very same mixing sequence, specifically, the elongational to shear flow ratios. It is remarkable, nevertheless, that good morphology/properties of nanocomposites are attained in a relatively small window of experimental conditions and using an extremely sensitive polymer, which represents a solid starting point to continue doing new and promising research on this field.
A very acceptable performance of the RMX (good distributive and dispersive mixing) was evidenced, to obtain PLA/EG nanocomposites with remarkable reinforcement (high modulus) at low weight concentration of filler. An apparently wide distribution of particle size in a similar range for all samples suggests that selected mixing conditions could be below those of a possibly break-through morphology (high degree of exfoliation) as a function of the RMX capacity. However, for this particular system, PLA degradation is indeed a limiting factor to further expand the experimental design. On the other hand, thermo-mechanical degradation of PLA plays a key role during the compounding of PLA/ EG nanocomposites in the RMX. It seems that chain diffusion as a function of molecular weight reduction during the mixing sequence dictates at some extent the dispersion and reinforcement mechanisms and, thus, the efficacy of filler on final PLA physical properties. Even though it is difficult so far to establish punctual elongational and shear contributions to the RMX mixing efficiency, because of the drastic rheology changes during mixing, this proved to be higher than that of an internal mixer at values in the relatively low to medium RMX capacity. Furthermore, the specific influence of v and N in the final dispersion and reinforcement of nanocomposites strongly depends on the direct effect of the mixing parameters on PLA molecular weight.
The authors acknowledge Cathy Royer, Neurosciences Centre, University of Strasbourg, for all her kindly support on TEM characterization.
[1.] A.K. Geim and K.S. Novoselok, Nat. Mater., 6, 183 (2007).
[2.] S. Park and R.S. Ruoff, Nat. Nanotechnol., 4, 217 (2009).
[3.] R. Sengupta, M. Bhattacharya, S. Bandyopadhyay, and A.K. Bhowmick, Prog. Polym. Sci., 36, 638 (2011).
[4.] S. Stankovich, D.A. Dikin, G.H.B. Dommett, K.M. Kohlaas, E.J. Simney, E.A. Stach, R.D. Piner, S.T. Nguyen, and R.S. Ruoff, Nature, 442, 282 (2006).
[5.] T. Ramanathan, A.A. Abdala, S. Stankovich, D.A. Dikin, M. Herrera-Alonso, R.D. Piner, D.H. Adamson, H.C. Schniepp, X. Chen, R.S. Ruoff, S.T. Nguyen, I.A. Aksay, R.K. Prud'homme, and L.C. Brinson, Nat. Nanotechnol., 3, 327 (2008).
[6.] P. Steurer, R. Wissert, R. Thomann, and R. Mulhaupt, Macromol Rapid Commun., 30, 316 (2009).
[7.] F.M. Ohl, Q. Yao, and C.A. Wilkie, Polym. Adv. Technol., 16(7), 533 (2005).
[8.] M. Muriaru, A.L. Dechief, L. Bonnaud, Y. Paint, A. Gallos, G. Fontaine, S. Bourbigot, and P. Dubbois, Polym. Degrad. Stab., 95(5), 889 (2010).
[9.] A.A. Katbab, A.N. Hrymak, and K. Kasmadjian, J. Appl. Polym. Sci., 107, 3425 (2008).
[10.] H. Kim and C.W. Macosko, Polymer, 50, 3797 (2009).
[11.] Y.F. Zhao, M. Yiao, S.J. Wang, X.C. Ge, and Y.Z. Meng, Compos. Sci. Technol., 67, 2528 (2007).
[12.] C. Rauwendall, Plast. Add. Comp. 1, 4, 21 (1999).
[13.] I. Manas-Zloczower, Z. Tadmor, Eds., Mixing and Compounding of Polymers, Theory and Practice, 2nd ed., Carl Hanser Verlag, Munich (2009).
[14.] M. Meller, A. Luciani, A. Sarioglu, and J.A.E. Manson, Polym. Eng. Sci., 42(3), 611 (2002).
[15.] J.M. Hausman, U.S. Patent, 2,948,920 (1960).
[16.] R.F. Westover, Soc. Plast. Eng. Trans. 1, 14 (1961).
[17.] M.R. Mackley, R.T.J. Marshall, and J.B.A.F. Smeulders, J. Rheol., 39, 1293 (2005).
[18.] Y. Son, J. Appl. Polym. Sci., 112, 609 (2009).
[19.] M. Bouquey, C. Loux, R. Muller, and G. Bouchet, J. Appl. Polym. Sci., 119, 482 (2010).
[20.] K. Madhavan Nampoothiri, N. Rajendran Nair, and R. Pappy John, Bioresour. Technol., 101, 8493 (2010).
[21.] Q. Meng, M.C. Hezvey, and P.J. Carreau, Polym. Degrad. Stab., 97, 2010 (2012).
[22.] M. Tait, A. Pergoretti, A. Dorigato, and K. Kalaitzidou, Carbon, 49, 4280 (2011).
[23.] F. Cordari, S. Lazzari, M. Soos, G. Storti, M. Morbidelli, and D. Moscatelli, Polym. Degrad. Stab., 97(11), 2460 (2012).
[24.] F.N. Cogswell, Trans. Soc. Rheol., 16(3), 383 (1972).
[25.] T.D. Fornes, P.J. Yoon, H. Keskkula, and D.R. Paul, Polymer, 42, 9929 (2001).
[26.] H.M. Laun, Rheol. Acta, 43, 509 (2004).
[27.] R.A. Vaia, K.D. Jandt, E.J. Kramer, and E.P. Giannelis, Macromolecules, 28, 8080 (1995).
[28.] A. Yasmin, J.-J. Luo, and 1. M. Daniel, Compos. Sci. Technol., 66, 1179 (2006).
[29.] T.D. Fornes and R.D. Paul, Cien. Technol., 13, 212 (2003).
[30.] F.J. Rodriguez, A. Coloma, M.J. Galotto, A. Guarda, and J.E. Bruna, Polym. Degrad. Stab., 97, 1996 (2012).
[31.] M. Bousmina, Macromolecules, 39, 4259 (2006).
Rigoberto Ibarra-Gomez, (1) Rene Muller, (2) Michel Bouquey, (2) Jerome Rondin, (2) Christophe A. Serra, (2) Fatima Hassouna, (3) Yamna El Mouedden, (3) Valerie Toniazzo, (3) David Ruch (3)
(1) Centro de Investigacion en Materiales Avanzados S.C, Miguel de Cervantes 120, Complejo Industrial Chihuahua, Chihuahua 31109, Mexico
(2) Institut de Chimie et Procedes pour l'Energie, l'Environnement et la Sante (ICPEES), ECPM, Universite de Strasbourg, 25 rue Becquerel, 67087 Strasbourg, France
(3) Centre de Recherche Public Henri Tudor, 5 rue Bommel, ZAE Robert Steichen, L-4940 Hautcharage, Luxembourg
Correspondence to: Rigoberto Ibarra-Gomez; e-mail: firstname.lastname@example.org
Published online in Wiley Online Library (wileyonlinelibrary.com).
TABLE 1. RMX mixing conditions. v (mm/sec) N (cycles) Neat PLA 10 40 20 10 40 10 40 20 PLA/EG 10 10 10 20 10 30 10 40 20 10 40 10 TABLE 2. Estimated shear and elongational strain rates at different RMX mixing conditions. v [[??].sub.corr] [??] (mm/sec) N ([sec.sup.-1]) ([sec.sup.-1]) Neat PLA 10 40 2070 265 20 10 4140 530 40 10 8280 1060 40 20 8280 1060 PLA/EG 10 10 2070 * 265 * 10 20 2070 265 10 30 2070 265 10 40 2070 265 40 10 8280 1060 * assumed strain rates for compounds TABLE 3. Physical properties of PLA/EG nanocomposites. [T.sub.g] Sample (v/N) E' (MPa) % [DELTA]E' ([degrees]C) PLA ref 2750 -- 60 10/10 3600 30.9 65 10/20 3350 21.8 64 10/30 3400 23.6 65 10/40 3300 20 64 20/10 3400 23.6 63 40/10 3800 38.2 61
|Printer friendly Cite/link Email Feedback|
|Author:||Ibarra-Gomez, Rigoberto; Muller, Rene; Bouquey, Michel; Rondin, Jerome; Serra, Christophe A.; Hassou|
|Publication:||Polymer Engineering and Science|
|Date:||Jan 1, 2015|
|Previous Article:||Tri-(butanediol-monobutyrate) citrate plasticizing poly(lactic acid): synthesis, crystallization, thermal, and mechanical properties.|
|Next Article:||Temperature-responsiveness and sustained delivery properties of macroporous PEG-co-PNIPAAm-co-PCL hydrogels.|