Poly(lactic acid)/low-density polyethylene blends and its nanocomposites based on sepiolite.
Poly(lactic acid) (PLA) is a biodegradable aliphatic polyester, with poor thermal stability in its processing, and low tensile ductility [l]. The combination of requirements. mechanical properties and thermomechanical degradation in the processing, leads to the development of polyblends with PLA and other thermoplastics such as poly(butylene adipale-co-terephthalate) and different types of poly-ethylenes (PEs) [2-5]. In the literature, various methods to modify PLA such as the copolymerization of the lactide with other monomers have been reported [6. 7]. However, blending PLA with immiscible or miscible polymers is a more practical and economical way of toughening it [2-5, 8-12]. Blending PLA with low-cost commodity PEs as dispersed phase could be an alternative to find the desired requirements for different applications, although immiscible blends with less biodeeradability character and poor barrier properties to oxygen (02) and carbon dioxide (C02) will result. It is now well established that the phase morphology of immiscible polymer blends can be controlled by addition or in situ formation of compatibilizers. which can act as interfacial agents (13), (14).
To cope with the problem of immiscibility, compatibilizing agents such as copolymers containing segments miscible with the blend components [4, 15, 16], or polymers with reactive groups that can link the matrix with the dispersed phase via covalent bonds formed in situ during the melt-blending process, are used. They both reduce the interfacial tension between the immiscible phases [17, 18]. Considering the poor thermal stability of PLA during melt processing, the nanocomposite technology is also a beneficial alternative to the design of these blends. Additionally, extensive dispersion of the layered clay could improve their barrier properties to [O.sub.2] and [CO.sub.2], and increase their thermal stability and tensile modulus [19, 20]. On the other hand, some immiscible blends of clay nanocomposites show lower dispersed particles sizes than similar blends without the nanofiller and compatibilizing agents. Among the different mechanisms of compatibilization in those blends, a high viscosity of the nanocomposites and/or the migration of the clay through the blend interface where a solid barrier is formed that inhibits or prevents the coalescence of the dispersed polymer drops are noteworthy . Moreover, the material parameters that must be controlled and which can have a profound influence on the nature and properties of the final nanocomposite blends include the type of clay and clay pretreatment, the selection of the blend components, and the way in which the polymers are incorporated into the nanocomposite.
Most of the literature regarding PLA nanocomposites (nPLA) is devoted to lamellar-layered silicates [19, 20, 22, 23]. In several works [23, 24], the neat polymers compounded with sepiolite are considered to be a nanocomposite because of the excellent distribution of the unmodified inorganic filler in its finest elemental units, even at concentrations as high as 5 wt%, and due to the absence of particle aggregations. Although there are few studies concerning nPLAs based on sepiolite, a high level of reinforcement was found in this nanocomposite . In that sense, the factors that influence or determine the sepiolite dispersion in blends of PLA with polyolefins have been less investigated.
This article focuses on the study of the effectiveness of two grafted polymers as tensile toughening materials in ternary blends with PLA as matrix phase, low-density PE as dispersed phase, and sepiolite clay as filler. A styrene/ethylene-butylene/styrene rubber and a linear PE grafted with maleic anhydride (MA) were used as compatibilizing agents. The morphology of materials was evaluated in terms of transmission and scanning electron microscopy (TEM and SEM), and thermogravimetric analysis (TGA) and rheological and tensile properties determinations were also made. Blends without the clay were also evaluated for comparison purposes.
A PLA and a low-density PE (PEI) were used as the continuous and dispersed phases in the blends, respectively. Additionally, nanocomposite blends with sepiolite were prepared. Two commercial polymers functionalized with MA, a styrene/ethylene-butylene/styrene rubber (SEBS-g-MA), and a grafted PE (PE2-g-MA) were used as compatibilizer agents in the ternary blends with PLA. The clay used for the preparation of the nPLA blends was a commercial sepiolite (Pangel HV CDT-11) supplied by Tolsa. The characteristics of the neat polymers are reported in Table 1.
TABLE 1. Characteristics of the neat polymers: density ([rho]), melt flow index (MFI), and grafting degree (GD). Material Commercial Supplier [rho] (a) MFI (b) GD (a) name (g/[cm.sup.3]) ([degrees] (wt%) C/min) PLA PLA Cargill 1.24 10 -- 2002D -Dow PE1 PEBD Polinter 0.92 2.4 -- 0240 PE2 Fusabond Du 0.96 4.0 0.2 -g-MA EMB Pont 206D SEBS Kraton Kraton 0.91 1.5 1.6 -g-MA FG-1901 Polymers (a.) Reported by suppliers. (b.) Determined according to ASTM D-1238 (Condition E).
Preparation of the nPLA and Blends
The nPLA and blends were prepared in a Berstorff (ECS-2E25) corotating intermeshing twin-screw extruder at 210[degrees]C (die temperature) and 100 rpm. In the nPLAs preparation, all blend components were fed through the first port of the extruder and the sepiolite through the second port. The solid materials (pellets of polymers) were fed to the extruder by a solid feeder, and the extrudates were cooled in a water bath and pelletized afterward. The effective content of sepiolite in the nanocomposite blends was determined by ash residue at 900[degrees]C (see Table 2). The test specimens for determining the properties were compression molded for 2.5 min at 200[degrees]C. The PLA and its nanocomposites and blends were dried in a vacuum oven at 50[degrees]C for 24 h before mixing and testing. The sepiolite clay was also dried before mixing at 100[degrees]C for 4 h. In the residence time experiment, the neat PLA was starved-fed into the extruder by a solid feeder at 3.18 k/h of mass flow rate and 100 rpm. Once the steady state was reached, the mass flow rate of the polymer and the temperature at the die were measured. A yellow pigment was used as the tracer and added as a pulse just after the solid feeder and the mean residence time was obtained in a visual way at the maximum concentration of the pigment. The blends and composites prepared are shown in Table 2.
TABLE 2. Blends and compositions. Blend Content Content Type of Effective of of Compatibilizer content Of PLA (wt%) PEl (wt%) And content (wt%) Sepiolite (wt%) PLAPE1 80 17 PE2-g-MA (3) -- PLAPE1 60 34 PE2 -- 2 -g-M A (6) PLAPE1 80 17 SEBS -- S -g-MA (3) nPLA 95 -- -- 4.6 nPLAPE1 75 17 PE2 4.7 1 -g-MA (3) nPLAPE1 55 34 PE2 4.0 2 -g-MA (6) nPLAPE1 75 17 SEBS 5.5 S -g-MA (3)
The melting and crystallization behaviors of the neat polymers and the nPLA were determined by differential scanning calorimetry (DSC) using a Mettler Toledo DCS 821/400. 10 mg of the samples were heated to 280[degrees]C, held for 5 min at this temperature, then cooled to -- 20[degrees]C, and heated again to 280[degrees]C, at constant rates of l0[degrees]C/min. In all cases, second heating scans were used for analysis. Additionally, the neat polymers, the nPLA, and the blends were subjected to oscillatory shear in a Haake RS-600 Rheometer over a frequency range of 1-100 rad/s at 200[degrees]C. Isothermal time scans were also performed for up to 30 min at a fixed strain, at a temperature of 200[degrees]C and frequencies of 3.14 and 6.28 rad/s for the neat PLA and 0.5 rad/s for the PLA blends, nPLA and its blends because of the thermo-oxidative degradation of the PLA with time at high temperatures. The PLA and its blends were dried before each testing. As well, the intrinsic viscosities [[eta]] of the neat PLA and PLA processed under the same conditions of the blends were measured using an Ubbelohde viscometer at 25[degrees]C in chloroform. The viscosity molecular weights (Mv) of both materials were calculated according to the Mark-Houwink equation . Melt flow index values (MFI) were measured using a Davenport melt flow indexer according to ASTM D-1238 Standard Procedure at 190[degrees]C and 2.16 kg of load (condition E) for neat materials.
In order to analyze the morphology of the obtained materials, samples of the nPLA and its blends were observed through TEM, using a JEOL JEM 2000FX Electron Microscope, with an acceleration voltage of 200 kV. The specimens were prepared by ultramicrotomy (Ultracut S from Leica), and some samples were exposed to osmium tetroxide ([O.sub.s][O.sub.4]). On the other hand, the surface of cryogenically fractured specimens was observed by SEM in a Hitachi S-4700 Electron Microscope with 20 kV of accelerating voltage, after gold coating. In addition, TGA was carried out using a thermal analyzer Mettler Toledo TGA851 at 10[degrees]C/min from 50 to 900[degrees]C under nitrogen flow. The thermal degradation temperatures taken into account were the temperature at 5% of weight loss (T5%) and the temperature of maximum weight loss rate ([T.sub.max]).
The tensile tests were performed using a Lloyd instrument at a crosshead speed of 1 mm/min. at room temperature according to ASTM D-638 Standard Procedure. The test specimens for determining the tensile properties were compression-molded for 2.5 min at 200[degrees]C. The PLA, its nanocomposites, and blends were dried in a vacuum oven at 50[degrees]C for 24 h before testing.
RESULTS AND DISCUSION
PLA undergoes severe thermal degradation at its processing temperatures. In order to study this degradation, isothermal time scans were performed in oscillatory shear flow for up to 30 min at a fixed strain, at a temperature of 200[degrees]C and frequencies of 3.14 and 6.28 rad/s.
The strong decrease of molecular weight due to the thermal degradation of PLA can be attributed to hydrolytic, radical degradation, and/or to residual catalyst that can promote transesterification reactions during processing at high temperatures . Pillin et al.  proposed that the thermal degradation of PLA was mainly due to free radical degradation and not by hydrolysis, where residual catalysts are probably involved in the catalytic production of free radicals. A reduction in viscosity with time is obtained when the weight-average molecular weight decrease in the thermal-oxidative degradation of PLA (see Fig. 1a).
According to the model of thermal degradation in a random chain scission mechanism for PLA presented by Liu et al. , the intluence of time in its PLA molecular weight can be described according to the following expression:
M(0)/M(t)- 1=[M.sub.o][K.sub.x]t/W, (1)
where W is the molecular weight of the polymer repeating unit, [K.sub.x] is the thermal degradation rate, and M(0) and M(t) are the molecular weights of the PLA at zero time and t, respectively. On the other hand, the influence of the weight-average molecular weight ([M.sub.w]) in the Newtonian viscosity ([eta]0) for linear homopolymers is well known :
[[eta].sub.0]= A [M.sup.b][M.sub.w], (2)
where A and b are constants al isothermal conditions, and the b value is about 3.4-3.5. Combining Eqs. 1 and 2, the complex viscosity ([eta]*(t)), at time t can be expressed as a function of time:
[([eta] * (t)).sup.-1/b] = [([eta] * (O)).sup.-l/b] + [[M.sub.o][K.sub.x]([eta] * (O)).sup.-1/b] t/w (3)
The good linearity (regression coefficient of 0.997) of the [([eta]*(t)).sup.1/b] versus time curves at short time values (before 150 s) for the PLA at 3.14 rad/s of frequency showed that the kinetics of the PLA isothermal degradation reactions was mainly a random chain splitting process and the complex viscosity at zero time ([eta]*(0)) can be determined by the intercept of this linear relationship. In a similar way, the viscosity value at 200[degrees]C, 6.28 rad/s of frequency and zero time was calculated. These values are shown in Table 3. Regarding the relationships between the viscosities as a function of time, the expression [([eta]*(0)/[eta]*(t)).sup.1/b] as a function of time at 3.14 and 6.28 rad/s of frequencies and 200[degrees]C are presented in Fig. lb. This curve was practically independent of the frequency at times lower than 15 min, and two stages of PLA thermal degradation were obtained where a linear relationship is obtained at the first stage process from 0 to 150 s. Similar results were found by Liu et al.  in the PLA thermal degradation under isothermal conditions. Nonetheless, no correction was made to eventual degradation of the polymer during the specimen preparation (compression molded at 200[degrees]C) or by differences in the moisture content prior to measurement.
TABLE 3. Dynamic rhcological parameters at zero time and 200[degrees]C: storage modulus (C'(0)), loss modulus (G"(0)), and complex viscosity ([eta]*(0)). Material G' C" [eta]* [eta]* (0) (0) (0) (0) (Pa) (Pa) (Pa s) (Pa s) from thennal degradation model PLA 263 3592 1147 1148 at 3.14 rad/s PLA 870 7178 1151 1140 at 6.28 rad/s PLAPE1 1 84 362 743 -- at 0.5 rad/s PLAPE1 2 -- 359 725 -- at 0.5 rad/s PLAPE1 S 17 72 148 -- at 0.5 rad/s nPLA 218 728 1520 -- at [+ or -] [+ or -] [+ or -] 0.5 17 75 130 rad/s nPLAPE1 1 965 936 2693 -- at 0.5 rad/s [+ or -] 89 [+ or -] 32 [+ or -] 320 nPLAPE1 2 1923 2423 6191 -- at 0.5 [+ or -] [+ or -] [+ or -] rad/s 71 94 115 nPLAPE1 S 1351 1066 3442 -- at 0.5 [+ or -] [+ or -] [+ or -] rad/s 230 174 102
On the other hand, the influence of time in the complex viscosity curves for PLA blends nPLA and its blends at a frequency of 0.5 rad/s can be observed in Fig. 2a. In PLA blends without clay there is thermal degradation of the PLA matrix during its processing and thus the viscosity of the blend should decrease. In order to calculate the loss modulus (G'(0)) and the complex viscosity ([eta]*(0)) at zero time and a frequency of 0.5 rad/s for the PLA and its blends without clay after processing, the data obtained between the time limits of 0-30 min were fitted to high-order polynomial equations. High regression coefficients (0.999) were obtained in these high-order polynomial expressions with a reduction in the loss modulus (G") in the storage modulus (G'), and in the complex viscosity ([eta]*) with time for the PLA and its blends without clay where the dynamic rheological parameters found at zero time are shown in Table 3. The complex viscosity at zero time ([eta]*(0)) was calculated by the following expression:
[eta]*(0) = (([G".sup.2](0) + [G'.sup.2][(0)).sup.0.5])/[omega], (4)
where [omega] is the frequency, and G'(0) and G"(0) are storage and loss moduli at zero time, respectively.
No significant changes in the complex viscosity with time for the nPLA and its blends can be observed in Fig. 2a. This increase in the thermo-oxidative stability of the nanocomposites at 200[degrees]C could be attributed to the physical barrier to [O.sub.2] by the well dispersed sepiolite in these materials, as it will be seen later, and to the interactions between the hydroxyl groups of the sepiolite clay and the carbonyl groups of the different compatibilizer agents used in the blends preparation that provided a barrier effect to the PLA reactive groups in its thermal degradation. Zhou and Xanthos  obtained that the PLA degraded 41.2% on melt processing compared with the neat PLA. However, the addition of MMT-Na" or the organomodified clay limited this degradation to only 22.1 or 19.6%, respectively.
In order to obtain the viscosity and storage modulus values at a frequency of 0.5 rad/s for those samples, an average of the values at each time was calculated in the evaluated time range. The parameters from dynamical rheological analysis at a frequency of 0.5 rad/s for the nPLA and it blends are presented in Table 3. The viscosity values of the PLA obtained by the thermal degradation model and that for the high-order polynomial equations at 3.14 and 6.28 rad/s of frequencies are very similar. In addition, PE1 showed a high thermal stability at 0.5 rad/s of frequency and 200[degrees]C. The viscosity ratios of the blend components ([eta]PEI/[eta]PLA) at 200[degrees]C and 3.14 rad/s of frequency as a function of time are presented in Fig. 2b. An increase of these viscosity ratios was obtained as a function of time due to the low thermal stability of PLA at 200[degrees]C and low frequencies.
Thermal Properties (DSC) and Dynamic Rheological Behavior of the Blend Components
The thermal properties of the blend components are reported in Table 4 and the second heating DSC scans of neat PLA and its nanocomposite (nPLA) are shown in Fig. 3. The thermal properties of the low-density PE (PE1) and the grafted PE (PE2-g-MA) are attributed to their molecular characteristics. The PEl is a PE with long-chain branching and as a consequence it has lower melting peak temperature and crystallinity degree than those of the other types of PEs like the PE2-g-MA sample. On the other hand, neat PLA has a low crystallization rate because of its rigid backbone and the presence of sepiolite does not affect the glass transition temperature ([T.sub.g]) of the PLA matrix, which occurs at ~64[degrees]C. A weak crystallization exotherm was detected followed by a melting endotherm (melting enthalpy of 1.7 J/g) at 153[degrees]C. However, a broad cold crystallization (exothermic peak at 124[degrees]C) and an increase in the melting enthalpy were observed when sepiolite was added to PLA. A melting endotherm at 155[degrees]C with a melting enthalpy of 11.9 J/g was found. We assume that the high dispersion of sepiolite increased the nucleation density for the crystallization of PLA as has been previously reported by Fukushima et al. . Similar results were obtained by Tartaglione et al.  in their nanocomposites of polypropylene (PP) and poly(butylene terephthalate) (PBT) with sepiolite. Nonetheless, little influence on the cryslallinity degree of polyarnide-6 (PA-6) nanocomposites with sepiolite was found by Xie et al.  and Bilotti et al. .
TABLE 4. Thermal properties of the blend components: glass transition temperature ([T.sub.g]), melting peak temperature ([T.sub.m]), melting enthalpy ([DELTA][H.sub.m]). and melting temperature range [DELTA]T. Material [T.sub.g] [T.sub.m] [DELTA] [DELTA] T ([degrees]C) ([degrees]C) [H.sub.m] (J/g) [degrees]C PLA 64 153 1.4 145-170 nPLA 64 155 11.9 145-170 PE1 -- 113 100 50-131 PE2 -- 127 123 50-147 -g-MA
Moreover, an exothermic peak at 124[degrees]C (cold crystallization) and higher melting enthalpy with an endothermic peak at 153 C in the second heating scans was obtained by Signore el al.  in their PLA samples processed at 200 C. The exothermic peak was attributed to the reorganization of amorphous domains into crystalline ones, on account of the increased macromolecuiar flexibility and mobility on increasing temperature due to the reduction of the neat PLA molecular weight during its processing. Similar results were reported by Fukushima et al. . From the DSC results found here, the hydrolysis of the nPLA based on sepiolite and consequent reduction of neat PLA molecular weight during its processing can not be rejected. The sepiolite clay is a hydrophilic material that can absorb moisture during the nPLA preparation by extrusion. Nevertheless, thermomeehanical degradation of the nPLA was lower than that of neat PLA in the rheolog-ical test under isothermal conditions at 200 C, as it was said before.
The complex viscosity ([eta]*) as a function of frequency and the storage modulus (C') as a function of the loss modulus (G") of the neat materials at 200[degrees]C are presented in Fig. 4. The viscosities of the neat low-density PE (PE1) and the grafted PE (PE2-g-MA) decreased as the frequency increased, indicating a pseudoplastic behavior (Fig. 4a). Neat PLA shows a characteristic homopolymer-like terminal (low behavior, expressed by the Newtonian behavior and the power law G' [varies] [G".sup.2] (i.e., terminal zone slope is about 2) at 200[degrees]C, The complex viscosities at 200[degrees]C presented in Table 3 for the neat PLA without thermomechanical degradation are practically the same at 3.14 and 6.28 rad/s of frequencies. The lower shear thinning characters, viscosities, and storage moduli of neat PLA (without thermomechanical degradation) may be in agreement with its molecular characteristics (weight-average molecular weight and molecular weight distribution) [1, 8].
The relationship between G' and G" does not depend on the molecular weight, temperature, and blend composition (in homopolymers with narrow molecular weight distribution and in miscible blends) [29, 33]. These types of plots reflect the elasticity contribution of each individual component in miscible blends, and long-chain branching content or molecular weight distribution in homopolymers (Fig. 4b). Han and Jhon  have shown that in the terminal region (low frequencies) of linear, flexible, and entangled monodispersed homopolymers the following expression is:
log G' = 2 log G" + log(6[M.sub.e]/5[rho]T), (5)
where [M.sub.e] is the entanglement molecular weight, T is the absolute temperature, and [rho] is the density of the material, indicating that plots of log G' versus log G" are independent of molecular weights and very weakly dependent on temperature for entangled monodispersed homopolymers. On the contrary, this plot is very sensitive to the polymer molecular weight distribution in high-density PEs and to the presence of long-chain branching in low-density PEs. The higher storage modulus of PE1 at constant loss modulus (Fig. 4b) than that of PE2-g-MA could be attributed to the presence of long-chain branching content in the former and its wide molecular weight distribution. It has long been recognized that the degree of long-chain branching influences both the viscous and the elastic behavior of low-density PEs .
Blend Morphology by SEM and TEM
The dispersion of the minority phase in immiscible blends during melt blending involves the stretching of drop-like particles until fibers are drawn, followed by the rupture of these filaments in order to form smaller drops. The affine deformation of large drops indeed occurs if the shear stress of the matrix apparently dominates the interfacial stress completely when the drop is subjected to the flow. The coalescence of these drops would, in turn, create larger ones. The balance of these processes determines the final particle size, which is controlled by the viscosity of the components (the viscosity ratio), the melt elasticity, shear stresses and rales in the matrix, the mobility of the interface, and the interfacial tension, because a lower tension promotes the stretching of even smaller drops producing a very fine morphology. Furthermore, the time for breakup is determined by the viscosity ratio of the blend components. In immiscible blends, the lower interfacial tension is the result of the chemical reactions forming copolymers at the matrix/dispersed phase interface. Because of the presence of these copolymers, the coalescence rate decreases since they immobilize the interface, reducing the final particle sizes of the dispersed phase. Thus, the use of compatibilizing agents reduces the interfacial tension and improves the adhesion in polyblends without clay, affecting their final properties [13, 14].
The thermomechanical degradation of the PLA presented before (Figs. 1 and 2) was made in the dynamic rheometer under isothermal conditions at low frequencies. Nevertheless, the PLA thermomechanical degradation in the extruder must be different, because the processing conditions are nonisothermal and very different shear rates can be found in each element of the screw configuration. Therefore, hydrolysis reactions of the PLA matrix could be occurred because of the PLA moisture absorption in the extrusion process, although the PLA and blends were dried before blending. A reduction of 40% in intrinsic viscosity of the PLA processed under the same extrusion conditions of the blends preparation (117 s of mean residence time at 3.18 k/h of mass flow rate and 100 rpm) compared with that of neat PLA was obtained. Thus, its viscosity molecular weight ([M.sub.v]) decreased 49%. Without thermomechanical degradation of PLA in the extrusion process, the viscosity ratio of the blend components ([eta]PEI/[eta]PLA) is about 1 at 200[degrees]C and 10 [s.sup.-1] of shear rate. However, at higher shear rates (kneading elements in the extruder), this viscosity ratio must be lower than 1 at 200[degrees]C because of the higher pseudoplastic character of the dispersed phase (PE1) than that of PLA (see Fig. 4). Nonetheless, the reduction of the PLA viscosity by thermomechanical degradation and/or hydrolysis reactions in the extruder should increase this viscosity ratio. The influence of processing conditions during melt extrusion on the degradation of PLA has been already investigated [8, 27, 28, 34]. This degradation is influenced by the temperature, residence time in the extruder (screw rotation speed and mass flow rate), and the moisture content. Taubner and Shishoo  found a reduction in shear viscosity of about 50% at 210[degrees]C and 120 rpm for the samples conditioned at 65% of relative humidity (RH) prior to measurements.
In polymer blends research, the drop breakup phenomenon has been studied extensively, but the effect of the thermomechanical degradation (chain scission reactions) of the continuous phase (PLA) during the mixing process has not been fully explored [13, 14]. In the extrusion process, the shear viscosity of PLA (matrix) is reduced and a negative effect on the dispersion of the dispersed phase (PE1) should be expected. Then, there are an enhancement of this viscosity ratio ([eta]PEI/[eta]PLA) and a reduction of the matrix shear stress with time. Consequently, an increase of the droplest sizes droplets of the dispersed phase should be obtained.
The SEM morphology of the blends without sepiolile whose dispersed phase is constituted by 17 wt% of PE1 is shown in Fig. 5a and c (P LAPE1 1 and PLAPE1 S blends). Although two commercial grafted polymers (PE2-g-MA and SEBS-g-MA) were used as compatibilizer agents in these blends, no morphological evidence of good adhesion between the matrix and the dispersed phases can be seen. During the cryogenically fracture process used in the surface preparation for SEM characterization, many domains have been pulled away from their previous positions and they remain as empty holes. After compression molding at 200[degrees]C, the low-density PE phase (PE1) has formed domains with very different sizes and forms. Some particles are quasispherical and others exhibit elongated shapes with different aspect ratios of the ellipsoids. These elongated shapes could be ascribed to the low shear stresses in the PLA matrix (very low thermal stability of PLA in the extrusion process) that can not break some treads of the PE1 in those elements of the screw configuration with low shear rate. The number-average diameter ([D.sub.n]) and the ratio of weight-average to number-average diameter ([D.sub.w]/[D.sub.n]) of the dispersed phase are shown in Table 5. It is important to point out that the grafting degree of the PE2-g-MA is lower than that of the SEBS-g-MA (Table 1). Nonetheless, the PLAPE1 S blend has a higher number-average diameter ([D.sub.n]) and similar particle size distribution ([D.sub.w]/[D.sub.n]) of the dispersed phase than that of PLAPE1 1 as can be seen in Table 5. In that sense, the thermal properties of the blend components have some influence in the morphology of these blends ,
The mixing of semicrystalline polymers with different melting temperatures and rheological behavior in extruders is very complex. The thermal properties of the blend components are shown in Table 4. In all the studied blends, the majority phase is PLA with a melting peak temperature of 153[degrees]C, a narrow melting temperature range (145-170[degrees]C) and a low melting enthalpy. On the other hand, the minority phase is a low-density PE (PE1) with a broad melting temperature range (50-133[degrees]C). The PE2-g-MA is a linear PE that melts at 127 C and with also a broad melting temperature range (50-I47[degrees]C) and the SEBS-g-MA is an amorphous copolymer with high plasticating temperatures (180-230[degrees]C). In the first stage of the extrusion process, the polymer that has the lower melting temperature forms the continuous phase (PE1) and in subsequent stages, a phase inversion takes place where the majority phase melts and could forms the continuous phase . In the TEM micrograph of the PLAPE1 S blend without clay (Fig. 6a), the presence of inclusions in the interior of the large dispersed phase particles can be seen. These inclusions could be considered to be PLA and/or SEBS-g-MA particles trapped within the PEI like in "salami" morphology because of the differences between the melting and/or plasticating temperatures of the blend components. This type of morphology also affects the dispersion of the minor component in PLAPE1 S blend because of its influence on the dispersed phase rheological properties (SEBS-g-MA and PLA inside PEI droplets). Furthermore, there is less amount of SEBS-g-MA available for compatibilization. Nonetheless, in the PLAPE1 1 blend, the dispersed phase and the compatibilizing agent (PE2-g-MA) have an onset melting temperature of 50[degrees]C and melt together. The TEM micrographs in Fig. 6a and b were obtained with [O.sub.S][O.sub.4] as staining agent, because it cannot react with PLA, and hence, PE was stained in dark.
TABLE 5. Number-average diameter ([D.sub.n]), ratio of weight-average to number-average diameter ([D.sub.w]/[D.sub.n]); tensile properties: Young's modulus (E), yield stress ([[sigma].sub.y]), tensile strength ([[sigma].sub.b]), elongation at break ([[epsilon].sub.b]) and tensile toughness (EF); thermogravimetric analysis: temperature onset ([T.sub.onset] at 5% of weight loss) and temperature at maximum (first derivative of PEI). Material [D.sub.n] [D.sub.w]/[D.sub.n] E (MPa) ([micro]m) PLA -- -- 2893 [+ or -] 46 PE1 -- -- 108 [+ or -] 2 PEl (b) -- -- 127 [+ or -] 2 PE2-g-MA -- -- 555 [+ or -] 8 SEBS-g-MA -- -- 6.3 [+ or -] 0.2 PLAPE1 1 8.7 1.7 2305 [+ or -] 83 PLAPE1 2 -- -- 1820 [+ or -] 62 PLAPE1 S 20 1.4 2171 [+ or -] 13 nPLA -- 3636 [+ or -] 108 nPLAPE1 1 6.4 1.4 2581 [+ or -] 30 nPLAPE1 2 -- -- 1531 [+ or -] 45 nPLAPE1 S -- -- 3025 [+ or -] 18 Material [[sigma].sub.y] [[epsilon].sub.b] [[sigma].sub.b] (MPa) (%) (MPa) PLA -- 2.2 [+ or -] 0.5 42 [+ or -] 1 PE1 6.8 (a) [+ or -] -- -- 0.2 PEl (b) 6.6 [+ or -] 0.2 317 [+ or -] 15 9.1 [+ or -] 0.5 PE2-g-MA 24 (a)[+ or -] 1 -- -- SEBS-g-MA 1.1 (a) [+ or -] -- -- 0.1 PLAPE1 1 29 [+ or -] 1 52.2 [+ or -] 0.1 25 [+ or -] 1 PLAPE1 2 15 [+ or -] 1 23.9 [+ or -] 0.5 12 [+ or -] 1 PLAPE1 S 32 [+ or -] 1 71 [+ or -] 2 31 [+ or -] 1 nPLA -- 1.8 [+ or -] 0.1 55 [+ or -] 3 nPLAPE1 1 32 [+ or -] 1 20 [+ or -] 3 28 [+ or -] 1 nPLAPE1 2 28 [+ or -] 3 8 [+ or -] 1 18 [+ or -] 3 nPLAPE1 S 47 [+ or -] 3 25 [+ or -] 9 46 [+ or -] 3 Material EF [T.sub.onset] [T.sub.max] PEI (MJ/[m.sup.3]) ([degrees]c) ([degrees]c) PLA 0.48 [+ or -] 346 -- 0.04 PEI -- 443 482 PE1 (b) 24 [+ or -] 2 -- -- PE2-g-MA -- -- -- SEBS-g-MA -- -- -- PLAPE1 1 13 [+ or -] 1 328 483 PLAPE1 2 2.6 [+ or -] 1 355 484 PLAPE1 S 21 [+ or -] 2 338 480 nPLA 0.62 [+ or -] 346 -- 0.04 nPLAPE1 1 5.5 [+ or -] 0.7 356 487 nPLAPE1 2 1.4 [+ or -] 0.8 355 485 nPLAPE1 S 11 [+ or -] 1 353 483 (a.) Determined at 50% of elongation. (b.) Determined at a crosshead speed of 50 mm/min.
It has been reported that the SEBS-g-MA is an efficient emulsifying agent in PP/PA-6 blends and a number-average diameter of the dispersed phase lower than 1 [micro]m has been found in those blends. These results have been explained by the emulsifying process where the SEBS-g-MA compatibilizer agent migrates to the melted dispersed phase, and a copolymer is formed at the interface that reduces the coalescence of the dispersed phase droplets [35-37]. There was not a change either in the dispersed phase sizes when the amount of the compatibilizer agent used varied between 3 to 5 wt% [13, 14]. From the results obtained herein, the PE2-g-MA material seems more efficient than SEBS-g-MA in emulsifying PLAPE1 blends, although it would be expected that the material with the higher grafting degree will be more reactive to PLA (see Table 1). Therefore, the morphology and the largest sizes of the dispersed phase in the PLAPE1 S and PLAPE1 1 blends could be attributed to the following factors: the thermomechanical degradation and/or hydrolysis reactions of the PLA matrix phase in the extrusion process and the "salami" type morphology obtained. In addition, there are other two factors: the higher shear viscosity of the SEBS-g-MA than that of PE2-g-MA (see MFI of these materials in Table 1) and the content of MA of the SEBS-g-MA sample that could enhance the droplet sizes of the dispersed phase in the PLAPE1 S blend. The first factor increases the SEBS-g-MA diffusion times and reduces its migration toward the interface and the second one may induce a higher thermomechanical degradation of the PLA in the extrusion process The lowest complex viscosity ([eta]*(0)) and storage modulus (G'(0)) at zero time of the PLAPE1 S blend presented in Table 3 could be due to this highest thermomechanieal degradation of PLA in the extrusion process and/or to the larger particle sizes of the dispersed phase in this blend. Nevertheless, some evidence of interactions between the phases is found in these blends due to their higher tensile toughness than that of PLA, as it will be presented later.
However, sepiolile is nonswelling clay with a very high surface area (300 [m.sup.2]/g) . Regarding the morphology of PLA with sepiolite, the clay used for the preparation of the nPLA and blends was a commercial unmodified sepiolite with a fibrilar particle shape. The dimensions of a single sepiolite fiber vary between 0.2 and 3 [micro]m in length, 10-30 nm in width, and 5-10 nm in thickness, with an average aspect ratio of about 27. Nano-parlicles are particles having (one or more) dimensions below 100 nm ; in consequence, single fibers of sepiolile could be considered as nanoparlicles. In addition, nanomaterials can be defined as materials having structured components with at least one dimension of less than 100 nm. If high dispersion of the sepiolite clay is obtained in the PLA and their blends, these materials could be considered nanocomposites. in several works [23, 24, 40-42], the neat polymers with sepiolile were considered to be a nanocomposite because of the excellent distribution of unmodified inorganic filler in its finest elemental units, even at concentrations as high as 5 wt%, with no indication of particle aggregations.
In this research, all the components of the blends were fed through the first port of the extruder and the sepiolite (dried before mixing) through the second port. TEM micrographs with and without a staining treatment are shown in Fig. 6b and d for the nPLAPE1 S and nPLAPE1 1, in order to analyze the clay dispersion and where it was placed. As can be seen, the morphology of these composites are rather complex, because several kinds of particles are shown. These TEM micrographs seem to indicate that the clay resides in both phases (PLA and PE1) and in the interface, probably due to the sequence of addition of the components in the extruder. In Fig. 6b and c, single fibrils (about 20 nm of diameter) and some agglomerates of clay with variable dimensions can be seen in both phases (PLA and PE1). In both composites, the sepiolite exhibits nanodispersed structures with a similar aspect ratio of about 26. In Fig. 6c, single fibrils of clay seems to reside at the interface of the nPLAPE1 S sample and in Fig. 6d, a cloud of clay is located near the interface in the nPLAPE1 1 composite. It was reported that due to the discontinuity of the external silicate sheet, a significant number of silanol groups (SiOH) are present at the whole external surface of the sepiolite . Some particles could be located in this interface because of favorable polymer-particle interactions (hydroxyl groups of the sepiolite and carbonyl groups of the SEBS-g-MA compatihilizer agent). In addition, it has been reported that nanodispersed structures of sepiolite in PA-6 may be due to the strong interaction of the PA-6 chains with the Si--OH groups on sepiolite, together with the high shear stresses during compounding that tends to destroy agglomerated structures .
SEM micrographs of nPLA blends (nPLAPE1 1 and nPLAPE1 S) are presented in Figs. 5b and d. A reduction of the dispersed phase particles without elongated par-the dispersed phases can be seen compared with those in blends without clay (Figs. 5a and c and Table 5). The lower droplet sizes of the dispersed phase for the blends with sepiolite could be due to the higher viscosity of the nPLA (the sepiolite increases the PLA matrix viscosity, as it will be seen later) and/or higher thermal stability of the nanocomposite at 200[degrees]C (see Fig. 2a and Table 3). The viscosity ratio of the blend components could be lower than one at high shear rates and a reduction of the sizes of the dispersed phase could be obtained due to the high matrix shear stresses in the extrusion process. In addition, the high viscosity of the matrix and the localization of clay at the matrix-dispersed phase interface that can form a solid barrier can inhibit or prevents drops coalescence . Furthermore, the high shear stresses because of the high thermal stability and high viscosity of the nPLA blend (as will be seen later) produce the breakup of the treads of PE1 in the elements of the screw configuration and no elongated particles of dispersed phase were seen in these samples (see Fig. 5b). During the cryogenically fracture process used in the SEM characterization, many domains have been also pulled away from their previous positions and they remain as empty holes in the nPLAPF1 1 composite because of the low grafting degree of the PE2-g-MA compatibilizer material. The lower dispersed phase droplet sizes in the nPLAPE1 S than those of nPLAPE1 1 and lower amount of empty holes could be attributed to the higher grafting degree of the SEBS-g-MA that increases the blend component interactions in this blend. Chemical interactions might result from a transesterification reaction between sepiolite hydroxyl groups and anhydride groups on the grafted materials. Physical interactions are also possible through hydrogen bonds. Similar results were obtained by Orozco et al.  in their blends of PLA grafted with MA and starch.
Finally, when two immiscible polymers are compounded in mixing equipment, two types of blend morphologies are often observed: dispersed morphology and co-continuous morphology. This last morphology is an unstable intermediate morphology that eventually is transformed into a dispersed morphology at long times for asymmetric blend compositions. This transient co-continuous morphology is influenced by the blend components ratio, thermal properties, rheological behavior, and the processing conditions and equipment used [35, 43]. Co-continuous morphologies (not shown here) were observed for the PLAPE1 2 and nPLAPE1 2 blends because of the high concentration of dispersed phase (PE1) used, high PLA thermomechanical degradation and the thermal behavior of the PLA matrix and the dispersed phase (PE1) in the extrusion process.
The tensile properties, Young's modulus (E), tensile stress at yield ([[sigma].sub.y]), tensile strength ([[sigma].sub.b]), elongation at break ([[epsilon].sub.b]), and the energy consumed during deformation, represented by the area under the stress-elongation curve (EF) of neat materials and blends are presented in Table 5. The tensile properties values of the neat PLA are very similar to those obtained by Signore et al. . Nonetheless, the large standard deviation values obtained in elongation at break could be attributed to variations in the hydrolytic or thermal degradation and/or moisture absorption before or during the preparation of PLA films by compression molding, although the samples were dried before and after compression molding. The PLA tensile strength and elongation at break decrease when the weight-average molecular weight decreases due to PLA degradation . However, the tensile modulus is less affected, because this decrease of the molecular weight of PLA is balanced by an increase in the crystallinity of the sample [8, 27, 34]. High values of standard deviations in PLA mechanical properties were also obtained in other works [8, 27, 34].
On the other hand, all blends without clay have lower Young's modulus and tensile stress and higher elongation at break and tensile toughness (EF) than those of PLA due to the tensile properties of the PE1 added as dispersed phase. As this is a duclile material, the tensile properties were determined and reported at two different values of crosshead speed: 1 and 50 mm/min (Table 5). Blends of PLA as matrix and low-density PE as dispersed phase without compatibilizer agents have been prepared by Wang and Hillmyer . They found materials with lower modulus, tensile strength, and a little higher elongation at break (10%) and energy to fracture (EF) than those of PLA because of the lack of interactions between the phases. A lower deformation at break and consumed energy during the deformation in the blends might be due to the lack of adhesion between the phases or due to the size of the dispersed phase, as the latter might have an adverse effect as a stress concentrator. If there is no adhesion between the phases, there can be no stress transfer at the interface, and the failure of the material is induced when reaching a particular strain that only eliminates the physical union achieved during blending. In well-compatibilized blends, the stress transfer happens through the interface formed by the compatibilizing agent, which should ensure enough bonding to stop crack propagation. It has been reported that the sole presence of van der Waals-type bonds is enough to improve the tensile toughness of the PA-6 blends with rubber. According to this, an optimum particle size and a good adhesion to the matrix are necessary for effective improvements in tensile toughness to occur. Moreover, the degree of functionalization and tensile properties of compatibilizer agents also affect the mechanical behavior as a consequence of their influence on the adhesion between the phases [13, 14].
It is well known that the three mechanisms that contribute to the overall dilatation (damage) of polymer blends under tension are crazing, decohesion of the particles, and cavitations (a component of volume strain). The development of volume strain in polymer blends was expected, because many authors observed the initiation of different kinds of voids in such materials at a microscopic scale. Crazing is a very important form of damage in blends whose matrix is brittle. Decohesion of the particles from the matrix is another important damage mechanism in blends with ductile matrix and poorly adherent particles. Cavitations (a component of volume strain) were also identified as an active mechanism in systems where a rubber-like phase (particles or interphase) is susceptible to "implode" under the effect of the hydrostatic stress induced by the applied tension . However, Keskkula and Schwarz  working with high-impact polystyrene (HIPS) showed from detailed morphological observation that not only crazing in the PS matrix is the active source of damage but also decohesion at the PS/PB interface and cavitations in the PB nodules play a significant role as well.
Many authors have staled that dispersed rubber-like particles constitute preferential sites of cavitations due to the contrast of elastic modulus with the matrix. Such particles can be plain elastomer droplets (binary blends) and also core-shell particles (ternary blends) in which the rubber occupies a thin envelope. In ternary blends, the shell undergoes cavitations, while the core ensures sufficient rigidity. In addition, the thickness of the rubber interlayer in ternary blends is affected by the rubber cavitations. In blends with low compatibilizer content and/or compatibil-izer with a low grafting degree, the rubber interlayer is thin and the crack propagates easily across the section [44, 46]. A similar mechanism of rubber cavitations could be applied to the compatibilizer agents in the PLA/PE blends. The ideal compatibilizer agent is one with higher Poisson's modulus, lower Young's modulus, and tensile strength than the matrix in order to increase the effectiveness of the cavitations process, and higher adhesion capacity between the blend components [37, 44, 46-48].
The grafted PE (PE2-g-MA) has higher tensile strength at 50% of elongation and Young's modulus, and lower Poisson's modulus than grafted SEBS material with elas-tomeric characteristics (see Table 5). On the other hand, the SEBS-g-MA has a higher grafting degree than PE2-g-MA, which enhances the interactions between the blend components. Then, its effectiveness as tensile toughening agent in the PLAPE1 S blend could be higher and in consequence, the energy consumed during the deformation, represented by the area under the stress-elongation curve (EF) is higher for the PLAPE1 S blend than that of the PLAPE1 1 sample. Although, no evidence of interactions were seen in the blends morphology and the number-diameter of the droplets of the dispersed phase for the PLAPE1 S blend is higher than that of PLAPE1 1 sample (see Table 5), the highest tensile toughening was found for the PLAPE1 S blend. Similar tensile toughness and strength values than those of the PLAPE1 S blend were obtained by Ho et al.  in their blends of PLA with linear low-density PE (TPO) as dispersed phase and 5 wt% of TPO-PLA copolymer as compatibilizer agent. Nonetheless, similar tensile strength but lower elongation at break and tensile toughness were found by Wang and Hillmyer  in their blends of PLA/low-density PE compatibilized with 5 wt% of a PE-PLA block copolymer. Blends of PLA with several dispersed phases have been studied with and without compatibilizer agents in other works [2-5, 8, 11, 42-44]. The following polymers: low-density PEs (LDPE), linear low-density PE (LLDPE), metallocene PE (PEm), PP, poly (butylene adipate-co-terephtalate) (PBAT), PAs, and thermoplastic starch (TPS) have been used as dispersed phases in these blends without clay [2-5, 8, 49-51].
In the nanocomposite materials, the addition of sepiolite clay into the PLA allows obtaining nucleated PLA with a higher degree of crystallinity than that of neat PLA (see Fig. 3) that increases its Young's modulus ([pounds sterling]). The enhancement in modulus can be also explained considering that the filler, besides being incompressible and undeformable, provides a high contact surface area owing to the adequate dispersion in the matrix, promoting an increase in the material's stiffness . Higher Young's modulus of a PA-6 nanocomposite with sepiolite than that with montmorillonite was obtained by Xie et al.  and Bilotti et al. . The high tensile strength in the nanocomposite with sepiolite (nPLA) obtained in this work could be also due to the orientation of the single fibers during the tensile lest. Usually, the uniform dispersion of a montmorillonite layered clay results in an increase of the tensile strength and modulus of nPLAs [l9, 20, 36, 37]. The enhancements in storage modulus ([pounds sterling]') at 30[degrees]C obtained by Fukushima et al.  in nPLAs prepared with montmorillonite (Cloisite 30 B) and sepiolite were 17 and 25%, respectively. In this work, the increases in tensile strength and modulus of the nPLA based on sepiolite prepared by extrusion are 34 and 26%, respectively. In addition, no significant reduction in the elongation at break was found here for the nPLA.
On the other hand, the filler dispersion, concentration, and the blend morphology are important factors to be taken into consideration when discussing the final clay reinforcement in the nanocomposite blends. Balakrishnan et al.  showed that the incorporation of linear low-density PEs into PLA improved the tensile toughness of the nanocomposites with montmorillonite but at the expense of stiffness and strength. The addition of OMMT led to a substantial improvement in stiffness in both PLA and PLA/LLDPE nanocomposites. In the present study, only a slight increase in Young's modulus and tensile strength were obtained for the nPLAPE1 1 sample compared to those of the PLAPE1 1 blend (without clay) probably due to the low interactions between the blend components.
The degree of functionalizaiion and tensile properties of the compatibilizer agents also affect the mechanical behavior as a consequence of the adhesion between the phases and cavitation process, already mentioned. Hence, the higher elongation at break and tensile toughness. Young's modulus, and yield stress obtained for the nPLAPE1 S than those of nPLAPE1 1. The presence of sepiolite in the PE interface and/or in the compatibilizer agents for the nanocomposite blends reduces the ability to cavitate and the effectiveness of these compatibilizer agents, resulting in lower elongation at break than those of the blends without clay. However, a yield stress and higher elongation at break and energy to fracture (EF) were obtained in these blends than those of neat PLA and the nPLA. The lowest tensile properties were found in the blends with composition 60/40 (PLAPE1 2 and nPLAPE1 2 blends), due to their co-continuous morphology.
Although melt-blending via extrusion is the most common and industrially practical blending strategy for preparing clay nanocomposites, few research efforts have been made to optimize the extrusion process. Most scientific literature has focused on surface modification of montmorrillonite, and compatibilizer types and concentration to maximize compatibility between the organoclay (montmorillonite) and the polymer matrix [19-21]. Nevertheless, there are few studies in sepiolite modification to increase the compatibilization with PLA. Tartaglioni et al.  obtained that surface treatments of sepiolite with organic surfactants, both adsorbed and grafted, do not significantly improve the sepiolite compatibility with PP and PBT. Moreover, these sepiolite modifications will add complexity to the system, increase the costs of the final composite product and show poor thermal stability with polymers, such as PLA. The use of nonmodified sepiolite clays eliminates all these effects. An excellent dispersion of modified montmorillonite and nonmodified sepiolite in PLA and in polymide-6 were obtained from the work by Fukushima et al.  and Xie et al. , respectively.
Alternatively, investigations concerning the role of processing history, the order and site of addition of the feed streams, and the selection of the mixing screw configuration and operation conditions are relatively scarce. In several works [13, 14], it was found that the micro-structure in ternary nanocomposites was significantly influenced by the blending sequence, which influenced their mechanical properties [52, 53]. It was shown by Dasari et al.  that blending nylon 66 and organoclay initially and later mixing with SEBS-g-MA is the preferred blending sequence to maximize the notched impact strength. Although a two-step process increases the nanoclay dispersion in nylon 66, in PLA that procedure could be detrimental due to its poor thermal stability. High shear rates could increase the heat generated by friction or viscous dissipation. The site of filler addition into the extruder is a very important factor in order to obtain a uniform dispersion without significant variations in the composite properties.
Generally, a high variation in properties is obtained in polymer composites due to its nonuniform morphology and/or filler concentration. The modulus of particulate-filled composites does not offer much information about morphology and interactions. Modulus depends on the orientation of anisotropic particles, but it is not influenced by specific surface area or the strength of interaction and by particles aggregation. Properties measured at larger deformations, that is, elongation at break and tensile strength are strongly dependant on morphology (sizes and its distribution) and interactions between filler and polymer . The standard deviations of the Young's modulus, tensile strength and yield stress were less than 4 and 8%, respectively, for the PLA composites and blends. The higher dispersion in the tensile properties at break (elongation and energy to fracture) in the PLA composite blends could be attributed to a nonuniform dispersion of the sepiolite in these blends (localization of the sepiolile and drop size variations) and/or variation in their sepiolite content by the way of the addition of this clay into the extruder. Moreover, the large standard deviation values obtained in their elongation at break could be also attributed to variations in the hydrolytic or thermal degradation and/or moisture absorption before or during the preparation of the composite films by compression molding. High standard deviations in the elongation at break in PA-6 with sepiolite and montmorillonite clays and PLA composites blends with thermoplastic starch and montmorillonite were obtained in other works [24, 32, 49]. Nonetheless, the high standard deviations in the energy to fracture (EF) for the PLAPE1 2 and nPLAPE1 2 could be attributed to their co-continuous morphology.
TGA of Blends and Composites
TGA curves for PLA, the nPLA and their blends in an inert atmosphere ([N.sub.2]) are shown in Fig. 7. The weight loss due to the formation of volatile products is monitored as a function of temperature. The typical curves of thermal decomposition in one step of neat polymers (PLA and PE1) where the PE1 has a higher thermal stability than PLA can be seen in Fig. 7a. Many authors considered several molecular as well as radical mechanisms for explaining the PLA thermal degradation. According to Mc Neill and Leiper  and Zeng et al. , the main reaction route is a non radical, backbiting ester interchange reaction involving --OH chain ends. This reaction mechanism can, depending on the size of the cyclic transition state, produce lactide, olygomers, or acetaldehyde plus carbon monoxide. Kopinke et al.  proposed that the dominant reaction pathway is an intramolecular trans-esterification for neat PLA, giving rise to the formation of cyclic oligomers.
In addition, the incorporation of sepiolite clay to PLA (nPLA) did not show an effective barrier behavior al the initial stages of decomposition and merely acted as an inert filler with respect to the thermal decomposition of neat PLA (Fig. 7b). Nevertheless, higher thermal stability for its nanocomposite with sepiolite was reported by Fukushima et al. . Nonetheless, Dhanushka et al.  showed that the onset decomposition temperature was not changed with 10 wt% of sepiolite loading. The TGA curves of the nanocomposites of PLA seem to depend on the sepiolite dispersion because the same type of sepiolite was used by these researchers and by us. High sepiolite dispersion could increase the thermal stability, because it may act as a barrier to the mass transport of low molecular weight products formed on decomposition. Generally in the literature is reported that incorporation of montmorillonite layered clay in the polymer matrix improves the PLA thermal stability .
The typical curves of thermal decomposition in two steps of immiscible polymer blends are shown in Fig. 7. Nevertheless, the addition of 17 wt% of PE1 and 3 wt% of compatibilizer agents (PLAPE1 1 and PLAPE1 S blends) result in a negative effect to the onset decomposition temperatures of the PLA matrix when tested in [N.sub.2] (Table 5 and Fig. 7a). In addition, higher weight losses at temperatures below 350[degrees]C for the PLAPE1 1 blend and 375[degrees]C for the PLAPE1 S blend than that of PLA were obtained. The immiscibility of both phases and the interactions between the blend components appear to have a significant effect in the initial temperature of decomposition of the blends without sepiohte with 17 wt% of dispersed phase (PLAPE1 1 and PLAPE1 S blends, see Table 5). The higher thermal stability of the PLAPE1 S sample than that of PLAPE1 1 blend could be due to the higher interactions between the blend components in the PLAPE1 S material, as it was said before. Similar thermal stability behavior of PLA and blends with a grafted PLA with MA and starch was obtained by Orozco et al, .
Therefore, the thermal stability of the PLA matrix in nitrogen is improved in the blends with sepiolite as compared with those without sepiolite. Indeed, a significant increase in Tonset and little variations in [T.sub.max] of the dispersed phase can be observed (Fig. 7b and Table 5). In addition, a similar thermal stability of both nanocomposite blends with 17 wt% of dispersed phase (nPLAPE1 I and nPLAPE1 S blends) was obtained. The presence of sepiolite in the blends seems to increase their thermal stability at temperatures below 400[degrees]C, without this implying a significant improvement in thermal stability of the nanocom-posites with respect to the neat materials (PLA and PE1). At temperatures higher than 400[degrees]C, lower weight loss was found for all PLA blends due to the presence of PE1 as dispersed phase with a higher thermal stability than that of PLA. The lowest interactions between the blend components were found for the PLAPE1 1 blend and the highest for the nPLAPE1 S blend. No changes in thermal stability of the blend 60/40 PLA/PE1 with respect to its nanocomposite were observed.
Dynamic Rheological Behavior of Blends and Composites
The knowledge of the melt rheological behavior is needed to optimize polymer processing conditions. The rheological behavior of immiscible polymer blends is intimately related to its morphology (i.e., the state of dispersion). Furthermore, higher interactions between the blend components and/or finest dispersed phases particles increase the viscosity in the low frequency range (long relaxation times) and a solid-like behavior could be obtained in the log G' versus log G" plots in immiscible blends . Oscillatory shear flow with parallel plates was used to determine the rheological characteristics of the neat materials and blends at 200[degrees]C, because it has been demonstrated that this type of measurement does noi affect or modify the morphology of the blends obtained during mixing. The complex viscosity ([eta]*) as a function of frequency and the storage modulus (G') as a function of the loss modulus (G") at 200[degrees]C of the blends and composites studied are presented in Figs. 8 and 9.
Regarding the rheological behavior of the blends without clay (Fig. 8 and Table 3), the complex viscosities of that with SEBS-g-MA as compatibilizer agent (PLAPE1 S) are lower than those of the blend with PE-g-MA (PLAPE1 1). Furthermore, the blend viscosities are lower than those of the individual components (see Table 3 and Fig. 8a). From the elasticity point of view, both blends have similar elastic moduli (G') at the same loss moduli (G") and the storage modulus curves (G') as a function of loss moduli are located between those of the blend components (Fig. 8b). These results can be ascribed to the morphology of the blends without clay, the largest particle sizes of the PLAPE1 S blend, the reduced compatibili-zation effects of the different grafted materials used, and the thermal degradation of the PLA continuous phase, as it was said before. Nonetheless, the lowest complex viscosity and storage modulus at zero time and 0.5 rad/s of frequency (Table 3) of the PLAPE1 S blend are due to its highest particle sizes. The higher storage moduli (G') at constant loss modulus (G") of the blends than those of PLA continuous phase is due to the PE1 dispersed phase particles with a higher individual storage modulus. Nevertheless, an enhancement in storage modulus and viscosity was obtained in the PLAPE1 2 blend due to its higher content of PE1 particles.
However, the rheological behavior of the composites with sepiolite is quite different. The rheological properties of filled molten thermoplastics and elastomers depend on many factors: volume fraction of filler, particle size, particle shape, applied shear rate, or shear stress and interactions between the filler and polymer matrix. Solid-like behavior, higher shear thinning, viscosity and storage modulus have been reported for filled polymers and polymers nanocomposites [19, 20, 29, 58-60]. There are two factors that can contribute to the yield behavior observed experimentally in filled molten thermoplastics and elastomers. One factor is the particle-particle interactions and the other one is the filler-matrix interactions. When the fillers are well dispersed in the matrix phase, filler-matrix interactions may become predominant over particle-particle interactions. Such a situation can occur when particles are adsorbed on the polymer matrix or when some types of physicochemical interactions exist between the filler and polymer matrix. Additionally, the smaller the size of particles (thus the larger the surface area of particles) in a nanocomposite, the lower the concentration of the filler will be that may give rise to yield behavior and a higher shear thinning behavior .
The complex viscosity ([eta]*) as a function of frequency and the storage modulus (G') as a function o( the loss modulus (G") at 200[degrees]C of the nPLA and blends are shown in Table 3 at 0.5 rad/s of frequency and in Fig. 9. Neat PLA shows characteristic homopolymer-like terminal flow behavior, expressed by the Newtonian behavior and the power law G' [varies] G"2 (i.e., terminal zone slope is about 2) at 200[degrees]C, as it was reported before. However, the nPLA shows a solid like behavior, higher shear thinning character, higher viscosity at low frequencies, and very high storage modulus at low loss modulus than those of neat PLA (without thermomechanical degradation). These results can not be only attributed to the high dispersion of the clay in the nPLA with 5 wt% of sepiolite of high aspect ratio. Although detailed description of the surface chemistry was not made, some interactions between the sepiolite clay and the PLA matrix must exist. It was reported that due to the discontinuity of the external silicate sheet, a significant number of silanol groups (SiOH) are present at the whole external surface of the sepiolite . These interactions could be originated from the hydrogen bonding between the carbonyl groups of the PLA matrix and the hydroxyl groups of the unmodified sepiolite. It has also been reported that nanodispersed structures of sepiolite in PA-6 may be due to the.strong interaction of this polymer with the Si--OH groups on sepiolite, together with the high shear stresses during compounding that tends to destroy agglomerated structures .
In conventional filled polymer systems, the disappearance of the Newtonian plateau at low frequencies, the solid-like behavior and shear thinning usually take place at much higher filler loadings than that used in the nPLA based on sepioiite, indicating the significant effect of the sepioiite on the composite rheological behavior. This is in agreement with experimental observations of PLA/ layered clay nanocomposites, where the enhanced melt viscosity could be attributed to the flow restrictions of PLA chains caused by the strong interaction between the layered clay and PLA molecules [19, 20, 58-60]. It is well known that interconnected structures with anisometric fillers result in an apparent yield stress which is visible in dynamic measurement by a plateau of G' versus G" at low storage modulus . This effect is more pronounced in G' than in G". Gu et al.  also found a stronger shear-thinning behavior at all measured frequencies, higher absolute values of storage modulus and gradual changes of behavior from liquid-like to solid-like with increasing Ca[CO.sub.3] content in a PLA matrix. Their results were attributed to the formation of network structures in PLA/Ca[CO.sub.3] at high Ca[CO.sub.3] contents and to the destruction of the percolating network and the formation of the shear induced orientation of the dispersed Ca[CO.sub.3] particles.
Therefore, it is worth noticing that complex viscosities curves of blends with clay are located within the viscosity curves of PE1 and nPLA. In addition, higher shear thinning character and storage modulus are found for the nanocomposite blends than those samples without clay, and for the nPLAPE1 S material compared to nPLAPE1 1 blend (Fig. 9). Furthermore, a solid-like behavior can also be seen in all blends with clay (Fig. 9). Small particles sizes of the dispersed phase, the presence of interactions between the blend components, and/or the high nanoparticle dispersion increase the viscosity and the storage modulus at low frequencies where G' becomes frequency independent (solid like behavior) [19, 29, 60]. Nonetheless, the solid-like behavior, higher viscosity, and storage modulus observed in the nPLAPE1 1 composite than those of PLAPE1 1 material at 200[degrees]C and 0.5 rads of frequency cannot be explained by a finer particle size of the dispersed phase in the former (see Tables 3 and 5 and Figs. 5b and 9b). These results could be attributed to the high dispersion of sepioiite and the interactions of the carbonyl groups of the PE2-g-MA compalibilizer agent in this blend with the Si--OH groups of the sepiolite (see Fig. 5c and d). The higher viscosity and storage modulus of the nPLAPE1 S material compared with those of the nPLAPE1 1 blend at zero time and 0.5 rad/s of frequency and its highest storage modulus at low loss modulus, the shear thinning and solid-like characters indicate strong interactions between polymer blend components due probably to the higher grafting degree of the SEBS-g-MA compalibilizer agent, as it was said before. On the other hand, a solid-like behavior and the highest shear thinning character, viscosity and storage modulus at zero time and 0.5 rad/ s of frequency were found for the nPLAPE1 2 sample with a co-continuous structure. A material (nPLAPE1 S composite) with similar thermal degradation and higher thermal stability in its processing, a tensile stress at yield, higher tensile strength, elongation at break and tensile toughness values, similar Young's modulus and higher viscosity and storage modulus values in the melt than those of pure PLA was obtained. However, the biodegradability in compost of all blends has to be studied. These results are related to the sepiolite dispersion and interactions between the blend components, to the type of morphology of the different blends and the location of the clay at the PE1 interface and/or the PLA matrix phase, to the thermome-chanical degradation of the PLA phase and the grafting degree of the compalibilizer materials.
On the other hand, one of the most important aspects in the development of new materials in thermoplastics engineering is to achieve a good combination of mechanical properties, processability, and PLA degradability at a moderate cost. As far as mechanical properties are concerned, the main target is to obtain a balance of stiffness and toughness. Table 5 shows the effect of sepiolite clay on stiffness and toughness of PLA and its blends. As previously discussed, the incorporation of the PE1 has improved the toughness of the blends without sepiolite but at the expense of stiffness and strength. The addition of sepiolite led to a substantial improvement in stiffness at room temperature and at 200[degrees]C (see Fig. 9), and in tensile strength in neat PLA and nPLAPE1 S composite. This last composite has also a lower complex viscosity at high shear rates and high tensile toughness but high dispersion in the values of elongation at break and energy to fracture. The other composite blend with 17 wt% of PE1 (nPLAPE1 1) has lower tensile properties and storage modulus at 200[degrees]C than those of nPLAPl S but lower standard deviations in elongation at break and energy to fracture, but this composite has higher complex viscosity at 200[degrees]C and high shear rates. Nonetheless, a lower weight loss than that of neat PLA was obtained at temperatures higher than 400[degrees]C for the PLA composites (see Fig. 7).
The effectiveness of the grafted materials employed as tensile toughening agents in the studied blends was confirmed by the presence of a yield peak in the tensile stress curves and the increase of the elongation al break and the energy consumed during the deformation. The large sizes of the dispersed phases in the PLA blends without clay could be attributed to the thermal degradation of the PLA matrix during the extrusion process. The sepiolite clay reduced the thermo-oxidative degradation of the nPLA in the evaluated time range in isothermal conditions. The presence of sepiolite at the PE interface and in the PLA matrix phase reduces the effectiveness of these compatibilizer agents, resulting in lower elongation at break than those of the blends without clay. Nevertheless, the nanocomposite blends exhibited similar thermal degradation, higher thermomechanical degradation, lower tensile strength, and Young's modulus values and increased elongation at break, tensile toughness, complex viscosity, and storage modulus compared with those of the nPLA. The blend prepared with SEBS-g-MA as compatibilizer agent (PLAPE1 S) was the toughest in the tensile test. The highest storage modulus at low loss modulus, shear thinning and solid-like characters of the nPLAPE1 S blend indicate high dispersion of the sepiolile clay and strong interactions between the polymer blend components. A material (nPLAPE1 S composite) with similar thermal degradation and higher thermal stability in its processing, a tensile stress at yield, increased tensile strength, elongation at break and tensile toughness values, similar Young's modulus and higher viscosity and storage modulus values in the melt than those of PLA was obtained. These results are related to the sepiolite clay dispersion and interactions between the blend components, to the type of morphology of the different blends, the localization of the clay at the PE interface and in the PLA matrix phase, the thermomechanical degradation of the PLA matrix phase in the extrusion process and the grafting degree of the compatibilizer materials employed.
NOMENCLATURE Ca[C0.sub.3] Calcium carbonate C[O.sub.2] Carbon dioxide [D.sub.n] and Average particle diameters [D.sub.w] of the dispersed phase (number and weight. respectively) DSC Differential scanning calorimetry E Young's modulus EF Area under the stress-elongation curve GD Grafting degree G' Storage modulus G" Loss modulus HIPS High-impact polystyrene MA Maleic anhydride MFI Melt How index values MMT Montmorillonite [M.sub.w], Average molecular [M.sub.n] and weights (weight, number, [M.sub.v] and viscosity, respectively) [N.sub.2] Nitrogen nPLA Nanocomposite of PLA [o.sub.2] Oxygen OMMT Organophilic montmorillonite (modified with octadecyl-ammonium) Os[O.sub.4] Osmium tetroxide PS Polystyrene PA-6(s) Polyamide-6(s) PB Polybutadiene PBAT Poly (butylene adipate- co-terephtalate) PEm Metallocene polyethylene PE(s) Polyethylene(s) PE- Polyethylene g- grafted with MA Maleic anhydride PLA Poly(lactic acid) PP(s) Polypropylene(s) PP- Polypropylene g- grafted with MA Maleic an hydride RH Relative humidity SEBS- Styrene/ethylene- g-MA butylene/styrene rubber SEM Scanning electron microscopy TEM Transmission electron microscopy [T.sub.g]] Glass transition temperature TGA Thermogravimetric analysis [T.sub.m] Melting peak temperature [T.sub.max] Temperature at maximum (first Derivative of PEl). TPO Linear low-density polyethylene TPS Thermoplastic Starch [T.sub.onset] Temperature onset [DELTA] Melting [H.sub.m] Enthalpy [DELTA] Melting T Temperature Range [[epsilon]. Elongation sub.b] at break H Intrinsic viscosity H* Complex viscosity [rho] Density [[sigma]. Tensile sub.b] Strength [[sigma]. Tensile sub.y] stress at yield [omega] Frequency
(1.) D. Garlotta,.J. Polym. Environ., 9, 63 (2001).
(2.) M. Coltelli, S. Bronco, and C. Chinea, Polym. Degraci Stab., 95, 332 (2010).
(3.) K. Anderson, S. Lim, and M. Hillmyer, J. Appl. Polym. Sci., 89, 3757 (2003).
(4.) Y. Wang and M. Hillmyer,.J. Polym. Sci. Part A: Polym. Chem., 39, 2755 (2001).
(5.) C.H. Ho, C. Wang, C. Lin, and Y. Lee, Polymer, 49, 3902 (2008).
(6.) R. Auras, B. Harte, and S. Selke, Macromol. Biosci., 4, 835 (2004).
(7.) R. Kulkarni, K. Pani, C. Neuman, and F. Leonard, Arch. Surg., 93, 839 (1966).
(8.) F. Signori, M. Coltelli, and S. Bronco, Polym. Degrad. Stab., 94, 74 (2009).
(9.) Y. Li and H. Shimizu, Macromol. BioscL, 7, 921 (2007).
(10.) I. Noda, M. Satkowski, A. Dowrey, and C. Marcott, Macromol. Biosci., 4, 269 (2004).
(11.) H. Balakrishnan, A. Hassan, M.U. Wahit, A.A. Yussuf, S. Bahri, and A. Razak, Mater. Des., 31(7), 3289 (2010).
(12.) J. Lu, Z. Qiu, and W. Yang, Polymer, 48, 4196 (2007).
(13.) D. Paul and S. Newman, Polymer Blends, Academic Press, USA (1978).
(14.) W. Baker and G. Scott, Reactive Polymer Blending, Hanser Publisher, Cincinnati (2001).
(15.) C. Zhang, L. Feng, X. Gu, S.S. Hoppe, and G. Hu, Polymer, 48, 5940 (2007).
(16.) J. Kim, H. Zhou, S. Nguyen, and J. Torkelson, Polymer, 47, 5799 (2006).
(17.) M. Diaz, S. Barbosa, and N. Capiati, Polymer, 48, 1058 (2007).
(18.) A. Bhattacharyya, A. Ghosh, A. Misra, and K. Eichhorn, Polymer, 46, 1661 (2005).
(19.) S. Sinha Ray and M. Okamoto, Prog. Polym. Sci., 28, 1539 (2003).
(20.) S. Paviidov and D. Papaspyrides, Prog. Polym. Sci., 33, 1119 (2008).
(21.) F. Fenouillot, P. Cassagnau, and J. Majeste, Polymer, 50, 1333 (2009).
(22.) S. Bourbigot and G. Fontaine, Polym. Test., 27, 2 (2008).
(23.) K. Fukushima, D. Tabuani, and G. Camino, Mater. Sci. Eng., 29, 1433 (2009).
(24.) S. Xie., S. Zhang, F. Wang, M. Yang, R. Seguela, and J.M. Lefebvre, Compos. Sci. Technol., 67, 2334 (2007).
(25.) M. Kurata and T. Tsunashima, Polymer Handbook, Wiley Interscience, New York (1999).
(26.) F. Kopinke, M. Remmler, K. Mackenzie, M. Moder, and O. Wachsen, Polym. Degrad. Stab., 53, 329 (1996).
(27.) I. Pillin, N. Monrelay, and Y. Grohems, Polym. Degrad. Stab., 93, 321 (2008).
(28.) X. Liu, Y. Zou, W. Li, G. Cao, and W. Chen, Polym. Degrad. Stab., 91, 3259 (2006).
(29.) C.D. Han, Rheology and Processing of Polymeric Materials. Volume 1: Polymer Rheology, Oxford University Press, Oxford (2007).
(30.) Q. Zhou and M. Xanthos, Polym. Degrad. Stab., 94, 327 (2009).
(31.) G. Tartaglione, D. Tabuani, G. Camino, and M. Moisio, Compos. Sci. Technol., 68(2), 451 (2008).
(32.) E. Bilotti, R. Zhang, H. Deng, F. Quero, H.R. Fischer, and T. Peijs, Compos. Sci. Technol., 69(15-16), 2587 (2009).
(33.) C.D. Hand and J.K. Kim, Polymer, 34, 2533 (1993).
(34.) V. Taubner and R. Shishoo, J. Appl. Polym. Sci., 79, 2128 (2001).
(35.) C.D. Han, Rheology and Processing of Polymeric Materials. Volume 2: Polymer Processing, Oxford University Press, Oxford (2007).
(36.) C. Rosales, V. Contreras, M. Matos, R. Perera, N. Villarreal, D. Garcia-Lopez, and J.M. Pastor, J. Nanosci. Nanotechnol., 8, 1742 (2008).
(37.) G.M. Kim, G.H. Michler, J. Rosch, and R. Mulhaupt, Acta Polym., 49, 88 (1998).
(38.) A. Helmy and S. de Busetti, Appl. Surf. Sci., 255. 2920 (2008).
(39.) A.P. Kumar, D. Depan, N.S. Tomer, and R.P. Singh, Frog. Poiym. Sci., 34, 479 (2009).
(40.) T. Dhanushka Hapuarachchi and T. Peijs, Composites, 41(8), 954 (2010).
(41.) R. Grim, Clay Mineralogy, McGraw-Hill, New York (1962).
(42.) V.H. Orozco, W. Brostow, W. Chonkaew, and B.L. Lopez, Macromol. Symp., 277, 69 (2009).
(43.) L.R. Ming, Y. Wei, and Z. ChiXing, Polym. Bull., 59, 545 (2007).
(44.) C. G'Sell, S. Bai. and J. Hiver, Polymer, 45, 5785 (2004).
(45.) H. Keskkula and P.D.R. Schwarz, Polymer, 27. 211 (1986).
(46.) S.L. Bai, C. G'Sell, J. Hiver, and G. Malhieu, Polymer, 46, 6437 (2005).
(47.) A. Wilkinson, M. Clemens, and V.M. Harding, Polymer, 45, 5239 (2004).
(48.) S.L. Bai, G.T. Wang. J.M. Hiver, and C G'Sell, Polymer, 45, 3063 (2004).
(49.) O.H. Arroyo, M.A. Huneaull, B.D. Favis. and M.N. Bureau, Polym. Compos., 3, 114 (2010).
(50.) W. Zhang, L. Chen, and Y. Zhang, Polymer, 50, 1311 (2009).
(51.) N. Reddy, D. Nama, and Y. Yang, Polym. Degracl. Slab., 93, 233 (2008).
(52.) D. Wu, Ch. Zhou, W. Yu, and F. Xie, J. Appi. Poiym. Sci., 99, 346 (2006).
(53.) D. Garcia-Lopez, S. Lopez-Quinlana, I. Gobernado-Mitre, J.C. Merino, and J.M. Pastor, Poiym. Ben., Sci., 47, 1033 (2007).
(54.) A. Dasari, Z.Z. Yu, and Y.W. Mai, Polymer, 46, 5986 (2005).
(55.) I.C. McNeill and H.A. Leiper, Poiym. Degrad. Stab., 11. 309(1985).
(56.) Q.H. Zeng, A.B. Yu, G.Q. Lit, and D.R. Paul, J. Nanosci. Nanotechnol.,5, 1574(2005).
(57.) F. Bcllucci, G. Camino, A. Frache, and A. Sarra, Poiym. Degrad. Stab., 92, 425 (2007).
(58.) G. Galgali, C. Ramesh, and A. Lcle, Macromolecules, 34, 852 (2001).
(59.) A. Durmus, A. Kasgoz, and C.W. Macosko, Polymer, 48, 4492 (2007).
(60.) S. Sinha Ray, K. Yamada, M. Okamoto, and K. Ueda. Poly-mer, 44, 857 (2003).
(61.) Y. Di, S. lannacc, E. Di Maio, and L. Nicolais, J. Poiym. Sci. Part B: Poiym. Phys., 43, 689 (2005).
(62.) S. Gu, C. Zou, K. Zhou, and J. Ren, J. AppL Poiym. Sci., 114, 1648 (2009).
Correspondence to: K. Nunez; e-mail: firstname.lastname@example.org
Contract grant sponsor: Simon Bolivar University: contract grant number: Grupo G-014.
Contract grant sponsor: Minislerio de Educacion y Ciencia/Spain; contract grant number: MAT2008-06379.
Contract grant sponsor: Junta de Castilla y Leon; contract grant number: GIE 104.
Published online in Wiley Online Library (wileyonlinelibrary.com). [c] 2011 Society of Plastics Engineers
K. Nunez, (1), (2) C. Rosaies, (1) R. Perera, (1), (2) N. Villarreal, (3) J.M. Pastor (2), (3)
(1.) Departamento de Mecanica, Universidad Simon Bolivar, Apdo 89000, Caracas 1081, Venezuela
(2.) Departamento de Fisica de la Materia Condensada E.I.I. Universidad de Valladolid, Paseo del Cauce 59. 47011, Valladolid, Spain
(3.) CIDAUT, Foundation for Research and Development in Transport and Energy, Parque Tecnologico de Boecillo 47151, Valladolid, Spain
|Printer friendly Cite/link Email Feedback|
|Author:||Nunez, K.; Rosales, C.; Perera, R.; Villarreal, N.; Pastor, J.M.|
|Publication:||Polymer Engineering and Science|
|Date:||May 1, 2012|
|Previous Article:||Synergistic effect of polyaniline and multiwalled carbon nanotube on the microstructure and the electric property in PP/PANI/MWNT composites.|
|Next Article:||Cavitation-induced rupture in high-density polyethylene copolymers.|