Solid-state polymerization of poly(ethylene terephthalate): effect of organoclay concentration.
Polyethylene terephthalate) (PET) is a semicrystalline engineering polymer of affordable cost and high performance. In view of its excellent transparency and good barrier properties, it is used in a large variety of applications such as containers, films, bottles, and fibers (l). One of the areas of growing interest for using PET is for food and beverage packaging. Substantial growth in this area requires improvements in the barrier to [O.sub.2] and C[O.sub.2], along with a reduction of the weight of the final products. Recent studies show that the presence of organoclay platelets in PET can lower permeability to oxygen [2-5] and water vapor [6, 7]. Moreover, improvements of mechanical properties of PET by incorporation of organoclays have been reported [3, 5, 8, 9], The preparation of polymer nanocomposites by melt-mixing at high temperature, however, faces challenges mainly related to the degradation of both polymer and nanoclay modifiers at high processing temperatures. In the case of PET nanocomposites, several efforts have been made to lower the rate of degradation and enhance dispersion of organoclays in PET, using a more stable clay modifier [6, 10, 11]. Other approaches use a clay-supported catalyst [12, 13] or a chain extender [9, 14],
Solid-state polymerization (SSP) of PET nanocomposites is a common process in the production of PET. It is also a practical route to compensate for polymer [M.sub.w] reduction that occurs during the melt-mixing process . Although SSP reactions are slower than those involving chain extenders, the former do not lead to significant changes in the chemical structure of PET. Moreover, some chain extenders can cause side reactions with undesirable by-products. Different researchers have demonstrated that SSP of PET in the presence of nanoclays and nano-Si[O.sub.2] was feasible [15-17]. These studies have also shown a reduced rate of the SSP, compared to the neat PET. SSP is usually carried out under moderate temperature conditions in the reaction temperature range of 200-230[degrees]C. Thus, SSP could raise the molecular weight of PET with less thermal degradation than melt phase polymerization, while reducing the contents of byproducts, such as acetaldehyde and oligomers, to acceptable levels [18-21]. SSP of PET involves two reversible equilibrium reactions, esterification and transesterification  and it is controlled by two types of diffusion. The diffusion of reaction byproducts (physical diffusion) that determines the rate of the forward reactions and the diffusion of end-groups (chemical diffusion) that allows the reaction to proceed. These two types of diffusion are controlled by different parameters, such as [M.sub.w] of the prepolymer, particle size, temperature, polymer end groups, catalyst, heat stabilizer, and crystallinity of PET , At temperatures below 200[degrees]C, as the mobility of the chain ends is quite low, the reaction rate is determined by diffusion of chain ends. However, at higher temperatures where PET particles can be fused together, the diffusion of by-products (water and ethylene glycol) out of the reactor becomes more difficult; thus byproduct diffusion controls SSP. Inert gas plays a significant role in SSP by removing the by-products, excluding oxygen and preventing polymer oxidation [20, 23-25],
Rheological methods are useful to characterize and analyze the microstructure of polymer nanocomposites as well as their processability . Several researchers studied the state of organoclay dispersion in polymer matrices using small amplitude oscillatory shear rheometry [27, 28]. A transition from liquid-like to solid-like behavior at low frequencies was observed from the storage modulus and complex viscosity in nanocomposites and filled composites [29-31], Conversely, polymer degradation and [M.sub.w] reduction were also detected from changes in the storage modulus and complex viscosity data at high frequencies, where these parameters were found to be lower than those of the neat polymer [32-35], However, the rheological behavior of polymer nanocomposites is quite complex and dependent on so many parameters including the organoclay dispersion, the aspect ratio, nanoclay orientation, interactions between nanoclay and polymer chains, and nanoclay-nanoclay interactions, as well as [M.sub.w] of the polymer matrix [26, 36],
This work considers SSP of PET and PET nanocomposites incorporating different concentrations of Cloisite 30B (C30B) in a batch reactor for different reaction times. Moreover, the effect of polymer particle size on the rate of SSP is analyzed. Rheological measurements, intrinsic viscosity, and titration measurements are used to characterize the samples after SSP. More specifically, the changes in the rheological behavior of PET nanocomposites with increasing [M.sub.w] of the matrix are also presented. SSP was carried on PET and PET nanocomposites of different C30B concentration prepared first by water-assisted melt-mixing. The novelty of this work compared to our previous investigations is to report SSP data for different particle sizes and different reaction times as well as to establish the relationships between rheological and thermal properties and molecular weight and end groups of neat PET and PET nanocomposites. It should be pointed out that, in our previous work [4, 37], we obtained better dispersion and distribution of C30B in the PET nanocomposites prepared using water-assisted melt-mixing compared to the preparation without water.
A general purpose PET (PET 9921, Eastman Co, Kingsport, TN) was utilized as the matrix; it has a molecular weight of 65,000 g/mol, a melting point of 243[degrees]C, and a glass transition of 77[degrees]C. The organoclay C30B (Southern Clay Products, Gonzales, TX) was used at nominal concentrations: 2, 3.5, and 6 wt%. C30B is based on the modification of sodium montmorillonite by ion exchange with methyl, tallow, bis-2-hydroxyethyl quaternary ammonium cations , This organoclay has platelet shape. The modifier concentration of this organoclay is 90 meq/ 100 g clay. Based on the information presented by the supplier, the typical dry particle size of this organoclay is 2 [micro]m (less than 10%), 6 [micro]m (less than 50%), and 13 [micro]m (less than 90%).
Phenol, 1, 1, 2, 2-tetrachloroethane, o-cresol, dichloro methane, methanol, and potassium hydroxide (KOH) supplied by Fischer Scientific (ON, Canada) were used without additional purification for the determination of the intrinsic viscosity and of carboxyl groups.
PET and the nanocomposites were processed using a corotating twin screw extruder (TSE; Berstorff ZE25, Hannover, Germany) with a 25-mm diameter (D) screw and length-to-screw diameter ratio of 28 at a screw speed of 200 rpm. The screw configuration is presented in Ref. , The temperature profile was 245, 265, 260, 255, 255, and 255[degrees]C from the hopper to the die.
We used liquid nitrogen to cool the PET granules and then by using a grinder supplied by Waring Commercial (Heavy duty blender model), the PET powder and the organoclay were put in a box and dry mixed. The mixture was fed into the TSE at 3 kg/h. A 2-L pressure vessel (Pair Instrument, Moline, IL) was utilized to produce and feed steam in the second zone of the TSE at a rate of 0.3 L/h, temperature of 160[degrees]C, and pressure of 550 kPa. The temperature of the second zone of the extruder was 265[degrees]C. Finally, water was removed from the TSE using a vacuum pump in the last zone of the extruder. The extrudate was collected after reaching steady state, dried, and ground using the Waring blender.
SSP was carried out to rebuild the reduced [M.sub.w] of hydrolyzed PET and PET nanocomposites. Before SSP, the ground PET and PET nanocomposites were sieved to a powder of size less than 200 and 400 [micro]m. SSP was performed in a 1-L stainless steel stirred reactor equipped with a heating jacket provided by Supercritical Fluid Technologies (Newark). The reactor was operated at 215[degrees]C under N2 for 4, 8, and 12 h with the powder size of 400 [micro]m. This temperature was the maximum value at which we could operate without encountering experimental difficulties. The effect of smaller powder size (200 [micro]m) was also investigated for 8 h SSP. The flow rate of [N.sub.2] into the reactor was 2 L/min. After SSP, the heater was turned off, but the nitrogen purge was continued to lower the reactor temperature to 100[degrees]C, prior to collecting the resulting polymer or nanocomposites. Some of the sample codes are presented in Table 1.
After SSP, the samples were compression molded at 270[degrees]C under nitrogen to obtain 25-mm disks, used for rheometry. All the samples were dried under vacuum for 24 h at 80[degrees]C prior to melt processing and molding.
Intrinsic viscosity, [[eta]], measurements were performed using an Ubblehode capillary viscometer at 25[degrees]C in a mixture of phenol/1, 1, 2, 2-tetrachloroethane (60/40, w/w). The samples were maintained in the mixture of the above solvent at 110[degrees]C for less than 30 min to be completely dissolved. The solution was then cooled to room temperature. Intrinsic viscosity was calculated using the Solomon-Ciuta equation :
[[eta]] = (2[([[eta].sub.r] - 1 1n [[eta].sub.r]).sup.0.5]]/c (1)
where [[eta].sub.r] is relative viscosity obtained from the ratio of average the solution flow time (/) to the average solvent flow time ([t.sub.0]) in an Ubblehode capillary viscometer, c is the polymer solution concentration, g/dL. The weight-average molecular weight [M.sub.w] was calculated using the following equation :
[[eta]] = 4.68 x [10.sup.-4] [M.sup.0.68.sub.W] (2)
and the number-average molecular weight was estimated based on the Ulgea equation :
[[eta]] = 2.52 x [10.sup.-4] [M.sup.0.8.sub.n] (3)
Carboxyl end group (CEG) content in the samples was determined by titrating a solution of the PET and PET nanocomposites in o-cresol/dichloro methane according to the ASTM D7409-07. KOH (0.005 M) solution in methanol was used as a titrator solution and bromophenol blue as indicator. The CEG content is then from the following equation:
CEG =[([V.sub.s] - [V.sub.b]) x M x 1000]/w (4)
where [V.sub.s] and [V.sub.b] are the volumes of KOH to titrate the sample and the blank, respectively. M is molarity of the KOH/methanol solution and w is the weight of the PET sample or the weight of the PET in the nanocomposites by subtraction of the weight of the nanoclay.
Rheological measurements were carried out at 265[degrees]C under nitrogen using a Bohlin Gemini HR rheometer (Malvern Instruments, Worcestershire, UK) and an Advanced Rheometric Expansion System (ARES, TA Instruments, New Castle, DE), both with a 25-mm parallel plate geometry. The samples were dried under vacuum at 80[degrees]C for 24 h before the rheological tests. Time sweep tests were performed at 0.1 Hz, and frequency sweep tests were done in the linear viscoelastic region for each sample determined by strain sweep tests. Frequency sweep tests over an angular frequency range of 0.3-100 rad/s were performed from low to high and high to low frequencies to assure the repeatability of the data.
The thermal properties of the neat PET and its nanocomposites were determined by differential scanning calorimetry (DSC Q1000, TA instruments, New Castle, DE) under N2 atmosphere. The samples were heated from room temperature to 300[degrees]C at 10[degrees]C/min and held at that temperature for 3 min, then cooled to 30[degrees]C at a constant rate of 10[degrees]C/min.
RESULTS AND DISCUSSION
Molecular Weight Changes During SSP
The intrinsic viscosity ([[eta]]) and molecular weight ([M.sub.w]) of PETs obtained after different SSP times are presented in Table 2. It shows that hydrolysis during melt-mixing in the presence of steam causes a reduction of the [M.sub.w] of PET from 65,000 to 32,000 g/mol, while the polydispersity (PD) becomes slightly smaller. This can be related to the more favorable hydrolysis of the long chains compared to shorter ones. During SSP, [M.sub.w] of PET increases significantly, thus, after 4 h of SSP, [M.sub.w] increases from 32,000 to 57,000 g/mol. The effects of time and particle size are also observed in Table 2. The particle size has a significant effect on [M.sub.w]. During 8 h of SSP, the PET with particle size of 400 [micro]m reaches a [M.sub.w] of 68,000 g/mol, whereas in the case of 200 [micro]m particles the [M.sub.w] attains 84,000 g/mol, which is close to the [micro]w of the PET exposed to SSP for 12 h for the 400 [micro]m particles. The SSP rate of PET increases with decreasing particle size, suggesting that SSP is controlled by diffusion of the by-products; as the particle decreases, the larger interfacial area and the shorter diffusion path to the particle surface facilitate by-products diffusion.
During the initial stage of SSP, the polymerization reaction is believed to occur primarily near the pellet surface because the by-products can be removed easier. As SSP progresses, however, end-group concentrations at the pellet surface are depleted and the SSP reactions proceed at greater depths within the pellets; then, by-products diffusion becomes increasingly important, limiting the SSP rate.
Table 2 shows that the PD increases significantly during SSP, from 2.18 for hydrolyzed PET before SSP to 2.52 after 12 h of SSP. As the forward reactions (esterification and transesterification, shown in Fig. 1) continue, diffusion limitations of byproducts (water and ethylene glycol) become noticeable, leading to a radial concentration gradient of by-products inside the particles with a lower concentration in outer regions of the particles , Therefore, the effective polymerization rate in outer regions is greater resulting in a gradient in the local [M.sub.w], which in turns leads to an increase of the overall PD , Moreover, a distribution of polymer crystallite sizes could cause a microscopic nonuniformity of the solid reactant, which consequently will lead to a larger PD as observed in SSP for PA-6 .
Rheology-Molecular Weight-Nanostructure Relations
The rheological results of small-amplitude oscillatory frequency scans are reported in Fig. 2. Figure 2a illustrates the effects of SSP and particle size on the complex viscosity ([[eta].sup.*]) versus angular frequency for the neat PETs. Higher complex viscosity values are observed with increasing SSP time and decreasing particle size. The zero-shear viscosity ([[eta].sub.m0]) changes from 60 Pa.s in hydrolyzed PET before SSP to 2100 after 12 h SSP. Conversely, by reducing the particle size of the neat PET, the zero-shear viscosity becomes larger: [[eta].sub.m0] is 800 Pa.s for the 400 [micro]m particle size sample, whereas it is 2200 Pa.s for the 200 [micro]m particle size sample. The increases of the complex viscosity for the PET samples after SSP are in accordance with the increasing [[eta]] and related [M.sub.w]. It is also observed that as [M.sub.w] increases, the behavior of neat PET changes from pseudoNewtonian to non-Newtonian, which is due to the increased number of chain entanglements with increasing chain length. The zero-shear viscosity, [[eta].sub.m0], can be used as a scaling factor to obtain a master curve for the complex viscosity data of all the PETs, independent of [M.sub.w], as shown in Fig. 2b. All the curves are superimposed to one line, indicating that the structure of all of the PET chains before and after SSP is the same.
Figure 2c reports the storage modulus versus the loss modulus for the different PET samples. All PET data are superimposed on each other, except those of the PET before SSP (S-0), which exhibit very low G' values. Generally, at the same [M.sub.w], polymers with larger PD show higher G' as the high molecular chains contribute significantly to the elastic modulus [41, 42], Hence, the narrower PD PET exhibits a lower G'. However, the slopes of all the curves are the same, which is another signature of the linear chain structure of PET even after SSP. In our previous work, we verified by [sup.1]H NMR and [sup.13]C NMR that after SSP that the structure of PET chains did not change and the linear structure was preserved after SSP ,
The reduced storage modulus (G'/[omega] x [[eta].sub.m0]) as a function of [omega] x [[eta].sub.m0] for the different PETs is also presented in Fig. 2d. We observe that [[eta].sub.m0] is not a good scaling factor for the storage modulus compared to the complex viscosity as illustrated in Fig. 2b. As explained for Fig. 2c, G' is highly sensitive to PD due to the more important contribution of the tail of the high [M.sub.w] polymer chains to the elasticity of the polymer , As depicted in Fig. 2e the value of the phase angle ([delta]) at low frequencies is 90[degrees] and it decreases with shifted frequency; this is the typical behavior of linear polymer chains , As [M.sub.w] increases, G' becomes more significant compared to the loss modulus (G"), and [delta] decreases with increasing frequency. These results show no indication that branching or crosslinking occurred during SSP and that the linear structure of PET was maintained. Finally, the effect of PD can be noted as the reduction of the phase angle with frequency is more rapid for the larger PD (more elastic) PET.
Figure 3 presents the relationship between the zero-shear viscosity and the weight-average molecular weight of PET in the range of 25,000-84,000 g/mol at a reference temperature of 265[degrees]C, using the formula;
[[eta].sub.m0] = k[M.sup.a.sub.W] (5)
The values of a and k are found to be 3.6 and 2.95 x [10.sup.-15] Pa x s x [mol.sup.3.6]/[g.sup.3.6], respectively. The value of a demonstrates that the molecular weight of the PETs studied here are above the critical molecular weight for entanglements ([M.sub.c]) and [[eta].sub.m0] of PET scales with [M.sub.w] within the range as expected for linear flexible polymer chains. The value of 3.6 compared to the usual 3.4 could be attributed to the PD larger than 2, and also the presence of comonomer in the structure of the as-received PET ,
The complex viscosity at 100 rad/s can also be scaled with [M.sub.w] using:
[[eta].sup.*.sub.m(100 rad/s)] = k[M.sup.a.sub.W] (6)
The values of a and k are found to be 3.14 and 3.94 X [10.sup.-13] Pa x s x [mol.sup.3.14]/[g.sup.3.14].
The linear viscoelastic data of the neat PET and PET nanocomposites before SSP have been presented elsewhere , However, we present in Fig. 4a plots of the complex viscosity versus complex modulus that illustrate better the effect of clay concentration on the rheological behavior of the PET nanocomposites. Above 2 wt% C30B, the complex viscosity is seen to rapidly increase with decreasing complex modulus (or frequency), indicative of an apparent yield stress for the concentrated nanocomposites, due to the stronger particle-particle and/or polymer-particle interactions. Consequently, these results show that the percolation threshold is somewhere between 2 and 3.5 wt%. Plots of the storage modulus versus loss modulus (Fig. 4b), analogous to Cole-Cole plots, are used to investigate changes in the microstructure of the PET nanocomposites. We observe that for a given loss modulus, the storage modulus increases significantly with increasing nanoclay content. The slope of the storage modulus versus loss modulus (log-log plot) is reduced significantly with increasing nanoclay content, illustrating changes in the microstructure with a transition from liquid to solid-like behavior.
Figure 5 reports the complex viscosity and storage modulus data for the PET nanocomposites after SSP for different reaction times and two particle sizes. Figure 5a-c presents the data for the 2, 3.5, and 6 wt% organoclays, respectively. As expected, the complex viscosity in the whole range of angular frequency increases with SSP time and decreased particle size in nanocomposites with 2 wt% C30B (Fig. 5a). In the linear viscoelastic regime, the behavior of the matrix is prominent at high frequencies, whereas the low-frequency data are quite sensitive to the organoclay interactions and structure formation in the nanocomposites. The increased complex viscosity values at high frequencies are due to the significant [M.sub.w] increases of the PET after SSP. It is worth noting that for the larger nanoclay contents (3.5 and 6 wt%) the complex viscosity at low frequencies does not change much with SSP time (see Fig. 5b and c), but the complex viscosity at high frequencies increases with SSP time. The corresponding storage modulus data are reported in Fig. 5 d and e, which shows that G' increases in the whole frequency range with increasing SSP time and decreasing particle size. At low frequencies, the nonterminal behavior (tendency toward a plateau) observed for the storage modulus data is another indication of solid-like behavior due to the nanoclay-nanoclay and/or polymer-nanoclay interactions. The increases at high frequencies are more important than those at low frequencies, due to the sensitivity of high-frequency data to [M.sub.w] of the matrix.
Figure 6 compares the linear viscoelastic data of the samples before and after SSP as plots of the complex viscosity (overall flow resistance), [[eta].sup.*], against the overall deformation resistance [G.sup.*]. For the nanocomposites, the behavior becomes more shear-thinning as the nanoclay content is increased, in the case of the nanocomposites containing 3.5 and 6 wt% C30B a significant yield stress is observed; however, at lower extent than for the samples prior to SSP. This is due to the stronger interactions between C30B and hydrolyzed PET in the melt state compared to the SSP samples, as discussed below.
The following modified Herschel-Bulkley model (Eq. 7) was used to determine the apparent yield stress of the PET nanocomposites containing 3.5 and 6 wt% C30B before and after SSP.
[[eta].sup.*] = [G.sup.*.sub.0]/[omega] + k[([[gamma].sup.0][omega]).sup.n-1] (7)
where [G.sup.*.sub.0] is the magnitude of the complex viscosity at the lowest frequency, [[gamma].sup.0] is the strain amplitude, k is a constant, and n is the flow index. Then, the apparent yield stress is [[sigma].sub.0]=[G.sup.*.sub.0][[gamma].sup.0]. [[sigma].sub.0] and n are related to the microstructure of the nanocomposites and their values, determined from the best fits of the data of Fig. 6b and c are presented in Table 3. With increasing [M.sub.w] of the PET during SSP, n increases, whereas [[sigma].sub.0] decreases. This suggests that the hydrodynamic interactions become more dominant compared to the particle-particle interactions.
In all our work [4, 37] so far on PET/clay nanocomposites, there was an evidence of matrix degradation in the presence of organoclays during processing. The PET degradation in nanocomposites was quantified by Ghanbari et al. , We follow the same approach here to calculate the apparent molecular weight of the PET in the nanocomposites after SSP, using the empirical model of Maron-Pierce, written for the complex viscosity as:
[[eta].sup.*]/[[eta].sup.*.sub.m] = [(1 - [phi]/[[phi].sub.m]).sup.-2] (8)
where [[eta].sup.*] and [[eta].sup.*.sub.m] are the complex viscosities of the nanocomposite and the matrix, respectively. <p is the volume fraction of the clay and [[phi].sub.m] is the maximum packing volume fraction. [[phi].sub.m] is estimated based on the following formula [43, 44]:
[[phi].sub.m] = 3.55/p (9)
where p is the clay aspect ratio determined from TEM analysis, taken as 40, as reported in our previous work . As at high angular frequencies, the hydrodynamic interactions are dominant, the complex viscosity at 100 rad/s is used in Eq. 8 to determine the apparent molecular weight of the PET matrix for each nanocomposite. The apparent [M.sub.w] of the PET nanocomposites was estimated using Eq. 6, whereas [M.sub.w] of the neat PETs were determined from the intrinsic viscosity measurements using Eq. 2 and reported in Table 2. The variations of [M.sub.w] with SSP time for the neat PET and the different nanocomposites are presented in Fig. 7. We observe strong nonlinear increases of [M.sub.w] with SSP time, but markedly less for the nanocomposites. The reduction of the reaction rate for the neat PET with time, as observed by the decreasing slope in the figures, is attributed to the increasing crystallinity of PET as SSP proceeds at 215[degrees]C. The [M.sub.w] changes with SSP time are reported in Table 4. For the neat PET with particle size of 400 pm, the increase of [M.sub.w] during the first 4 h of SSP is 6250 g/mol, whereas it is only 1750 g/mol for the nanocomposite containing 6 wt% C30B. Conversely, for the neat PET, the Mw increase during the first 8 h of 4500 g/mol jumps to 6500 g/mol by reducing the particle size from 400 to 200 [micro]m. However, for the nanocomposite containing 6 wt% C30B, the corresponding rise is only from 1750 to 1800 g/mol. Hence, the lower increases of the apparent [M.sub.w] of PET in the case of the nanocomposites suggest that the organoclay platelets restrict the mobility of the reactive groups as well as increase the diffusion path length of the by-products, thus reducing the rate of the SSP reactions. Another contributing factor could be related to the changes of the crystallinity in the presence of C30B .
Carboxyl Group Concentration During SSP
The reduction of the carboxyl groups of PET chains during SSP is an indicator of the rise of [M.sub.w] of PET. Conversely, an increase of these groups before SSP is an indication of more polymer degradation. Figure 8 shows the variations of the CEG concentration with the reaction time for the different samples. The CEG concentration decreases sharply during the first stage of SSP. For example, the carboxyl content of PET before SSP is 75 [micro]mol/g, whereas after 4 and 12 h of SSP it is reduced to 44 and 9 [micro]mol/g, respectively. These results indicate that SSP of PET is a useful method to obtain PET products with a considerably low concentration of CEGs. This decrease is in accordance with the corresponding increase of the intrinsic viscosity and molecular
weight (Table 2). It is also found that before SSP the concentrations of the carboxyl groups of the nanocomposites are larger than for the neat PET, for example: 90 [micro]mol/g for the nanocomposite containing 2 wt% C30B compared to 75 [micro]mol/g for the neat PET. The carboxyl content increases when the concentration of C30B is raised to 6 wt%. The higher carboxyl group concentrations in the nanocomposites before SSP are attributed to the accelerated degradation of PET in the presence of C30B during melt-mixing, which is intensified with increasing C30B concentration. Conversely, it is observed that the rate of the carboxyl group reduction (slope in Fig. 8) for the nanocomposites during SSP is less than for the neat PET, indicating a restricted mobility of the reactive groups and a lower diffusion rate of the by-products as it is well-known that polymers that contain nanoparticles exhibit a lower permeability to gases. Figure 8 also shows that the carboxyl group content for the 200-pm particle sample is less than for the 400-pm particle sample. It is another indication of the significant effect of reducing the particle size on the rate of SSP. Overall, the reduction of carboxyl group content is in agreement with the rise of Mw reported in Fig. 7.
The reduced viscosity ([eta]*/[[eta].sup.m0]) and storage modulus (G'/([[eta].sub.m0] X 100rad/s))of PET nanocomposites as functions of to x [[eta].sub.m0]) are reported in Fig. 9. The angular frequency of 100 rad/s was arbitrary chosen, and [[eta].sub.m0] is the zero-shear viscosity of the matrix (PET) calculated from Eq. 5, using the determined apparent [M.sub.w] values reported in Fig. 7. Figure 9a-c show that reasonable master curves for the complex viscosity of the PET nanocomposites for the three different C30B contents. The deviations observed could be attributed to the mild influence of the carboxyl groups on the complex viscosity. However, the effect of the carboxyl group contents on the reduced storage modulus as illustrated in Fig. 9d-f. We observe that G'/([[eta].sub.m0] x 100rad/s) increases markedly with increasing carboxyl group content (values indicated in the figure).
In our previous paper , we reported based on XRD and TEM analysis that SSP did not have any effect on the d-spacing of the clay galleries and on the organoclay dispersion. Conversely, the rheological results show a more solid-like behavior and larger apparent yield stress values (Fig. 6b and c) for the nanocomposites before SSP (see also Dini et al. [4, 37]) compared to the samples after SSP. Therefore, based on the results of Fig. 9 d-f, it can be concluded that the presence of more carboxyl groups (before SSP) causes more polymer chain interactions with the hydroxyl groups of C30B leading to a stronger structure.
SSP Effects on Crystallization of PET and Nanocomposites
As PET is a semicrystalline polymer, its microscopic properties are strongly dependent on its degree of crystallinity and crystal structure. Conversely, crystallization studies are very useful to understand the thermal properties of PET with different molecular weights and PET nanocomposites. Therefore, DSC studies will provide important information about macroscopic characteristics of the samples. DSC experiments were carried out on the samples after SSP, and the results of the second ramp (cooling) are presented. The cold crystallization temperature ([T.sub.c].) and degree of crystallinity ([X.sub.c]) of PET and PET nanocomposites are reported as functions of [M.sub.w] in Fig. 10. The crystallinity of the neat PET and PET nanocomposites was calculated using the following formula:
[X.sub.c](%) = [DELTA][H.sub.c]/[DELTA][H.sub.0] (1 - w) X 100 (10)
where w is the weight fraction of clay, [DELTA][H.sub.c] represents the enthalpy of crystallization, and [DELTA][H.sub.0] refers to the heat of fusion of 100% crystalline PET, which is 140 J/g ,
Both [T.sub.c] and [X.sub.c] of PET (Fig. 10) decrease with increasing [M.sub.w]. For the neat PET, as Mw increases from 32,000 to 83,000 g/mol, [X.sub.c] decreases from 33 to 22.5%, and [T.sub.c] decreases from 187 to 156[degrees]C. High [M.sub.w] PETs start to crystallize at a lower temperature and the reduction of [X.sub.c] with Mw is an indication of thinner and/or less perfect crystals , We note [X.sub.c] is a unique function of Mw of the PET, and Tc is slightly higher for the nanocomposites. As Fig. 10a shows at a same [M.sub.w], for example, 67,000 g/mol, [T.sub.c] of the 2 wt% nanocomposite is 174[degrees]C but for the PET without organoclay it is 166[degrees]C. Increases of [T.sub.c] in the presence of C30B can be attributed to the reduction of the chain mobility and decrease of the chain entropy to nucleate crystals. Various researchers [6, 47] have reported increases of Tc in nanocomposites after melt-mixing compared to neat PETs, which they explained by the nucleating role of the nanoclay, but they did not account for the PET degradation in the presence of organoclays. Therefore, our results show that [M.sub.w] is a much more important parameter than the presence of C30B as a nucleating agent.
At the same [M.sub.w] of PET, the presence of C30B restricts the motion of polymer chains and results in the imperfect crystallites with different shape and size compared to the neat PET; therefore, the changes of the final crystallinity compared to the neat PET is negligible. Marginal effects of nanoclays on the crystallinity of PET have been also reported [6, 48]. Figure 11 presents the DSC curves of the neat PET and PET nanocomposites containing 2 wt% of C30B for different PET molecular weights. The differences observed in the DSC exothermic peaks are mostly attributed to the differences in the Mw of the various samples.
Figure 12 reports the variations of the degree of crystallinity as functions of time for the cooling crystallization cycle of different samples. The half-time for crystallization ([t.sub.1/2]), defined as the time required to reach half of the final crystallinity, can be easily obtained from the figure. For the data prior to SSP (Fig. 12a), [t.sub.1/2], is shown to decrease from 101 to 60 s when 6 wt% clay is added to the PET. Therefore, the presence of organoclay increases the rate of the crystallization. However, as shown in Fig. 10b, the degree of crystallinity raises only due to the decreased [M.sub.w], hence, degradation of the PET in the presence of the organoclay. Moreover, Ghasemi et al.  reported that the rate of crystallization in the presence of organoclay increases, whereas the required work for chain folding and activation energy of crystal growth increases. However, their measured enhanced crystallization rate could have been due to the PET degradation (lower [M.sub.w]) in the presence of organoclays as well as nucleating role of organoclays.
Figure 12b and c depict the effect of SSP and, consequently, increasing Mw on [t.sub.1/2], for the neat PET and PET nanocomposites containing 6 wt% C30B. For the neat PET, as Mw is increased from 32,000 to 84,000 g/mol, [t.sub.1/2], rises from 101 to 168 s, whereas in PET nanocomposites containing 6wt% C30B, [t.sub.1/2]. increases from 60 s before SSP to 89 s after 12 h of SSP. The increase in polymer chain mobility, either by [M.sub.w] reduction during compounding or by a plasticization effect coming from low-molecular weight degradation products may considerably speed up the crystal growth rate. Therefore, with decreasing [M.sub.w], [t.sub.1/2] decreases, whereas [T.sub.c] and [X.sub.c] increase.
PET/C30B nanocomposites of different organoclay contents were prepared using water-assisted melt-mixing. The reduction of the molecular weight of the PET matrix, caused by hydrolysis during the water-assisted extrusion, was compensated by subsequent SSP. SSP of PET was carried out at a temperature below the melting point but above the glass transition of PET using two particle sizes for different reaction times. The zero-shear viscosity was found to vary with the weight-average molecular weight to the 3.6 power for the linear PETs. The Maron-Pierce model was used in this work to determine the molecular weight of the PET in the nanocomposites. Significant increases in molecular weight ([M.sub.w]) were found for the neat PETs and PET nanocomposites using SSP, which were accompanied by substantial reductions of the CEGs. Lower Mw was observed for the nanocomposites compared to the neat PETs. This could be attributed to a restricted mobility of reactive groups and diffusion of by-products formed during SSP (i.e., water and ethylene glycol) due to the presence of nanoparticles. The linear viscoelastic data for the neat PETs and PET nanocomposites could be correlated using the zero-shear viscosity of the matrix. However, our results showed enhanced rheological properties (mostly for the storage modulus) for samples containing more carboxyl groups, suggesting that the interactions between PET chains and C30B for hydrolyzed samples in the melt state was stronger than for SSP samples due to the larger content of carboxyl groups. DSC results showed reductions in the crystallization temperature and degree of crystallinity as well as increases in half-time of crystallization with increasing Mw of PET. Conversely, the presence of C30B increased [T.sub.c] and reduced [t.sub.[1/2]] due to the nucleating role of the organoclay.
The authors would like to gratefully thank CREPEC members, Dr. T. Mousavand and Mrs. M. Hamdine for their technical help.
[1.] J.S. Lee, J. Leisen, R.P. Choudhury, R.M. Kriegel, H.W. Beckham, and W.J. Koros, Polymer, 53, 213 (2012).
[2.] H. Ghasemi. P.J. Carreau. M.R. Kamal, and S.H. Tabatabaei, Polym. Eng. See, 52, 420 (2012).
[3.] X.F. Xu, A. Ghanbari, W. Leelapornpisit. M.C. Heuzey, and P.J. Carreau, Int. Polym. Process., 26, 444 (2011).
[4.] M. Dini, T. Mousavand, P.J. Carreau, M.R. Kamal, and M.T. Ton-That, Polym. Eng. Sci., 2013, doi: 10.1002/pen.23685.
[5.] H. Ghasemi, P.J. Carreau. M.R. Kamal, and N. Chapleau, Int. Polym. Process., 26, 219 (2011).
[6.] S. Hayrapetyan, A. Kelarakis, L. Estevez, Q. Lin, K. Dana K, Y.L Chung, and E.P. Giannelis, Polymer, 53(2), 422 (2012).
[7.] S. Li, K. Auddy, P. Barber, T.J. Hansen, J. Ma, H.-C. zur Loye, and H.J. Plochn, Polym. Eng. Sci., 52, 1888 (2012).
[8.] Y. Shen, E. Harkin-Jones, P. Hornsby, T. McNally, and R. Abu-Zurayk, Compos. Sci. Techno!., 71, 758 (2011).
[9.] A. Ghanbari, M.C. Heuzey, P.J. Carreau, and M.T. Ton-That, Polymer. 54. 1361 (2013).
[10.] H. Ghasemi. P.J. Carreau, M.R. Kamal, and J. Uribe-Calderon, Polym. Eng. Sci., 51, 1178 (2011).
[11.] K. Stoeffler, P.G. Lafleur, and J. Denault, Polym. Degracl. Stab., 93, 1332 (2008).
[12.] T.Y. Tsai, C.H. Li, C.H. Chang, W.H. Cheng, C.L. Hwang, and RJ. Wu, Aclv. Mater., 17, 1769 (2005).
[13.] W.J. Choi, H.L. Kim, K.H. Yoon, O.H. Kwon, and C.I. Hwang, J. Appl. Polym. Sci., 100, 4875 (2006).
[14.] X. Xu, Y. Ding, Z. Qian, F. Wang, B. Wen, H. Zhou, S. Zhang, and M. Yang, Polym. Degracl. Stab., 94, 113 (2009).
[15.] D.W. Litchfield, D.G. Baird, P.B. Rim, and C. Chen, Polym. Eng. Sci., 50, 2205 (2010).
[16.] D.S. Achilias, D.N. Bikiaris, V. Karavelidis, and G.P. Karayannidis, Ettr. Polym. J., 44, 3096 (2008).
[17.] D. Bikiaris, V. Karavelidis, and G. Karayannidis, Macromol. Rapid Commun., 27, 1199 (2006).
[18.] S.A. Jabarin and E.A. Lofgren, J. Appl. Polym. Sci., 32, 5315 (1986).
[19.] B. Duh, Polymer, 43, 3147 (2002).
[20.] B. Gantillon, R. Spitz, J.L. Lepage, and T.F. McKenna, Macromol. Mater. Eng., 289, 88 (2004).
[21.] B. Gantillon, R. Spitz, J.L. Lepage, and T.F. McKenna, Macromol. Mater. Eng., 289, 119 (2004).
[22.] Y. Ma, U.S. Agarwal, DJ. Sikkema, and P.J. Lemstra, Polymer, 44, 4085 (2003).
[23.] S.N. Vouyiouka, E.K. Karakatsani, and C.D. Papaspyrides, Prog. Polym. Sci., 30, 10 (2005).
[24.] T.Y. Kim, E.A. Lofgren, and S.A. Jabarin, J. Appl. Polym. Sci., 89, 197 (2003).
[25.] C. Shuya, S. Ming-Fa, and C. Shu-May, J. Appl. Polym. Sci., 28, 3289 (1983).
[26.] S. Sinha Ray and M. Okamoto, Prog. Polym. Sci., 28. 1539 (2003).
[27.] P. Cassagnau, Polymer, 49, 2183 (2008).
[28.] E.P Giannelis, R. Krishnamoorti, and E. Manias. "Polymer-Silicate Nanocomposites: Model Systems for Confined Polymers and Polymer Brushes" in Polymers in Confined Environments, Vol. 138, S. Granick, K. Binder, P.G. Gennes, E.P. Giannelis, G.S. Grest, H. Hervet, R. Krishnamoorti, L. Leger, E. Manias, E. Raphael, and S.Q. Wang, Eds., Springer, Berlin, 107-147 (1999).
[29.] M.Y. Gelfer, C. Burger, B. Chu, B.S. Hsiao, A.D. Drozdov, M. Si, M. Rafailovich, B.B. Sauer, and J.W. Gilman, Macromolecules, 38, 3765 (2005).
[30.] S. Abbasi, P. Carreau, A. Derdouri, M. Moan, Rheol. Acta, 48, 943 (2009).
[31.] G. Hu, C. Zhao, S. Zhang, M. Yang, and Z. Wang, Polymer, 47, 480 (2006).
[32.] H. Zhao, Z. Cui, X. Wang, L.S Turng, and X. Peng, Compos. B Eng., 51, 79 (2013).
[33.] A. Ghanbari, M. C. Heuzey, P.J. Carreau, and M. T. Ton-That, Rheol. Acta, 52, 59 (2013).
[34.] N. Najafi, M.C. Heuzey, P.J. Carreau, and P.M. Wood-Adams, Polym. Degrad. Stab., 97, 554 (2012).
[35.] D.H.S. Souza, C.T. Andrade, and M.L. Dias, Mater. Sci. Eng. C, 33, 1795 (2013).
[36.] R. Krishnamoorti and T. Chatterjee, Rheology and Processing of Polymer Nanocomposites. Applied Polymer Rheology, John Wiley & Sons, Inc., Hoboken, N.J., 153-177 (2011).
[37.] M. Dini. T. Mousavand, P.J. Carreau, M.R. Kamal, and M.T. Ton-That, Polym. Eng. Sci., 2013, doi 10.1002/pen.23736.
[38.] H.R. Dennis, D.L. Hunter, D. Chang, S. Kim, J.L. White, J.W. Cho, and D.R. Paul, Polymer, 42, 9513 (2001).
[39.] B. Fox, G. Moad, G. van Diepen, I. Willing, and W.D. Cook, Polymer, 38, 3035 (1997).
[40.] G.P. Karayannidis, D.E. Kokkalas, and D.N. Bikiaris, J. Appl. Polym. Sci., 50, 2135 (1993).
[41.] J. Vlachopoulos and N. Polychronopoulos, Basic Concepts in Polymer Melt Rheology and Their Importance in Processing. Applied Polymer Rheology, John Wiley & Sons, Inc., Hoboken, N.J., 1-27 (2011).
[42.] J. Dealy and J. Wang, Linear Viscoelasticity. Melt Rheology and its Applications in the Plastics Industry, Springer, Netherlands, 49-89 (2013).
[43.] L. Sun, W.J. Boo, J. Liu, A. Clearfield, HJ. Sue, N.E. Verghese, H.Q. Pham, and J. Bicerano, Macromol. Mater. Eng., 294, 103 (2009).
[44.] T. Wan, B. Wang, S. Liao, and M. Clifford, J. Appl. Polym. Sci., 125, E27 (2012).
[45.] S.G. Kim, E.A. Lofgren, and S.A. Jabarin, J. Appl. Polym. Sci., 127, 2201 (2013).
[46.] F.J. Medellin-Rodriguez, R. Lopez-Guillen, and M.A. Waldo-Mendoza, J. Appl. Polym. Sci., 75, 78 (2000).
[47.] H. Ghasemi, P.J. Carreau, and M.R. Kamal, Polym. Eng. Sci., 52, 372 (2012).
[48.] T. Wan, L. Chen, Y.C. Chua, and X. Lu, J. Appl. Polym. Sci., 94, 1381 (2004).
Maryam Dini, (1) Pierre J. Carreau, (1) Musa R. Kamal, (2) Minh-Tan Ton-That, (3) Babak Esmaeili (1)
(1) CREPEC, Department of Chemical Engineering, Ecole Polytechnique, Montreal, Quebec, Canada H3T 1J4
(2) CPEPEC, Department of Chemical Engineering, McGill University, Montreal, Quebec, Canada H3A 2B2
(3) Automotive and Surface Transport Portfolio, National Research Council of Canada, Boucherville, Quebec, Canada J4B 6Y4
Correspondence to: Pierre J. Carreau; e-mail: firstname.lastname@example.org Contract grant sponsor: The Natural Sciences and Engineering Research Council of Canada (NSERC), National Research Council of Canada (NRCC), and Canada Development Bank (CDB).
Published online in Wiley Online Library (wileyonlinelibrary.com).
TABLE 1. Sample description. Nominal Duration of Particle size Sample code C30B (wt%) SSP (h) ([micro]m) S-0 0 0 -- S-4,400 0 4 <400 S-8,400 0 8 <400 S-8,200 0 8 <200 S-C2-4,400 2 4 <400 S-C3.5-4,400 3.5 4 <400 S-C6-4,400 6 4 <400 S-C6-8,200 6 8 <200 In the codes, 4, 8, and 12 represent duration of SSP and 0 indicates before SSP. The size of the particles is shown as 400 or 200. The letter "S" indicates SSP, whereas "C" refers to a nanocomposite incorporating Cloisite 30B. TABLE 2. Intrinsic viscosity ([eta]), weight average molecular weight ([M.sub.w]), and PD of the neat PETs. PD ([M.sub.w] Sample [eta] (dL/g) [M.sub.w] (g/mol) /[M.sub.n]) As received 0.87 [+ or -] 0.02 64,500 [+ or -] 2200 2.42 PET S-0 0.54 [+ or -] 0.01 31,800 [+ or -] 900 2.18 S-4,400 0.8 [+ or -] 0.02 56,700 [+ or -] 2100 2.38 S-8,400 0.9 [+ or -] 0.04 67,500 [+ or -] 4400 2.44 S-12,400 1.03 [+ or -] 0.01 82,300 [+ or -] 1200 2.52 S-8,200 1.04 [+ or -] 0.04 83,500 [+ or -] 4700 2.52 TABLE 3. Herschel-Bulkley parameters of PET nanocomposites containing 3.5 and 6 wt% C30B. [[sigma].sub.0] [G.sup.*.sub.0] k (Pa x Sample [[gamma].sup.0] [s.sup.n]) n S-C6-0 39 78.0 0.68 S-C6-4,400 21.5 150 0.75 S-C6-8,400 16 219 0.75 S-C6-12,400 4.5 355 0.77 S-C6-8,200 11 275 0.77 S-C3.5-0 7.5 72 0.76 S-C3.5-4,400 2 303 0.9 TABLE 4. Changes of [M.sub.w] with time during SSP. In first 4 h Front 4 to In first Sample of SSP (g/mol) 12 h (g/mol) 8 h (g/mol) S,400 6250 3250 4500 S-C2,400 4750 2500 3250 S-C3.5,400 3800 1750 2750 S-C6,400 1750 1500 1750 S,200 -- -- 6500 S-C2,200 -- -- 4875 S-C3.5,200 -- -- 3250 S-C6,200 -- -- 1800
|Printer friendly Cite/link Email Feedback|
|Author:||Dini, Maryam; Carreau, Pierre J.; Kamal, Musa R.; Ton-That, Minh-Tan; Esmaeili, Babak|
|Publication:||Polymer Engineering and Science|
|Date:||Dec 1, 2014|
|Previous Article:||Flame retardant ethylene-vinyl acetate composites based on layered double hydroxides with zinc hydroxystannate.|
|Next Article:||Surface modification of a calcium fluoride filler and the effect on the nonisothermal crystallization behavior of poly(ethylene terephthalate).|