In-situ polymerization of polyamide 66 nanocomposites utilizing interfacial polycondensation, Part 1: organoclay nanocomposites.
Nylon 66 (PA66) is conventionally manufactured by polymerization invented by Carothers from a salt solution that can be optionally followed by solid-state polymerization. In that process, aqueous solutions comprising about 40-60% concentration of hexamethylene adipamide salt are heated over long periods at temperatures above 200[degrees] C to evaporate the water of solution and later the water of reaction to drive the composition to a low-to-moderate molecular weight (1). Solid-state polymerization continues to build the molecular weight by exposing the preformed polymer to heated dry inert gas for several hours (2). An alternative to this salt solution polymerization is the interfacial polycondensation (IPC) process (3), also attributed to Carothers, for polyamides and polyesters, which features operability at room temperature, the applicability of simple lab-bench scale equipment and the avoidance of thermal stability issues.
IPC involving two different phases, aqueous and organic, is a diffusion-controlled process. According to Morgan et al. (3) the reaction is assumed to progress near the interface in the organic phase. Although IPC reactions are generally considered to be reversible, addition of a scavenger for the by-product HCI allows the acid dichloride and diamine to react in an irreversible fashion according to the Schotten-Baumann reaction mechanism (4). The important reaction variables of concentration level, concentration ratio of diamine to acid chloride, rate of removal of polymer film (in the case of unstirred reactors) or stirring speed, and stirrer geometry (in the case of stirred reactors) (3) determine the diamine and diacid chloride transfer rates. It is additionally important to prevent the hydrolysis of the acid chloride to the less reactive diacid through contact with the aqueous phase. These parameters, along with the reactant and solvent purities and the solubility of the forming polymer affecting its precipitation from solution, affect the characteristics of the end product. The large number of operational variables lends itself to a statistical design of experiments (DOE) analysis in order to identify main effects and interactions leading to optimal conditions for obtaining a high molecular weight nanocomposite with desired properties. Furthermore, the statistical analysis enabled by the DOE approach leads to more reliable conclusions by accounting for and minimizing the effects of experimental error. Modified factorial designs (5) are used for the analysis of the experimental data in this work.
Layered silicate nanocomposites (NC) have been extensively studied because of their enhanced thermal, mechanical, barrier, and fire retardant properties over pure polymers (6), (7). The level of dispersion of the silicate layers ranges from swelling to exfoliation and finally to homogeneous distribution within the polymer matrix. Interfacial bonding between the silicate and the matrix is also important for the development of end properties.
Each lamella of the commonly employed sodium montmorillonite (NaMMT) silicate comprises two silica tetrahedral layers sandwiching an octahedral alumina. In nature, cations such as Fe, Mg, and Al present in the octahedral lattice are substituted by lower-charged cations in an isomorphic replacement, resulting in the generation of a negative charge on the clay layers. This negative charge is compensated by the adsorption of cations such as sodium, calcium or potassium on the gallery walls. These inorganic cations can then be exchanged with ammonium salts having alkyl or functionalized chains to enhance the compatibility between the polymer and the clay (8), resulting in increased basal spacing to facilitate dispersion. The teachings of Vaia and Giannelis (9) derived from lattice-based mean field theory that enthalpy of mixing is the primary driving force for polymer intercalation of clay galleries guide the selection of clay treatments to tailor the interactions between the nylon chains and the gallery walls of the layered silicates. Such ammonium-modified organoclays are prone to thermal degradation at high processing temperatures (10): other surfactant types are under development to alleviate this drawback (11). Unfortunately, at present they are not cost effective.
We have previously reported the feasibility of synthesizing layered silicate NC in PA66 via IPC (12), (13). This approach has been recently confirmed by Tarameshlou et al. (14). We now present a broader study in Part I that investigates various processing and compositional parameters and presents evidence of mechanical reinforcement. Statistically designed experiments are directed at uncovering the effects of reaction variables such as reagent concentration ratio, suspended clay concentration and reactor stirring on reaction yield, molecular weight, and nanoscale structure and properties. In addition, this work evaluates the dispersion of MMT organoclays treated with both nonpolar and polar surfactants in PA66. The Part I of this study shows the preparation of organoclay NC by dispersing organoclay in the organic phase of the interfacial polymerization. The Part II of this study covers the preparation of untreated layer silicate NC by dispersing the clay in the aqueous phase (15). Therefore, Part II is a continuation of this study and deals with the effect of different mixing schemes on the polymerization and dispersion of untreated layered-silicates (15).
Materials and Equipment
Layered silicate polymer NC were prepared from the following components: 60 wt% aqueous solution of 1,6 diaminohexane (HMDA) from Acros Organics, adipoyl chloride (AdCl) from Sigma-Aldrich (purum [greater than or equal to] 99%), toluene from Fisher Scientific, distilled water, Cloisite[R] 20A MMT treated with dimethyl dihydrogenated tallow, and Cloisite[R] 30B MMT treated with methyl, tallow, bis-2-hydroxyethyl quaternary ammonium salts from Southern Clay Products, [Na.sub.2][CO.sub.3] (anhydrous, pure) from Acros Organics.
Two types of reactors were employed: a stirred reactor comprising a 1 L Waring laboratory blender operating at 23,000 rpm and a 1000-ml Pyrex[R] beaker serving as an unstirred reactor for the purpose of comparison with the stirred reactor. An Omni GLH Homogenizer (5000 rpm) and a Misonix Sonicator[R] 3000 Ultrasonic Liquid Processor were used to prepare clay suspensions. A DV-II+ Pro programmable viscometer manufactured by Brookfield Engineering Laboratories with a small sample adapter was used to measure the viscosity of polyamide solutions dissolved in formic acid, 88%, at different concentrations. Sample ashings were conducted in a BI Barnstead Thermolyne 1300 Furnace. The dynamic mechanical analyzer (DMA) instrument manufactured by Perkin-Elmer and the X-ray diffractometer manufactured by Rigaku were used. A JEOL 1200 EX II electron microscope was used for transmission electron microscopy along with a Leica Reichert Model Ultracut S/FC S Cryo-ultramicrotome.
PA66 is produced from two reactants. HMDA and AdCl, dissolved in immiscible water and toluene phases, respectively. Concentration (C) of AdCl in the organic solvent phase comprising one half of the 550 ml total reaction volume was fixed at 0.15 M, the other half being distilled water. The concentration of HMDA in the aqueous phase was varied in order to change the concentration ratio while keeping the AdCl concentration constant. Conditions for material preparation were selected according to a half-factorial design including center points. The nomenclature NC20A_X represents NC prepared form Cloisite[R] 20A MMT containing X weight percent silicate mineral. The three factors incorporated in the design (clay suspension concentration, monomer concentration ratio, and reactor type) and the measured responses (reaction yield, intrinsic viscosity (IV), and the concentration of the clay incorporated in the nanocomposite) are shown in Table 1. Yield is calculated by the ratio of final polymer mass to the theoretical amount that could be produced stoichiometrically by complete conversion of the monomers.
TABLE 1. Type 20A nanocomposites prepared according to the experimental design. Designation [C.sub.HMDA]/[C.sub.AdCl] [C.sub.clay Reactor Yield in solvent] type g/dl NC20A_0 1.00 0.00 Stirred 57 NC20A_4 3.00 0.12 Stirred 92 NC20A_0 3.00 0.00 Unstirred 90 NC20A_2.3 2.00 0.06 Unstirred 82 NC20A_3.7 1.00 0.12 Unstirred 65 NC20A_2.6 2.00 0.06 Stirred 82 Designation [[eta].sub.intrinsic] dl/g [C.sub.clay] wt% NC20A_0 0.55 0.00 NC20A_4 0.38 4.00 NC20A_0 1.30 0.00 NC20A_2.3 1.27 2.30 NC20A_3.7 0.78 3.70 NC20A_2.6 0.15 2.60
A composition of 8 ml 60 wt% aqueous HMDA solution (in the example case of a 1:1 concentration ratio), 264 ml of distilled water, and 4.64 g of sodium carbonate was prepared by stirring in the blender for 15 sec. Separately, 5.9 ml of AdCl, for nanocomposite samples, and the layered silicate were similarly mixed with 275 ml of toluene. The weight of organoclay needed for the complete incorporation at a targeted silicate mineral concentration in the nanocomposite composition was calculated according to (16):
Organoclay(wt%) = [[0.935SilicateMineral(wt%)]/[1 - LoI]] (1)
where Lol stands for the fractional Loss on Ignition, which is 0.38 for Cloisite[R] 20A MMT. The clay slurry was dispersed by using the homogenizer at a speed of 5000 rpm for 10 min followed by ultrasonication for an additional 30 min. The AdCl/clay solution was then poured from a separatory funnel into the HMDA solution over a 30-sec time period under intense stirring. This mixture was blended for about 2 min at 23,000 rpm. The precipitated polymer collected on filter paper in a Buchner funnel was washed on a frittered glass filter with 1 L of 30 vol% methanol in water and 1 L of hot water (80[degrees]C). It was vacuum dried at 100[degrees]C until the excess liquid was removed. The polymer collected was placed into a paper extraction thimble and further washed with boiling methanol overnight in a Soxhlet extraction apparatus, followed by drying at 100[degrees]C overnight to constant weight. The preparation of 30B NC employed a single [C.sub.HMDA]/[C.sub.AdCl] ratio of 1:3 under slightly varied experimental conditions in an attempt to enhance the deficient dispersion attributes of that nanoclay. Sodium carbonate (4.64 g) and HMDA (8 ml) were dissolved in 264 ml of distilled water and placed into the blender. AdCl (18 ml) was dissolved in 275 ml of either chloroform to form NC30B1 or benzene to form NC30B2. Clay suspensions were prepared by dispersing 0.08 g Cloisite[R] 30B for 30 min with the homogenizer at 7500 rpm and then for 15 min of ultrasonication. The AdCl phase was added to the HMDA solution from a separatory funnel after addition of the clay suspension. The reaction continued at 29,000 rpm for 2 min. Since these experiments were done on a very small scale the performance evaluation was limited by the amount of polymer synthesized.
Clay slurry was prepared as described above. After preparing the two different solutions in different containers, the less dense solution was poured into the denser solution very slowly down the wall of the beaker. The film forming at the interface between the two solutions was pulled out of the solution from the center of the beaker with the help of tweezers and attached to a motorized wind-up device, which operated at a speed of 4 cm/sec. After collection, the polymer was washed and dried similarly to that of the samples from stirred reactors. The characterization techniques used for the analysis are described in the next sections.
Characterization of Molecular Weight
Measurements of solution viscosity in the Brookfield viscometer at 150 rpm spindle speed were used to evaluate the polymer molecular weight both in its neat form and as the matrix of the NC. Relative viscosity, the ratio of the viscosity of a polymer solution to that of the solvent (4), was calculated from the ratio of the corresponding torque values. IV [eta] was then calculated from the relative viscosities according to:
[[eta].sub.specific] = [[eta].sub.relative] - 1 (2)
[[eta].sub.reduced] = [[eta].sub.specific]/C] (3)
[eta] = lim([[eta].sub.reduced])c[vector]0 (4)
where C represents the polymer concentration in the solution, by extrapolating [eta]reduced to zero concentration (4).The Mark--Houwink equation
[[eta]] = K[bar.[M.sup.a]] (5)
where K is a proportionality constant that is characteristic of the polymer and solvent combination, and a is a function of the shape of the polymer coil in solution, could then be applied to relate [[eta]] to the average molecular weight of the polymer, M.
It should be noted that the effect of clay presence on the solution viscosity has been previously estimated by Kalkan and Goettler (12) through application of the Einstein equation to the suspension of clay platelets and found to be negligible in comparison to the polymer-solvent interactions for the low clay loadings employed in these investigations.
Ashing was conducted for 7 h in a 900[degrees]C furnace. The residue amount was calculated according to:
%residue = ([[m.sub.f] - [m.sub.c]]/[[m.sub.t] - [m.sub.e]]) * 100 (6)
where [m.sub.f] is the weight of the crucible after ashing, [m.sub.c] is the weight of the empty crucible, and [m.sub.t] is the initial weight of the crucible filled with polymer or nanocomposite, was taken as an indication of the inorganic mineral content in the NC as the residue from neat polymer was negligible.
Differential scanning calorimetry (DSC) has been used to characterize the crystalline structure and phase transitions in the nylon polymer with and without nanoclay reinforcement under [N.sub.2] gas flow with a ramp rate of 10[degrees]C/min for both the heating and cooling cycles. Samples weighing about 5 mg contained in aluminum pans were heated up to 275[degrees]C and cooled down to 10[degrees]C. At the end of each ramp the sample was allowed to anneal isothermally for 5 min before continuing to the opposite heat flow. Heating and cooling cycles were run twice sequentially; % crystallinity (X) was calculated from the exothermic curves according to:
%Crystallinity = [[([DELTA][H.sub.m])]/[[DELTA][H.sub.m.sup.0]]] * 100 (7)
where [DELTA][H.sub.m] (J/g) represents the actual heat of melting and [DELTA][H.sub.m.sup.0] is a reference value corresponding to the heat of melting for 100% crystalline polymer. For PA66 [DELTA][H.sub.m.sup.0] is taken to be 188 J/g (1).
The PerkinElmer DMA employed in a tensile mode at a ramping rate of 2[degrees]C/min over the range from room temperature to 130[degrees]C measured viscoelastic deformation behavior of the polymers produced under a sinusoidal stress at 1 Hz in terms of a storage modulus (indicating the extent to which elastic energy can be stored), a loss modulus (showing the viscous dissipation of energy into heat) and mechanical damping, which is the ratio of loss modulus to storage modulus. Also, glass transition temperature of the polymer matrix can be determined by DMA as a peak in the damping coefficient as a function of temperature (17). Nitrogen gas was circulated in the furnace during the experiments.
The following procedure was employed for the preparation of test specimens: samples were prepared from powdered resin dried overnight under vacuum at 90[degrees]C and molded with a Haake ram type mini-injection molding machine at a cylinder temperature of 290[degrees]C, a mold temperature of 40[degrees]C, and a melt pressure of 0.55 MPa. After melting the polymer or composite powder in the cylinder for about 2-3 min, the material was injected by the plunger into the mold cavity, which was in the shape of a 60 x 12 x 2 (height x width x thickness) mm rectangular bar. Smaller rectangular specimens were then cut to 4 cm length and 0.95 cm width for analysis. Rectangular bars prepared as described above were kept under vacuum at room temperature overnight before analysis to eliminate moisture absorption by the samples.
Static tensile tests were run on similarly prepared small tensile specimens and tested on an Instron machine at room temperature at a rate of 1.0 mm/min. Elongation was derived from jaw separation for calculation of the Young's modulus. Ultimate properties could not be characterized due to premature failure outside of the gauge section of the specimens.
Morphology and Clay Dispersion
TEM was performed under an acceleration voltage of 120 kV to observe the clay morphology in injection molded samples microtomed at room temperature into 70-nm slices that were oriented normal to the molding flow direction. As polymerized samples were first embedded in Buehler Epo-Thin Low Viscosity Epoxy Resin with Buehler Epo-Thin Low Viscosity Epoxy Hardener at a 3:1 weight ratio and cured for 24 h at room temperature to facilitate the microtoming of 70 nm slices at -85[degrees]C with a Leica Reichert Model Ultracut S/FC S Cryo-ultramicrotome. Samples were collected on dry copper grids with aid of an eyelash attached to a thin wooden rod.
Wide Angle X-Ray Diffraction (WAXD)
Diffraction was performed at room temperature with nickel-filtered CuK[alpha] radiation ([lambda] = 0.154 nm) at 40 kV and 150 mA to characterize the layer spacing in any residual intercalated clay tactoids. Cross-sectional slices cut from flexural bars perpendicular to the molding flow direction were analyzed in the transmission mode. To compare the layer spacing in as-polymerized NC before injection molding, powder specimens prepared from reactor grindings were analyzed under the same conditions; neat-layered silicate organoclays in powder form were also analyzed as controls. The [d.sub.001]-spacing between the silicate layers was calculated according to Bragg's law (18):
[lambda] = 2d sin [theta] (8)
where [lambda] is the wavelength, d is the basal spacing between the clay layers, and [theta] is the diffraction angle.
For thermogravimetric analysis (TGA) a TGAQ500 manufactured by TA Instruments was run under air with a ramp rate of 10[degrees]C/min. Samples weighing from 5 to 7 mg were put in platinum pans and were heated from room temperature to 950[degrees]C. Purge gas flow rate was set to 40 ml/min. Percent weight loss and its derivative with temperature were recorded in the TGA thermograms.
Optimization of Reaction Variables
As mentioned in the experimental description, DOE analysis was employed to determine the effects of diamine:acid chloride concentration ratio and reactor type on both the nylon polymerization and the efficiency of clay incorporation in the IPC process. The results correlated by the regression equations below reveal that the primary factors determining the nylon molecular weight are the type of reactor and the concentration ratio. As seen in the equation below, there is an interaction between
the reactor type and the concentration ratio on the IV, and hence the molecular weight.
[[eta].sub.intrinsic] = 0.738 - 0.187*[[C.sub.HMDA]/[C.sub.AdCl]](reactortype) (9)
where 1 and - 1 are inserted for the stirred and unstirred reactor, respectively. These equations, plotted in Fig. 1, show lower molecular weights generated in stirred vs. unstirred reactors under the conditions of the experimental design. While there is a rise in molecular weight with increase of concentration ratio in the unstirred reactor, the reverse relationship is observed for the stirred reactor.
[FIGURE 1 OMITTED]
DOE analysis also delineates the effects of the composition and process variables on the reaction yield. As shown in Fig. 2, yield increases strongly with the diamine:acid chloride reagent concentration ratio. On the other hand, reactor stirring and clay concentration have no significant effect on the yield of the polymerization reaction. The regression equation:
Yield = 48 + 15[[C.sub.HMDA]/[C.sub.AdCl]] (10)
[FIGURE 2 OMITTED]
as plotted in Fig. 2 fits the data with 4% standard error.
Figure 3 shows that it is possible to effectively incorporate layered silicates from the suspensions into the growing polymer film. Final clay content depends only upon the clay concentration in the suspension, as seen in Fig. 3, with no significant effect of the concentration ratio or the reactor type. The relationship between clay concentration and final clay content can be summarized by the regression equation:
Final[C.sub.clay](wt%) = 49.6[C.sub.clay] - 146[([C.sub.clay])sup.2] (11)
[FIGURE 3 OMITTED]
Dispersion of Layered Silicate
The nanoscale morphology of injection-molded NC was investigated via WAXD and TEM. The NC containing type 20A organoclay prepared in unstirred reactors show an intercalation peak at 6.2[degrees] 2[theta], indicating a 14 E basal plane spacing that is actually lower than that in the neat organoclay, as shown in Fig. 4, where the curves were shifted vertically to separate the data for clarity. A higher clay concentration further increases the peak intensity. Type 20A NC prepared in stirred reactors, on the other hand, display only a small shoulder (see Fig. 4), indicating a high degree of clay tactoid swelling with disorganization of the clay structure.
[FIGURE 4 OMITTED]
The WAXD results are verified by TEM micrographs. NC synthesized in the stirred reactor (see Fig. 5) have better dispersion than the ones from the unstirred reactor (see Fig. 6). In fact, exfoliation was also prominently observed along with the intercalated structures in the NC from the stirred reactor. In contrast, NC prepared in the unstirred reactor feature a uniform distribution of residual unswollen tactoids along with the intercalated structures.
[FIGURE 5 OMITTED]
[FIGURE 6 OMITTED]
DMA measurements of the viscoelastic character of the NC prepared in the unstirred reactor are shown in Fig. 7. The storage modulus increase in the NC vs. pure polyamide ranges from 19% at 30[degrees]C to 60% at 130[degrees]C. Mechanical damping of the NC as measured by the height of the tan [delta] peak decreases vs. pure polyamide, largely due to the increased storage modulus. The tan [delta] peak also broadens with the addition of clay. Glass transition temperatures ([T.sub.g]) of these NC are about the same as for pure polyamide.
[FIGURE 7 OMITTED]
Figure 8 shows the DMA analysis of NC prepared in stirred reactors. The increase in the storage modulus vs. pure polyamide ranges from 17% at 30[degrees]C to 88% at 110[degrees]C due to the incorporation of the stiff clay layers at higher 4 wt% clay content and better clay dispersion.
[FIGURE 8 OMITTED]
The wide range of IV (molecular weight) achieved in the nylon matrix of the NC considering both reactor types together affords the opportunity to analyze for its potential effect on the nanocomposite storage modulus, along with that of the nylon crystallinity. Statistical analysis of the data in Table 2 utilizing the DOE approach reveals that viscosity average molecular weight can be as significant a factor as the addition of clay layers under its large variation from 3000 to 33,000 g/mol. On the other hand, the smaller changes in percent crystallinity do not affect the storage modulus significantly, according to the DOE analysis. The relationship between the storage modulus and the factors of IV and wt% silicate mineral can be described by the equation:
E'(GPa) = 1.03 + 0.238IV + 0.074ClayContent (12)
TABLE 2. Factors affecting storage modulus. Reactor type Sample name IV (a) % wt clay % X (b) E' GPa (c) Stirred Control 0.55 0.0 29 1.23 Unstirred Control 1.30 0.0 23 1.32 Unstirred NC20A_2.3 1.27 2.3 27 1.46 Stirred NC20A_2.6 0.15 2.6 29 1.17 Unstirred NC20A_3.7 0.78 3.7 28 1.56 Stirred NC20A_4 0.38 4.0 23 1.44 (a) Intrinsic viscosity calculated by solution viscosity measurements. (b) Crystallinity calculated from DSC. (c) Storage modulus at 30[degrees]C.
In contrast to the findings from the unstirred reactor, a slight decrease in glass transition temperature, [T.sub.g], of about 3[degrees]C results from the addition of layered silicates in the lower molecular weight NC prepared in the stirred reactor, as shown in Fig. 8.
Table 2 also shows no significant effect of clay concentration on the percent crystallinity of the nylon matrix; nor did the DSC results indicate any effect on its crystallization temperature. However, the degradation temperature characterized as the temperature of maximum rate of weight loss in a TGA did consistently increase from 435 to 454[degrees]C over the range of clay contents is shown in Table 2.
Polar-Treated Organoclay Nanocomposite
Figure 9 showing WAXD of NC prepared by suspending type 30B organoclay in chloroform (Sample NC30B1) and benzene (Sample NC30B2) before polymerization indicates that neither of these selected solvents, while found to be the most compatible, have sufficient affinity to swell clay layers and disperse them. No improvement in the basal spacing of the clay layers was evident with either sample type. A TEM micrograph of NC30B2 however shows agglomerated clay layers that are more open and diffuse compared to NC30B1, as shown in Fig. 10.
[FIGURE 9 OMITTED]
[FIGURE 10 OMITTED]
IPC represents an efficient route for the generation of NC that incorporates nanoscale fillers into a polymer film as it is being formed by polymerization at a reaction zone close to the interface between two immiscible liquid phases containing the monomeric reagents, as shown in Fig. 11. No model currently exists for a mechanism to explain the incorporation of exfoliated clay platelets suspended in one of the reactive phases into the polymer film. One possibility is that the suspended clay layers can be transported by the net diffusive flux of monomer molecules toward the reaction zone, or by convectional flow resulting from natural circulation or mixing in the reactor. Since the IPC reaction for PA66 is known to be very fast, the overall polymerization rate would be controlled by the diffusion of the reagents rather than by the reaction kinetics (3), (19). While the reaction may initially occur right at a newly formed liquid-liquid interface, as the molecular weight grows over time the reaction zone can either become enmeshed in the previously formed swollen polymer gel with its reactive end groups, which could further reduce the diffusion rates in that region, or move to either side of that region accordingly as the diffusion rate of one of the monomers may be more encumbered than that of the other. The molecular weight would then suffer, as new polymer chains would be initiated in lieu of building length to those already located in the swollen polymer gel.
[FIGURE 11 OMITTED]
Due to the relative diffusivities and solubilities of the HMDA and AdCl molecules (higher solubility of diamine in organic media as compared to acid chloride in water), the reaction zone is believed to initiate slightly into the organic phase (3). It is reasonable to envision the suspended clay layers becoming entrapped amongst these growing polymer chains. A further simplification in the overall reaction scheme is to allow the reaction to be considered infinitely fast, which implies that the two reacting monomers cannot coexist at any point in the film. The reaction zone is then reduced to a plane (as schematically shown in Fig. 11).
The molecular weight of the polymer obtained via IPC can thus be maximized by maintaining an equivalent mass transfer rate for the two reacting monomers into the polymerization zone (3), (19). Johnson et al. (19) explain that any rate imbalance resulting in an excess of either monomer type will cap the growing polymer chain ends (in the growing swollen polymer mass), thus favoring the initiation of new chains (at a location to either side of said formed polymer) and thus lowering the average molecular weight of the polymer in the product film.
The mass transfer rate for each component can be considered to be a product of its diffusion coefficient and concentration gradient, which according to the film theory model can be described as the ratio of its bulk concentration to the thickness of a hypothetical stagnant film that depends on the degree of convection in the bulk phase. The concentration of each component is taken to vanish at the reaction plane because of the very high kinetics of the polymerization reaction.
The observed occurrence of lower molecular weights in the stirred vs. the unstirred reactor can be attributed to the stirring causing an imbalance of reagents at the reaction zone. Stirring thins the diffusion film in both phases around the reaction plane in comparison to the unstirred reactor, in which the only convection reducing diffusional resistance is generated only in the organic phase by the polymer film withdrawal through that phase. Such reduced diffusional resistance faced by the diamine in the aqueous continuous phase due to the formation of a thinner film in the stirred reactor then necessitates a more dilute diamine phase compared to the unstirred reactor to obtain high molecular weight polymer. There was thus a negative interaction of the IV with the HMDA/AdCl bulk concentration ratio observed for concentration ratios above unity studied in this work (see Fig. 1). An increase in the reagent concentration ratio resulted in a lower molecular weight using the stirred reactor, but a higher molecular weight using the unstirred reactor. That is, an increase in relative hexamethylenediamine concentration favors the balance of reagent delivery to the interface in the case of the unstirred reactor (due to its higher diffusional resistance there), whilst vice versa for the case of the stirred reactor. According to the results, the presence of organoclay dispersed in the organic phase does not disrupt these mechanisms.
Yield increases in both stirred and unstirred reactors at higher concentration ratio of hexamethylenediamine to AdCl, as shown earlier in Fig. 2, since the polymerization reaction is then driven further to completion. The absence of any effect of reactor type on yield indicates that the reagents are equally consumed in the allotted reaction times in both reactors.
Stirring results in better dispersion of the layered silicates in the polymerized PA66, as seen by comparing the micrographs of NC prepared by both methods, due to the shear stresses generated that serve to separate the clay layers and prevent reaggregation, especially as viscosity increases due to molecular weight building over time.
The greater modulus enhancements observed in Fig. 8 at higher temperature where the polymer is softer derive from the greater difference between the reinforcement and matrix moduli (20). Other potential mechanisms are reorientation of the clay platelets into the stress direction and progressive platelet separation under the higher strains generated during deformation by the greater polymer ductility at elevated temperature.
Various combinations of solvent and organoclay were screened to achieve a clear clay suspension. Some of the organoclays trialed in methylene chloride and toluene were the Cloisite[R] types 30B, 20A, 15A, 25A, 10A, and 93A from Southern Clay Products Company. The nonpolar 20A type of clay treatment was selected for its clear suspension in toluene. For dispersing a polar-treated organoclay (Cloisite[R] 30B), a more polar solvent, such as benzene or chloroform that is also not miscible with water, would be expected to provide a clear suspension. Still, the clay layers could not be perfectly dispersed in either chloroform or benzene, most likely due to the still insufficient affinity between the clay and the solvent. While not achieving full clarity, these two solvents were nevertheless adopted as the best candidates for 30B nanoclay suspension. Of these, benzene proved to be superior (viz. sample NC30B2) to chloroform (sample NC30B1). Predispersion of the clay in the solvent before the onset of polymerization was always essential.
IPC affords the opportunity to generate NC at room temperature as opposed to compounding or in-situ polymerization from a salt solution, for which the application of high temperature is necessary. The detrimental effect of high temperature on platelet dispersion during melt flow can be seen by comparing the finer morphology of as-polymerized IPC NC in Fig. 12 to their injection molded counterparts previously presented in Fig. 6. Even so, a comparable level of dispersion of type 20A nanoclay in molded PA66 NC is achieved by this method as compared to melt compounding (21).
[FIGURE 12 OMITTED]
Stiffening enhancement via IPC is also equivalent to that achievable in a compounding preparation, as shown in Fig. 13. The data point for "melt compounding" represents a 5-min mix of the type 20A organoclay into a melt of nylon previously produced by the IPC process (22). Measurements of the dynamic storage modulus in this figure were made at a constant stress level. Figure 13 compares the static modulus reinforcement achieved in this work via IPC with published values achieved via melt compounding reported by Goettler et al. (21) and by Chavarria and Paul (23) that are indicated on the graph as G-L-T and C-H, respectively. It is evident that IPC results in a slightly improved nanocomposite, even after melt processing via injection molding, attributed to its advantageous clay dispersion. The use of the IPC process as a means to disperse the more stable pristine MMT-layered silicates in PA66 is the subject of a forthcoming publication, Part II.
[FIGURE 13 OMITTED]
IPC is an effective means for producing PA66 NC on a small scale without detrimental effects by suspending organically modified MMT in the organic phase of the reaction media. PA66 molcular weight is significantly impacted by the state of mixing in the polymerization reactor. Highest molecular weight is achieved in unstirred reactors under favorable conditions for satisfying the reaction stoichiometry, although the reaction rate and yield are lower under those conditions.
Increased storage modulus is expected in NC since the layered silicate reinforcement is stiffer than the polymer matrix. In line with conventional NC, the percentage increase in storage modulus is greater at temperatures above the glass transition temperature of the polymer matrix, and mechanical damping diminishes with the addition of clay layers as compared to pure polyamide due to the reinforcement effect.
Better clay dispersion can be achieved by stirring the clay suspension in the reaction vessel. Even so, polar-treated organoclay such as type 30B produces a poorly dispersed PA66 nanocomposite via in situ IPC, despite favorable interactions between such a polar surfactant on the treated layered silicate and the polar polyamide chains, due to the absence of an optimized organic dispersing solvent. Thus, this work demonstrates that clay platelets can be incorporated into nylon in a generally exfoliated state via in-situ IPC as long as they can be well dispersed in one of the solvent phases before the initiation of polymerization.
(1.) M.I. Kohan, Nylon Plastics Handbook, Vol. 17-23, Hanser Publishers, Munich, 142 (1995).
(2.) R.E. Kirk and D.F. Othmer, Kirk-Othmer Encyclopedia of Chemical Technology. Vol. 19, Wiley, New York, 491 (1996).
(3.) P.W. Morgan, Condensation Polymers: By Interfacial and Solution Methods. Vol. 7-11, 19-21, Interscience Publishers. New York, 31 (1965).
(4.) C. Carraher, Seymour|Carraher's Polymer Chemistry, An Introduction, Marcel Dekker, New York, 261 (1996).
(5.) D.C. Montgomery, Design and Analysis of Experiments, Wiley, New York, 160 (2000).
(6.) S.S. Ray and M. Okamoto, Prog. Polym. Sci., 28, 1539 (2003).
(7.) M. Alexandre and P. Dubois, Mater. Sci. Eng., R 28. 1 (2000).
(8.) E.P. Giannelis. Appl. Organomet. Chem., 12, 675 (1998).
(9.) R.A. Vaia and E.P. Giannelis, Macromolecules. 30, 7990 (1997).
(10.) W, Xie, Z. Gao, W.P. Pan, D. Hunter, A. Singh, and R. Vaia. Chem. Mater., 13, 2979 (2001).
(11.) R.D. Davis J.W. Gilman, T.E. Sutto, J.H. Callahan, P.C. Trulove, and H.C. Trulove, Clays Clay Miner., 52, 171-179, (2004).
(12.) Z.S. Kalkan and L.A. Goettler, "Society of Plastics Engineers" in Annual Technical Conference, Conference 62, 3, 2902-2906 May 18, (2004).
(13.) Z.S. Kalkan and L. A. Goettler, "Society of Plastics Engineers" in Annual Technical Conference, Conference 64, 538 (2006).
(14.) M. Tarameshlou, S.H. Jafari, H.A. Khonakdar, M. Farmahini-Farahani, and S. Ahmadian, Polym. Compos., 28, 733 (2007).
(15.) Z.S. Kalkan and L.A. Goettler, Polym. Eng. Sci., In Press.
(16.) S. Bourbigot, D.L. Vanderhart, J.W. Gilman, S. Bellayer, H. Stretz, and D.R. Paul, Polymer, 45, 7627 (2004).
(17.) L.E. Nielsen, Mechanical Properties of Polymers, Vol. 1, Marced Dekker, New York, 139 (1974).
(18.) B. H. Stuart, Polymer Analysis, Wiley, Chichester, 166 (2002).
(19.) E.D. Johnson, A study of the Kinetics and Mechanism of Interfacial Polymerization, Ph.D. Thesis, Carnegie-Mellon University, Pennsylvania (1985).
(20.) L.E. Nielsen, Mechanical Properties of Polymers and Composites, Vol. 2, Marcel Dekker, New York, 454 (1974).
(21.) L.A. Goettler, K.Y. Lee, and H. Thakkar, Polymer Reviews, 47, 291 (2007).
(22.) Z.S. Kalkan, The Generation and Thermo-Mechanical Characterization of Advanced Polyamide-6,6 Nanocomposites Using Interfacial Polycondensation, Ph.D. Thesis, The University of Akron, Ohio (2006).
(23.) F. Chavarria and D.R. Paul, Polymer, 45, 8501 (2004).
Zehar S. Kalkan, Lloyd A. Goettler
Institute of Polymer Engineering, The University of Akron, Akron, Ohio 44325-0301
Correspondence to: Zehra S. Kalkan; e-mail: email@example.com
Zehra S. Kalkan is currently at Baxter Healthcare Corp., Round Lake, IL 60073.
Lloyd A. Goettler is currently at 305 Hollythorn Dr., Copley, OH 44321.
Published online in Wiley InterScience (www.interscience.wiley.com).
[c]2009 Society of Plastics Engineers