Melt processing and rheology of an acrylonitrile copolymer with absorbed Carbon Dioxide.
Fibers based on acrylonitrile (AN) copolymers having AN content of 85 mol% or greater are typically spun via a solution spinning process to avoid the intramolecular cyclization and intermolecular crosslinking reactions that would take place in a melt spinning process before fiber formation (1), (2). Although lower AN content copolymers (around 65 mol%) are melt processable due to the disruption in chain order and consequent decrease in melting temperature caused by higher amounts of comonomer, a high AN content is desirable because the resulting textile and carbon fibers have better mechanical properties, and the higher AN content accelerates the stabilization process before conversion to carbon fiber (1). At 220-230[degrees]C, the viscosity of these higher AN content materials is suitable for melt spinning, but at this temperature the crosslinking and cyclization reactions become too rapid, and the materials become intractable (1). A method to produce melt-processable high AN content copolymers was patented by BP (3), but the resultant fibers required long stabilization times and were impractical for use as carbon fiber precursors (4).
There are several significant drawbacks in using a solution spinning process. Most of the solvents used are environmentally unfriendly and require recovery and recycling. These solvents include N,N-dimethylformamide (DMF), N,N-dimethyl acetamide (DMAC), and dimethyl sulfoxide (DMSO) as well as aqueous inorganic solvents, including sodium thiocyanate (5). In addition to the environmental concerns, an economical concern is that the prepared dope contains only 7-30% solids depending on whether the materials are dry jet or wet spun (6). Hence, there is a need for a less expensive and nonpolluting alternative to the AN fiber solution spinning process. Melt processing these materials could potentially solve these problems by increasing the throughput of solids on a per pound basis and eliminating the need for solvent use and recovery (2).
To melt process these polymers, one must slow the kinetics of the crosslinking and cyclization reactions. Once copolymers with AN molar ratios of 85% and greater reach 220[degrees]C, the crosslinking kinetics become significant to the point where the viscosity increases so rapidly that the materials cannot be extruded through the spinneret pack (7). However, it is at this temperature, that the viscosity is low enough for melt processing. The steady shear viscosity of these materials was stable for 30 min when the temperature was reduced by at least 20[degrees]C, indicating that the kinetics of the crosslinking reaction were sufficiently low, but the viscosity was too high for extrusion (2), (8). This suggests that if the processing temperature of these AN copolymers can be reduced by at least 20[degrees]C, the crosslinking and cyclization reactions can be sufficiently slowed to permit melt processing, as long as a suitable viscosity can be achieved.
Plasticization has been studied as a method to reduce the processing temperature of polyacrylonitrile copolymers for melt spinning (9-18). The plasticizing medium was typically water, and was often water in conjunction with other plasticizers to form a lower boiling azeotrope. The fact that water could hydrate the pendant nitrile groups of PAN was discovered in 1952 by Coxe (13). The hydration decouples the nitrile-nitrile associations that cause cyclizing and thermal degradation well below the melting temperature of pure PAN (320[degrees]C). With the melting point significantly lowered, hydrated PAN could be melt extruded without significant degradation, but the material would foam on leaving the extruder. Porosoff (15) addressed the foaming problem with the use of a steam pressurized solidification zone. A single phase fusion melt of up to 40% water in PAN was extruded into the steam pressurized solidification zone, which prevented rapid water evaporation. This process eliminated the need for environmentally hazardous solvents, which eliminated their recovery cost and the associated pollution. By stretching the fiber in conjunction with the extrusion step, the need for a separate stretching step and the associated energy requirements were eliminated. Differences in porosity between the sheath and core were reduced by the controlled diffusion of water out of the fiber, and the presence of water in the fiber helped to maintain a stretchable state after solidification.
Several studies were performed to determine the feasibility of implementing the Porosoff process for carbon fiber production. Determining the structure and mechanical properties of such melt spun PAN fibers was necessary. Grove and coworkers (16) reported that a process similar to that of Porosoff could be used to create carbon fibers with reasonable strength and modulus, but micron-sized holes and broken filaments were observed, and it was believed that this would prevent the implementation of such a process in place of wet or dry spinning. Similar results were published by Min and coworkers (9) for precursor fibers before stabilization and carbonization. In an effort to produce high strength carbon fiber precursors, Daumit and coworkers (10), (11) at BASF looked at the use of mixtures of water (12-28% by weight of the polymer), a C1 to C4 monohydroxy alkanol (0-13%), and either a C1 or C2 nitroalkane (3-20%) or acetonitrile (5-20%) to plasticize PAN for melt spinning. The C1 to C4 monohydroxy alkanol was found to beneficially influence the filament internal structure such that carbon fiber mechanical properties were enhanced, and when it was combined with acetonitrile it lowered the temperature at which the polymer hydrated and melted. The homogeneous melt was extruded at temperatures between 160 and 185[degrees]C into a chamber which was pressurized with steam in a range of 69-345 kPa (10-50 psig) above atmospheric pressure and heated to between 90 and 200[degrees]C. The release of the plasticizer and void formation was controlled by the pressure chamber during initial drawing of the fibers. The remaining plasticizer was removed to <1 wt% upon leaving the pressure chamber and entering an oven. Post spin drawing in a steam atmosphere at 124 kPa (18 psig) helped to collapse and remove the majority of voids resulting from plasticizer removal.
The BASF process showed that carbon fibers could be produced using a pseudo-melt spinning process. Various cross-sectional configurations, e.g., trilobal configuration, could also be produced and still yield a suitable fiber for carbonization (10). However, despite the successes of the BASF process, it was never commercialized. Solvent recovery was still required to handle the high plasticizer content. The solvents, besides water, were extremely toxic and required very careful use, especially at elevated temperatures. For example, in the case of acetonitrile, it could degrade into cyanide. Minimal environmental benefits were obtained due to the use of a high amount of plasticizers other than water, and the recovery requirements produced little financial gain over solution spinning processes. The process provided no economic benefit over solution spinning once commercial production levels were reached (9 X 105 kg/year). However, it is important to note that the BASF process used considerably less solvent than the amount required for wet (80%) or dry (70%) solution spinning.
[CO.sub.2], which has been studied as a replacement for organic solvents in a number of processes (19), was found to significantly plasticize AN/methyl acrylate (MA) copolymers by Bortner and coworkers (20), (21). Differential scanning calorimetry (DSC) and pressurized capillary rheometry were used to determine the effect [CO.sub.2] had on the glass transition temperature ([T.sub.g]) and viscosity relative to the pure copolymers. Copolymers with 65, 85, and 90 and higher mole percent AN were studied. Increasing the wt% of absorbed [CO.sub.2] yielded a greater reduction in the viscosity. For example, in the 65 mol% AN sample, viscosity reduction went from 7 to 60% over the increase in [CO.sub.2] content of 2.7-6.7 wt%, and in the 90 mol% AN sample, viscosity reduction went from 31 to 56% over the increase in [CO.sub.2] content of 2-3 wt%. Each of the 65, 85, and 90 mol% samples were found to have their viscosity reduced by a similar amount (55-60%) compared to the pure copolymer when saturated with [CO.sub.2] at 17.2 MPa, despite a decreasing amount of absorbed [CO.sub.2] obtainable with increasing AN content (6.7, 5.6, and 3.0 wt% [CO.sub.2] when saturated at 120[degrees]C and 17.2 MPa for the 65, 85, and 90 mol% copolymer, respectively). This suggests that as the AN content is increased, a smaller amount of absorbed [CO.sub.2] is required to plasticize the polymer and obtain a comparable reduction in viscosity. However, above 90 mol%, the copolymers did not absorb a significant amount of [CO.sub.2] under the conditions tested. It was explained that this represented a critical copolymer ratio, at which point the dipole-dipole interactions of the pendant nitrile groups generate strong intermolecular forces and significantly decrease the free volume of the copolymer.
The work by Bortner and coworkers (20), (21) successfully showed that [CO.sub.2] could absorb into and reduce the viscosity of AN copolymers, but it was done in a strict batch operation which utilized a static pressurized bomb to saturate small quantities (less than 500 g) of material for several hours with [CO.sub.2]. Although the solubility of [CO.sub.2] in the materials was significant, the low permeability of the AN copolymers necessitated the long absorption times. The time required to achieve the necessary uptake would preclude the use of this method in a commercial-scale melt spinning process. If, instead of batch saturation, the AN copolymers could be rapidly plasticized with [CO.sub.2] to achieve the necessary process temperature reduction, the process might be viable.
Although the work of Bortner and coworkers (20), (21) laid the groundwork for the ability to process AN copolymers with [CO.sub.2], no one has documented the ability to rapidly absorb [CO.sub.2] into an AN copolymer and measure its viscosity in an extrusion process that resembles that of a true melt processing scheme. Additionally, although the thermal stability of AN copolymers and terpolymers has been studied (22), no one has compared the thermal stability and chemorheology of melt processable AN copolymers (65 mol% AN) with that of higher AN content copolymers. Low AN content copolymers that are currently melt processable without the use of plasticizers will still undergo the crosslinking and cyclization reactions at elevated temperatures where the kinetics of these reactions becomes significant. By comparing the chemorheology of low and high AN content copolymers, the required temperature for safely melt processing high AN content copolymers can be determined. Finally, no one has documented the ability to suppress the foam growth of a polymer containing [CO.sub.2] with the use of a pressurized step-down chamber following extrusion.
The purpose of the present work is to determine if AN copolymers can be rapidly (i.e., within the residence time of the polymer in the injection zone) plasticized by [CO.sub.2] in a melt extrusion process to facilitate melt processing of high AN content copolymers. A 65 mol% AN copolymer was used as a model system to avoid significant crosslinking during plastization, [CO.sub.2] injection, and viscosity measurement that might have occurred with a higher AN content material. The effect that extrusion with [CO.sub.2] had on the 65 mol% AN copolymers was translated to the 85 mol% copolymer by comparing the viscosity reduction results with those previously obtained (21) using pressurized capillary rheometry. Finally, the potential to suppress foaming of the extrudate was evaluated using a pressurized chamber attached to the end of the extruder.
Two different AN copolymers containing 65 and 85 mol% AN were used in this study. The 65 mol% copolymer, Barex, is a commercial extrusion grade material made by BP. Barex contains 65 mol% AN, 25 mol% MA, and 10 mol% elastomer and is designed to be processed in the temperature range of 180-200[degrees]C. An 85/15 mol% AN/MA copolymer was produced via a heterogeneous free radical (emulsion) polymerization by Monomer Polymer, of Feastersville, PA, following a procedure developed by Bhanu and coworkers (7), The 85/15 material's [M.sub.n] and [M.sub.w] were previously determined (GPC) to be 26,500 and 65,900 g/mol, respectively (21). About 100 g of the 85/15 material were available in powder form for our measurements in this study.
Air Products medium grade (99.8% pure) [CO.sub.2] was used for absorption into the Barex copolymer. Air Products 99.998% pure (excluding Argon) nitrogen ([N.sub.2]) was used to pressurize the chamber designed for suppression of foaming.
Sample Preparation for Torsional Viscosity Measurements
To avoid significant crosslinking, specific steps were followed before the torsional rheological measurements. Samples of the 85 mol% copolymer powder were compression molded into circular samples of 25.0 mm diameter and 1.0 mm thickness using a pressure at room temperature, thereby imparting no thermal history before loading the sample into the rheometer. The 65 mol% copolymer pellets were compression molded using the same mold at 170[degrees]C, which causes the viscosity to increase less than 1% due to crosslinking during the molding process.
Dynamic Oscillatory Rheology
Dynamic oscillatory rheological measurements were performed using a Rheometrics RMS 800 Mechanical Spectrometer. Frequency sweeps were made over the range of 1-100 rad/s using 25.0 mm parallel plates and 5% strain. To minimize degradation during measurements of the complex viscosity, angular frequencies below 1 rad/s were not used because of the length of time required to obtain data. Time sweeps were performed over at least 1800 s using 25.0 mm parallel plates and 5% strain at an angular frequency of 0.1 rad/s. Temperatures in the range of 180-220[degrees]C were used for each test type. To ensure measurements were performed in the linear viscoelastic region, strain sweeps were performed at 200[degrees]C using an angular frequency of 10 rad/s over the range of strain of 0.6-100% (23). Samples were found to be linear up to at least 5% strain. During loading, heating, and equilibration of the samples in the rheometer, an inert nitrogen atmosphere was used in the rheometer oven to slow the cross-linking reaction.
Steady Shear Rheology
Steady shear viscosity measurements were carried out using a Rheometrics RMS 800 Mechanical Spectrometer. Measurements were made using 25.0 mm parallel plates and a sample thickness of 1.0 mm over the temperature range of 180-220[degrees]C at an angular frequency of 0.1/s. Measurements were conducted for at least 1800 s, which was estimated to be significantly longer than the residence time in an extruder (23)
Dynamic Mechanical Thermal Analysis
Dried polymer pellets for dynamic mechanical thermal analysis (DMTA) were injection molded at 180[degrees]C into 7.6 cm by 7.6 cm square plaques using an end-gated mold at 70[degrees]C. The molded samples were cut into 8-9 mm wide strips. The strips were dried, weighed, and then clamped between metal plates and saturated with [CO.sub.2] in a sealed, constant-volume, pressurized bomb. The bomb was initially charged with [CO.sub.2] at room temperature in the form of a high-pressure gas at 5.86 MPa (850 psi), then heated to 95[degrees]C, and then further pressurized with [CO.sub.2] to 17.2 MPa (2500 psi) using a Trexel model TR-1-5000L supercritical fluid system. The samples were allowed to equilibrate for 4 days. A forced convection cool down back to room temperature was followed by a controlled decompression over the course of 12 min. Samples were removed and weighed again using a fine mass balance to measure [CO.sub.2] uptake. One was tested immediately, and another was weighed and tested after 1 week to allow a portion of the gas to diffuse out for the purpose of obtaining an intermediate amount of [CO.sub.2] within the sample. A third sample was not exposed to [CO.sub.2], and was tested in the pure state. The three sample strips were weighed and then immediately loaded into a Rheometrics RMS 800 Mechanical Spectrometer for DMTA testing. The samples were tested from 40 to 100[degrees]C at 2[degrees]/min using a frequency of l rad/s at 0.2% strain.
Slit-Die Rheometer, Extruder, and Experimental Procedure
A 12.7 mm (0.5 inch) wide, 10:1 width-to-height ratio slit die designed for operation at high pressures (up to 50 MPa at entrance) was used to measure the viscosity of pure and saturated copolymer at 180[degrees]C. A schematic of the slit-die is shown in Fig. 1. This rheometer was based on the extrusion rheometers first developed by Han and Ma (24), (25), and later used by Royer and coworkers (26), (27). The rheometer had three pressure transducer taps fitted with two Dynisco model PT462E pressure transducers and one model TPT463E pressure transducer set one inch apart to measure temperature, pressure drop, and the linearity of the pressure drop. The pressure transducers were mounted flush with the die walls. A linear pressure drop is required to ensure both fully developed and single-phase flow within the measurement region (26), (27), A Keithley model KUSB-3100 data acquisition module and a PC were used to record pressure data. At the end of the die, a threaded section was integrated to allow the attachment of nozzles or additional equipment (26), (27). Attaching a nozzle or capillary is required to elevate the pressure within the measurement area of the slit die above the point at which [CO.sub.2] will come out of solution. Two different attachment configurations were used to increase pressure within the die. The first (called Nozzle 1) was an adapter with a diameter of 3.175 mm and L/D of 8, and the second (Nozzle 2) was the same adapter with a capillary die attached to it that had a diameter of 1.4 mm and L/D of 10. In the slit die, a lead-in section before the center of the first transducer with length equal to 20 die-heights was used to allow for flow rearrangement in the entrance region, and a lead out section after the third pressure transducer with length equal to twenty die-heights was used to allow for flow rearrangement due to any nozzle attachments. The slit die was heated with 12 100-W Watlow Firerod cartridge heaters.
[FIGURE 1 OMITTED]
Viscosity measurements were performed on the 65 mol% AN material containing an estimated 4-12 wt% [CO.sub.2]. Measurements were made at 180[degrees]C and compared to the pure copolymer viscosity measured at 180[degrees]C. The following equations were used to calculate the viscosity of the melt (28-32):
Shear stress [[tau].sub.w] = ([ - [DELTA]P]/L)[H/2] (1)
Apparent shear rate [[gamma].sub.a] = ([6Q]/[WH.sup.2]]) (2)
Apparent viscosity [[eta].sub.a] = [[[tau].sub.w]/[[gamma].sub.a]] (3)
True shear rate [gamma] = [[[gamma].sub.a]/3](2 + [dln[[gamma].sub.a]]/[dln[[tau].sub.w]]) (4)
where [[tau].sub.w] is the wall shear stress, [DELTA]P is the pressure drop, L is the length of the die, H is the slit height, [[gamma].sub.a] is the apparent shear rate, W is the slit width, Q is the volumetric flow rate, [[eta].sub.a] is the apparent viscosity, and [gamma] is the true wall shear rate . These equations are valid for fully developed flows of incompressible fluids. The well-known Rabinowitsch correction (Eq. 4), used to account for the non-Newtonian nature of the velocity profile in the polymer melt, was performed to correct the apparent shear rate to a true shear rate. The viscosity was then calculated from Eqs. 1 and 4. To obtain a viscosity curve over several different shear rates, the process was repeated at different screw speeds. To obtain volumetric flow rates for calculating the apparent shear rate, an equation of state was used to estimate the density of the melt mixture. For this analysis, the Sanchez-Lacombe equation of state (33-35) was used with the mixing rules described by Park and Dealy (36), No interaction parameter was used for our estimates. Incorporation of an interaction parameter caused the density values to differ by 1% or less over the range of pressures and concentrations studied when compared to the density when calculated without using the interaction parameter. The Sanchez-Lacombe parameters used for the 65% AN copolymer and [CO.sub.2] are shown in Table 1. The slit die was fed by a Killion KL-100 extruder with a 25.4 mm (1 inch) diameter 30:1 L/D two-stage screw. An increase in the channel depth in the two-stage screw at the point of injection of [CO.sub.2] facilitated a drop in the melt pressure, allowing the gas to be more easily injected. The metering section before this drop was a region of higher pressure that allowed the formation of a melt seal to help prevent [CO.sub.2] from escaping up the channel and out of the hopper. A 1.752 cc/rev gear pump (Zenith, HPB-5556) was used between the extruder and slit die to stabilize the flow of extrudate through the die. A high pressure conduit was used to attach the gear pump to the extruder, and an Omega model FMX8461S static mixer was placed in the conduit to help homogenize the melt. An injection port in the extruder at the beginning of the second stage of the screw was used to add [CO.sub.2] to the melt. The [CO.sub.2] was pressurized and metered with a Trexel model TR-1-5000L supercritical fluid unit. [CO.sub.2] flow was measured using a MicroMotion Elite CMF010P Coriolis mass flow meter and an RFT9739 transmitter. A Tescom 26-1700 Series back pressure regulator was used to control the [CO.sub.2] pressure on the downstream side of the pump.
TABLE 1. Sanchez-Lacombe parameters. Material [rho] * (g/[cm.sup.3]) T * (K) P * (MPa) Reference 65 mol% AN (a) 1.273 700.1 497.7 Carbon Dioxide 1.725 319 727.4 36 (a) Obtained by linear interpolation of parameters obtained for pure polyacrylonitrile (37) and pure poly (methyl acrylate) (38).
The extruder was used to melt, pressurize, and deliver the polymer to the slit die. First, dried polymer pellets were placed in the hopper to be fed to the rotating screw where they would melt and become pressurized. Next, for materials that were to be plasticized with [CO.sub.2], [CO.sub.2] was injected into the system at a measured rate and was mixed with the melt. The gas flow rate was controlled by regulating the back pressure to be ~3.4 MPa (500 psi) above the melt pressure within the extruder and adjusting the [CO.sub.2] pump's stroke length and speed. At least 5 min were allowed after each change in flow rate for the system to reach steady state. Next, the pressure generated by the turning screw forced the melt into the gear pump and then through the slit die. Finally, the pressure drop was recorded, and samples of the melt were taken at the exit of the die to measure the mass flow rate. The estimated wt% of [CO.sub.2] in the mixture, neglecting losses, was determined by measuring the mass flow rate of extrudate and comparing to the mass flow rate of [CO.sub.2] injected into the system. The rates at which [CO.sub.2] was injected into the polymer (referred to as injected rates or injected wt%) were not necessarily the amounts absorbed into the polymer (referred to as wt% or absorbed [CO.sub.2]) due to the loss of [CO.sub.2] out of the system.
It was assumed that the effects of pressure on the compressibility of the fluids within the slit die were insignificant compared to measurement error due to transducer sensitivity. This assumption has been made by several other researchers for other [CO.sub.2] and polymer mixtures and has been shown to be adequate [24-27, 39-42]. For up to 7 wt% [CO.sub.2] content, the density of the mixture within the slit die was calculated to change by only as much as 2% across the range of pressures in the die, which is within the range of error due to transducer sensitivity. It is important to note that for the low concentrations of [CO.sub.2] used in the study, the incompressibility assumption is valid, but higher concentrations can lead to significant errors (27).
To study the suppression of the nucleation and growth of bubbles following [CO.sub.2]-assisted extrusion of the 65% AN copolymer, a 1 m long, 6.35 cm inside diameter tubular pressure chamber from High Pressure Equipment Company was used to pressurize and collect the extrudate. The chamber cap was machined so that it could be screwed onto the adapter that was attached to the extruder. No gear pump or slit die was used in this setup, but the other process features as described above remained. The chamber cap had a 6.35 mm diameter hole bored through it and was heated with band heaters to 180[degrees]C to allow the extrudate to enter the chamber. The opposite chamber cap was machined to accept a high pressure gas valve for connection to a gas cylinder. Extrusion of copolymer and [CO.sub.2] was begun through the chamber cap, and once steady-state was obtained the chamber was closed and pressurized with nitrogen to 8.27 MPa (1200 psi). The process was run for ~20 min to collect sample. The extruder was then stopped, and the pressure chamber and contents were cooled down to room temperature via forced convection. The contents were then removed and analyzed with scanning electron microscopy. This part of the work was carried out as a first attempt at showing it is possible to suppress bubble growth in the extrudate at the exit of the die under moderate gas pressures.
Field Emission-Scanning Electron Microscopy
Samples of the 65 mol% AN copolymer extruded into atmospheric conditions and into a pressurized chamber were evaluated using a Leo 1515 FE-SEM instrument. Freeze-fracture surfaces were prepared by submersing a sample of the extrudate in liquid nitrogen for ~2 min and then bending with forceps until fracture occurred. Samples were platinum-coated before SEM examination.
RESULTS AND DISCUSSION
In this section, the thermal stability of the two AN copolymers and the effect of [CO.sub.2] on the viscosity of the 65 mol% copolymer in a continuous process are discussed. The ability to predict viscosity and processing temperature reduction in these materials for a particular amount of [CO.sub.2] absorption is also discussed. On the basis of these findings and through comparison to previously obtained measurements of viscosity reduction of the 85% AN copolymer due to [CO.sub.2] obtained using pressurized capillary rheometry (21), the ability to continuously melt process the 85% AN copolymer with [CO.sub.2] is evaluated. A discussion of the use of a pressurized chamber to form a homogenously solid extrudate of polymer with absorbed [CO.sub.2] concludes the section.
Because of the crosslinking and cyclization reactions that take place in AN copolymers in the melt state, the viscosity will increase as a function of temperature and residence time in the extruder. A chemorheological model was used to quantify the effects of time and temperature on the 65 and 85% AN copolymers. The model is similar to that used to describe the curing of epoxy resins (43) and the crosslinking of high AN content co- and ter-polymers (35), and is represented by Eqs. 5-7:
ln[eta](t) = ln[[eta].sub.r] + kt (5)
[[eta].sub.r] = [[eta].sub.[infinity]]exp([[DELTA][E.sub.[eta]]]/[RT]) (6)
In these equations, [[eta].sub.r] is the temperature dependent viscosity, k is the temperature dependent rate, [DELTA][E.sub.[eta]] and [DELTA][E.sub.k] are the activation energies associated with the temperature dependence of [[eta].sub.r] and k, respectively, T is temperature and t is time. These functions were applied to the time dependent dynamic oscillatory (complex viscosity) data (|[eta]*|) of each material at 180, 200, and 220[degrees]C in air and at an angular frequency ([omega]) of 0.1 rad/s. The tests were performed in air to better simulate the conditions the polymers experience in an extrusion process. Plots of In |[eta]*| versus time were made at each temperature, and the slopes and intercepts of these plots were used to obtain k and [[eta].sub.r], respectively, at each temperature and composition. Linear fits of the data in plots of In k and In [[eta].sub.r] versus 1/T were then used to obtain the constants in Eqs. 6 and 7. The linear fits of In |[eta]*| versus time for the 65% AN copolymer are shown in Fig. 2. A similar plot, not shown here, was used to obtain k and [[eta].sub.r] for the 85% AN copolymer. As shown in Figs. 3 and 4, linear fits of the plots of the natural logarithms of [[eta].sub.r] and k versus 1/T (obtained from Fig. 1) were used to obtain values for [[eta].sub.[infinity]] and [k.sub.[infinity]] from the intercepts and [[[DELTA][E.sub.[eta]]/R] and [[[DELTA][E.sub.k]]/R] from the slopes, respectively, for the 65 and 85% AN copolymers. Correlation coefficients of 0.95 or better were obtained for all linear fits of the data. The values of the constants obtained for Eqs. (5-7) are listed in Table 2 for the 65 and 85% AN copolymers.
[FIGURE 2 OMITTED]
[FIGURE 3 OMITTED]
[FIGURE 4 OMITTED]
In addition to dynamic oscillatory tests, time dependent steady shear viscosity tests were performed on the 85% AN copolymer to compare the two available test methods used to measure the stability. The results are shown in Fig. 5. The 65% material could not be satisfactorily tested in steady shear due to the formation of gels over time which would roll between the plates and force material out of the gap, leaving the dynamic oscillatory measurement method as the only way to compare the effects of time and temperature on the two copolymers. The dynamic and steady shear data agree reasonably well, but there is a small difference in the viscosity data between the two methods which grows with temperature. This may be due to the rapid crosslinking that occurs at higher temperatures, which can cause the materials to be at a different initial viscosity once the temperature equilibrates and the test begins. Also, the viscosity increase is slightly more rapid in the steady shear test at 220[degrees]C. This may be because the motion of the polymer in steady shear exposes more of the reactive sites to one another to form cross links as opposed to the dynamic test where comparatively little long range motion and site-to-site exposure occurs. Despite this occurrence, the difference between dynamic and steady shear data is sufficiently small to make the use of dynamic data for the chemorheological analysis reasonable.
[FIGURE 5 OMITTED]
TABLE 2. Parameters obtained from chemorheological analysis. AN content (mol%) [k.sub.[infinity]]/s [[eta].sub.[infinity]] (Pa s) 65 1.65E+4 7.44E-03 85 4.36E+8 5.58E-10 AN content (mol%) [DELTA][E.sub.k]/R(K) [DELTA][E.sub.[eta]]/R(K) 65 -3965 7569 85 -10,773 13,866
To compare the stability of the 65 and 85% AN copolymers, the chemorheological analysis was used to create plots of viscosity increase over time at 180, 200, and 220[degrees]C and viscosity as a function of temperature at times between 0 and 30 min. These plots are shown in Figs. 6 and 7. It can be seen from Fig. 6a that the viscosities of the two materials increase at almost identical rates at 220[degrees]C to a 28% increase after 30 min. At 200[degrees]C, the 65% AN copolymer viscosity actually increases faster than the 85% material to a 10% increase after 30 min, whereas the 85% material only increases by 4%. No change is discernable in the 85% AN copolymer at 180[degrees]C, and the 65% material only increases by about 4%. A comparison of the model predictions with experimental data is shown in Fig. 6b and c. There is good agreement between the experimental data and the model prediction. Some nonlinearity of the experimental data can be seen, and there is a considerable discrepancy at low times, but this is due to the overshoot at the beginning of the test.
[FIGURE 7 OMITTED]
The reason that the lower AN content material's viscosity increased faster and to a greater extent over 30 min than the higher AN content material may be the polymers' molecular weight. The 65% AN material has a higher molecular weight as seen from comparing the viscosity data of the two materials, which may cause its viscosity to increase faster due to crosslinking, because fewer crosslinks are necessary to fully crosslink the higher molecular weight material. However, Fig. 7 shows that once the temperature where the crosslinking and cyclization reactions begin to take off is reached, residence time has a much greater effect on the viscosity of the 85% copolymer than on that of the 65% material. At temperatures near and above 220[degrees]C, the viscosity of the 85% AN copolymer increases much faster over time than the lower AN content copolymer. This indicates that temperature and residence time must be carefully controlled to prevent the reactions from taking off. It should be noted that during the time dependent viscosity tests, the color of the sample discs went from translucent yellow to orange and then to dark orange-brown after 30 min at 220[degrees]C, indicating significant cylclization and crosslinking took place . Even at 200[degrees]C, the 65% sample turned orange after 30 min. The samples that increased in viscosity by 4% or less only changed in color to a slightly darker yellow after 30 min. These observations and the model predictions for viscosity increase as shown in Figs. 5 and 6 suggest that viscosity increase should be kept below 5-10% to ensure that no significant crosslinking or cyclization takes place. The recommended processing temperature range for the 65% AN copolymer is 180-200[degrees]C, and the viscosity of this material increases faster over time at these temperatures than the 85% material. These results indicate that for residence times less than 30 min, the 85% AN copolymer will be suitably stable if the processing temperature is maintained at or below 200[degrees]C.
[FIGURE 6 OMITTED]
To determine the effect of absorbed [CO.sub.2] on the glass transition temperature of the 65% AN copolymer, DMTA was performed at atmospheric pressure and from 40 to 100[degrees]C on samples with and without absorbed [CO.sub.2]. The loss modulus (G") versus temperature data is plotted in Fig. 8 for the 65% AN copolymer with 0, 5.5, and 8.8 wt% [CO.sub.2]. The [T.sub.g] peak is seen for the 0 and 5.5 wt% samples at 85 and 54[degrees]C, respectively. The 8.8 wt% sample appears to be near the reduced [T.sub.g] when the test begins at 40[degrees]C. A shoulder in the data curve can be seen between 50 and 55[degrees]C for the 8.8 wt% sample, which is similar to the shoulder seen between 60 and 70[degrees]C for the 5.5 wt% sample. The similar shape of these two curves suggests that the [T.sub.g] of the 8.8 wt% sample is near 41[degrees]C.
[FIGURE 8 OMITTED]
The [T.sub.g] reduction versus wt% [CO.sub.2] data are shown in Fig. 9 along with data for the same system obtained previously (20) using DSC and thermogravimetric analysis. There is good agreement between the two methods. The line is a linear fit (slope of 4.93), which accurately describes the [T.sub.g] reduction due to absorbed [CO.sub.2] ([R.sup.2] > 0.95). This relation allows one to determine the amount of [CO.sub.2] required to achieve a reduction of [T.sub.g] for low amounts of absorbed [CO.sub.2], which can be related to viscosity reduction.
[FIGURE 9 OMITTED]
Viscosity Reduction and Shifting
To help determine the feasibility of continuously melt processing AN copolymers by rapidly plasticizing them with [CO.sub.2], the viscosity versus shear rate of the model 65% AN copolymer with injected [CO.sub.2] was measured online in an extrusion process using a slit-die rheometer. In Fig. 10, the apparent viscosity versus shear rate data is shown at 180[degrees]C with [CO.sub.2] injection rates of 0, 4, 7, and 12% by weight, neglecting losses. The lines represent power law fits of the data. At [CO.sub.2] injection rates of 4 wt%, the average viscosity reduction compared to the pure polymer was 22%. At injection rates of 7% and 12%, the viscosity was reduced by 42 and 56%, respectively. The pressure drop in the slit die was linear for injection rates up to 7%, indicating fully developed and single phase flow, but at 12% there was a slight nonlinearity to the pressure drop at lower shear rates (lower average die pressure), which was due to [CO.sub.2] beginning to bubble out of solution when the pressure dropped below the equilibrium pressure for that composition. This represented the upper limit of pressure and concentration that can be achieved for viscosity measurements with this slit-die rheometer and not necessarily a limit of the achievable viscosity reduction for this material in an extrusion process.
[FIGURE 11 OMITTED]
The effect of pressure on the viscosity was analyzed by observing the change in viscosity over the measured range of shear rates due to increased pressure. Increases in pressure were obtained by adding nozzles to the slit die with different LID values. This effect can be seen in Figs. 11 and 12 for the 65% AN copolymer with [CO.sub.2] injection rates of 0 and 7 wt%, respectively. The data in these figures are plotted on axes with linear scales to better show the relative values of the data as opposed to using logarithmic scales. Average die pressures were calculated by taking the mean of the pressure data obtained from each of the three transducers (1000 Hz sample rate). No data above 22 1/s for Nozzle 2 was obtained because the system pressure limit was reached at this condition. In Fig. 11, the viscosity data for each nozzle configuration overlap, even though the average pressure from Nozzle 1 to Nozzle 2 is doubled at each shear rate. Similar results can be seen in Fig. 12 for the copolymer with an estimated 7 wt% [CO.sub.2]. The scatter seen at the higher end of the shear rate range for each nozzle is due to pressure fluctuations within the process, and represents the range of error in the data. Although not shown, the viscosity data of the copolymers with [CO.sub.2] injected at 4 and 12 wt% also overlap at each shear rate for all pressure ranges. These results indicate that for our viscosity measurements, there is no need to correct for pressure effects because viscosity reduction is determined relative to the pure copolymer viscosity at similar pressures. Pressures beyond the range of those encountered in this study may have an effect on the viscosity, as seen in the case of HDPE and [CO.sub.2] (36), but the effect is indistinguishable in the polymer and pressure ranges of interest in this study.
[FIGURE 12 OMITTED]
Viscosity shifting was performed by determining shift factors via least-squares fits of the viscosity data at each concentration to generate a master curve shifted to the pure copolymer viscosity. Recent studies have determined shift factors for concentration and pressure by using pure component qualities obtained from PVT data and model fitting, often in conjuction with a Williams-Landel-Ferry (WLF) or an Arrhenius analogue for [T.sub.g] depression and elevation due to [CO.sub.2] concentration and pressure (27), (36). For the polymer of interest in this study, PVT data and pure component parameters were not readily available. Instead, the DMTA results were used to correlate glass transition temperature depression and absorbed amount of [CO.sub.2], and the experimental viscosity reduction data was used to determine the shift factors.
The master curve generated by shifting the viscosity reduction data is shown in Fig. 13. The least squares method was used to determine the shift factors that would make the data at each injection rate of [CO.sub.2] best fit the pure copolymer data. The apparent shear rates were then corrected by applying the Rabinowitsch correction as shown above in Eq. 4. An Arrhenius analogue for the effect of concentration and pressure on the viscosity (27) was used to evaluate the ability to predict the shift factors obtained using the least squares method. For temperatures above the applicable range of the WLF equation ([T.sub.g] + 100[degrees]C), an Arrhenius expression (43) can be used to describe the effect of temperature on viscosity, as shown in Eq. 8:
[FIGURE 13 OMITTED]
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (8)
where [a.sub.T] is the time-temperature superposition factor, [[eta].sub.T] is the viscosity at temperature T, [E.sub.a] is the activation energy for viscous flow, R is the ideal gas constant, [T.sub.0] is an arbitrary reference temperature, and [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] is the viscosity at that temperature. Royer and coworkers (32) referenced the effect of pressure and concentration as shifts in [T.sub.g] in the Arrhenius expression, as shown in Eqs. 9 and 10:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (9)
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (10)
where [a.sub.P] is the pressure dependent shift factor, [a.sub.C] is the concentration dependent shift factor, [P.sub.0] is atmospheric pressure, P is the system pressure, and the subscript mix refers to mixture properties. Parameters for determining [a.sub.C] in Eq. 10 were obtained from experimental data. The flow activation energy, [E.sub.a], was determined by using Eq. 8 to perform a least squares fit of the dynamic oscillatory viscosity data of the pure 65% AN copolymer at 180, 200, and 220[degrees]C. The value for [E.sub.a] was found to be 93,575 J/mol. The glass transition temperature at atmospheric pressure for the pure and [CO.sub.2]-containing copolymer was determined from DMTA results and the linear relation between [T.sub.g] and [CO.sub.2] content. It is assumed that the pressure and concentration dependent shift factors can be combined to yield a shift factor for both pressure and concentration, [a.sub.[P,C]]. This combined shift factor is what was determined experimentally by shifting the viscosity data obtained from slit die measurements. The shift factor, [a.sub.C], determined from Eq. 10, was compared to the experimentally determined shift factor [a.sub.[P,C]]. This result is shown in Fig. 14. Lines in the figure are linear fits of the data. There is a large difference between the two shift factors. This may be largely due to the reference pressure used in Eq. 10, which is atmospheric pressure, whereas the pressures range of the experimentally determined shift factor, [a.sub.[P,C]], varied from about 5 to 35 MPa. The viscosity and pressure data obtained from the slit die showed that pressure had little effect on the viscosity for the ranges of pressure encountered, but the Arrhenius analysis suggests that pressure has a large effect on [T.sub.g] and viscosity reduction over the range from atmospheric pressure to test pressures. To correct the discrepancy, the [T.sub.g] reduction for the copolymer at slit-die test pressures and concentrations as opposed to atmospheric conditions can be calculated by dividing [a.sub.[P,C]] by [a.sub.C] for each mixture to obtain [a.sub.P] and then solving for the [T.sub.g] of the mixture at test pressure using Eq. 9. For the copolymer with 4% [CO.sub.2], neglecting losses, [T.sub.g] was calculated to be 5[degrees]C less than that of the pure material. For the copolymers with an estimated 7 and 12% [CO.sub.2], the [T.sub.g] reductions were calculated to be 10 and 16[degrees]C, respectively. While this analysis helped to show the difference between the [T.sub.g] reductions due to absorbed [CO.sub.2] obtained at atmospheric pressure and the [T.sub.g] and viscosity reductions obtained in the extrusion process, experimentally determined data is required to apply and correct the model. It may be possible to calculate the model parameters from more general physical property and PVT data rather than using experimentally determined viscosity and [T.sub.g] data, and a more rigorous application of the theory may provide a better model prediction.
[FIGURE 14 OMITTED]
Processing Temperature Reduction
To assess the ability of [CO.sub.2] to provide a processing temperature reduction for the AN copolymers, the viscosity of the pure and [CO.sub.2]-containing 65% AN copolymer was compared to the viscosity of the pure material shifted to higher temperatures. The data shifts were performed by using the value of [E.sub.a] (93,575 J/mol K) determined from the analysis of the dynamic oscillatory data and Eq. 10. The resulting plot is shown in Fig. 15, where lines are power law fits of the pure and [CO.sub.2]-containing viscosity data, and symbols are the pure 65% AN material viscosity data obtained from the slit-die at 180[degrees]C shifted to higher temperatures. The overlap of the concentration and temperature effects shows that for injection rates of [CO.sub.2] set to provide 4 wt% [CO.sub.2], an equivalent processing temperature reduction of at least 10[degrees]C can be obtained. For 7 and 12% [CO.sub.2], the pure data overlap at 200[degrees]C and 210[degrees]C, respectively. These results suggest that [CO.sub.2] can be used to process the material at temperatures up to 30[degrees]C lower and achieve the same viscosity as the pure material at the original temperature.
[FIGURE 15 OMITTED]
Predictions for Higher AN Content
To predict the effect of [CO.sub.2] on the 85% AN material in the extrusion process and its ability to facilitate melt processing, comparisons were made between data and results obtained in the slit die for the 65% AN material to data previously obtained using pressurized capillary rheometry for the 65% and 85% AN copolymers (20), (21). Table 3 summarizes these findings. For the 65% material, viscosity and potential process temperature reductions were in good agreement, but the [CO.sub.2] content and [T.sub.g] reduction data were off by 44 and 47%, respectively. The methods used to measure [T.sub.g] reduction due to [CO.sub.2] content showed good agreement (see Fig. 9). This indicates that the [CO.sub.2] content data is where the disparity lies and that the measured [CO.sub.2] injection rate in the extrusion process was not necessarily the actual amount of [CO.sub.2] that was absorbed into the material. During extrusion, a portion of the injected [CO.sub.2] likely escaped from the system by leaking through the melt seal and up the screw channel and then out of the hopper. This loss could be prevented by using a pressurized hopper and feed system, a different screw design, or by using a twin-screw extruder. However, for this study, the prevention of loss of [CO.sub.2] was not necessary to measure the viscosity reduction because comparisons of the uptake and viscosity reduction data can be made between the two methods to determine the amount absorbed in the extrusion process. Comparing the [CO.sub.2] uptake and viscosity reduction data between the two methods used to generate it suggests that the value of the injected amount of [CO.sub.2] should be decreased by up to 44% to give the actual amount of [CO.sub.2] absorbed into the polymer. For the estimated 12% uptake, the actual amount of gas absorbed into the polymer is ~7 wt%, corresponding to a [T.sub.g] reduction due to concentration at atmospheric pressure of ~35[degrees]C as determined by the DMTA results. Despite the loss of [CO.sub.2] from the system, the agreement of the viscosity and processing temperature reduction values for the 65% AN material between this study and the previous study might suggest that the 85% AN material could also be processed at lower viscosity and temperature in the extrusion process. The processing temperature reduction of 26[degrees]C for the 85% material obtained by Bortner and coworkers (21) can be applied to the chemorheo-logical analysis. Because the ideal viscosity for processing is obtained at 220[degrees]C for the pure material, when reduced by 26[degrees]C to 194[degrees]C, the analysis shows that the polymer will remain stable during extrusion, with minimal viscosity increase. If the 26[degrees]C reduction can truly be achieved with the 85% AN copolymer in an extruder as it was in the batch process, as was the case with the 65% AN copolymer, the 85% AN copolymer will be continuously melt processable with [CO.sub.2] at ideal viscosities and remain stable. Unfortunately, because of the cost and lack of sufficient amounts of the 85% AN material for extrusion processing with [CO.sub.2], it is not possible to confirm this, but it is certain that viscosity data obtained with the slit-die rheometer would corroborate these results.
TABLE 3. Summary of [CO.sub.2] absorption experiment data. AN content [CO.sub.2] content [T.sub.g] reduction (at [P.sub.0]) (mol%) (wt%) ([degrees]C) 65 12 59 65 6.7 31 85 5.6 37 Viscosity reduction (%) Potential process T reduction Source ([degrees]C) 56 30 This work 60 30  61 26 
To study the feasibility of using this process to produce homogenous extrudates and to determine what pressure was needed to suppress foaming, materials were extruded into a pressurized chamber attached to the end of the die. This chamber was operated in batch mode, but could be adapted to run continuously, similar to the steam pressurized chambers used to pressurize PAN plasticized with water and organic solvents (10), (15). For [CO.sub.2] injection rates of 7 wt% (around 3-4 wt% absorbed [CO.sub.2] after accounting for losses), extrudates were collected in the chamber at atmospheric pressure and then with nitrogen pressure at 8.3 MPa, which was easily obtained by filling the chamber with nitrogen from a gas cylinder. Samples extruded with and without nitrogen pressure were analyzed using scanning electron microscopy (SEM) as shown in Figs. 16 and 17. Figure 16a depicts an image of the material extruded into atmospheric pressure conditions. The formation and rupture of bubbles due to expansion of the absorbed [CO.sub.2] is readily apparent in this figure with a magnification of 59 X. Under nitrogen pressure, the sample shown in Fig. 16b shows no sign of bubble growth at 77 X magnification. This indicates that the pressure chamber, when operated at a reasonable pressure, can be used to successfully inhibit the growth of gas bubbles in the melt extrudate. At higher magnifications (50,000 X), the presence of nanometer sized voids in the material extruded into atmospheric conditions can be seen in Fig. 17a. These voids are also present to a lesser extent in the material extruded into the chamber under nitrogen pressure (Fig. 17b). These voids may be smaller than those observed in solvent-plasticized melt spun PAN fibers (9), (16).
[FIGURE 16 OMITTED]
[FIGURE 17 OMITTED]
The ability to rapidly plasticize and reduce the viscosity of a 65% AN copolymer in an extrusion process has been shown. The viscosity was reduced by up to 56% at 7 wt% [CO.sub.2] when corrected for loss of [CO.sub.2] through the extruder. A chemorheological analysis of the 65 and 85% AN copolymers was used to determine that the 85% AN copolymer would be suitably stable for over 30 min of extrusion at temperatures up to 200[degrees]C. Scaling with respect to [CO.sub.2] concentration and pressure was used to generate shift factors for the combined effect of pressure and concentration on the viscosity of the 65% AN material. These were then compared to pressure and concentration dependent shift factors generated from an Arrhenius analysis to determine if it could be used to predict the required [T.sub.g] reduction for a specified viscosity reduction. The comparison showed that concentration dependent shift factors calculated from [T.sub.g] versus uptake data at atmospheric pressure did not agree with the combined shift factors and, therefore, the viscosity reduction determined experimentally using the slit die. This was likely due to a large pressure effect arising between data at atmospheric pressure and at extrusion pressures, which was not encountered in the pressure and viscosity data obtained from the slit die. The pressure effect was ruled out for the viscosity data measured in the slit die because no change in viscosity was seen over the range of pressures generated in the die. Comparison of the viscosity and [T.sub.g] reduction data of the 65% AN copolymer to similar data obtained previously (20, 21) for the 65% and 85% AN copolymer along with comparisons of the stability data showed that the 85% AN copolymer may be melt processable at a viscosity suitable for melt spinning and remain stable for 30 min at 194[degrees]C, but direct evidence of this was not attainable. Foaming that generally occurs due to the rapid release of absorbed [CO.sub.2] at the die exit was greatly reduced in the 65% AN material by extruding the material into a pressurized chamber at a reasonable pressure. These results suggest that the potential to melt spin high AN content (at least 85 mol%) copolymers may be possible using [CO.sub.2] as a plasticizer and pressuring the extrudate to suppress bubble growth.
The authors thank the U.S. Environmental Protection Agency (EPA), Science to Achieve Results (STAR) program for financial support.
(1.) J.P. Riggs, Encyclopedia of Polymer Science and Engineering, Vol. 2. Carbon Fibers, Wiley, New York (1985).
(2.) P. Rangarajan, J. Yang, V. Bhanu, D. Godshall, J. McGrath, G. Wilkes, and D. Baird, J. Appl. Polym. Sci, 85, 69 (2002).
(3.) R.C. Smierciak, E. Wardlow, and L.E. Ball, U.S. Patent 5,602,222 (1997).
(4.) M.C. Paiva, P. Kotasthane, D.D. Edie, and A.A. Ogale, Carbon 41, 1399 (2003).
(5.) M.K. Jain and A.S. Abhiraman, J. Mat. Sci., 22, 278 (1987).
(6.) A.K. Gupta, D.K. Paliwal, and P. Baja, J.M.S.-Rev. Macromol. Chem. Phys. C31, 1 (1991).
(7.) V.A. Bhanu, P. Rangarajan, K. Wiles, M. Bortner, M. Sankarpandian, D. Godshall, T.E. Glass, A.K. Banthia, J. Yang, G. Wilkes, D. Baird, and J.E. McGrath, Polymer, 43, 4841 (2002).
(8.) P. Rangarajan, V.A. Bhanu, D. Godshall, G.L. Wilkes, J.E. McGrath, and D.G. Baird, Polymer, 43, 2699 (2002).
(9.) B.G. Min, T.W. Son, B.C. Kim, and W.H. Jo, Polym. J., 24, 841 (1992).
(10.) G.P. Daumit, Y.S. Ko, C.R. Slater, J.G. Venner, and C.C. Young, U.S. Patent 4,921,656 (1990).
(11.) G.P. Daumit, Y.S. Ko, C.R. Slater, J.G. Venner, C.C. Young, and M.M. Zwick, U.S. Patent 4,981,752 (1991).
(12.) S.K. Atureliya and Z. Bashir, Polymer, 34, 5116 (1993).
(13.) CD. Coxe, U.S. Patent 2,585,444 (1952).
(14.) B.G. Frushour, Polym. Bull., 7, 1 (1982).
(15.) H. Porosoff, U.S. Patent 4,163,770 (1979).
(16.) D. Grove, P. Desai, and A.S. Abhiraman, Carbon, 26, 403 (1998).
(17.) F. DeMaria and C.C. Young, U.S. Patent 4,303,607 (1981).
(18.) R.E. Pfeiffer and S.E. Peacher, U.S. Patent 4,318,680 (1982).
(19.) A. Cooper and S. Howdle, Mater. World, 8, 10 (2000).
(20.) M.J. Bortner and D.G. Baird, Polymer, 45, 3399 (2004).
(21.) M.J. Bortner, V.A. Bhanu, J.E. McGrath, and D.G. Baird, Polymer, 45, 3413 (2004).
(22.) M.J. Bortner, V.A. Bhanu, J.E. McGrath, and D.G. Baird, J. Appl. Polym. Sci., 93, 2856 (2004).
(23.) M.J. Bortner, Melt Processing of Metastable Acrylic Copolymer Carbon Precursors, PhD Dissertation, Virginia Tech, Blacksburg, VA (2003).
(24.) C.Y. Ma and CD. Han, J. Cell. Plast., 18, 361 (1982).
(25.) CD. Han and C.Y. Ma, J. Appl. Polym. Sci., 28, 851 (1983).
(26.) J.R. Royer, Y.J. Gay, J.M. DeSimone, and S.A. Khan, J. Polym. Sci. Part B: Polym. Phys., 38, 3168 (2000).
(27.) J.R. Royer, J.M. DeSimone, and S.A. Khan, J. Polym. Sci. Part B: Polym. Phys., 39, 3055 (2001).
(28.) J.L.S. Wales, J.L. denOtter, and H. Janeschitz-Kriegl, Rheol. Acta, 4, 146 (1965).
(29.) R. Eswaran and H. Janeschitz-Kriegl, Rheol. Acta, 3, 83 (1963).
(30.) R.B. Bird, R.C. Armstrong, and O. Hassager, Dynamics of Polymer Liquids, Fluid Mechanics, Vol. I, Wiley-Inter-science, New York (1987).
(31.) C.W. Macosko, Rheology: Principles, Measurements and Applications, VCH, New York (1994).
(32.) D.G. Baird and D.I. Collias, Polymer Processing Principles and Design, Wiley, New York (1998).
(33.) R.H. Lacombe and I.C. Sanchez, J. Phys. Chem., 80, 2568 (1976).
(34.) I.C. Sanchez and R.H. Lacombe, .J. Phys. Chem., 80, 2352 (1976).
(35.) I.C. Sanchez and R.H. Lacombe, Macromolecules, 11, 1145 (1978).
(36.) H.E. Park and J.M. Dealy, Macromolecules, 39, 5438 (2006).
(37.) P.P. Gan, D.R. Paul, and A.R. Padwa, Polymer, 35, 1487 (1994).
(38.) B.M. Hasch, M.A. Meilchen, S.H. Lee, and M.A. McHugh, J. Polym. Sci. Part B: Polym. Phys., 31, 429 (1993).
(39.) Y.C. Bae and E. Gulari,.J. Appl. Polym. Sci., 63, 459 (1997).
(40.) L.J. Gerhardt, C.W. Manke, and E. Gulari, J. Polym. Sci. Part B: Polym. Phys., 35, 523 (1997).
(41.) L.J. Gerhardt, A. Garg, C.W. Manke, and E. Gulari, J. Polym. Sci. Part B: Polym. Phys., 36, 1911 (1998).
(42.) C. Kwag, C.W. Manke, and E. Gulari, J. Polym. Sci. Part B: Polym. Phys., 37, 2771 (1999).
(43.) R.K. Gupta, Polymer and Composite Rheology, Marcel Dekker, New York (2000).
(44.) N. Grassie and J.N. Hay, J. Polym. Sci., 56, 189 (1962).
Matthew D. Wilding, Donald G. Baird
Department of Chemical Engineering, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061-0211
Correspondence to: D.G. Baird; e-mail: firstname.lastname@example.org
Contract grant sponsor: U.S. Environmental Protection Agency (EPA), Science to Achieve Results (STAR) program; contract grant number: R-82955501-0.
|Printer friendly Cite/link Email Feedback|
|Author:||Wilding, Matthew D.; Baird, Donald G.|
|Publication:||Polymer Engineering and Science|
|Article Type:||Technical report|
|Date:||Oct 1, 2009|
|Previous Article:||Numerical investigation and experimental validation of the performance of a tubular packed bed reactor for hydrogenation of diene-based polymers.|
|Next Article:||Polyamide-6/high-density polyethylene blend using recycled high-density polyethylene as compatibilizer: morphology, mechanical properties, and...|