Volume 6, Issue 4

This paper gives a review of existing scale-up rules for single screw plasticating extruders. It is shown that there were and still are rather varied trends. The main three are: – scale-up rules under the boundary conditions of an invariant shear rate, – scale-up rules including the power law, – scale-up rules including the zones of operation. In Europe, at least, the scale-up rules, which take the power law-flow characteristics into account, seem to have been established in practical terms. For economical reasons conventional extruders are designed in such a way, that the output increases with the square of the diameter. This however, results in the problem that the melting rate does not follow the same scale-up rules as the pumping rate and the solid conveying rate. Thus the discussion continues, to calculate the operating section lengths according to varying scale-up rules. Finally the scale-up problem is discussed once more with respect to the dominant dimensionless numbers, and despite several similarity theoretical difficulties a standardized solution will be attempted to be found.

In the development of a new polymer and a new compound, mixing equipment plays an important role. However, the properties of a compound tend to vary depending on the size and shape of the equipment used and a means to obtain the same compound by a small scale laboratory mixer as the one obtained by a large scale mixer has been longed for a long time. As a rubber compound is characterized by the physical properties, they are affected both by the dimensions and shape of the equipment and the mixing conditions. Recently, we developed a laboratory mixer generally following the designs of FH Banbury with exchangeable rotors and mixing chamber blocks which enable us to investigate the influences of the shape of equipment as well as the mixing conditions. In analysing rubber mixing, we found that the following factors should be taken into account, that is, unit-work which is the applied energy to the unit volume of the material during mixing, Mooney viscosity of the compound, bound rubber which is the amount of polymer unextractable from the compound by a solvent, and weight average molecular weight of polymer extractable by a solvent. If the values of the four factors are close enough for the two compounds obtained by different mixers of different size and shape, one may regard them as the same compounds. Furthermore, we experimentally measured the four factors for the compounds of typical formulations of three species of commercially available rubbers, that is, styrene butadiene rubber, ethylene propylene rubber and butadiene, and expressed the values of the factors as functions of mixing conditions and the parameters of rotor shapes by means of multiple regression analysis. Among the rotor parameters, the following showed significant effects: tip clearance, tip width, total bulkiness and wing overlap ratio. As for the mixing conditions, mixing time, rotor speed and mixing temperature were dominant. Using the functions obtained by the above mentioned method for the laboratory scale mixer, we tried to find the combinations of rotor shapes and mixing conditions that reproduced the compound mixed by large scale mixers. The optimum parameters of a laboratory mixer for the reproduction of the mixing of an industrial mixer were found to be larger rotor tip clearance, larger rotor tip width, larger wing overlap ratio, higher mixing temperature, and higher rotor speed than those of proportionally reduced dimensions and comparable conditions.

This work is an investigation on causes of extrusion problem of a high silica- and all silica-filled natural rubber compounds containing a coupling agent. The 67% silica and 33% carbon black, N110 compound (Compound 3) gave a wavy surface of extrudate from a profile die. The 100% silica-filled compound (Compound 4) was too stiff to be extruded. Reference compounds, 100% carbon black (Compound 1) and 33% silica with 67% carbon black (Compound 2), gave smooth extrudates. The amplitude dependence of storage modulus measurements indicated Compound 4 had a poorer dispersion of the filler. Scanning electron microscopy also showed a poorer dispersion for this compound. However, Compound 3 could not be distinguished from Compounds 1 and 2 in the above investigation. Transient shear measurements at very low deformation rate indicated a reaction between natural rubber and the coupling agent, resulting in gel and long-branch formation. Differential scanning calorimetry indicated the coupling reaction was completed at 140°C within the mixing time. By holding the maximum temperature of mixing as low as possible, the extrudate of Compound 3 became smooth, but Compound 4 still did not perform well. Improvement of dispersion is necessary to make Compound 4 extrudable.

Based on the classic Tadmor model, a number of mathematical models have been developed for calculating the melting profile in single-screw machines. The most comprehensive of these is the five-zone model from Lindt. All these models can only be solved by numerical methods. Parallel to this, attempts have been made to find analytical solutions which describe the melting profile with sufficient accuracy. In this report a comprehensive analytical solution for the whole premelting- and melting process on barrel and screw with respect to the power law will be given.

Dimensionless calculations for throughput and pressure gradients are reported for flow of non-Newtonian fluids in screw and kneading disk elements of a modular intermeshing self-wiping corotating twin screw extruder. These calculations are applied to compute pressure and fill factor profiles along the axis of commercial twin screw extruders of varying scale. The used twin screw geometries are found to allow scaling for both the double threaded/tipped and triple threaded/tipped screw/kneading disc families. Dimensionless calculations are presented and applications to modular machines design are described.

The isothermal flow of plastic fluids in coathanger dies is studied. A general three-dimensional finite element code is developed for the flow analysis purpose. The pressure distribution, the velocity distribution and the transverse flow rate distribution are calculated. The yield value of filled compounds gives rise to a mobile plug in the flow channel. The development of plug zone along the flow direction is also shown by the finite element analysis. The effect of the die geometry on the transverse flow rate distribution is critically discussed. It is found that the cross sections of the flow channel in a coathanger die has the shape of a dog bone or a dumbell which make a nearly even transverse flow rate distribution possible.

This paper demonstrates the usefulness of frequency analysis in monitoring melt flow instabilities. The flow behavior of a film-blowing grade of high density polyethylene has been studied using a capillary rheometer at constant piston velocities. The pressure in the reservoir was measured by a transducer fitted to the driving end of the piston. The signal was amplified and processed in a Fast Fourier Transform (FFT) frequency analyzer. The pressure fluctuations (noise) have a very low intensity at low flow rates in the sharkskin region. The noise signal increases dramatically at a critical shear stress of about 1.5 × 10 5 Pa, where the flow assumes a pulsating character (spurt defect). Above the pressure oscillation region the melt goes through a pseudostable region, followed by a secondary pressure oscillation region, eventually reaching a chaotic stage. The noise spectrum intensity seems to be related to the severity of the surface disturbances of the extrudate in this latter branch of the flow curve.

This paper demonstrates the usefulness of frequency analysis in monitoring flow behavior in a single-screw extruder fitted with a capillary die with a length to diameter aspect ratio of 20. The melt pressure in the barrel and die was measured by a transducer, and the pressure fluctuations processed in a Fast Fourier Transform (FFT) frequency analyzer. The pressure fluctuations were studied under various processing conditions for high density polyethylene, linear low density polyethylene and polypropylene. In general, stable extrudate flow generates a noise spectrum with low intensity, the intensity increasing with the level of instability. The machine vibrations appeared to have a direct effect on flow stability. Also, unstable melting of the resin along the screw resulted in an increase in noise level. It seems probable that the frequency of the surface distortions can be related to the shape of the noise spectrum. In general, although more complex in nature, the noise results related to melt flow in the extruder were in agreement with data reported for capillary flow using a piston rheometer.

An experimental study of structure development in melt spinning polyetherimide (PEI) and polyarylate (PAR) is presented. PEI was melt spun at 320° and 350°C and PAR at 310° and 340°C to drawdown ratios of up to 1 000. The melt spun fibers were found to be totally amorphous but to have significant levels of birefringence which increased with draw-down ratio. The birefringence was found to vary linearly with the spinline stress for both polymers and the data was independent of the melt spinning temperature. The proportionality constants are interpretable as equivalent to stress optical constants for the corresponding melts and have values of 9 400 Brewsters for the PEI and 7 400 Brewsters for the polyarylate. The mechanical properties of the melt spun fibers were determined in uniaxial extension. It was found that Young's modulus and tensile strength increased with spinline stress or birefringence while elongation to break decreased. The mechanical properties were found to be similar to polyether ether ketone, poly-p-phenylene sulfide, and polyethylene terephthalate in similar correlations, but as compared to polypropylene at the same spinline stress had higher Young's modulus and tensile strength but lower elongation to break. We were not able to successfully draw these materials over a hot plate at temperatures above their glass transition temperature.

In this paper the conditions for the manufacture of high content β-form crystal isotactic polypropylene fiber in the presence of β nucleation and the β to α form transition during processing were studied. The results showed that the melt temperature had little effect on the relative content of β-form (K value) presented in as-spun fibers within a certain temperature range (210 to 230°C) while the cooling rate and cooling medium had a significant effect on it. In the case of a constant pump deliver (Q), the K value in the as-spun fiber decreased from 0.95 to 0 as the spinning speed increased from 3.25 m/min to 70 m/min. The smaller the thermal conductivity of cooling medium, the greater was the K value. After stretching the β-form PP fiber, there were many crevices at the fiber surface. It was confirmed by the mercury porosimetry that the drawn fiber possessed microvoid structure. The size and quantity of microvoids were closely related to the K value, stretching temperature and stretching ratio. Compared with PP fiber without nucleating agent, the drawn fiber had a lower apparent density and its moisture absorption was improved.

Biaxially oriented PE films were prepared by compression of uniaxially oriented elongational-flow injection-molded precursors. The films so prepared, having different extents of biaxial stretching were compared on the basis of two derived functions, biaxial draw ratio (ƛ) and degree of biaxiality (B). The former gives a quantitative measure of the average draw ratio of the film surface. The latter describes the strain distribution along the principal directions of the film plane. It is shown that ƛ and B depend on the final applied pressure and on the type of pressing surfaces used. The dependence of ƛ and B on the initial orientation and on the initial anisometry of the precursors is also discussed in the light of structural arguments.

As a first step for the modelling of the processing of semi-crystalline polymers, it is proposed a new overall kinetic model which allows to calculate the evolution of the transformed volume fraction whatever the cooling conditions are. This model derives from initial Avrami-Evans' theory. The transformed volume fraction is calculated as a function of parameters which are measured during crystallizations at a constant cooling rate. This model is easy to use and can be introduced into a more complete thermomechanical calculation. For this purpose we propose to express the energy equation in terms of the enthalpy function. Two applications of our model are presented: the cooling of a polymer sheet (non-flowing polymer) and the cooling of the film on the chill-roll during the cast film extrusion process (flowing polymer).

A numerical model for non-isothermal mold filling and curing simulation in thin cavities with preplaced fiber mats was developed based on the control volume method. Both lumped temperature system (i.e. local thermal equilibrium between the resin and the fiber) and unlumped temperature system (i.e. thermal non-equilibrium locally) were considered. A Lagrangian coordinate system was used in the flow front region to improve the energy transfer calculation. Several molding simulation results were presented to show the effect of fiber mat presence (in the mold cavity) on the inlet pressure and temperaturedistribution.

Over the past few years a technique for relating the quality and properties of sub-components to those of real injection moldings has been developed. A major aspect of this is the correlation of microstructure in the two situations. This paper describes the computer prediction of structure development in injection molded polypropylene. This is done on a personal computer using a finite difference method and it is shown that accurate predictions can be made in relatively low computing time. The characteristic skin in polypropylene moldings is shown to be controlled by the filling phase and is dependent on the shear stress and temperature. The computer model is able to predict the onset of skin formation and hence the skin thickness. These predictions have been shown to agree well with experimental observations. The skin-core boundary which has been found to have a major effect on the mechanical properties of molded polypropylene has also been shown to lie between the no-flow isothermal and the maximum shear rate locus. This has important implications for computer simulations of the injection molding process.