Skip to main content

Postpone polymer processing instabilities (3PI)

Deliverables

In tyre manufacturing, rubbers (SBR) are mixed with a lot of ingredients, fillers as reinforcing agents, processing aids and curatives. SBR based compounds with silica as filler, are known to be hardly processable. Flow defects and poor rubber strength occur when compounds are extruded or injected. Flow and aspect defects can be only slightly postponed using: - Processing promoters that reduce compound viscosity as well (the most effective are fatty acid Zn salts based); - Covering the inner die surface that is in contact with the polymer with a layer that favours slippage. The addition of 2phr of DYNEON fluoroelastomer as slip promoter is completely inefficient, increasing flow defects and getting final compound properties worse. The best way to postpone compound flow instabilities is: - Use of polymer blends (two or more SBRs, blend with BR) to enlarge molecular weight distribution; - Addition of a certain quantity of Carbon Black. It was found that 30phr of CB are enough to improve considerably elongational properties of green compound. Surface aspect is even better, very fair and smooth. Silica compound ageing was found very important in determining flow defects: the more you store the compound, the more it generates flow instabilities and surface defects. Elongational compound properties and viscosity measured in extensional rheometry seem to be more predictive than common shear viscosity. It is going to be proposed as a lab test for new compound development.
In this part of the project the occurrence of slip near the die wall has been thoroughly studied. A program has been written to numerically simulate the dynamical behaviour of the polymer molecules in a thin layer near the wall. This yields the so-called slip law, i.e. the dependence of the slip velocity on the shear stress at the wall. Special attention has been paid to the interaction between polymer chains attached onto the wall and bulk chains that may are entangle with them. It is shown that the entanglement/disentanglement mechanism dominates the slip law. One of the findings of this study is that the slip law is usually a nonmonotonous curve. This remarkable result is used to develop a model that includes this type of slip law. The model clearly shows that a nonmonotonous slip law inevitably leads to the occurrence of spurt. Moreover, the model allows prediction of the critical extrusion rate at which spurt will manifest itself and makes it possible to calculate its characteristics, such as period and amplitude. This part of the project has led to a program, which may predict spurt behaviour as a function of the extruder geometry and the slip law. Furthermore, a simulation program has been developed to calculate the slip law. A start is made with unraveling the mechanism behind the sharkskin instability. It could be that the slip law is also an important ingredient for modeling sharkskin.
It has been shown that the use of a smooth convergent geometry was very efficient to postpone the volume defects encountered in polystyrene extrusion. Instead of a flat entry, a small angle (30°, 45°) or even best a smooth profile (trumpet-shape profile) permits to increase the flow rate by a factor 8 before the onset of any volume instability. This effect was also observed for the polypropylene. By using numerical simulation, it has been evidenced that the stress field was highly modified by the reduction of the entry angle and that the use of a trumpet-shape profile allowed one to largely reduce the extensional stresses at the entry of the die channel.
A broad selection of analytical and rheological techniques was used to characterise a variety of polymers. The result contains a round-robin comparison of the results of the various techniques. It provides new insights into the value and discriminating power of each technique. The results were summarised in a toolbox, which serves as input for further evaluation and calculation of the performance of these polymers. It was found that GPC analysis for molecular mass distrubtion determination can lead to large differences between partners, if not handled according to the same protocol. Rheological tests were better reproducable between labs. The combined use of creep and creep recovery and oscillatory shear and capillary rheometry provides a good tool to determine the entire shear relaxation spectrum and viscosity. Uniaxial extension was found to be a good test to distinguish non-linear viscoelastic effects. The result has introduced new tools to various partners. It has enabled determining differences between polymers that were hard to find with previous methods. It has stimulated the discussion on well-defined protocols.
Different experimental techniques have been used to characterize the flow conditions before and during the instabilities. They are mainly based on the use of transparent tools, i.e. extrusion dies with glass windows. Such tools have been developed in LSP Erlangen, CEMEF and University of Cambridge. They allow one to perform two kinds of measurements: - Flow Induced Birefringence (FIB): a laser or optical light is passed through the flow between polarizer and analyser. The optical anisotropy created by the flow leads to the observation of extinction bands, related to the stress field. Thus, FIB is able to provide pertinent information on the stress field, both in stationary or time-dependent flows. This technique was particularly powerful for the characterization of spurt conditions and volume distortions. Moreover, it is also a very efficient tool for assessing the parameters of constitutive equations or to check the validity of numerical simulations. - Laser Doppler Velocimetry (LDV): LDV allows one to measure the local velocity in the flow field. Depending on the system, one or two components of the velocity vector can be measured simultaneously. This method was particularly efficient to identify the boundary conditions (slip or no-slip at the wall) and to detect the presence and the size of vortices in abrupt contraction geometries. As for FIB, results are also very useful for controlling the validity of rheological laws or finite element simulations. Among the other experimental techniques used in 3PI project, we should include the Multipass Rheometer (MPR) developed by Cambridge. This specific rheometer with two pistons allows one to perform successive experiments with a limited amount of material and to characterize flow conditions under controlled pressure.
Linear polyethylenes exhibiting processing instabilities were modified by blending. Various combinations of low and/or high molecular weight linear polyethylene fractions were added to the main polymer. A small-scale processing study involving a capillary rheometer enabled estimating the dependence of onset and severity of various extrudate distortions on molecular weight. It was shown that various combinations lead to higher onset rates, whereas others lead to lower onset rates. Several promising candidate blends were produced in larger quantity. Blown films were made under different conditions. It was concluded that blending enables postponing the occurrence of the instabilitites to higher production rates. In general, the final mechanical properties of the film will be worse than for the base material. However, the effect is small enough to make this a viable route for various applications.
The aim of the Work package 4 was to develop numerical tools for polymer processing. The first step was the development of specific model for the draw resonance and the spurt flow instabilities. - Specific models. -- The linear stability of the spinning process in non-isothermal conditions has been studied. This model allows determining the critical draw ratio for the onset of the draw resonance instability. The influence of cooling due to convective exchange with air and different mechanical behaviour for the molten polymer are considered in this model. -- The model for the spurt flow instability assumes both compressibility in the reservoir and slip at the wall (in the capillary) above a critical shear stress. A complex stick-slip mechanism at the wall allows predicting stable and unstable flows. Strong hypothesis on the flow geometry are used for these two models and it results in efficient and relatively easy to use numerical codes. These tools are fitted for the analysis of laboratory experiments. The price to pay for this simplicity is the difficulty to consider more complex flows encountered in industrial process. These models of the first generation were completed by a more complex to use but more adjustable finite element code. - General models. The second task of Work package 4 was to develop general numerical tools allowing to compute numerically polymer flow obeying a generalized Newtonian (Carreau-Yasuda) or a viscoelastic (multi-mode Phan-Tien Tanner or multi-mode Pom-Pom model) behavior. We have used the general framework of the finite element code MEF++ (version 2.0) developed by the GIREF at the Laval University in Québec (Canada). It can be used to compute numerical problems in solids and fluids mechanics, heat transfer and coupled problems (see for example http: //www.giref.ulaval.ca). Specific modules was developed for the computation of molten polymer flow obeying realistic constitutive equations. These modules can now be used as bricks to built models for realistic geometries. This code uses classically balance and constitutive equations and boundaries conditions. It leads to a complex set of non linear equations solved using iterative techniques. The important steps are the following: - Meshing of the flow geometry, - Choose of a numerical path, - Iterations until convergence. It is necessary to precise that each of these steps requires some knowledge in numerical computations. - Conclusion. The numerical modules necessary for the study of the stability of polymer flows in industrial geometries are available. These modulus have been used to study the helical instability (C. Combeaud, CEMEF) and the flow in the Multipass Rheometer (R. Valette, Cambridge).
How to speed up the characteristic coating time for suppressing sharkskin defects: When a fluoropolymer-based ‘Polymer Processing Additives’ (or PPA) is added to a plastic formulation that is sensitive to surface defects like sharkskin (e.g. polyethylene), it is known that these defects are not eliminated immediately. The fluoropolymer creates a thin coating at the surface of the metal parts in the extruder (typically at the zones where high shears are like in the extruder head). This process of forming the coating takes some time. It is clear that the faster this coating is formed, the less production time is lost due to lower quality plastic film or off-spec material. Often a fluoropolymer-based PPA exists as a combination of a synergist and a fluoropolymer (fluoroplastic or fluoroelastomer). It has been proven in earlier tests in the Dyneon laboratories that the characteristic coating time of the PPA is strongly depending of the formulation. Chemical (reactions) or physical (abrasion, adsorption) interaction can slow down the formation of the coating. Recently Dyneon developed new fluoroelastomers that have molecular properties (molecular weight – viscosity) that are able to eliminate significantly faster the surface defects (sharkskin).
TIP Driving mechanisms and their consequences - Mechanism for upstream instability The experimental observations carried out by CEMEF helped establish that the upstream instability was associated with the viscoelastic flow of the material within the entrance region of the die. The instability originated in the entry region of the die and propagated through the die to the final extrudate. In some cases a symmetric pulsing action was observed and in other cases as periodic asymmetric instability was seen. The experimental data provides a sound basis for comparison with numerical simulation and also provides commercial design rules in order to avoid this instability. - Stick spurt instability Systematic experiments on the stick spurt instability were carried out at both CEMEF and LSP. Additional experiments were done by Repsol, Dow, LMPL, Dyneon, Pirelli. Both CEMEF and LSP developed experimental facilities to measure point wise velocities using LDV techniques. Both groups showed self consistent data relating to velocity profiles within the parallel section of a die. Below the stick spurt regime ‘power law’ velocity profiles were observed where either no slip or in some cases, partial slip was observed. In the stick spurt regime a substantial change in the velocity profile was observed. During the spurt part of the flow a near plug flow response was seen. These results unambiguously and for the first time show that stick spurt is related to wall slip. This result is of great significance in terms of an identified mechanism and is of relevance to essentially all partners. -- The mechanism for stick spurt The experimental observations have shown that stick spurt is related to a slip condition at the parallel section of the die. Clearly compressibility is also involved and the form of the instability is caused by an interplay between compressibility and wall boundary conditions. The result highlights the region of the flow where attention needs to be given in order to avoid this instability. - Sharkskin instability Systematic experiments were carried out by CAMB and LSP with additional work due by Dow, Dyreon, Argo and Repsol. CAMB established that side stream gas injection did not diminish sharkskin, however observation helped to confirm that the instability is related to a stress concentration at the exit of the die. CAMB showed that surface finish surface material and exit curvature influenced sharkskin. CAMB carried out experimental flow birefringence studies that indicated the instability originated from the local stress concentration at the exit. Both LSP and CAMB showed that Dyneon additivewas effective in eliminating sharkskin and LSP showed that sharkskin elimination could occur even without a velocity profile modification. -- Mechanism for sharkskin From the experiments carried out it is clear that the sharkskin instability is associated with the local stress concentration at the exit of the die. Partial slip can reduce sharkskin but other factors also play a part in the mechanism. The addition of fluorocarbon additives was found to be the most effective way to eliminate sharkskin. - Draw resonance Experiments were carried out by Dow and CEMEF using Rheotens apparatus. The critical draw ratio for a given draw length L was established for a range of different materials. -- Mechanism for draw resonance It is generally accepted that the mechanism for downstream draw resonance is an interplay between the draw ratio, geometry and viscoelastic forces and the experimental results obtained supported this view. - Additional experiments Systematic experiments on the time dependant viscoelastic flow of compressible polymer melts were carried out in CAMB in order that they can be compared with a numerical code developed at CEMEF/CAMB. The numerical capability that has been developed has the future potential to simulate and quantify, upstream, stick spurt , sharkskin and draw resonance instabilities. This code should provide a powerful tool for the future.