«Viscoelastic flow effects in multilayer polymer coextrusion Dooley, J. DOI: 10.6100/IR555718 Published: 01/01/2002 Document Version Publisher’s ...»
The fact that the multilayer stream at the die inlet is narrow (~2.5-10 cm) compared to die width makes it relatively easy to meter thin surface or adhesive layers.
The versatility of the feedblock has made it the most popular flat-die coextrusion method.
Large numbers of layers may be coextruded, layer structure may be readily altered with interchangeable modules, and thermally sensitive polymers may be protected by encapsulation.
It is estimated that over 95% of flat-die coextrusion systems use a feedblock.
Figure 1-6. Exploded view of modular feedblock and single-manifold die for three polymers forming a five-layer coextrusion One limitation of feedblocks is that polymer viscosities must be matched fairly closely because the combined melt stream must spread uniformly within the die. Severe viscosity mismatch results in nonuniform layers; the lower viscosity material tends to flow to the die edges. A crude rule of thumb is that polymer viscosities must be matched within a factor of three or four, which is a reasonably broad range for many commercially important coextrusions.
Layer uniformity may be adjusted by varying melt temperature within limits dictated by heat transfer. Increasing temperature decreases viscosity, and material moves from the center to the edges; decreasing temperature has the opposite effect. Typically, the individual polymer melt temperatures differ by as much as 30-60°C. Beyond that, heat transfer tends to nullify further adjustment by temperature variation.
Often polymers are intentionally selected with a mismatch in viscosities to avoid flow instabilities. Viscosity mismatch of a factor of ten or more may be necessary. Layer nonuniformity expected with the mismatch is compensated by using shaped feedport geometry;
i.e., the layers are introduced into the die non-uniformly, so that uneven flow within the die results in a satisfactorily uniform final distribution. Considerable art has been developed to extend the range of viscosity mismatch that can be accommodated in a feedblock system by using compensating feedport geometry (15, 16).
Some feedblocks are reportedly capable of coextruding polymers with viscosity mismatch of 100 or more (17, 18, 19). This style of feedblock has movable vanes that partition individual layers prior to combining (Fig. 1-7). The vanes may be freely floating, automatically seeking their equilibrium position on the basis of flow rates and viscosities. This self-adjusting feature can accommodate wide ranges in relative flow rates and viscosities while maintaining layer uniformity in the final product. The vanes may be rotated manually and locked into a nonequilibrium position to adjust uniformity further.
Often distribution pins are used, or the vanes are profiled to compensate for nonuniform layers in this style of die. Distribution pins are cylinders that are placed across the flow stream that can be rotated during processing. These distribution pins are normally cut so that one side is shaped into a specific geometry (such as a deeper slot). By rotating these pins to expose a specific geometry during processing, more or less flow will occur in a specific channel thus allowing flexibility in determining the layer thickness in the individual layers.
Figure 1-7. Feedblock with movable vane partitions that adjust positions for different polymer viscosities and/or layer-flow rates
1.6 Combined feedblock/multimanifold dies Combinations of feedblocks and a multimanifold die are also used commercially. The multimanifold die can incorporate the same design principles as the feedblock: i.e. vanes separating individual manifolds within the die. In a sense, the multimanifold die is a wide feedblock. A feedblock may be attached to one or more manifold inlets as shown in Figure 1-8.
With this system polymers with widely different polymer viscosities and processing temperatures may be coextruded. A very viscous or high temperature polymer may be extruded through one or more die manifolds, while a thermally sensitive or much lower viscosity polymer is coextruded with adhesive layers through a feedblock feeding another manifold. Combining of all layers occurs prior to the final die land.
1.7 Layer multiplication Layer multiplication using a feedblock 1.7.1 The feedblocks and dies discussed in the previous section are suitable for producing multilayer structures with typically ten layers or less, which is adequate for many industrial applications. However, there have been multilayer structures developed in recent years that require hundreds or thousands of layers to produce unique properties (20-23). Among these properties are enhancements in mechanical and optical properties. For instance, films and sheet with tailored optical properties that reflect or transmit specific portions of the electromagnetic spectrum have been developed.
In order to produce film or sheet structures with hundreds or thousands of layers, new techniques had to be developed to produce those structures. Standard feedblock techniques would not allow this many layers because of mechanical difficulties in joining so many layers.
One method to produce hundreds of layers was developed by Schrenk (24). This technique is shown schematically in Figures 1-9 and 1-10.
Figure 1-10. Feedblock for producing many layers – side view The technique shown in Figures 1-9 and 1-10 involves feeding two melt streams into opposite sides of a feedblock that distributes the melt streams radially to the edges of the feedblock. At this point, each of the two melt streams is subdivided into multiple substreams.
These substreams are then interdigitated near the outer periphery of the feedblock in a manner similar to that shown in Figure 1-11. These interdigitated streams are then fed to a central cavity where they exit the feedblock as a multilayered structure.
The feedblock described above is one solution to producing multilayer structures with many layers. However, this type of feedblock also offers other advantages. This type of feedblock was designed to produce structures with uniform layers but its geometry lends itself to easy manipulation of the layer thicknesses. For instance, Figure 1-12 shows a feedblock modification in which the feedblock has been subdivided into three sections labeled I, II, and III.
In each of these sections, the joining slot geometry can be modified to vary the layer thickness gradient discretely in each section (25). This would be useful if by modifying the layer thicknesses in each section the mechanical or optical properties could be modified substantially.
Figure 1-12. Feedblock for producing layer thickness gradients
Figure 1-13 shows a different modification of this feedblock. In this modification, the layer thickness gradient is continuous from one side to the other rather than changing discretely as it was in Figure 1-12. This is done by changing the land length, L, for each of the feed slots and thus changing the flow rate through each slot. This feedblock would produce a multilayer structure in which the layer thickness would progressively change from the top to the bottom of the structure.
Figure 1-13. Schematic diagram of layer thickness gradient feedblock channels 1.7.2 Layer multiplication using Interfacial Surface Generators Another method that is discussed in the patent literature to increase the number of layers in a multilayer structure is the use of interfacial surface generators (26, 27). A schematic diagram of a simple interfacial surface generator is shown in Figure 1-14. This diagram shows how a melt stream can be split and stacked to increase the number of layers in a structure.
Figure 1-14. Schematic diagram of a two-channel interfacial surface generator
Figure 1-15 shows the operative segments and the final flow for a two channel interfacial surface generator being fed with a two-layer structure. The two-layer structure is represented as white on the top and gray on the bottom as is shown in the first segment. Note how the original two-layer structure is split, spread, and stacked to form a four-layer structure. This type of interfacial surface generator will always produce a structure with twice as many layers at the exit as at the beginning of the channel.
One method to further increase the number of layers in a multilayer structure would be to place a number of the two-channel interfacial surface generators in series. Each new twochannel interfacial surface generator would double the number of layers from the previous structure. This means that placing “n” number of two-channel interfacial surface generators in series would produce 2n+1 layers when starting with a two-layer structure.
Figure 1-15. Layer pattern produced in an interfacial surface generator
Numerical simulation of devices similar to interfacial surface generators has been discussed previously (28, 29). In these simulations, a device termed a multiflux static mixer was modeled. A finite element mesh of a multiflux static mixer is shown in Figure 1-16. Note how this device has all of the functional characteristics of the interfacial surface generator since the flow is split, spread, and stacked to increase the number of layers.
Figure 1-16. Finite element mesh for a multiflux static mixer
Figure 1-17 shows the velocity contours from one of the simulations. Note how the flow is not balanced between the different channels in the mixer, which would lead to non-uniform layers in the final structure. These simulation results show how crucial the design of interfacial surface generator is in order to achieve good layer uniformity.
Figure 1-17. Flow patterns in a multiflux static mixer
Figure 1-18 shows the finite element mesh for a new design for a multiflux static mixer (28) that was developed by van der Hoeven to improve flow uniformity in the mixer. The improvement that was introduced was adding extra straight elements before and after the mixing section. These sections add additional resistance to the flow and help balance out the flows in the different channels.
Figure 1-18. Finite element mesh for an improved multiflux static mixer
Figure 1-19 shows the numerical results from the improved multiflux static mixer design.
These results show that the flow balance between the channels is significantly improved by the addition of the straight channels. This improvement in flow balance should relate directly to improved uniformity of layers in a multilayer system.
The experimental device used by van der Hoeven had dimensions of 10 mm x 10 mm x 150 mm. This allowed the insertion of different geometrical sections to determine their effect on the final layer distribution.
Interfacial surface generators can be used to create a variety of structures. Figure 1-20 shows the steps involved when starting with a two-layer structure with equal thicknesses in each layer (shown in the figure as grey and white). Step A splits the structure in the center of the channel and then the two structures are stacked vertically in Step B. In steps C and D, the layers are spread horizontally and then rejoined to form a four-layer structure with equal layer thicknesses.
Figure 1-20. Layer multiplication starting with two layers of equal thickness The process of layer multiplication with interfacial surface generators can also be used with other initial layered structures. Figure 1-21 shows an initial three layered structure with unequal layer thicknesses shown in black, white, and grey. The layer multiplication follows the same steps as before but the final structure now contains six layers with dissimilar layer thicknesses.
Figure 1-21. Layer multiplication starting with three layers of unequal thickness A similar layer multiplication technique could be used on almost any layered structure.
An example would be an asymmetric four-layered structure composed of three materials (A, B, and C) with individual layer thicknesses of 20% A / 5% B / 70% C / 5% B where resin B could be an adhesive material. This initial four-layer structure could flow through an interfacial surface generator and layer B would adhere the layers together keeping the structure intact after the layer multiplication.
Another way to increase the number of layers in a multilayer structure would be to use a four-channel interfacial surface generator rather than a two-channel interfacial surface generator.
A four-channel interfacial surface generator is shown schematically in Figure 1-22. This device has been described (26) as a more efficient way to multiply the number of layers in a multilayer structure.
Figure 1-22 shows how a four-channel interfacial surface generator performs the same functions of splitting, spreading, and stacking that a two-channel interfacial surface generator does. However, this device will quadruple the number of layers at the exit compared to the number at the entry while the two-channel device will only double the number of layers.
Figure 1-22. Schematic diagram of a four-channel interfacial surface generator
The layer multiplication done in a four-channel interfacial surface generator is shown graphically in Figure 1-23. This diagram shows how a two-layer structure shown in the lower left hand corner of the diagram is split into four two-layer structures which are repositioned, spread horizontally, and stacked to form an eight-layered structure.
Figure 1-23. Layer patterns in a four-channel interfacial surface generator One improvement that has been developed for the four-channel interfacial surface generator is the addition of vanes to help control layer thickness (27). This addition is shown in the diagram in Figure 1-24. This diagram shows how the vanes have been added in the section at which the layers are rejoined. Movement of these vanes has similar results to those discussed earlier in conventional feedblocks and dies. The movement of these vanes will preferentially allow more or less material to flow through a specific channel that will change the final layer thickness in the structure.