Computer flow analysis troubleshoots film coextrusion.

Computer Flow Analysis Troubleshoots Film Coextrusion

Flow problems are encountered in all types of coextrusion dies. The two most important types of problems are uneven layer thickness across the width of multilayer fims, and interfacial flow instability, manifested by poor optical properties and/or poor appearance. To help solve these problems, a computer program for multilayer flow analysis, recently developed by Quantum Chemical Corp.'s USI Division, enables various resin candidates for coextrusion applications to be evaluated quickly and inexpensively. A typical analysis requires only about five minutes to run. A portable personal computer can run the program, thus permitting on-site analysis and optimization.

The Quantum program works for two- for seven-layer cast or blown films, whether made with multimanifold dies or feedblocks. The program can calculate interfacial parameters - such as shear stress and shear rate, velocity viscosity, and temperature - based on resin properties and processing conditions.

The choices for solving a coextrusion processing problem are to either: 1) fine-tune process conditions; 2) change the resin(s) used; 3) adjust the die gap; or 4) modify the die or feedblock design. The program helps exhaust the first three types of solutions - by far the least expensive and time-consuming to implement - by quick modeling of "what-if" scenarios.

The program is proprietary and is only available as a technical service to customers of Quantum's USI Div.

VISCOSITY, LAYER UNIFORMITY

The layer-thickness uniformity problem is caused by the difference in melt viscosities of the layers. This difference can be seen in Fig. 1, where two different melts are fed side by side into a coextrusion die. A cross-section of the downstream development of the interface between melts #1 and #2 is shown at the bottom of Fig. 1. The less viscous melt #2 tends to displace the more viscous melt #1 from the high-shear region near the die wall, and eventually completely encapsulates the more viscous melt. It is a natural tendency for a less viscous fluid to migrate in a high-shear region so as to minimize energy requirements, since energy requirements for flow are proportional to viscosity.

However, encapsulation of a more viscous fluid by a less viscous fluid is a slow process and would require an extraordinary die length to reach completion. In practice, dies are short, and therefore only a portion of the encapsulation process occurs before the materials exit the die. The cross-section at the second point from the left is typical of what is observed in practice - i.e., layer #2 has just started to displace layer #1 from the side walls, thereby causing a nonuniform thickness across the width of the die.

There are a couple of ways to alleviate the layer uniformity problem in coextrusion: 1) increase or decrease viscosity of the skin layer so as to reduce the viscosity mismatch; and 2) modify the feedblock or die design to compensate for the viscosity mismatch. Our computer program can help with alternative resin selection to achieve the former.

INTERFACIAL STRESS & CLARITY

Interfacial instability manifests itself by a waviness, which in severe cases becomes chaotic. The magnitude of this interfacial waviness is in the size scale of the wavelength of light (about 0.5 micron), thus causing the loss of optical clarity.

Interfacial instability does not affect contact clarity. As can be seen in Fig. 2 top both the "good" and "bad" films have the same contact clarity, as is evident by the readability of the text through the film. However, when the text is moved some distance away from the film (Fig. 2, bottom), the difference in see-through clarity between the two films becomes quite apparent. The text is still readable through the "good" film, but not through the "bad" film.

In Fig. 3, micrographs show cross-sections of the interfacial area in the "good" and "bad" films of the previous example. In the "good" film, the interface is a distinct, smooth line between the two layers. In the "bad" film, the interface is irregular and much less well defined.

SHEAR STRESS IS CRITICAL

What is responsible for interfacial flow instability? Figure 4 shows a cross-section of a coextrusion die and the shape of the interface at the onset of instability in the die land. Note also that the interface waviness extends only over the die land and not farther upstream. The die land is the narrowest passage the material flows through, where shear stresses are the highest.

The conclusion, therefore, is that interfacial instability occurs when a critical shear stress at the interface of melt layers is exceeded. Although such a statement is very difficult to prove or disprove, there are a few experimental results available in the literature that indicate that the onset of interfacial instability does correlate with shear stress at the interface.

Based on the foregoing conclusion, we assume that the controlling factor for interfacial instability is a critical interfacial shear stress. Viscosity ratio and layer-thickness ratio contribute indirectly to this instability through their effect on interfacial shear stress. (It is not clear, at present, what the effect of melt elasticity is, if any.)

COMPUTER PREDICTION OF

SHEAR STRESS

Interfacial shear stress cannot be measured. The only way to estimate it is through computer simulation. Quantum's coextrusion computer program requires the following material and process data:

* For each resin - density, heat capacity, thermal conductivity, flow rate, process temperature, and viscosity as a function of shear rate and temperature;

* Die geometry - gap, length and width.

Profiles of shear stress (Fig. 5A) velocity (Fig. 5B), viscosity (Fig. 5C), temperature (Fig. 5D), and shear rate at various locations in the die can be viewed graphically on the computer screen. Figure 5A shows the shear stress of a three-layer coextrusion, with the dotted lines indicating the interface positions. The bottom interface is at a low shear stress, while the top interface is in a higher stress region.

An actual example of using the coextrusion computer program is shown in the accompanying table. The coextruded structure was a two-layer film. An extrusion trial with the original configuration (#1) showed interfacial instability, as manifested by the characteristically poor opticals (the "bad" film in Fig. 2), and this configuration was unacceptable.

A computer simulation of the original configuration (#1) showed that the interfacial shear stress was 4.9 psi. To eliminate the instability, the stress had to be reduced. The maximum acceptable stress limits cannot be predicted a priori, at least for the time being. However, all that needed to be known in this example was that the stress was over the (unknown) limit, and it had to be reduced.

First, as shown in #2, changing the output rates for each layer changed the skin-layer thicknesses and moved the interface toward the center, into a region of lower shear stress (1.6 psi. This alternative was not practical, since changing the layer thicknesses was not acceptable.

The second alternative, as shown in #3, was to open up the die gap. The interfacial shear stress dropped to 3.5 psi. However, this alternative also was not acceptable: opening the die gap would have required greater drawdown to maintain the same final film thickness.

The third alternative, #4, was to replace resin "A" with resin "a," having a higher melt flow rate (MFR) and a viscosity about one-fourth that of resin A. The computer simulation showed an interfacial shear stress of 3.2 psi. This solution was considered the best, as it achieved the desired result (reducing the interfacial shear stress, from 4.9 to 3.2 psi) without changing anything else, and the film had better optical characteristics.

Quantum's computer program does not directly handle the layer uniformity problem. Only indirect conclusions can be drawn. For a given problem, the computer simulation provides the viscosity variation along the die gap. If the viscosities of the two adjacent layers are very different at the interface, then a layer uniformity problem is likely at that interface. Computer simulation assists what is essentially a qualitative prediction of a layer uniformity problem; at present, a quantified prediction is not possible.

'VISCOSITY MATCHING' NOT

NECESSARY

Another example of using computer simulation of interfacial shear stress dispels the common misconception that "viscosity matching" is the appropriate strategy in coextrusion. This example (based on a personal communication from C.R. Finch) involves a two-layer decorative sheet of high-impact polystyrene (HIPS) with an MFR of 3 g/10 min for more than 95% of the structure, and clear, general-purpose PS (GPPS), available in the 3 to 15 MFR range, for less than 5% of the structure. Matching of viscosities (using a GPPS of 3 MFR) caused severe interfacial instability.

The instability problem disappeared when a 15 MFR GPPS was used, although in this case the viscosities of the two layers differ by about a factor of five (assuming viscosity to be inversely proportional to MFR).

Computer simulation shows (Fig. 6) that the interfacial shear stress is reduced by a factor of two when using 3 MFR HIPS with 15 MFR GPPS, compared with the original 3 MFR HIPS/3 MFR GPPS configuration. Interfacial shear stress is the critical factor in this example, and not the matching of layer viscosities.

PHOTO : FIG. 1 As they flow through the die, the less viscous melt #2 tends to displace the more viscous melt #1 from the high-shear region near the die wall, and eventually completely encapsulates #1.

PHOTO : FIG. 2 Interfacial instability does not affect contact clarity, as shown by this comparison of "good" and "bad" coextruded films. When the text is moved some distance away from the films, the difference in see-through clarity caused by interfacial instability becomes obvious.

PHOTO : FIG. 3 Micrograph of cross-section of "good" film from Fig. 2 shows the interface as a distinct, smooth line between the two layers. In the "bad" film, the interface is irregular and less well defined.

PHOTO : FIG. 4 Schematic derived from photomicrographs shows onset of interfacial instability in the die land, where shear stress is highest, and not earlier in the die. (Source: W.J. Schrenk, N.L. Bradley, T. Alfrey and H. Maack, Polymer Engineering and Science, SPE, June 1978)

PHOTO : FIG. 5A Computer output of shear stress in a three-layer coextrusion. Dotted lines indicate interface positions. Botton interface is at lower shear stress than upper interface.

PHOTO : FIG. 5B Velocity output from coextrusion computer model.

PHOTO : FIG. 5C Viscosity output from coextrusion computer model.

PHOTO : FIG. 5D Temperature output from coextrusion computer model.

PHOTO : FIG. 6 A "viscosity mismatch" may be beneficial. For the HIPS/GPPS sheet, reducing interfacial shear stress - not matching viscosity - was the key to eliminating instability.