Just Because The Plastic Part Has Not Failed Yet Does Not Mean It Was Molded Properly

By Michael Sepe, Materials Analyst

You can't count on a robust design to save the plastic part anymore. The molecular weight must be preserved.

Just because a plastic part has not failed yet does not mean it was molded properly. Ask most processors to name the most important task a molder faces in processing and you will get a lot of discussion about cycle time, meeting cosmetic requirements, or controlling dimensional tolerance variation to a particular Cpk requirement. While all of these are important to customer satisfaction and profitability, there is one critical requirement that gets very little attention and renders all of these others moot. It is the task of preserving the molecular weight of the polymer during processing.

We keep returning to the topic of molecular weight because it remains one that is crucial to product performance but gets very little attention until something goes wrong. When processes are first developed, most of the focus is placed on making a full part that is free of flash, meets the print requirements, and looks good. In some cases, a physical test may be devised to ensure that the part meets certain minimum strength or impact requirements. The implicit assumption is that if the part passes these tests, it has been properly molded. In the two case studies profiled below, the process may never have been under control from the standpoint of molecular weight preservation. But due to the benefits of robust design, a lot of product was made before the process finally drifted far enough out of control to result in failures.

AN ALLOWABLE CHANGE IN MFR
It is important first to understand how we evaluate the change in molecular weight that occurs in a polymer during processing. While there are several different tests that can be performed, the simplest and most accessible of these is the melt-flow-rate (MFR) test. Most material suppliers produce their compounds to a particular MFR specification. This is intended to provide controls on the range of average molecular weight values. It is not, as many molders believe, intended as a tool for producing resin that processes consistently.

MFR measurements are made at very low shear rates that are orders of magnitude away from the realm of the real world of injection molding. Shear thinning, which occurs when the velocity of a polymer flow front increases, renders materials with significant differences in MFR as nearly equivalent during first-stage filling as long as the process is not pressure limited.

However, large changes in MFR can spell trouble for the performance of the molded part because this property is often closely tied to the average molecular weight of the polymer. Higher MFRs are associated with lower-molecular-weight materials, and lower-molecular-weight materials lack the performance characteristics of their higher-molecular-weight cousins.

My colleague John Bozzelli has a slide in one his seminars that says, in effect, melt flow is related to molecular weight-sometimes! And he's right. If a compounder adds a lot of mineral oil to a polystyrene or a TPO, it can significantly increase the MFR of the material without altering the average molecular weight of the polymer. As usual, things are more complicated than we might prefer.

However, when we judge the integrity of a molding process, we compare the MFR of a raw material to that of related molded parts. In these instances, the composition of the material has not been altered; therefore, we can attribute any changes in the MFR to the molecular weight of the polymer.

This technique works best for unfilled materials that do not contain large amounts of modifiers that can complicate the mechanism of viscosity reduction. Materials that contain glass fibers or impact modifiers may undergo changes in MFR that appear to be catastrophic but are in fact quite normal. But for simple systems like polyethylene, polypropylene, and polycarbonate, the rules for an allowable change in MFR from pellets to parts are well established.

Most material suppliers will agree that an increase in MFR from pellets to parts of less than 40% represents an acceptable level of molecular weight preservation. This number is not universally accepted, and some prefer a dividing line of 35% or even 30%. Many will talk about levels of quality, with up to 30% being optimal, 30-40% being borderline, and more than 40% being out of bounds.

Everyone agrees that the smaller the change, the better. And the dividing line is somewhat artificial, as it usually is when a hard number is proposed. It does not mean that every part molded with a shift of 39% will work and every part with a shift of 41% will crumble in the hands of your quality department. What it does indicate is that as the change in MFR that occurs during processing increases, the chances for performance problems also increase.

This number also provides a tool that is useful in guiding the failure analysis process. If a part that is breaking in the field exhibits an increase of only 25% in MFR relative to the raw material, we do not focus our attention on polymer degradation as the root cause. Instead, we are more likely to be concerned about part design, the application stresses relative to the inherent properties of the material, or mechanical weaknesses caused by features such as weldlines. But if the increase is 100%, we need to go right to the molding machine and address the process before spending time on other issues.

WHEN GOOD ISN'T GOOD
In our first case, the part was being produced from an unfilled, general-purpose PBT polyester. Raw material was submitted along with “good” parts and failed parts. As is usually the case with plastic product failure, failed meanscracked or broken under conditions at which the material is expected to perform. Good simply denotes a part that has not failed. The results of the MFR tests performed on all three samples appear in Table 1. Each value represents an average of six determinations and the standard deviation for each data set represents 1-2% of the average value.

The failed parts show clear evidence of significant degradation. The increase in MFR for these parts is 128%, well above the upper limit of 40% discussed above. The loss in toughness is a direct result of the loss in molecular weight represented by the higher MFR number. But the “good” parts are also problematic with an increase in the polymer MFR of 89%. While these parts are better than those that are failing, they still represent a poorly controlled process. Only a robust design or an application environment that does not test the properties of the resin keeps these parts from also failing.

A finding like this often leads to a look back through the history of the part. The essential questions that we are trying to answer with an exercise like this are, “Has the process always been like this?” and “If not, when did it go wrong?” Unfortunately, archiving practices for old production are not always as they should be, and such a review will be impossible to perform if parts have not been retained. However, in some cases we get an opportunity to trace the production history.

ONLY THE RAW MATERIAL IS TRUE
The second case presented such an opportunity and involves a clear polycarbonate part. Here again two sets of parts were submitted for an evaluation after breakage was noted during drop testing of an assembly. An examination of the current molded parts inventory revealed a slight color difference between two lots of parts and the assumption was made that the product with a slight straw-colored cast represented the parts that were failing. Samples from each lot were provided, initially without the raw material. However, the material supplier gave a nominal MFR specification of 10 g/10 min for the grade being used to produce the parts.

Despite the color difference, the MFR results showed that the two sets of parts were essentially identical and in both instances exhibited an unacceptably high MFR. The parts with good color had an MFR of 29.37 g/10 min and the parts with the straw tint measured 29.69 g/10 min. A larger sample of drop tests demonstrated that both sets of parts produced about the same failure rate.

The temptation here is to immediately blame the process. However, without a sample of the raw material, this conclusion is based on the assumption that the raw material is within specification and, just as important, that the correct grade of material was brought to the molding press when the parts were produced.

Subsequently, a sample of raw material was provided and measured at 9.97 g/10 min, essentially right on the nominal value for the grade. With this established, the conclusion was clear that both sets of parts were made from polymer that had been degraded during processing with MFR increases of nearly 200%.

The next step was to go back to first-article product from approximately 18 months earlier. When these parts were tested, they too were found to be degraded, but not as severely. The MFR for the first-article parts was 22.68 g/10 min, an increase over the nominal value for the raw material of 127%. It now appeared that the process may never have been under control, but again the demands on the application did not produce failures until the process had drifted to a particular point.

In this case the molder regrouped, examined all the processing conditions that could contribute to this type of degradation, and produced a sample run of parts with an MFR of 11.02 g/10 min, an increase of only 10%. Table 2 summarizes the findings of the study.

DIMINISHING DESIGNS
Both of these cases show the potential hazards of molding parts without knowledge of how the polymer is being affected by the process. Performance testing is certainly a good screening tool for establishing fitness for use, but it does not guarantee that the processing conditions result in a healthy polymer. If the normal process results in degraded material and the molder is unaware of the condition, it is only a matter of time before the process drifts off target sufficiently to produce failures.

In addition, in both of these cases the part was clearly designed with enough forgiveness to accommodate a material that had not been processed well. Everyone in the plastics industry knosws that this type of overdesign is quickly becoming a thing of the past. In many cases, parts are now being designed at the other end of the spectrum, where the material must possess its optimum level of performance to function reliably.

It is for this reason that a check of MFR on raw material and molded parts should be as much a part of first-article qualification as all the capability studies and inspections that are currently part of most quality assurance procedures. Without satisfactory preservation of molecular weight, none of the other traditional measurements of product quality matter very much.

October , 2006 - Reprinted with permission from Injection Molding Magazine. Copyright © Canon Communications LLC.