The Materials Analyst, Part 13: It really is the molecular weight

On two occasions in this series, we have looked at the use of the melt flow test to detect degradation in a molded part. The technique consists of comparing the melt flow rate of pellets to the melt flow rate of parts. Molecular weight and melt viscosity are related; as one increases the other increases.

Molecular Weight and Melt Flow Rate
Because melt flow rate has a reciprocal relationship to melt viscosity, the relationship of molecular weight to melt flow rate (MFR) is an inverse one; as the MFR increases, the molecular weight drops. In the pellet-to-part comparison, a small change in the MFR indicates that the process preserved the molecular weight at a satisfactory level. Large changes, however, indicate that the average polymer chain length has been reduced to a point where performance may begin to suffer.

Adequate molecular weight is a fundamental property of polymers. Polymers are, by definition, made up of large molecules. More than anything else, it is this property that defines the behavior of the polymers that are the main ingredient in our plastic materials. The extended chains allow for a molecular interrelationship that is unique to polymers: chain entanglement. This entanglement provides the capability to sustain high levels of stress without failure.

Consider a rod of polyethylene. The rod can be bent into any number of shapes, even tied into a knot if it is thin enough, without producing a break in the rod. Now try the same thing with a candle. The candle snaps abruptly with almost no deflection. Chemically, the material making up both the polyethylene rod and the candle is identical. The distinguishing property is chain entanglement, a result of high molecular weight.

In stressing the candle, we need only disrupt a small number of links between the short molecules. To break the polyethylene rod, we need to disturb hundreds of these relationships, and the energy transfer throughout the matrix is very efficiently managed. The first property to suffer when molecular weight declines is the ability to elongate. We usually perceive this as brittle behavior, and it is most easily detected using an impact test.

So the concern over molecular weight preservation is appropriate. A material with the inherent ability to perform in an application may be rendered worthless if, during processing, the molecular weight is reduced by too great a degree. How much is too much? Over the years, many empirical studies of failed products have helped material suppliers come up with guidelines, using the melt flow rate test as a relative indicator of molecular weight preservation. If the MFR of a part molded in an unfilled material increases by no more than 30 to 40 percent from the MFR of the pellets, the processor is considered to have done a good job of taking care of the material. Once the shift exceeds 40 percent, the processor becomes the focus of any discussions regarding part failures.

In discovering the root cause of a performance problem, it is important to have raw material from the lot actually used to make the parts. If a material supplier produces a polycarbonate to a release specification of 10 to 14 and a molded part comes back from the field at 17, no conclusion can be drawn without specific lot data. If the pellets are a 10 MFR, then a 17 constitutes an unacceptable 70 percent increase. But if the pellets are a 14 MFR, the parts represent only a 20 percent increase. It is also important to determine when the degradation took place. If the part has been in the field for five years and has experienced the rigors of an aggressive environment, the melt flow test may tell us the polymer has been damaged, but it will not tell us the root cause.

Filled vs. Unfilled
While the guidelines are straightforward for unfilled materials, they become less exact when we attempt to apply the same techniques to filled compounds. Most filled materials contain glass fiber. Breakage of these fibers is an unavoidable consequence of injection molding. In the pellet state, the fiber length may approach the diameter of the orifice in the melt flow rate test apparatus. Under the low-shear conditions of the melt flow test, there will be little fiber alignment, and the glass will significantly increase the measured MFR of the material.

After molding, the average fiber length is usually reduced by 35 to 50 percent, and the material will flow more easily through the melt flow tester independent of any changes in the polymer. For this reason, the allowances for glass-filled materials must be more liberal; the more filler present the more viscosity may shift before signaling a problem.

The best rules are those relating viscosity changes directly to the performance of specific parts. Good design can increase these allowances, and poor design will reduce them. However, the rule of thumb is generally that a 10 to 15 percent glass-filled material can sustain a doubling of the pellet MFR without unduly compromising its performance. Materials at the high end of 30 to 50 percent filler may be able to undergo a tripling or even a quadrupling of the pellet MFR before trouble sets in. In these materials, the melt flow rate shift is a composite response of the polymer and the filler, and the greater the filler content the greater the allowance must be for changes in viscosity. Table 1 summarizes these "rules of thumb."

If reductions in molecular weight during processing are of concern to part performance, it is reasonable to assume orchestrated reductions in the form of easier flowing virgin materials would also bring some penalty in properties. Anyone who has worked with materials that are graded primarily according to flow rate--such as polypropylene, polystyrene, or polycarbonate--knows this to be the case. Even though the notched Izod values won't show it, a polycarbonate with an MFR of 5 will have better impact properties than a grade with an MFR of 15.

However, the trends toward greater part complexity, higher cavitation, and thinner walls all drive the processor to seek lower viscosity materials in order to improve processing. Often, these intentional changes are the difference between a part that succeeds and one that fails. Through no fault of the processor, the molecular weight of the product has been reduced, and the decreased toughness shows up in the field as an increased failure rate.

TABLE 1

Field Failures
Four recent field failures we have worked on stem from the simple problem of using a lower molecular weight material. In the first case, a thin-walled part in high-density polyethylene was being made in a 25 MFR material using a 16-cavity mold. To reduce part cost, a new 32-cavity mold was constructed. The additional flow length made the use of the 25-melt material difficult, and the application was changed to use a 40 MFR grade. Within a few months, scattered field failures began to come back. When the failed parts were analyzed, they had a melt flow of 44.11 based on a virgin material with an MFR of 39.37, a shift of only 11 percent. Using our guidelines, we concluded that the processor was taking good care of the material. But the initial melt flow rate was simply too high to ensure good performance.

In the second case, a molder specified a 40 percent calcium-carbonate-filled polypropylene for an application with a good potential for incidental abuse. Often with filled materials, the melt flow specification becomes secondary to other considerations such as filler content and type. The listed MFR range for the grade being used was 8 to 12. The molded parts were placed in outdoor applications, and failures began to occur regularly in situations where the temperature dropped below 50F. One batch of parts, however, performed extremely well. Even an unscientific comparison of a good and a bad sample showed the good part was much tougher and broke in a more ductile manner while the bad part provided less resistance and failed in a brittle fashion.

No differences were found except in the melt flow rate. The bad part had a melt flow of 15.22 based on a batch of material with an MFR of 10. While this is a 52 percent shift, the high filler level accounts for some of this change, and the process was considered to be appropriate. The good parts, however, checked out at 4.07. Somehow, a batch of material had been shipped that was out of specification. In this case, it turned out to be the key to solving the problem. The higher molecular weight material was the only material to produce good product. It is interesting to note that the processing people had not noticed any significant changes in their machine settings when they processed the "stiffer" material.

In the third case, a 10 percent glass-reinforced polycarbonate was being produced without a stated melt flow specification. The molded parts had to be able to withstand a considerable amount of incidental abuse. According to notched Izod tests, adding glass fiber to polycarbonate drastically reduces the toughness of the material. However, instrumented falling dart tests have shown, at the low loading of 10 percent, it is still possible to make a product with about 50 percent of the energy-absorbing characteristics of the unfilled material. This performance is very dependent upon the molecular weight of the material.

In this case, most of the initial lots fell in a very tight melt flow range of 7 to 9. Occasionally, batches showed up that tested at 16. With no melt flow specification, no flag signalled the possibility of a problem lot. In addition, no study had been performed to determine if this variation constituted a quality problem. The molder had the material tested by falling dart impact and found the total energy at break fell from 14 to 3.7 ft-lb. Most of the decline occurred when the MFR climbed above 12. This illustrates how quickly the properties can deteriorate as the melt flow rate increases.

The final example is our own. We started using a special color-stable grade of a PET polyester for a high-heat application. Although it is customary for us to screen materials using the melt flow rate test, we skip this test on some highly reinforced materials, particularly when the supplier uses an alternative method such as a high-shear capillary rheometer test. Because the color stable formulation was in every way the same as the standard 30 percent glass-filled material, we assumed the MFR would be in the same range of 7 to 11.

When isolated broken parts came back from the field, we performed the melt flow test. For 30 percent glass-filled PET, guidelines have been developed stating if the melt flow rate of the molded part increases by a factor of four from pellet to part, the processing is acceptable. Therefore, when we tested the returned parts, we were expecting a number no higher than 40. To our surprise and dismay, the values were in the high 80s. The immediate conclusion was the material was degraded. We had not yet checked the pellets. We were surprised again to find this material was being made to a flow rate of 25 to 30, and our particular batch was 27--not the 7 to 11 range we expected.

In none of the above cases did the processor abuse the material and create a quality problem by reducing the molecular weight in processing. However, either knowingly or unknowingly, the molecular weight of the incoming material had been altered, which resulted in a dramatic change in part performance. While the melt flow test is only an approximation, it has real value in providing guidelines for good performance and is a simple test that can be performed rapidly on an inexpensive piece of equipment, ideal for the injection molding industry.