Nonsensical Plastic Part Analysis

By Michael Sepe, Materials Analyst

To achieve the right result from a material test, it’s important to start with the right test.

Performing analytical testing to arrive at a root cause for a manufacturing or performance problem is a lot like medical practice. When a patient comes in with a complaint, the practitioner asks questions to determine symptoms and then uses a combination of training, experience, and intuition to decide on a course of testing to determine the cause for the malaise. Sometimes, the first path turns out to be a dead end; things that might be wrong turn out to be in line. These items are checked off the list so that additional work can be done to find the problem.

Often the first tests do not result in a root cause, but they suggest a new course of inquiry that ends up providing the answer. When a project is complete, hindsight sometimes shows that fewer tests might have been required, but an analyst should always be able to point to sound reasons for conducting the tests that he or she ran based on sound fundamentals. Unfortunately, some testing is done that does not adhere to these fundamentals. In this article, we look at three of these.

Thermogravimetric Analysis
When brittle parts are submitted for analysis, one of the possible culprits is the presence of excessive moisture in the raw material during processing. This is particularly detrimental to certain classes of condensation polymers such as polyesters, polycarbonates, polyamides, and polyurethanes, where the presence of water in the raw material rapidly results in polymer degradation at typical melt temperatures.

A brittle condition in a polymer should always focus early on this possibility of degradation with some type of viscosity measurement that will provide an assessment of average molecular weight. On occasion, a more sophisticated evaluation of molecular weight distribution using gel permeation chromatography (GPC) may be called for. However, sometimes analysts attempt to show that moisture was present in the raw material at the time of processing by measuring the moisture content of the material in the failed part. This is often done using thermogravimetric analysis (TGA).

This approach is fundamentally incorrect on several levels. First, TGA, while a very good test for providing information on composition and thermal stability, is a weight-loss technique. It monitors changes in the mass of a sample as a function of temperature, time, and atmosphere. Normally it is used to decompose a sample in a controlled environment in order to examine the weight losses associated with organic constituents and to measure the level of noncombustibles like glass fiber, mineral fillers, and inorganic pigments like titanium dioxide.

In the early stages of heating, any absorbed moisture in the sample will certainly be driven off; however, it will be indistinguishable from other volatile material in the compound such as lubricants, antioxidants, and other additives. The TGA is not capable of determining how much of the lost weight is actually attributable to moisture unless the gases coming off at this stage of the test are directly routed to other instruments capable of separating the various substances, identifying them, and quantifying them.

This step is seldom undertaken. The fact that a modern TGA instrument can cost $75,000-$100,000 does not make it any more precise in this respect than a loss-in-weight moisture analyzer that costs $9000. It is fundamentally unsuited for this task.

But even if it were suitably accurate, there is another problem with this approach. The moisture content in the part at a point days, weeks, or months removed from the molding date has nothing to do with the moisture content of the raw material at the time of molding. Any hygroscopic material begins to absorb moisture from the atmosphere as soon as the molded part falls from the machine. The longer the part is exposed to the environment, the farther the moisture content will drift from whatever it was at the time the part was produced. The only chance that an analyst might have to infer the approximate moisture content of a raw material based on the moisture content of the molded part would be at the moment of manufacturing. And even then, a part molded of wet material will not contain all of the moisture that was in the pellets because some of that water will have been consumed in the reaction that resulted in the hydrolytic degradation.

Figure 1 shows a TGA scan of a polycarbonate part that indicates a weight loss of .96% prior to the onset of polymer decomposition. The analyst interpreted this as evidence of processing with wet material. In reality, there is no way of showing that this weight loss is due to moisture. In addition, the part had been sitting in a humid warehouse for six months. So although some of the observed weight loss was almost certainly due to water, there is no connection between that moisture content and the moisture content at the time of molding.

Differential Scanning Calorimetry
The second common mistake involves determining the degree of crystallinity in a molded part. Parts molded from semicrystalline polymers will achieve a certain degree of crystallinity depending upon the composition of the polymer and the molding conditions. Some polymers, like high-density polyethylene (HDPE) and acetal, naturally crystallize more efficiently than PET polyester or PEEK.

Cooling rate also has an influence on the crystallinity of the molded part. The crystallinity of a material can be established, at least on a relative scale, by heating a sample in a differential scanning calorimeter (DSC). As the polymer approaches the crystalline melting point, a plot of heat flow vs. temperature provided by the DSC deviates from its baseline. The area under the curve for the resulting endotherm is a measure of the energy required to melt the crystal structure and is related to the crystallinity of the material in the molded part.

Many analysts are accustomed to putting the sample through a thermal preparation that involves heating the material through its melting point at a controlled rate, cooling it at the same rate, and then heating it a second time. They use the second heating as an illustration of the inherent properties of the material.

While this may be true, it says nothing about the state of the molded part because the first heat melting process erases the crystal structure created by the molding process. You are now starting over, and the crystallinity obtained on second heat is the result of the slow and controlled cooling in the instrument. It seldom is the same as the state of the polymer in the molded part, which is cooled much more rapidly.

Figures 2-4 show first heat, cooldown, and second heat results for HDPE. All three of these thermograms provide potentially important information about the state of the polymer, and a good DSC analysis should almost always include all three data sets. However, only the first heat scan in Figure 2 tells us anything about the polymer structure in the as-molded part.

Another bewildering practice involves grinding up the molded part in question, molding standard test specimens from the regrind, and taking the DSC sample from one of the newly molded specimens. The test result may reveal something about the crystallinity of the standard test specimen, but it says nothing about the state of the polymer in the original part.

Melt-flow-rate Test
Finally, we have the simple melt-flow-rate (MFR) test. The perceived value of the MFR test has risen and fallen through the years. Molders initially tried to use it as a predictor of processibility, which ignores the fundamental relationship between viscosity and shear rate in polymers. When many molders noted that they could not connect processing conditions to the MFR, they discarded their instruments. However, material suppliers have always known that the principle value of the MFR result was its connection to the average molecular weight of the polymer.

This in turn is a key determinant of physical performance, particularly for those properties associated with toughness.

Therefore, MFR is a quality control tool for the material supplier. The molder can use the test to verify that the material it receives from the supplier is within specification. It can also use this gauge to compare the raw material to the molded parts it produces.

The desired relationship between the MFR of the raw material and the molded parts has been covered in previous articles. The key point here is that sample preparation is important to good results. The most important consideration in this sample preparation is proper drying. If a material is not hygroscopic, then this is not a necessary step. Polyethylenes, polypropylenes, and polystyrenes do not absorb appreciable amounts of moisture and therefore they are not affected by this parameter. However, a polycarbonate or a nylon, if not properly dried, will do the same thing in the melt-flow tester that it does in the cylinder of the injection molding machine: It will foam and ultimately degrade. This will play havoc with accurate measurements of MFR.

Since MFR is often the test that determines the occurrence of polymer degradation, it is important that samples be dried appropriately. Ideally, the samples will even be tested for moisture content using a moisture-specific analyzer before the MFR test is conducted.

Table 1 shows the results of MFR tests conducted on polycarbonate raw materials and brittle molded parts. One set of results was obtained with samples tested without drying and the other set was obtained from dried specimens. The results for the raw material were similar because the raw material was tested soon after it was removed from the drying hopper. But the molded parts had been in the field for a few months and had the opportunity to absorb a considerable amount of moisture from the surrounding atmosphere.

The initial tests obtained on the undried parts show more than a 100% increase in MFR from pellets to parts and point to poor processing by the molder. The new tests on dried parts show only a 14% increase and focused attention on other possible root causes. Ultimately, the cause of brittle behavior was related to composition, but those tests might never have been conducted if the initial MFR tests had been allowed to stand.

So when you go out to obtain testing services, provide as much information as you can to your analyst about what ails your parts; material type, mold construction, process conditions, and field application environment are all important.

But then also ask a lot of questions about what is going to be done and why. To the uninitiated, testing terminology looks like a bewildering alphabet soup, just like in the medical profession. But you are more likely to know what you are getting into, and to get what you want, if you insist on a two-way dialogue with your analytical services provider.

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