The Materials Analyst, Part 42: Combining techniques to find contamination

Last month we highlighted an application where a semicrystalline material had contaminated parts molded in PPO, causing a problem with a hot plate welding assembly process. Finding semicrystalline contaminants in an amorphous polymer is relatively simple because semicrystalline materials have melting points that dominate a DSC test. Amorphous resins show only a glass transition, an event that is much less energetic. 

But what happens when there is contamination of one semicrystalline material by another semicrystalline polymer and the two materials have similar melting points? In cases like this more work is often needed, particularly if the client wants to know the exact source of the contamination. Multiple techniques can complement one another in reaching a final answer. 

In this case the client was a user of fans for small motors. The fans were molded in a 33 percent glass-filled nylon 6/6, but a series of contaminations with nylon 6 had plagued the product line with failures. These had been readily identified by DSC since nylon 6 and nylon 6/6 have melting points that are about 40° C apart. A new problem involved failures in a nylon 6/6 that contained an impact modifier along with the glass fiber. Again the failure mode was essentially brittle behavior; fins would break away from the main body of the fan. The first suspicion was that the parts had been molded in a grade that did not contain the impact modifier. 

First Clue: Heavy Char 
We began with thermogravimetric analysis (TGA), since this is often one of the easiest ways to detect an impact modifier in a nylon product. Figure 1 shows the result of that test. This graph points out the importance of plotting both the weight loss and the weight loss rate as a function of temperature. If we look only at the weight loss process (see gray bar), there is no clue that the material is not homogeneous; the eye cannot pick up the small change in the rate of decomposition. But the derivative shows that there are two overlapping steps to the polymer breakdown. The first one is associated with the nylon while the second one shows the presence of a material with a higher degree of thermal stability. 

Decompositions in this temperature region are usually associated with materials based on polyethylene. Impact modifiers for nylon are frequently based on modified polyethylenes or elastomers with a similar chemistry. So TGA gives us at least an indication that the impact modifier is present. TGA also gives us the filler content at 34.7 percent. For a material with a nominal filler content of 33 percent this is certainly within reason. 

But this test also suggests that there is a problem. When most nylons are decomposed in nitrogen they form a small amount of char that can then be burned off in air. Because of the chemical structure of nylon 6/6, this char amount is typically near 2 percent. Even a grade heavily loaded with carbon black will seldom produce more than 3 percent. This sample showed more than 5 percent, which indicated the possible presence of another polymer. 

Recrystallization Discrepancies 
The DSC test was next. Figure 2 shows the result for initial heating in the DSC. For contrast, a DSC scan for a known nylon 6/6 with the same glass content is also shown. There is an obvious difference here. The suspect part has a broadened melting point with a shoulder at 254C, just 7° C below the main melting point for the nylon 6/6. Unfortunately, this is not conclusive; nylon 6/6 can exhibit a lower secondary melting point when it is cooled slowly. Results from reheating a slow-cooled nylon part and the suspect part revealed this fact. While the fingerprint for the good material was not precisely the same, it was close enough to our failed part to be ambiguous. 

Nevertheless, the circumstantial evidence was starting to add up. The TGA gives an indication that a second polymer may be present. And because of the higher char content, this second polymer would have to have certain characteristics. It would have to degrade at the same temperature as the nylon and melt near 254C, making it very difficult to separate from the nylon 6/6. But it would also have to contain structural elements in the polymer chain backbone known as aromatic rings. Chemists also sometimes refer to these constituents as benzene rings or phenyl groups. A search of general information about other commercial polymers shows that PET polyester fits this profile. 

This is where a full use of the capabilities of the DSC is so important. When two polymers have similar melting points, they seldom have the same recrystallization temperature. Therefore, a lot can be learned by monitoring the DSC test through a controlled cooling run. 

Figure 3 shows the result for this phase of the test, and there is a substantial amount of information here. First, we see the recrystallization of the nylon 6/6 in the peak at 224C. Next, we see another strong recrystallization at 208.5C. This is not related to the nylon 6/6, but it corresponds perfectly to the recrystallization temperature for certain PET polyesters. We also get a bonus from the cooling run. Near the end of this segment of the test we see a broad and weak recrystallization with a peak at 42.5C. This is a signature for the impact modifier that we had first detected in the TGA. 

TGA+DSC+IR=PET 
At this point we had several pieces of information that pointed to the presence of PET polyester. However, in many cases like this there is a desire for ironclad proof. As persuasive as all of this work might be, there was no positive identification of the contaminant. Creative interpretation could develop alternate theories, and in cases where the problem-solving process becomes adversarial, alternate theories are plentiful. 

This is where infrared spectroscopy is invaluable. Last month we ended up with an ambiguous identification of the contaminant in the PPO based on a melting point—it could have been PBT polyester or nylon 6. In this month's analysis, the client had a need to know the exact culprit; it was critical to tracing the problem. 

Figure 4 shows the result of the infrared work. The top spectrum belongs to our troublesome fan blade. Beneath that are library spectra for PET polyester and nylon 6/6. While the peaks for the nylon are clearly present in our part, there are additional peaks that do not belong there. Those extra bands are accounted for by the polyester, particularly the ones highlighted near 1750 cm^-1 and 1250 cm^-1. When these two spectra are added, the result is a perfect match with our sample. This closes the loop on the results that the DSC and the TGA uncovered. 

It is important to note that the infrared result by itself would not have been sufficient to identify either the primary polymer or the contaminant. This is because most nylon materials look the same by infrared. Nylon 6, 6/12, 6/6, 6/10, and 4/6 all provide virtually the same IR fingerprint. By the same token, PBT and PET polyester produce the same infrared scan. This is where the melting points and recrystallization temperatures are indispensable in finishing the picture, illustrating the complementary nature of these techniques.