The Materials Analyst, Part 2: Brittle PPO
December 4, 1998
This series of articles is designed to help molders understand how a few analyticaltools can help diagnose a part failure problem. Michael Sepe is our analyst and author. Heis the technical director at Dickten & Masch Mfg., a molder of thermoset andthermoplastic materials in Nashotah, WI. He has provided analytical services to materialsuppliers, molders, and end users for the last 10 years. He can be reached at (414)369-5555, Ext. 572.
Poor final properties in a molded part are often caused bydegradation of the polymer during processing. As discussed in the last article (September1997 IMM) this degradation can come from excessive heat and long barrel residence times,or it can be the result of exposure to high levels of moisture. However, in someinstances, both of these conditions can be carefully controlled and a product still doesnot perform as expected.
The discovery of the problem is often madeaccidentally. A person involved in assembling a product will notice that a particular bosscracks with surprising ease. This typically leads to a more quantitative evaluation by thequality control department. Comparison of the new lot of product with retained parts froma previous lot convince the quality and engineering departments that something unusual ishappening, and before long, a sample of the problem product ends up in an analyticallaboratory for testing.
When a molded plastic part suddenly fails toperform as expected, most end users suspect incorrect composition. Therefore, the requestfor analysis usually involves some verification that the material from which the suspectpart is fabricated matches the print specification. Composition testing can be carried outby a number of analytical methods. One of these is a method called differential scanningcalorimetry. For those who spend a lot of time around the lab, it is simply referred to asDSC. DSC has a number of uses, but one of the things it does especially well is measurethe change in energy content or heat content of a material as it is heated or cooled.
Figure 1. A typical glass transition event for polycarbonate, an amorphous material. Figure 2. DSC scan for HDPE, which is a highly crystalline material. Notice the sharpdeparture and return to the baseline marking the melting process. Figure 3. Results of DSC test performed on small fragments of the two good PPO parts. Figure 4. DSC result for the bad PPO part. |
When a material begins to soften or melt, asudden change in the heat content of the material occurs. If the material happens to beamorphous (as in the case of polycarbonate or acrylic), then the most important event thatDSC measures is known as the glass transition.
In amorphous thermoplastics, the temperatureof the glass transition is closely related to the more familiar heat deflectiontemperature, and it can be thought of as the softening temperature for this class ofmaterials. Figure 1 shows a typical glass transition event for polycarbonate. Mostmaterials have a glass transition temperature that is sufficiently unique, and with theproper library of standards, it is usually possible to match up an unknown material withsomething the analyst has seen before.
Materials such as acetal and polypropyleneare semicrystalline. While the glass transition of these materials can be difficult tomeasure by DSC, this second class of materials has something that amorphous plastics donot have. The materials contain crystals. Melting changes the heat content of a plasticdramatically because it takes a significant amount of additional energy.
Figure 2 shows a DSC scan for high-densitypolyethylene, one of the most crystalline thermoplastics. The sharp departure and returnto the baseline marks the melting process. As with the glass transition temperatures, themelting points and the shapes of the curves associated with the melting process aredistinct for each thermoplastic. With a sufficient library of known results, matchingresults from an unknown lab sample can be a relatively simple process.
The sample submitted by our client wassupposed to have been produced from a 30 percent glass-reinforced PPO, familiar to mostpeople in the industry under the name of
Noryl. The color of the material was black.Three parts were sent in; two of the parts appeared to be performing properly. The thirdpart, while normal from a cosmetic standpoint, did appear to be much less resistant to anyapplied stress and broke in a brittle manner when different areas of the part were placedunder even a mild load.
The customer had already done a little ofhis own detective work, burning off all of the plastic from segments of both"good" and "bad" parts to determine whether or not the glass fiberloading had changed. The results showed that the material from both types of partscontained an amount of glass consistent with the normal tolerance range of 28 to 32percent.
DSC tests were run on small fragments of allthree parts. Figure 3 shows the results for one of the two good parts. The scan for theother good part looked virtually identical. Heating the material all the way to 300C(572F) revealed only one event, a glass transition at 141C, right where it is supposed tobe for this particular PPO grade.
Figure 4 shows the DSC result for the badpart. The PPO glass transition appears at approximately the same temperature, 137C. Byitself, this subtle difference would not be sufficient evidence for anything out of theordinary. However, a melting event with a peak at 252C (485F) does not belong to the PPO.
A match with our in-house library revealedthat the most likely contaminant was PET polyester. The most common commercial grades ofPET polyester are made with a glass fiber loading of 30 percent, which provided a possibleexplanation for the similar filler content in the bad part. It also explained the reasonfor a noticeable increase in part weight for the bad product since PET polyestercontaining 30 percent glass has a much greater density than does the PPO.
By comparing the energy associated with themelting point in Figure 4 with that of a pure PET, it was even possible to estimate thedegree of contamination at 15 to 20 percent. The increase in part weight for the bad partswas 6 percent. This provided a simple means of separating the good and bad product withoutrequiring additional lab tests.
This case provides an excellent illustrationof how DSC testing can readily identify contamination in a molded product.
You May Also Like