This series of articles is designed to help molders understand how a few analytical tools can help diagnose a part failure. Michael Sepe is our analyst and author. He is the technical director at Dickten & Masch Mfg., a molder of thermoset and thermoplastic materials in Nashotah, WI. Mike has provided analytical services to material suppliers, molders, and end users for 15-plus years.

Polypropylene the part is not. Polycarbonate it is, according to forensic analysis. Then why does it keep failing?

This case study involves a molded article that is part of a consumer product. Production of parts like this one has been carried out in Asia for some time now. In this particular case, however, the responsibility for production had only recently been shifted from North America to China. A number of field failures involving this component reportedly occurred soon after the product was purchased by the consumer. This led to some product testing with a series of unexpected outcomes.

The subject part was specified in a polypropylene and the initial request was for an examination of composition. Whenever an evaluation of polymer composition is requested and funds to perform the work are limited, the analyst often has to decide between two important techniques: infrared spectroscopy (FTIR) and differential scanning calorimetry (DSC).

FTIR provides a definitive fingerprint based on the presence of specific chemical bonds. Therefore, a nylon can be readily distinguished from a polyester or a polypropylene by this method. But when it comes to determining the difference between materials with the same chemical makeup but different molecular arrangements, FTIR may fall short. For example, FTIR cannot readily distinguish between a PBT and a PET polyester, or between an acetal homopolymer and a copolymer. In these cases, the important feature that separates these materials is the melting point of the polymer, a job well-suited to the capabilities of DSC.

In polypropylenes, DSC can easily determine the difference between homopolymers and copolymers and can further distinguish between random copolymers and block copolymers. In addition, the part in question was green. Whenever colorants are added to polypropylene there is a concern that the dyes or pigments may act as nucleating agents. Nucleation alters the crystalline structure of a polymer, making the molded article more rigid but also less impact resistant. The effect of a colorant on the crystallization rate can be assessed by examining the behavior of the material as it cools from the melt to the solid state.

What Happened to the PP?

In this case, the DSC test revealed a surprise. Figure 1 shows a typical DSC result for a polypropylene homopolymer. Figure 2 shows the result that was obtained for one of the failed parts. You do not need to be a polymer chemist to tell that the part in question is not made from the specified material.

The DSC for the failed part shows a very small melting event at 118°C and a step transition at 142°C. This step represents the glass transition, which in practical terms represents the softening point for an amorphous material. Unfortunately, DSC is not the best technique for identifying amorphous polymers because too many amorphous resins have similar glass transition temperatures. In this case, a search of library results returned two possible matches: a high-heat PPO/HIPS alloy and an impact-modified polycarbonate. FTIR would need to be performed to make a positive identification. Initially, the result was considered an anomaly. But tests on additional failed parts, as well as some nonfailed parts, continued to show the same result. Even original first-article pieces from this particular supplier that had gone through extensive qualification testing showed a similar fingerprint. The only noticeable difference between these initial pieces and the ones that were failing was that the good parts had a slightly higher glass transition temperature (Tg) and did not exhibit the small melting event.

Just as the puzzling over the meaning of these results had begun, the vendor came through with an answer. A certification for the material they were using showed that the resin was a polycarbonate with a nominal melt-flow rate (MFR) of 6.5 g/10 min. This type of information alters the entire course of an investigation. When a polypropylene part fails, the focus is placed on details such as polymer structure (homopolymer vs. copolymer), molecular weight, the presence of nucleating agents, and oxidative stability. Not all of these factors apply to polycarbonate. Furthermore, why would polycarbonate, a strong and tough material, fail in an application originally designed for polypropylene?

The most obvious reason for polycarbonate to fail in a part specified in polypropylene is chemical resistance. A long list of solvents and cleaning agents that do not affect polypropylene will either attack or stress-crack polycarbonate. But even a cursory examination of the failed parts showed no evidence of such mechanisms. The problem appeared to be a case of simple mechanical overload.

Color Concentrate Failure

The answer came, as it so often does, from an MFR test. The first failed part, tested using the appropriate conditions for polycarbonate, produced a value of 55 g/10 min—an increase over the nominal MFR of the raw material of almost 750%! Subsequent tests showed less severe but still alarming levels of polymer degradation.

Based on these results, it would be easy to blame the molder for poor processing technique. But in a world where nothing is certain, the possibility cannot be overlooked that the raw material being used is not within specification. A request went out for a sample of the raw material. When it arrived there was an additional surprise. Instead of containing a sample bag of green polycarbonate, it held two bags. One contained transparent pellets of natural PC; the other contained green color concentrate.

Figure 3 shows a DSC scan for the natural polycarbonate. Note that there are two differences between this result and the scans associated with the failed parts. First, the glass transition temperature is a few degrees higher. Second, the natural polycarbonate did not contain the small melting event detected in the failed parts. In other words, it matched the characteristics of the first-article parts. In addition, the MFR of this raw material was 6.86 g/10 min—well within the correct specification range for this material.

The color concentrate was another story. Figure 4 shows the DSC scan for the green colorant, a result consistent with low-density polyethylene (LDPE). In addition, an MFR test performed on the concentrate at conditions appropriate for polyethylene produced a very high value, indicating that the carrier resin molecular weight was very low. When this colorant was added to the base resin, the MFR of the resulting parts immediately tripled, even with the best of process control.

With a little lack of attention to proper drying, the numbers quickly climbed into the 30s, 40s, and 50s that had characterized the failed product. Even parts that were not failing showed levels of change in the MFR that are technically consistent with polymer degradation. Given the fact that this part was designed to be functional in polypropylene, it should not be surprising that parts in mildly degraded PC functioned properly.

Success in Collaboration

The story behind the change from polypropylene to polycarbonate is as fascinating as unraveling the mystery of the product problem. When the mold used to make the parts was first designed, the mold shrinkage was erroneously figured at .007 in/in instead of .017 in/in. Obviously, all of the critical dimensions were undersized.

With no time to make the extensive tool changes, the vendor recovered by making the parts in polycarbonate. The initial material was a fully compounded resin purchased direct from the material supplier. This ensured that there would be no contamination from a low-cost carrier resin and that the color package selected for the application would be correct. Parts were molded and passed all qualifying tests, even though later analysis would show problems with molecular weight retention. Later, in an effort to recoup some of the expenses associated with running polycarbonate instead of polypropylene, the molder switched to natural with an inexpensive color concentrate. This added enough problems to tip the scales and cause part failure.

The endgame is that in a relatively short time after the problem was first discovered, the OEM, the molder, and the material supplier worked together to recover. The material supplier assisted with selection of a compatible color concentrate and trained the molder’s personnel on proper drying and molding techniques for their material. Subsequent tests showed that the efforts were successful.

This is a cautionary tale for domestic molders who believe they can rely indefinitely on the inferior know-how of foreign competitors. The OEM learned the lesson that trusting the paper trail is not sufficient to guarantee compliance and quality. Occasional ongoing testing of incoming products is an essential part of ensuring that a price reduction truly becomes a cost reduction.

Contact Information Dickten & Masch Mfg. Co. Nashotah, WI Mike Sepe (262) 369-5555, ext. 572 dmlab@execpc.com www.dicktenplastics.com