|This series of articles is designed to help molders understand how a few analytical tools can help diagnose a part failure problem. 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. He has provided analytical services to material suppliers, molders, and end users for the last 11 years. He can be reached at (414) 369-5555, Ext. 572.|
A client called one day with a difficult problem. The company had developed a gear application in a glass fiber-reinforced PPS compound that also contained some PTFE (Teflon) to lower the coefficient of friction and improve wear resistance. First parts had been made, measured, and run through a series of qualification tests. Everything appeared to be satisfactory, and production was started.
After three runs of product spanning a period of four months, none of the production parts were working as anticipated. All the parts made after the initial sampling were failing due to abrasive wear. As is very often the case in these situations, no raw material was available. However, molded parts from the original sampling and all three production dates were sent in for analysis. Accompanying the parts was the obligatory specification sheet for the material. It was listed as a 30 percent glass-reinforced PPS with an unspecified amount of Teflon.
Typically, the biggest mistake we see with materials like PPS is the failure to use a hot mold during processing. PPS is one of those semi-crystalline materials that need a lot of help to crystallize properly. If the temperature of the material drops below a certain temperature too quickly, the polymer cannot arrange itself properly into crystals.
This critical temperature for PPS is 275F, and therein lies the problem. Most molders use water to regulate the temperature of their molds, and obviously water turns to steam somewhere near 212F. Even high-pressure systems will top out at about 250F. Consistent steel temperatures above 220 to 230F are very difficult to achieve with water, and molders must rely on electric or oil heat to achieve the mold temperatures required for proper crystallization. Many molders just do not want to contend with the higher cost of the alternate heating methods and the perceived safety problems of operating at such high temperatures. So when they get their first PPS job, they hook up the water heaters, turn them up all the way, and hope for the best.
The best is not what they get. Without the proper crystallinity, PPS will lack strength and stiffness, particularly at high temperatures. It will also lack chemical resistance, dimensional stability, creep resistance, and, yes, even wear resistance.
|Figure 1. DSC comparing PPS molded in a cold mold versus a heated mold.
Finding the Problem
It is quite easy to spot this problem. If PPS does not crystallize completely during the molding process, it will do so in the solid state the first time it is heated above 275F (135C). This solid-state crystallization is easily detected by a technique we have discussed extensively known as DSC. Figure 1 shows a DSC test result for a PPS product molded in a cold mold and in a properly heated mold. Note the part produced in the cold mold, set at about 190F, undergoes a significant recrystallization as it is heated. The properly molded part reveals no such rearrangement. We decided to run DSC on all four samples, certain we would find differences in the degree of crystallinity.
|Figure 2. DSC on one set of failed gears showing melting points of the two components.
We were wrong. Figure 2 shows the results of the DSC test for one of the failed parts. While there is a very slight deviation from the ideal baseline as the material approaches 100C, it is far too small to be significant. In any case, all four parts gave the same result.
Some molders have learned the trick of eliminating the secondary crystallization by running a cold mold and then annealing the parts in an oven for several hours afterwards. However, parts produced in this manner will tend to have a mottled, grainy appearance out of the mold that the annealing process cannot improve. These parts all looked excellent. The molder had done a good job, and crystallization was not the problem. However, a close look at the DSC results for all four parts gave us additional information that did point to the problem.
|Figure 3. DSC showing heating and cooling of PPS gear.
PPS and Teflon are both semi-crystalline polymers with well-defined melting points. PPS typically melts between about 280 and 285C and Teflon has a very sharp melting point at 327C. When the materials are cooled from the melt, they provide equally distinct recrystallization events: Teflon at 316C and PPS at 235C. Figure 3 shows the DSC of the material during first heating and during cooling. The area under the curve associated with a melting point is governed by the degree of crystallinity in the polymer. Assuming the crystallinity of a material is consistent, this area under the curve can be used to estimate the relative concentration of a polymer in a mixture. It is not advisable to use the data from the initial heating because the molding process can introduce slight variations in crystallinity that can be mistaken for differences in concentration.
Once the thermal history of the material has been erased by melting the sample, the cooling cycle will give reliable numbers that can be used to compare composition. When we compared the areas under the curves for the good product with that of the bad product, we saw a difference right away.
|Figure 4. Cooling DSC showing good and bad gear.
Figure 4 compares the good product to one of the bad parts. Visually, it is easy to see that in the good material the Teflon peak is much stronger relative to the PPS peak. Without samples of the pure polymers, it is not possible to produce exact percentages for each material in the mixture. But it is possible to come up with relative ratios. Table 1 shows the results.
The results were clear. The amount of Teflon in the original samples was much higher than in any of the production batches. This would be expected to significantly influence the wear characteristics of the material and make it more likely to fail at the lesser amount.
To double-check our results, we used the TGA technique as verification. These tests allowed us to examine the glass fiber content in each sample. We were also able to obtain additional data on the relative amounts of Teflon and PPS present in each sample.
This is how it works: PPS and Teflon are both very stable materials. When heated to very high temperatures, they both decompose at approximately the same temperature. So the concentration of each material cannot be determined by weight loss. However, if we heat the materials in a relatively inert atmosphere such as nitrogen gas, PPS and Teflon do behave differently. PPS, as it decomposes, produces a large amount of char, carbon that can later be burned away in air. In fact, PPS produces so much char that for every 1 percent of polymer lost in nitrogen another 1 percent burns away later as char. Teflon, on the other hand, produces no char at all. Therefore, if the amounts of PPS and Teflon were varying, the ratio of char to total polymer should change. More char would indicate a higher amount of PPS; less char would signal the presence of more Teflon.
The first thing we discovered was the glass fiber content was not very consistent. The good sample contained 32.25 percent glass which was close to the target of 30 percent. The bad parts varied from 33.77 to 36.04 percent. Once this was deducted from the total mass of the sample, the total resin content was compared to the char. Table 2 shows the results.
|Figure 5. TGA comparing weight loss in good and bad gear samples.
Remember, a higher resin-to-carbon ratio means more Teflon. Once again, the original material used to produce the first samples stands out from the rest of the batches. The TGA tests confirmed the Teflon concentration had dropped between the time the original samples were made and the time production began. While the production lots were quite comparable to each other, they represented a distinct departure from the original. Figure 5 shows a comparison of the weight loss process for the good product and one of the bad parts.
From this analysis, we were able to show the problem with part performance was related to material composition and not to processing. In addition, finding high glass contents also helped to explain part of the problem with wear resistance. While glass fiber improves strength and stiffness and therefore increases the force limits of an application, the glass also increases the tendency for abrasive wear. Additives like Teflon and silicone fluid are incorporated to counteract this negative effect. By increasing the glass content while at the same time reducing the Teflon content, the product was being penalized twice.