Sponsored By

The Materials Analyst, Part 23:Stress cracking: How to avoid this killer (part 2)

August 15, 1999

12 Min Read
The Materials Analyst, Part 23:Stress cracking: How to avoid this killer (part 2)

In last month’s article we defined the terms environmental stress cracking and solvent stress cracking, briefly explained the mechanisms involved, and provided a couple of examples of actual case studies. This month we want to conclude the discussion with a few more case studies and then discuss tests and material selection techniques that can keep us out of trouble with this challenging problem.

One of the most difficult things to explain to a client who has a stress cracking problem is why every part does not fail. However, if we keep in mind that both a certain stress level and a certain degree of chemical exposure are needed to produce the effect, the variability becomes easier to understand. This principle is illustrated in our next case study.

The problem involved some handsets for telephones molded in ABS. Two small threaded brass nuts were molded into the part, and after the customer received the parts he noticed cracks that appeared to be emanating from the area around the inserts. The problems appeared in some lots but not in others.

We were baffled until we started to ask the client questions about their assembly process. We discovered that the end user of this handset was very particular about the appearance of any marks such as fingerprint smudges. So, to clean the surface, all of the parts were wiped down with a solution that was a mixture of ethyl alcohol and ethyl acetate.

Ethyl alcohol will not cause any problems for ABS, but ethyl acetate can be very aggressive. This still did not explain the lot-to-lot variability until we obtained a specification for the cleaning solution. It turned out that the solution could vary in composition from 75 percent ethyl alcohol/25 percent ethyl acetate to 25 percent ethyl alcohol/75 percent ethyl acetate. When the solution was predominantly ethyl alcohol, the parts could withstand the cleaning process. But when the ethyl acetate concentration began to increase, the combination of the solvent and the molded-in stress produced by the shrinkage around the brass inserts was enough to produce cracking.

Detecting Stress
This instance leads us to a discussion of one of the methods that can be used to test for the presence of stresses in a part that might lead to stress cracking. In transparent parts, the easiest way to detect stress is by viewing the part with polarized light. Stresses appear as birefringence, a condition manifesting as a series of multicolored bands. Qualitative evaluations can be made visually, but a variety of more sophisticated instruments allows the measurements to be more quantitative. But in many parts, geometry makes such viewing difficult. More importantly, the addition of modifiers such as colorants, fillers, reinforcements, and tougheners render many parts opaque.

In these cases, one of the most effective ways to evaluate a part is through a technique known as the mixed solvent method. In this test, two materials are mixed in varying concentrations. One of the ingredients is relatively inert to the polymer of interest while the other one is meant to serve as the stress crack agent. The idea is to start with a solution that is weak in the stress crack agent and move progressively to increasingly aggressive mixtures. A part that can stand a relatively high level of stress crack agent is considered to be relatively stress free. But if a part fails while the stress crack agent concentration is still quite low, then the part will likely have problems in the field.

For some materials such as polycarbonate and polysulfone, researchers have even been able to calculate the approximate stress level associated with failures in certain mixtures.

Without realizing it, this type of test was being conducted in a haphazard way by our telephone handset client. Sometimes the operators would use a mild mixture that was mostly ethanol with a small amount of ethyl acetate and the parts would work well. At other times, the ethyl acetate concentration would be at the high end of the tolerance, and the parts would fail. Many people find it surprising that a simple wiping of the plastic part is sufficient to produce the problem, but the mixed solvent technique usually involves relatively brief exposure times. Typically, a 90-second dip followed by a waiting period of a few minutes is all that is necessary to reveal the problem. If cracks appear, they characteristically will show up in the traditional areas of high stress: near gates, at sharp corners, in areas of suddenly changing wall thickness, and along knitlines.

Elusive Solution
The next two case studies show how the mixed solvent method uncovered unexpected problems related to stress cracking. The first involved some polycarbonate tubes. These tubes were gated at one end with a ring gate that was machined away after molding. At some point during secondary operations the gate end of the part was swabbed with a mixture of 1 percent isopropyl alcohol and 99 percent water. Stress cracks would occur occasionally in this area. Our client’s managers were looking for a cause and a possible solution.

They kept unusually good records and had saved retains from several runs going back five years. They sent us samples from a variety of these runs and asked us to evaluate the parts.

The mixed solvent test for polycarbonate is well developed. It involves varying concentrations of n-propyl alcohol as the inert ingredient and toluene as the stress crack agent. The mildest mixture is a 10:1 mix of n-propanol and toluene. This is followed by a 5:1 blend. And if the parts are good enough to withstand that, a 3:1 mixture is applied. The exposure time is 3 minutes. If problems are going to occur, they usually appear within minutes of removing the part from the solution, although it is prudent to check the parts periodically for at least an hour to detect cracks that may develop slowly.

Now you may be wondering to yourself, if we use the n-propyl alcohol as the inert agent in the stress crack solution, how in the world is a mixture of 1 percent alcohol and 99 percent water causing the parts to stress crack? Well, that’s what we wanted to know. We tested all of the various lots that were submitted. Every lot but one passed the exposure to the first solution.

Normally this is sufficient to expect that field performance will be satisfactory. Most of the sample lots made it past the second solution and one group even passed the most aggressive solvent mixture. But what about this one lot that did not make it past the first screening? What was different about it?

Well, all of the parts except the ones from this lot were submitted as molded. But this one lot had been decorated with a black ink designed to indicate the fluid level in the tube. And the failures in this group were remarkable. We had received four parts. One part went into each of the three solutions and promptly failed. The fourth part was never removed from the bag that it came in. But when the failed parts were returned to the bag, the fourth part also failed just from the casual contact with the other three parts. It was the ink! A check of the composition revealed a relatively high level of an aromatic petroleum distillate.

The Right Material
A second case study involved a roller bearing housing made of high impact polystyrene. It was gated at one end and the material had to flow approximately 8 inches to fill the entire part. At the end of the part was a cylindrical detail into which a bearing was pressed as a secondary assembly operation. This assembly was conducted at the end user’s plant and was therefore not done until days or even weeks after the part was molded.

A significant number of parts developed cracks within days or weeks of having the bearing inserted. At first it looked like a straightforward case of weldline failure. Because of the flow path, two weldlines were present on the diameter that was accepting the bearing and the failure appeared to propagate along these weldlines.

But there was a problem. First, the failure always occurred in a delayed fashion. No one ever saw a part crack outright at the moment the bearing was inserted. Second, every part we received for review had a greasy residue on the polystyrene in the vicinity of the bearing. The bearing manufacturer performed a finite element analysis on the region of the part where the bearing was inserted and showed that the peak stresses were very close to the yield stress of the polystyrene. That was the stress part of the equation. The solvent appeared to come from the grease. The identification of the mechanism as stress cracking was confirmed by some photographs of the crack region taken at very high magnification. Two of these are shown in Figures 1 and 2.

Figure 1 shows the crack magnified by a factor of 20. A close look at this shows that the crack does not start at the tip of the weldline at the end of the part. Instead, the crack is actually discontinuous and stops in several places and it does not even extend to the end of the part where we would expect the weldline to be the weakest. This indicates that the failure is starting in several places simultaneously and that the cracks actually start on the outside diameter where the stresses are the highest and propagate inward from multiple points of origin.

Figure 2 provides a view of the surface exposed by the crack magnified by a factor of 200. The top of the photograph is the outside diameter of the part and the multiple cracks travel in parallel lines as the failures move from the outer to the inner diameter. The lack of any evidence of chemical attack on the crack surface confirms that this is not simple solvent damage. Once the stress is relieved by the initiation of the crack, the material does not undergo any further chemical attack.

Two solutions were posed to this problem. The first option was to change to a filled polypropylene. Polypropylene is semicrystalline and therefore has much better chemical resistance than polystyrene. It is much less likely to stress crack in the presence of the grease. The use of fillers is primarily designed to match the shrinkage of the amorphous polystyrene. Although this was not actually implemented in production, tests run on samples indicated that this material change would solve the problem.

The second fix was to move up the amorphous polymer performance ladder to ABS. Since the ABS could still encounter problems from the grease, we also wanted to reduce the stress component from the bearing insertion. So we ran some parts in ABS and inserted the bearings after the parts had a chance to cool. Other bearings were installed as soon as the parts were out of the mold. The part was then allowed to shrink around the bearing. Now the question was, which method developed less stress on the plastic? For this we turned to a mixed solvent test.

Stress Crack Resistance
One of the best chemicals for checking stress crack resistance in ABS is concentrated acetic acid, called glacial acetic acid by chemists. It is very aggressive and usually produces cracks in all but the most stress-free ABS parts. Vinegar, on the other hand, is 5 percent acetic acid and the rest is mostly water. It rarely causes stress cracks in ABS moldings. So our mixed solvents are glacial acetic acid and water. We used 95 percent, 65 percent, 35 percent, and 5 percent solutions of glacial acetic acid to test the two sets of assemblies.

The 95 percent acetic acid solution caused both sets of parts to crack, although a close examination showed that the cracks were more severe and occurred more rapidly in the parts where the bearing had been installed after the part was cool. This was confirmed at the 65 percent solution. At a 65 percent acetic acid concentration, the parts assembled hot showed no problems while the parts assembled cold still produced some cracking. We also noticed that the grease was actually coming from the bearing. When the bearing was inserted into the cold part, the interference between the outside diameter of the bearing and the inside diameter of the molded part cause some of the bearing grease to squeeze out onto the plastic. This did not occur when the bearings were inserted into warm parts.

Rules of Prevention
Now the good news is that all of these field problems had a solution or at least an obvious cause to the problem. The bad news is that a lot of time and money were spent uncovering the problem, recalling product, and arriving at these solutions. It is much better to learn the rules for preventing stress cracking in the first place.

The first rule to remember is that amorphous materials are more likely to have problems with this type of failure than semicrystalline polymers. If you cannot use a semicrystalline material because of other considerations, a blend containing a semicrystalline and an amorphous component may be a good compromise. All things being equal, an alloy of polycarbonate and polyester will work better than a straight polycarbonate.

Second, reduce the stress inherent in the design of the part. Eliminate drastic changes in nominal wall thickness, sharp corners due to insufficient radiuses, and gates and weldlines located in regions where high levels of external stress will be applied.

Third, avoid incorporating part features that force the processor into bad decisions regarding molding conditions. A thick wall placed at the end of the flow path will lead to the use of excessive packing pressure in an attempt to reduce sink marks. This will build high levels of stress near the gate.

Fourth, when checking chemical resistance data, look for tests that have incorporated some applied stress into the evaluation. If the addition of 1 percent strain causes a material to fall several notches in the rankings when exposed to a chemical in your application, look for another material or find a way to protect your chosen material from that chemical.

Fifth, add reinforcement to improve the performance of a material. As little as 10 percent glass fiber can reduce problems with stress cracking in materials like polycarbonate and polysulfone. And finally, as is the case with so many properties, molecular weight is critical. The higher molecular weight materials will resist stress cracking for much longer and at higher critical stress levels. Once you have selected the appropriate molecular weight, make sure your process conditions preserve it.

Stress cracking is a frequent problem in plastic applications. But there are methods for getting to the root of the problem. To avoid problems, select your material with care and employ the fundamental rules of good design and processing.

Use this link to read Part 1 of this article.

Sign up for the PlasticsToday NewsFeed newsletter.

You May Also Like