The Materials Analyst, Part 53: Where does the moisture go? (Part 3)

The original purpose of this third segment was to follow various levels of moisture in a material through the molding process to provide an understanding of moisture's role in property development. However, the first two articles in this series generated such an interesting mixture of responses that a detour seemed advisable. Instead, this month's segment will be devoted to clarifying some of the comparisons that have been made between various moisture measurement techniques.

The essential argument for using simpler moisture measurement techniques such as loss-in-weight is that the more rigorous chemical methods are simply too difficult for most molders. It is a curious line of reasoning since it essentially puts forth the notion that a fundamentally inaccurate method of measuring moisture can substitute for a fundamentally accurate one simply because it is easier to perform.

To repeat a point from the March article, a system that heats a material and considers the entire resulting weight loss as moisture lacks the ability to distinguish between water and other volatile materials that are driven off by the heating process. If the heating process is conducted in an oxygen-containing atmosphere such as ambient air, then this mixture can also include degradation byproducts if the test temperature is sufficiently high. The amount of the weight loss that can be attributed to water is unknown unless a method that is moisture specific is also run for corroboration. This method is Karl Fischer, a chemical technique that is moisture specific. Every manufacturer of an alternative moisture measurement system makes use of Karl Fischer as a verification of its instrument.

Mr. Fischer: Myth and Reality
The warnings against Karl Fischer are familiar, and we have seen them in the responses from some of the manufacturers of alternative instruments. The most common is that the Karl Fischer method employs toxic chemicals. The reference to toxic chemicals is actually a holdover from the composition of the original Karl Fischer reagents, which contained pyridine—definitely toxic. But about 20 years ago, chemical manufacturers began to introduce pyridine-free Karl Fischer reagents and today, they are the industry standard.

The second adjective mentioned regarding chemicals is "expensive." The expense is somewhat dependent on how much the instrument is used. Our system gets very heavy use because, in addition to the technicians on each of the three shifts, the lab staff uses it to verify the moisture content of the materials that must be dried prior to melt-flow-rate testing. Even with this around-the-clock duty, the annual cost is approximately $1000. In contrast, our moisture-specific loss-in-weight system that was purchased several years ago as a backup uses a disposable septum that costs $.85 and must be replaced every three to five tests. If this instrument were used to the same degree as the Karl Fischer analyzer, the annual costs for the septums would be just more than $2000.

The third assertion is that Karl Fischer instruments are fussy, demand constant attention to work properly, and require a trained chemist on staff. Most of this is again a throwback to the days of the older instruments. But some of it comes from a problem with method of operation. Karl Fischer instruments work best when they are in continuous operation. When the instrument is on, the reagents are constantly being stirred and equilibrium is maintained with the surroundings. If the instrument is turned off every night, it takes a while in the morning before everything comes back into balance.

But in a production setting that runs around the clock this should not be an issue. In our facility, there are at least 20 people who can come into the lab with a sample at any time and run a test. In the 16 years that we've had the instrument, it has been down for a total of six days. The routine maintenance of the instrument is handled by one of our lab technicians (who is not a trained chemist), just as the injection molding machine maintenance in our manufacturing facility is handled by technicians trained in hydraulic and electrical circuitry.

The fourth argument against Karl Fischer is all that glassware. It is true that Karl Fischer instruments are lab instruments. You cannot wheel the apparatus around the plant, and you have to leave the "use a bigger wrench" mentality at the door. But the idea of portability for the alternate systems is oversold. The reality is that the accuracy of any instrument is affected if placed in an area where the ambient air humidity and temperature vary. Our moisture-specific loss-in-weight instrument was initially used in the general plant environment with very poor results. It is now located in an office on the plant floor with a controlled relative humidity, and the results are much improved.

Problems with Preprogramming
A point has also been made that with simple loss-in-weight systems the user does not need to worry about method development; everything comes preprogrammed. Unfortunately, it is not that simple. For whatever reason, the list of test conditions provided by moisture analyzer instrument suppliers does not meet the test of reality, and this includes suppliers of Karl Fischer systems. Often temperatures are too high or too low, and some specifications seem to ignore the fact that many materials do not absorb enough moisture to warrant testing.

The other problem with predetermined test temperatures goes back to something we discussed in the first section of this series. An inert atmosphere is essential to accurate measurements because a temperature that removes all the moisture from a material will also remove a lot of other volatile components and may begin to degrade the material. Loss-in-weight systems that run in air attempt to get around this problem by reducing the test temperature.

For instance, consider the table above. Under these test conditions, the three instruments provide reasonably good agreement with each other with a couple of exceptions. The ABS results for the moisture-specific loss-in-weight system come in consistently high by a factor of two or three, suggesting that the 230C condition might be a bit aggressive for ABS. But the real concern comes in the results for the PET. The standard loss-in-weight instrument finds only 68 ppm using the reduced temperature of 155C while the Karl Fischer detects 126 ppm, almost twice the amount.

It is easy to conceive of a situation where the actual moisture content for the PET could rise higher than the critical threshold of 200 ppm (.02 percent) and still test as dry using the loss-in-weight system. And this is precisely the problem we see when we review these instruments. In order to compensate for the fact that loss-in-weight systems count all volatiles and that the presence of oxygen results in some degradation, test temperatures must be turned down on these instruments. But the reduced temperature does not ensure that only moisture is being monitored. It simply reduces the total volatile content to the point where it agrees with the actual moisture content. If you are working with noncritical materials like ABS and PPO, this is probably an uncertainty you can live with. If you are running materials that can hydrolyze such as polyurethane (PUR), PET, or PC, then you cannot.

Field Reports of Inaccuracies
In working on processing and part performance problems with various clients, we have encountered loss-in-weight systems that were reporting both high and low values. In one case, a molder was convinced that its PUR was being dried to 30 ppm based on the results from a loss-in-weight system. In spite of this, the molded parts were degraded. A Karl Fischer measurement showed that the actual moisture content was more than 300 ppm.

In another case, a company was actually using TGA (thermogravimetric analysis) to determine the moisture content of mineral-filled nylon parts that were brittle and failing during assembly. TGA is a very sophisticated technique compared to a moisture analyzer. But for all its sophistication, it is still a loss-in-weight system. The TGA was measuring a weight loss of 1.53 percent, and the end user was using this value to conclude that the parts had been fully moisturized and therefore must be breaking because of poor molding practices. A Karl Fischer determination showed that moisture content was actually only .40 percent, and with proper moisture conditioning the parts achieved their required ductility.

These failures do not discount the fact that for a given grade of material, careful correlation of results between techniques can result in usable test values. But loss-in-weight systems cannot be moisture specific, and, consequently, the agreement has to be established for every grade of material using the time-tested chemical techniques as the final arbiter.

Capable Processors
The statement has been made that loss-in-weight systems were created to answer a demand in the market for something simpler than chemical techniques like Karl Fischer. This is a bit of rewriting of history, because it suggests that the plastics industry, after struggling with the more sophisticated chemical techniques for years, finally turned to the simpler methods. In fact, most companies that use loss-in-weight systems had probably not even heard of Karl Fischer, much less tried it. If they had, and if they had received the proper training and support, they would have discovered what we learned in the mid-1980s: It is less challenging than learning how to operate an injection molding machine or programming your VCR.

The notion that processors cannot master the principles of scientifically correct moisture measurement underestimates the technical acumen of processors and strengthens the belief that they represent the weak link in the supply chain.

In addition, in a world where the threat of job migration is a constant, it should be understood that technological awareness is one of the best defenses against becoming a commodity. If something that is technologically sound looks a little harder or more challenging than what everyone else is doing, then it should be thought of as a potential competitive advantage, and not a cost line item to be avoided.