The Materials Analyst, Part 89: The myth of overdrying—Part 1

Moisture content affects viscosity, which in turn can cause processing problems. But that’s only the beginning of the complex concept called overdrying.

Some years ago, one of the process engineers I was working with brought me some parts being molded in a 33% glass-fiber-reinforced nylon. The parts had an unusually poor surface appearance, with glass fibers evident on the surface. The typical samples archived in the quality control department showed a resin-rich surface, indicating that something was not normal.

In these types of situations the debate usually ensues about material quality vs. process as the root cause. The engineer was expecting a full-blown composition analysis of the material and was almost certainly expecting that we would find that the material was out of specification on the high side for glass. However, our first and last test was a check of the moisture content in the material. A Karl Fischer titrator showed that the moisture content of the resin was 187 ppm, or about 0.019%.Why the focus on moisture content? At the next machine over, the same lot of material was being used to mold a similar product. It was running with no evident cosmetic issues. A check of the moisture content for the material in that dryer gave a result of 905 ppm. When this wetter material was introduced into the process on the first machine, the cosmetic problem immediately cleared up.

The shift supervisor told the process engineer that the material in the first dryer had been overdried and would have to be discarded. I suggested that the material simply be removed from the dryer for a period of time and allowed to rehydrate. The material was returned to the gaylord and in about 3 hours a moisture check produced a value of 761 ppm. This material was reintroduced to the process and produced parts with good appearance.

Less moisture, harder to flow

This type of situation occurs hundreds of times each week in molding plants around the country. It is driven by a lack of understanding about what happens to a material during the drying process. If you refer to processing manuals published by manufacturers of nylon materials, they usually set an upper limit of 0.2%, or 2000 ppm, for the allowable moisture content during processing.

This actually refers to unfilled material, and since a lot of the nylon that the industry processes is filled, it is important to adjust this value to account for the filler content. Glass fiber and most other fillers do not absorb moisture to an appreciable degree. If 33% of the compound is glass, which displaces the polymer, then the allowable moisture content for that compound must be reduced by that same 33%. This gives us a maximum allowable moisture content of 0.134% or 1340 ppm. So what was happening out on the floor? The “problem” material had been dried to a very low level, less than 15% of the allowable maximum. The other two samples were running a little above the midpoint between completely desiccated material and the maximum allowable level.

Figure 1 shows a plot of melt viscosity vs. moisture content for a 33% glass-filled nylon tested at a shear rate of 1000 sec-1. Note that the melt viscosity increases as the moisture content drops. Note also that the relationship is not completely linear. Once the moisture content gets down to about 250 ppm (0.025%), the rate of viscosity increase changes dramatically. At a little over 100 ppm (0.01%) the melt viscosity is almost 20% higher than it is at 800 ppm (0.08%). The higher melt viscosity results in greater difficulty in achieving a resin-rich finish on the part.

The right response to viscosity increase

This relationship between melt viscosity and moisture content is not limited to nylons. It exists for a whole class of materials that chemists refer to as condensation polymers. These include polyesters such as PET and PBT, as well as materials like polycarbonate. The reason the phenomenon gets so much attention in nylons is that these materials have a relatively large range of allowable moisture content at which they can be molded. In most other condensation polymers, the upper end for good processing is 200 ppm (0.02%), or only one-tenth of the value permissible for nylon compounds. But the pattern of behavior is the same.

Figure 2 (p. 52) shows a plot of melt viscosity vs. moisture content for a 30% glass-fiber-reinforced PET polyester tested at a shear rate of 400 sec-1. It shows the same general relationship between moisture content and melt viscosity that is observed in Figure 1. The difference is that, in the case of PET, the maximum allowable moisture content for good processing is 200 ppm, as indicated.

It is probably not a surprise that as the moisture content increases to levels above 200 ppm, the viscosity of the material drops. Water reacts with PET polyester in a mechanism known as hydrolysis. The polymer chains actually break and the shorter chains flow with less resistance than the longer chains. This can be expected to result in reduced performance in the molded part. We will come back to this relationship in Part 2 of this topic.

However, if the moisture content drops below the benchmark value of 200 ppm, we can see the opposite effect—the melt viscosity increases. At a moisture content of 50 ppm (0.005%), the melt viscosity at this particular shear rate rises by 25%. Now imagine a scenario where a process is set up in which the first-stage injection pressure is limited. When a drier material enters the process and the melt viscosity increases, the machine encounters greater resistance to flow and the injection velocity slows down. Most molders who work with glass-filled materials know that if you reduce injection velocity, the part appearance worsens. If the molder is not observant and does not realize the nature of the change in the process, the response to this cosmetic problem is to make other adjustments to counter the viscosity increase. If these adjustments fail to yield a solution, then the material comes under scrutiny as it did that day in our plant.

Second-stage problems

The processing problem that occurred that day is a real one. And it is the reason that many nylon suppliers advise that the processing of their materials is most effectively conducted with moisture content values not lower than 0.05% (500 ppm). This value typically pertains to an unfilled material. There’s a practical reason for this guideline: Many molding machines lack the available injection pressure required to maintain constant velocity in a demanding flow path as material viscosity fluctuates by 25-30%. So to keep life simple, material suppliers have learned that the process is more controllable and repeatable if the material moisture content does not get too low.

The problem can actually get worse once the process switches over to the second stage of injection. Figure 3 shows the relationship of melt viscosity to moisture content in a 33% glass-filled nylon when the measurement is made at a shear rate of 10 sec-1. This result refers to the same material shown in Figure 1.

Note first of all that the melt viscosity values are approximately three times higher than those in Figure 1. This is to be expected since melt viscosity declines with increasing shear rate. At 1000 sec-1, the material moves much faster than it does at 10 sec-1. This change is similar to the one that occurs when the molding machine transfers from first- to second-stage injection and the screw abruptly slows down. But while the viscosity increased by 25-30 Pa-sec when the material became very dry at high flow rates, at the slower flow rates the increases are now closer to 100 Pa-sec.

Now imagine a molding process that either profiles the latter stages of first-stage injection to a slower speed, or transfers from first to second stage earlier than it should. The material’s viscosity increases when the injection speed drops, but if the material is very dry, that viscosity increase will be exaggerated and processing problems will be more likely to occur. If a weldline that is critical to the part’s strength forms near the end of flow, it is easy to see why the strength of this weldline might be reduced if the material’s viscosity increases and there is no compensating influence from the process.

But in the industry at large, this relationship between viscosity and weldline strength is blamed on “overdrying.” The assumption here is that very dry material produces weaker parts than material that is not as dry. Many molders can cite case after case where the properties of a nylon part got worse when the material was dried to a very low moisture content. They seldom look beyond the phenomenon to the root cause of an unmanaged viscosity increase.

True overdryingIf the part geometry is sufficiently intricate, it may be impossible to achieve the desired part configuration or performance when a very dry material is injected at the slower speeds associated with second stage. Some years ago I was involved in developing insert-molded parts where the bond between the metal component and the nylon being injected into the mold was critical. A leak path test between the metal and the polymer was part of the quality control process and the intricate knurled geometry of the metal insert was designed to prevent the leak from occurring.

One day some months after a process had been developed that produced good parts, leak tests suddenly started to fail. After looking at a number of variables, it finally became clear that when the moisture content of the resin fell below a certain value, leaks became a problem. When we cracked open the parts to expose the area where the polymer filled in the intricate knurls in the metal insert, we saw that the parts molded with a material of moderate moisture content exhibited a crisp, well-defined, filled-out detail. However, the parts molded in the very dry material displayed a fuzzy and poorly formed geometry. No amount of additional pressure would resolve the problem. Because this area did not develop its shape until the process entered the second-stage packing phase of injection, we had no choice but to regulate the moisture content so that it stayed above a certain critical value.

But the mythology surrounding overdrying is much more complex than the relationship between moisture content and viscosity. There are other things that change in a material during the drying process that can cause real and irreparable damage to the polymer. Experience with these types of changes is what caused the shift supervisor to want to throw out the very dry material. Next month, we will address those other issues and explain why the notion of overdrying has become so entrenched in the molding industry. We will also show that drier material produces stronger, tougher parts as long as the process can compensate for the fluctuation in melt viscosity.