The Materials Analyst, Part 47: I can't believe it's not crystalline (Part 4)

By now, we are starting to see some patterns in the way chemists put polymers together. We should also be gaining an appreciation for the importance of the specific arrangement of the constituents within the polymer chains. Without the aid of polarity, we have shown how a material can be produced that has a glass transition and a melting point that qualify it for service in the world of engineering applications. This month we will look into the techniques used to push beyond this performance level. 

We said earlier that the bonds that make up polyethylene and polypropylene are nonpolar. As a result the attractions that develop between the individual chains tend to be relatively weak. This gives rise to a variety of distinctive properties including low melting points, excellent electrical properties, resistance to moisture absorption, and that characteristic waxy feel. 

If we introduce elements into the polymer chain that produce a polar bond within the polymer, then we get a certain degree of polarity between the polymer chains. These polar attractions are harder to disrupt and this means higher temperatures are needed to melt the material. 

Polyamides: The Good and the Bad 
A classic example is the family of materials we refer to as nylons, or polyamides. If we were to look at a very small segment of a nylon chain, say three or four carbon units in length, we would see that it looks like a polyethylene chain. But as we pulled back to get a broader view, we would notice that at regular intervals there is an interruption in the pattern, a strange combination of carbon, hydrogen, oxygen, and nitrogen that chemists call an amide group (hence the name polyamide). 

Because of the way the atoms are arranged in this group, the amide linkage is highly polar. This results in individual chains that are more strongly attracted to each other. Figure 1 shows the alignment of short sections of nylon 6/6 chains. The dotted lines represent these strong attractions, which are known as hydrogen bonds. These have a strength many times greater than that of the weak attractions that arise between nonpolar molecules. 

This is both good news and bad news. The good news is that as a result of these stronger attractions the melting point of nylon 6/6 is 260C (500F). In addition, the strength at yield of nylon 6/6 is approximately four times that of an HDPE and the modulus has doubled. The bad news is that now we have a material that will absorb water. Even worse, we have introduced chemistry into the system that will actually break down if the material is heated in the presence of that water. Enter the dryer as an indispensable part of the molding process. 

The numbers in the nylon product designation tell us how far apart the amide groups are. Nylon 6/6 is the best known of these variants, but it is possible to change the properties by changing the spacing between the amide groups. For years some applications have relied on specialty nylons that absorb less water than the mainstream nylon 6/6. Since the water goes where the amide groups are, increasing the spacing between amide groups reduces the moisture absorption. 

Nylon 6/12 is a good example of this. While a part molded in nylon 6/6 will readily absorb 2 to 3 percent by weight in water over time, the nylon 6/12 may only pick up a fraction of 1 percent. Not surprisingly, the melting point of nylon 6/12 is lower, only 218C (424F), because the hydrogen bonds occur less frequently along the chain. This reduction is a measure of the importance the amide group has in elevating the high-temperature performance of the material. 

If we continue to increase this spacing we will produce materials with lower moisture absorption properties and lower melting points. The properties will approach those of polyethylene, which can be thought of as nylon with the amide groups spaced infinitely far apart. 

High-temperature Nylon 
But suppose we want to go the other way with the properties, to a higher level of performance. Moving the amide groups closer together, as in nylon 4/6, will produce the desired effect. The melting point of nylon 4/6 is 290C (554F). So by simply manipulating the spacing of the amide groups in the chain, we can produce melting points that span 130 deg C. 

However, we find that as we make these structural changes, the glass transition temperature changes very little. The T_g for nylon 6/12 and nylon 4/6 differ by only 10 to 15 deg C (18 to 27 deg F). This is because the melting point is related to the strength of the attractions in the areas where the crystals form, while the T_g is related to the stiffness of the individual chains. Changing the spacing of the amide groups has a large effect on the strength of the attractions that bind the crystals together, but it has little effect on the stiffness of the individual molecules. 

The small changes in the T_g are important for two reasons. First, since materials form crystals for as long as they are above T_g, the mold temperature requirements do not change much for the various nylons. A mold temperature of 70 to 90C will produce the desired crystallinity in most cases regardless of the type of nylon being processed. 

Second, the load-bearing performance of the material at elevated temperatures is not improved appreciably as the melting point undergoes this dramatic rise. This is because even a well-crystallized nylon, in its unfilled state, will lose 75 to 80 percent of its stiffness and strength as it passes through the T_g. Even with a heavy loading of glass fiber the loss is still 50 percent of the room temperature properties. So to make substantive gains in elevated temperature performance, we need to elevate the T_g along with the melting point. 

This is the function of the newer high-temperature nylons, also known as polyphthalamides (PPA). The PPA materials extend the melting point of the polyamide family even higher than nylon 4/6, to levels of 300 to 315C (572 to 599F). But the real benefit of these materials is the increase in T_g, to 110 to 140C (230 to 284F). This delays the onset of the large property loss associated with T_g, extending the performance of this family of materials. 

Figure 2 (right) shows an example of how it is done. This figure shows a single repeating unit for nylon 6/6 and one type of polyphthalamide. The structures are very similar, but notice that one of the segments that used to contain a straight carbon chain in nylon 6/6 has now been replaced in the PPA with one of those large hexagonal benzene rings. We saw these in the polystyrene structure, but they were hanging off to the side of the main chain. Now we have inserted them directly into the backbone with great effect. Even if you are more inclined to think mechanically rather than as a chemist, you can see that the presence of such a ring would produce a stiffer structure than a straight chain. 

Figure 3 compares the modulus vs. temperature behavior of a nylon 6/6 and a PPA with the same glass fiber loading. The advantages of a material with a stable modulus up to 130C are obvious. Engineers who use traditional nylons in, for example, under-the-hood automotive applications, must negotiate a decline in load-bearing properties of more than 50 percent as the temperature rises from 20 to 120C. No such considerations are necessary when working with the high-temperature PPA. 

This insertion of benzene rings into the backbone is the most common way to extend the temperature-dependent properties of polymeric materials. Table 1 (right) gives a list of high-performance semicrystalline materials that employ a ring structure in the backbone. 

Trade-offs: Properties vs. Processing 
Now here is the bad news. The higher T_g and the limited motion permitted by the stiff chain make the process of crystallization much slower and raise the minimum temperature at which crystallization can occur. (In fact, if the ring structures become plentiful enough we wind up with materials like polycarbonate and polysulfone where no crystallization occurs at all.) Worse yet, the minimum temperature needed for crystallization tends to be above that attainable with hot water. Many processors attempt to get by with the hottest water they can produce, hoping to get close. But this process of crystallization can be an all-or-nothing proposition with many of these high-performance materials. 

Figure 4 shows a plot of modulus vs. temperature for a 40 percent glass-fiber-reinforced PPS run at mold temperatures of 80C and 140C. While the room temperature modulus values are similar, the performance at elevated temperatures is alarmingly different. The part produced in the hot mold loses the expected 50 to 60 percent of its stiffness as it passes through the glass transition at approximately 120C. The part molded at 80C forms very little crystalline structure. Consequently, it begins to lose properties at a lower temperature. But more importantly it loses almost 95 percent of its room temperature performance. 

Up to a point the PPS molded in a cold mold looks like a true amorphous material. If PPS were naturally amorphous the modulus would simply drop to near zero just as we saw in amorphous polystyrene in last month's article. But instead something curious happens; the modulus begins to increase as the temperature rises above the T_g. This increase occurs over an extended temperature range and at one point the cold-molded material almost catches up with the well-crystallized part. 

This behavior is associated with a process called solid-state crystallization. The natural state of the PPS is to be semicrystalline, but this state is suppressed by a cold mold. The crystallinity can only develop to a fraction of its intended level during the molding process. The first time the material reaches a temperature high enough to reactivate the required mobility, the crystallization process starts again. 

Unfortunately, once the part is in the application, this process occurs in an uncontrolled way. Crystallization involves shrinkage, and once the part is in use these dimensional changes can have serious consequences for the function of the entire product. In addition to the problems with strength and dimensional stability, an undercrystallized part will lack chemical resistance, fatigue resistance, and surface hardness often critical to good wear properties. 

But of even greater concern is the temperature region between the start of the glass transition and the recrystallization process. The modulus plot shows that the undercrystallized material comes extremely close to softening, and this is a compound with 40 percent glass fiber content! 

The results in Figure 4 drive home a very important principle that was suggested in Figure 5 from last month's article. The upper temperature performance limit of a semicrystalline material, at least for short-term use, is usually governed by the melting point of the polymer. But when the polymer fails to crystallize, the upper limit becomes the glass transition. As we have already pointed out more than once, the difference between T_g and T_m is typically 150 deg C (270 deg F). This is what is potentially sacrificed when molding conditions fail to produce the desired degree of crystallinity. 

This simple fact accounts for some very puzzling failures. Next month we will conclude this series by taking these chemistry lessons to the molding floor. We will cover a host of problems that all revert back to a failure of a high-performance polymer to crystallize to the intended level. 

Editor's note: Mike Sepe will speak on "If This Stuff is 50 Percent of Our Cost Structure, Why Are We Paying So Little Attention?" at IMM's technology conference, Oct. 1-2 in Chicago, preceding Plastics USA. For more information, go to www.immconference.com.