The Materials Analyst, Part 45: I can't believe it's not crystalline (Part 2)

Last month we laid some groundwork for a discussion on the structural details that determine whether a material is semicrystalline or amorphous. We spent most of our time discussing the shape of the molecule using polyethylene as a model. But to explain fully why crystals form in a polymer we must also consider the opportunity that the polymer molecules have for motion. The mobility of the individual polymer chains depends upon the strength of the attractions that form between them. These, in turn, are dependent upon the strength of the attractions within them. 

Let's return to the polyethylene. We mentioned last month that a polyethylene molecule was composed entirely of a chain of carbon atoms that formed the backbone of the polymer. Attached along the side of this backbone are hydrogen atoms. When carbon and hydrogen atoms combine, the electron pairs that form the bonds between the atoms are shared almost equally. From the outside looking in we see a system where the bonds have no strong positively or negatively charged poles. We call these systems nonpolar, and virtually every hydrocarbon, from gasoline and propane to polyethylene and polypropylene, is nonpolar. 

This is important to us as processors because nonpolar materials and polar materials, like water, will have nothing to do with each other. If you don't believe this, try mixing gasoline and water. Consequently, hydrocarbon-based polymers do not absorb water and therefore require no drying before processing. 

This lack of polarity also means that the attractions that form between polyethylene chains are relatively weak. This results in a material with a relatively low melting point. And yet when the branching is controlled and limited, polyethylene is the most crystalline of the semicrystalline polymers. 

Part of this is the result of the shape of the molecule, as we discussed last month. But it also comes from the fact that weakly bonded polymer chains can move independently until the temperature becomes quite low. This motion is important to the formation of crystals. If this motion stops, crystal formation ceases. 

It is a fundamental rule of polymer structure that crystal formation cannot begin until the material temperature drops below the melting point. As the cooling process continues the temperature keeps dropping until it falls below another key point called the glass transition temperature. At this point, crystal formation stops. The window of opportunity lies between the melting point (T_m) and the glass transition temperature (T_g). As it turns out, the two temperatures are related. For most semicrystalline materials the T_g will be approximately 150 deg C (270 deg F) below the melting point. 

Because the attractions between polyethylene chains are relatively weak, the melting point of even the high-density materials is quite low: 135C (275F). But how crystalline is the final product? To answer this question we use a familiar technique, DSC. With this test, we can measure the degree of crystallinity as the area under the curve associated with the melting of the crystals. We then compare this energy requirement to that needed to melt a theoretical polyethylene with a crystallinity of 100 percent. 

According to the literature, the energy requirement to melt a 100 percent crystalline polyethylene would be 290 J/g. To accurately measure the degree of crystallinity for a particular polyethylene, we compare the heat of fusion for the melting process to this ideal value of 290 J/g. Figure 1 shows the heat of fusion for the process of melting a low-density and a high-density polyethylene. 

First, we notice the difference in the melting point of the two materials: 23 deg C (42 deg F). But perhaps more significant is the difference in the heat of fusion, a relative measure of the degree of crystallinity. The LDPE has a heat of fusion of 102 J/g so the degree of crystallinity for this material is 102/290, or 35 percent. The HDPE, however, has a heat of fusion of 178 J/g, for a degree of crystallinity of 178/290, or 61 percent. 

The Role of Cooling 
When we dig into this a little deeper we discover that the injection molding process affects this degree of crystallinity. While we may say that a high-density material has a degree of crystallinity of 65 percent, we find that this value can be influenced somewhat by the rate at which the material cools from the melt. Cooling rate is everything in this business of developing crystallinity since it determines how rapidly we move through the temperature region between T_m and T_g. Thinner wall sections and lower mold temperatures tend to result in lower degrees of crystallinity because the material spends less time in this all-important temperature zone where crystals can form. We can observe this by comparing the heat of fusion in a molded part to the same sample after it has been cooled slowly in our DSC and then reheated. 

Figure 2 shows the DSC result for a sample cut from a molded HDPE part compared with the same sample after it has been cooled slowly and reheated. Although the melting point has changed only slightly, the heat of fusion on second heat is higher by 21 J/g. This translates to a degree of crystallinity of nearly 61 percent for the molded part vs. 69 percent for the reheated material. The second heat value is often thought of as representing the inherent degree of crystallinity while the first heat value gives the as-molded degree of crystallinity. 

What would it take to remove this 8 percent gap by producing the maximum crystallinity in the molded part? Well, consider that the cooling rate in the DSC for this test was 10 deg C/min (18 deg F/min). In the molding process, the polyethylene enters the mold at, let's say, 227C (440F) and is ejected 30 seconds later at a surface temperature of 52C (125F). At the part surface this is a cooling rate of 350 deg C/min (630 deg F/min), 35 times faster than in the DSC. So to produce the perfect part, our 30-second cycle would have to stretch out to several minutes and probably involve a very hot mold. It might be the perfect structure, but who is going to buy it? 

So the business of producing a crystalline product is a balancing act between ideal structure development and economics. We cannot afford to abandon all considerations of structural integrity for the sake of cycle time, however. When we specify a semicrystalline material, we are doing so because we need certain properties. Strength, stiffness, retention of structural properties at elevated temperatures, dimensional stability, chemical resistance, and fatigue resistance are all maximized when the crystal structure increases. 

In the case of polyethylene it is difficult to quench the material because the T_g is so low. Remember, the T_g is usually 150 deg C below the melting point. If the melting point for a highly crystalline HDPE is 135C, then by our rule of thumb the T_g would be -15C (5F). (It is actually much lower, polyethylene being one of those materials that bends some very important rules.) This is well below room temperature, requiring a cold mold to substantially suppress the crystallization process in a polyethylene. 

Getting Practical 
Nevertheless, different cooling rates can result in different degrees of crystallinity, which brings us to our first practical application of this principle. While the business of crystalline structure may seem somewhat remote from our everyday concerns, crystallization is directly related to an attribute of polymer behavior that is very much on our minds every day: shrinkage. Crystalline structures are more organized than amorphous regions and they take up less space. So the more a material crystallizes, the more it shrinks as it cools. Uneven cooling caused by either varying wall thicknesses or poor temperature control in the mold results in different levels of shrinkage in different areas of the part. This in turn leads to warpage or high levels of molded-in stress, or both. 

A client was experiencing warpage in a large panel molded of high-density polyethylene. The panel was large and flat with only a very shallow rim around the entire perimeter of the part. This perimeter lip was distorting despite being designed with a uniform nominal wall. Raw material and a part were submitted. In the DSC we heated a sample of the raw material above the melting point, cooled it at a controlled rate, and reheated it through its melting point. This second heating gave us the ideal degree of crystallinity for the material. Again, these are not practical conditions for manufacturing, but they provide a baseline for comparing actual molded parts. We then heated a section of the part cut from the interior where the part was flat and another section from the perimeter where the warpage was evident. In these cases we were interested in the first heating, since this gave us the state of the as-molded material. 

The results appear in Figure 3. It is easy to see that the central region of the part is composed of material where the degree of crystallinity is very close to the ideal state of the material. However, the edges where the warpage occurred had a substantially lower degree of crystallinity. The difference in crystallinity resulted in a difference in shrinkage, which caused the warpage. 

An analysis of the mold temperature profile showed that the center of the part had excellent cooling and the temperature of the steel in this region typically ran at about 40C (104F). However, at the edges no waterlines had been drilled and the temperature of the perimeter varied throughout the run, starting as low as 20C (68F) but eventually rising to near the temperature of the rest of the mold over the course of the first 8 to 12 hours. This was consistent with the pattern of poor-quality product; most of the bad parts came from the early part of a run or after a restart. This example shows some of the practical considerations attached to the question of crystallinity. 

While the behavior of a material like polyethylene is obviously subject to some variation, it is inevitable that most of the material that can crystallize will do so under normal processing conditions. This is not the case for all materials. Next month we will examine what happens when we venture into the world of higher-performance resins. Here the molder plays a much bigger role in developing properties using materials where the stakes are typically much higher. 

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, "Molding Technology 2001—Technology Investments that Pay Off," Oct. 1-2 in Chicago, preceding Plastics USA. For more information, go to www.immconference.com.