The Materials Analyst, Part 30: How usable is your regrind?

This series of articles is designed to help molders understand how a few analytical tools can help diagnose a part failure problem. Michael Sepe is our analyst and author. He is the technical director at Dickten & Masch Mfg., a molder of thermoset and thermoplastic materials in Nashotah, WI. Mike has provided analytical services to material suppliers, molders, and end users for the last 12 years. He can be reached at (414) 369-5555, ext. 572.

Two and a half years ago we began this series with an article on some polycarbonate parts that were brittle because poor processing, in this case inadequate drying, had reduced the molecular weight of the material. We documented the change in molecular weight with a relatively simple melt-flow-rate (MFR) test. Since then, we have come back to this theme of the connection between molecular weight, melt viscosity, and properties several times.

It is difficult to emphasize this correlation too much since it remains the number one problem with plastic product performance and it is perhaps the most fundamental tenet of polymer behavior. It is an 80-year-old idea that was revolutionary and controversial at the time of its advent, but by now we should all have gotten it.

Usually we back into this subject of molecular weight as a way of explaining a quality problem that has been sent to us by a client. But this month’s study focuses more on planning and forethought. The purpose of this investigation was to determine the progressive effect of recycling a material through the molding process several times.

This is a technique that material suppliers have used for many years to demonstrate the robustness of their materials. They will run a standard sample shape like a disk or a plaque in virgin material, grind some of the parts and make some more product in 100 percent first-pass regrind, and repeat this step until they have parts made from virgin material and five generations of regrind. Then they will lay these parts out on a white background and take some pictures so that everyone can see how good the color retention is. Normally they will also conduct the same trial on their competitor’s material, which of course performs much worse. Seldom, if ever, do we see any property information on these repeated passes.

Using Scrap Effectively
Our client had performed just such a trial on its product, a part molded in a 4-MFR polypropylene homopolymer, and wanted to review the melt flow rate of each pass to determine the change occurring in the material with each pass through the machine. In these competitive days, effective use of regrind is obviously a key factor in managing costs. Some processors, either because of regulatory dictates or out of habit, simply throw out any scrap generated in the form of runners and bad parts as though it were a thermoset. Others add it back to virgin material at a particular percentage like 20 or 25 percent. This spreads out any detrimental effects of processing and ensures that material from a second, third, or fourth pass will be watered down.

Still others adopt what has become known as a cascade process. They collect all the first-pass regrind from a given lot of material and then run it straight. While this first-pass material is being run they are collecting the second-generation scrap, which they then run at 100 percent and so on, just like the material suppliers in the tests we described. The supposed benefit of this method is increased traceability. Assuming very good handling procedures are employed to ensure cleanliness, it can be made to work for many materials, but under this system it becomes even more important to understand the effect that the process is having on the material.

Our client could see that by the third trip through the injection molding machine, its parts, which were being run in natural material, were beginning to yellow. We ran a simple melt-flow-rate determination on parts made from virgin material and all five generations of regrind. We had no virgin material, so the assumption was made that the particular lot the client was working with had the nominal melt flow rate of 4.0.

The results of the investigation appear in Table 1 and in graphical form in Figure 1. Remember that the guideline for good processing of an unfilled material is an increase in the melt flow rate from pellets to parts of no more than 40 percent and preferably less than 30 percent. While no single pass through the molding machine exceeded the 40 percent limit, the cumulative effect on the material is striking. The parts made from first-time regrind are already out of bounds, and the problem spirals rapidly out of control from there. By the time the last set of parts has been made, the 4-melt material has become a 16-melt material with color problems.

Often, these changes are also accompanied by a loss in critical additives such as antioxidants, ingredients that are designed to protect the polymer from the rigors of aggressive field service conditions. Fortunately, in this case the stabilizers were still intact. A test designed to assess oxidative stability was run on virgin parts and those made from fifth-generation regrind and the change was a surprisingly small 10 percent.

Nevertheless, the parts varied greatly in molecular weight. Now, there is nothing inherently wrong with a 10-melt or a 16-melt poly-propylene, but it simply will not have the properties of a 4-melt. If a 4-melt material is needed, then parts made from the higher-flow regrinds are likely to run into performance problems. The good news is that much of the change that we saw in this case is avoidable and usually arises from running the melt temperature too high. We were able to reproduce our client’s results using a simple tensile bar mold and running a melt temperature of 249C (480F). We then repeated the experiment using a melt temperature of 204C (400F).

The results appear in Table 2 and in Figure 2, and they are dramatic. Although each step changes the material to a slightly different degree, the overall shift in the material processed at 249C is 330 percent. But the material molded at 204C has only moved 25 percent after five passes through the molding process. In other words, it will still have the intended properties and after five runs through the press it is in better condition than after one run at 249C. The starting melt flow rate of this particular lot of raw material was 3.98g/10 minutes.

Improved Processing Through Material Analysis
If you engage most processors in a discussion about their selection of an optimum melt temperature, most of them will cite the need to fill the cavity as their rationale for running the material at a particular setting. In most cases, this selected temperature is higher than it needs to be because molders do not take full advantage of the other method for reducing viscosity—shear.

All plastic materials exhibit viscosity reduction with increasing shear rate, and one of the tasks of process optimization is to arrive at the best balance of shear and temperature. For some amorphous materials, notably polycarbonate and polysulfone, the balance is clearly on the side of temperature. But for most materials, and especially a material like polypropylene, shear rate provides a much more effective route to viscosity reduction.

Figure 3 shows a viscosity vs. shear rate plot for a 4-melt polypropylene at 204C (400F), 227C (440F), and 249C (480F). You don’t have to be a rheologist to understand the implications of this graph. Simply look at how closely spaced the viscosity curves are for the three temperatures. Even though this is a logarithmic plot, something many of us don’t use every day, a close look at the region between 1000 and 10,000 sec–1 is an education that we can take back to the molding machine.

For example, look at the viscosity of the material when the melt temperature is 204C and the shear rate is 2500 sec–1. You should come up with 55 Pa-sec. Now look at the same shear rate point for the curve generated at 249C. The viscosity has dropped to 40 Pa-sec. Now comes the good part. Go back to the curve for 204C and follow it from 2500 sec–1 to the point where it crosses the viscosity of 40 Pa-sec. What is the shear rate? That’s right, 4000 sec–1. How do we change the shear rate from 2500 to 4000 sec–1? We increase the injection speed by 60 percent, for example from 1.25 inches/sec to 2 inches/sec.

Processors that find themselves raising the melt temperature because they have run out of speed, or because the speed setpoint is not obtainable due to a pressure-limited process, should realize that the price they pay is huge. Not only does it require much more energy to put all that extra heat into the material (energy consumption), it takes more time to get it back out (cycle time). And as the data above have shown, you get degraded material in the bargain.

Right about now some of you are probably thinking that you have skipped a page in the magazine and are no longer reading an article about material analysis. But that is exactly the point. The best material analysis relates directly back to improved processing; it is not an activity that takes place in a vacuum. The practice of optimizing a process and coming up with test data that certify a healthy process are activities that should be joined at the hip.

We recommend that a viscosity comparison between pellets and parts be done as part of the first article inspection, because the part can be to print in every respect, but if it is made from degraded polymer, it’s still no good. And sound management of regrind is impossible if you don’t know what your process is doing to your material.