Click Image to Enlarge
How do you know if you are over-cooling? Cut coupon-sized samples of the extrudate and heat them for several minutes in an oven at approximately 200-250 F, then cool and measure the shrinkage of the cooled surface vs. the opposite surface. The shrinkage will look like this for a flat sample that was cooled unevenly during processing.
If a little cooling is good, then it stands to reason that a lot of cooling is better. But that’s not always true for heavy-wall extruded products made from crystalline polymers that are cooled from one side. This would be typical of pipe, many sheet products, blow molded parts, and hollow profiles.
Crystalline polymers have high shrinkage rates as they cool from their melted state. Almost all molten polymers have a completely amorphous structure lacking in molecular orientation. However, crystalline polymers develop a partially crystalline structure upon cooling, which is their natural or relaxed state. Crystallization causes the polymer to increase its density as the crystal structure forms. The molecules form a more ordered and more tightly packed structure than the amorphous structure. HDPE, for example, changes from a melt density of approximately 49 lb/ft³ to 59 lb/ft³ as it cools from its melt to semi-crystalline natural state. That’s an increase in density of 20% or overall shrinkage of 20%. Other crystalline polymers include LDPE, LLDPE, PP, nylon 66, PET, PLA, and PTFE.
Crystallization proceeds very rapidly at a temperature midway between the melting point and glass-transition temperature of the polymer. As the temperature is reduced below that maximum, the rate of crystallization decreases rapidly because the molecules have reduced freedom of movement to achieve their relaxed state. In most processes 60% to 80% of the crystallization takes place during the cooling phase of processing, and as much as 90% occurring within a week of processing. Remaining crystallization can take months to complete, depending on temperature. However, crystallization continues until a stable crystal structure is achieved.
Why is this important? An amorphous structure can be temporarily “frozen in” with rapid cooling. But since polymers are poor conductors of heat, one surface can be cooled while the opposite surface of the extruded product may remain hot even as it exits the production line. This results in a largely amorphous structure on one side and a partially crystalline structure on the other. Since the crystalline portion will have a higher density and more shrinkage than the amorphous portion, there is a great deal of internal stress developed within the part as one side shrinks more than the other.
In extruded sheet, rapid cooling of one side will cause warpage. In pipe this effect will cause a high stress in the wall that will reduce its physical properties, particularly impact and stress-crack resistance. In a blow molded product, it will cause distortion of the walls and unbalanced stress between the thinner and the thicker sections. This has a negative effect on stress-crack resistance and drop impact. In complex profiles it will cause the edges to warp or even cause sink marks.
Control of the crystallization or shrinkage and resultant stress depends on control of the cooling rate through the entire part. This can be achieved by reducing the overall cooling rate or interrupting the cooling so that the outer surface of the extruded part cools slower and allows heat from the other side to diffuse to the cooled surface. Since it’s usually necessary for extruded parts to quickly freeze the outer dimensions to conform to a certain shape, the initial cooling usually needs to be adequate to set the shape. With continued rapid cooling, there is the potential for developing internal stress in the part.
How do you know if you are over-cooling? A simple and quick way is to cut some small coupon-shaped samples of the extrudate and heat them for several minutes in an oven at approximately 200 to 250 F, then cool and measure the shrinkage of the cooled surface vs. the opposite surface. The exact heating conditions and time will vary with the polymer and thickness of the coupon. The shrinkage will look like the accompanying illustration for a flat sample with uneven cooling, before and after oven heating.
Samples that are curved can be measured using tape. Complex surfaces that are difficult to measure can be evaluated by the amount of distortion as compared with a fully annealed control sample.
These measurements should not be confused with orientation due to drawdown or post-extrusion stretching. Almost all heavy-wall extruded parts will have more machine-direction shrinkage than transverse-direction shrinkage, due to drawdown and die shaping. To evaluate cooling shrinkage by itself, only the differential of the opposite surface dimensions should be compared, not the overall shrinkage of the piece.
Cooling imbalance can cause many types of part failures. For example, post-extrusion cracking can occur as the amorphous layer progresses toward its stable crystallized structure, putting it under enormous stress. With severe cooling imbalance, the failures may even look like layer separation and be misdiagnosed as laminations due to poor mixing.
If you experience part failures at low stress levels, the cause could be as simple as cooling too fast. This is particularly prevalent when cooling from just one side. To test, simply shrink some coupons and if the surfaces vary significantly, reduce coolant temperatures or allow breaks in cooling to anneal the part.
Jim Frankland is a mechanical engineer who has been involved in all types of extrusion processing for more than 40 years. He is now president of Frankland Plastics Consulting, LLC. Contact firstname.lastname@example.org or (724)651-9196.