Extrusion Brittleness: When the Profile Cracks Instead of Bending

Brittle is a symptom, not a property

A customer reports that a profile cracked during installation, or that a length broke in the back of a van, or that a corner snapped when a screw was driven. All three are the same complaint: the material did not have enough toughness to absorb the energy of a normal handling event.

Toughness in a rigid PVC profile comes from the compound and from the way it was fused. The polymer itself is inherently notch sensitive, which is why impact modifiers exist, and the fusion level determines how well the structure has developed. A profile made from a correct compound that was under-fused is brittle; so is one made from an over-filled compound that was fused perfectly.

The distinction matters because the fixes are different and neither is in the die. A shop that receives a brittleness complaint and starts polishing the die is working on the wrong part of the process entirely, and will still have the complaint after the die is perfect.

There is a third cause that is easy to overlook: thermal history. A profile that was run too hot, or that sat in a hot die for a long time at start-up, has degraded material at the surface. That layer is brittle even when the core is sound, and it shows up as surface cracking rather than a clean break.

Filler takes toughness away quickly

The chart on the left plots notched impact strength against calcium carbonate loading in a rigid PVC window compound. At zero filler the impact is 48 kilojoules per square metre, at 10 phr 44, at 20 phr 36, at 30 phr 26, at 40 phr 17 and at 50 phr 11. The dashed line at 20 is the level below which handling damage becomes routine.

The curve is not linear and it is not gentle. Between 20 and 40 phr the impact falls by more than half, and the filler loading in that range is exactly where cost pressure tends to push a formulation. A compound that was developed at 15 phr and has drifted to 35 phr through a series of small cost decisions will be brittle and nobody will remember deciding it.

The reason is that filler particles are stress concentrators. Each one is a point where a crack can start, and the more of them there are, the shorter the path from an initiation site to a failure. A fine, surface-treated filler concentrates less stress than a coarse untreated one, which is why the grade of the filler matters as much as the amount.

The practical control is a specification with a filler range and a test. A compound specification that says calcium carbonate 10 to 20 phr, with a notched impact minimum, is a control. A specification that says filled PVC is an invitation to drift.

Under-fusion is the most common cause

Under-fusion accounts for 28 percent of brittleness investigations and it is the hardest to see, because an under-fused profile looks identical to a well-fused one. Fusion is the process by which the PVC grains lose their boundaries and become a continuous matrix, and it happens in the last part of the screw and in the die.

The variables that control it are temperature, shear and residence time. A profile run at a low barrel temperature with a high output rate has less fusion than the same compound run hotter and slower, and the difference in impact can be a factor of two. The profile dimensions will be identical.

The test that finds it without a laboratory is a dichloromethane soak. A small sample is immersed in solvent for a measured time, then examined: a well-fused sample swells and holds together, an under-fused one falls apart or shows a grainy, porous structure. It is a crude test and it separates good from bad reliably.

The other field test is a simple bend. A strip cut from the profile is bent around a former of a specified radius at a specified temperature. A tough profile bends and whitens; a brittle one snaps. The test is only useful with a fixed former and a fixed temperature, because both change the answer.

Lubricant balance and thermal history

Lubricant imbalance accounts for 18 percent. Lubricants control the friction between the polymer and the metal, and they also delay or accelerate fusion. Too much external lubricant and the compound never fuses properly; too much internal lubricant and the melt viscosity falls so far that there is not enough shear to develop the structure.

The symptom is a profile that processes beautifully and tests badly. Output is good, surface is good, dimensions are good, and the impact is half what it should be. A compound like that has been over-lubricated to make it easy to run, and the ease has been paid for in toughness.

Thermal history at 14 percent covers both over-heating and long residence. A profile run at 210 degrees instead of 190 has a degraded surface layer, and a profile that sat in the die during a twenty minute stop has a degraded section at that point. Both produce local brittleness, and the second produces a failure that appears at one position on one length.

Table 1 — Separating the causes of a brittleness complaint.

Regrind and the slow drift

Regrind overload accounts for 10 percent and it behaves like filler: a small amount is harmless, a large amount is not. Regrind has a shorter molecular chain than virgin material because it has been through the extruder once already, and a shorter chain means lower impact.

The effect is cumulative if the regrind is recycled repeatedly. A closed loop with no virgin make-up will see the molecular weight fall pass by pass, and the impact falls with it. A cap of 15 to 20 percent regrind with virgin make-up keeps the loop stable; a closed loop does not.

Regrind also carries contamination. A bin that has collected trim from several compounds, or from a run with a different filler level, delivers a mixture whose average is fine and whose local variation is not. That variation shows up as inconsistent impact across a production run, which is harder to diagnose than a uniformly brittle profile.

The control is a dosing unit with a recorded ratio, the same as for dimensional drift. Where dosing is manual, the ratio belongs on the process sheet and should be checked at the start of each shift, because the failure mode is an operator topping up a bin with whatever is nearest.

Table 2 — Notched impact against regrind share in a rigid PVC compound.

Proving the fix

A brittleness fix is only proved by a test that failed before and passes after. That sounds obvious and it is frequently skipped: a compound is changed, the profile looks better, and the complaint returns in three months because nothing was measured.

The minimum proof is a bend test on ten samples from three different lengths, at a fixed temperature, before and after. The dichloromethane soak adds the fusion evidence. Together they separate a compound fix from a process fix, which matters because the two are often applied at once and then nobody knows which one worked.

The second element of proof is a record. The filler level, the fusion result, the impact number and the regrind ratio should go on the process sheet for that profile, so that a future change is visible against a baseline. A shop without a baseline cannot tell whether it has improved or merely changed.

The third is the specification. If the customer's drawing says nothing about impact, the shop has no defence when a complaint arrives and no target when it tries to fix one. A single line on the drawing, with a test method and a temperature, prevents most of these arguments.