Tearing in Silicone Parts: Demoulding Force and the Undercut Nobody Drew

Tearing is a force problem

Silicone tears when the force needed to remove it from the tool exceeds its tear strength at the point where the force is concentrated. Both halves of that sentence matter: the force comes from the tool, and the concentration comes from the geometry.

The force has three sources. The interference fit over an undercut, which is the largest and the one that is usually not on the drawing. The friction on the walls, which depends on the shrinkage and the surface finish. And the adhesion to the tool, which is small unless the tool is dirty or the release agent is wrong.

The concentration has two sources. A thin section, which has less material to carry the load. And a sharp corner, which multiplies the stress. A part with a 0.5 mm section and a 0.2 mm radius will tear at a force that a 2 mm section would shrug off.

So the diagnosis starts with where the tear is. A tear at the edge of a thin section is a section problem. A tear at the base of a feature is an undercut problem. A tear that is random across the part is a temperature or a release problem.

Demould temperature is the first lever

A silicone part is weakest when it is hot. The tear strength of a typical 50 Shore A compound at 160 C is about 60 percent of its room temperature value, and at 190 C it is about 35 percent. So a part demoulded hot is a part that tears at a force it would survive cold.

The measured tear rate follows that. At 150 C it is 0.4 percent. At 160 it is 0.6, at 170 it is 1.5, at 180 it is 4.2, and at 190 it is 9.8. Between 160 and 190 the rate goes up by a factor of sixteen for 30 C.

The obvious fix is to demould cooler, and the cost is cycle time. A part that is demoulded at 160 C instead of 190 C needs 20 to 40 seconds more cooling, which on a 90 second cycle is a 25 to 45 percent throughput loss. That is why the compromise is usually 170 to 175 C, where the tear rate is 1.5 to 2.5 percent and the cycle is still competitive.

The other temperature to check is uniformity. A part that is cool at the edge and hot in the middle tears at the edge, because the edge is being pulled while the middle is still soft and moving. A tool with uneven cooling produces a tear pattern that follows the temperature map.

The undercut nobody drew

An undercut is any feature that the part has to stretch over to come off the core. A designed one is a snap fit or a sealing bead. An undesigned one is a tool feature that was added for another reason and happens to interfere.

The common undesigned undercuts are a draft that is reversed on one wall, a polishing mark that has been blended into an edge, a text engraving with vertical walls, and an ejector pin that stands 0.05 mm proud. None of them look like an undercut on the drawing and all of them cost force.

The force is large. With no undercut the demould force is about 12 newtons. At 0.2 mm it is 26, at 0.4 mm 48, at 0.6 mm 82, at 0.8 mm 131, and at 1.0 mm 198. The relationship is not linear: doubling from 0.4 to 0.8 mm nearly triples the force, because the silicone has to stretch further and its modulus rises as it does.

A 60 newton limit is a reasonable working ceiling for hand demoulding, and it corresponds to about 0.45 mm of undercut. Beyond that the part needs a stripping tool or a redesign, and a part that is stripped by hand at 130 newtons is a part that is torn by the operator on a bad day.

Finding the undercut is a matter of measurement, not inspection. A dial indicator run around the core at the point where the tear starts will show a step of 0.05 to 0.2 mm that the eye does not. A blue-marking compound rubbed on the core and then checked on the part shows exactly where it is gripping.

Table 1 — What an undercut costs in demould force

Geometry that concentrates the force

A tear needs a place to start. On a smooth, generous part the force is spread over an area and the stress is low. On a part with a thin section and a sharp corner the same force is concentrated into a few square millimetres and the stress is above the tear strength.

The rule is that every internal corner should have a radius of at least half the section thickness, and every external corner at least the section thickness. A 1 mm section with a 0.2 mm internal radius has a stress concentration of about 3.5; the same section with a 0.5 mm radius has about 1.8.

Thin sections are the other half. A silicone part with a 0.3 mm wall is a part that will tear at a demould force that a 1 mm wall would ignore, because the force is carried by a tenth of the material. A thin wall is often a design requirement, and the answer is a stripping tool rather than a hand pull.

The parting line is a third. A parting line that is not flush leaves a step, and a step is an undercut. A tool with a 0.05 mm mismatch at the parting line is a tool with a 0.05 mm undercut all the way round the part, and that is enough to double the demould force on a large part.

Release, and the things that are not the tool

A tear that is random across the part, rather than at a feature, is usually a release problem. The part is sticking to the cavity as well as the core, so the demould force is pulling against two surfaces and the part tears where it is weakest.

The causes are a tool that needs cleaning, a release agent that has been applied unevenly, an internal release agent that has been washed out of the compound, or a compound that is under-cured. An under-cured part is soft and tacky and it tears at a fraction of the force a fully cured part would take.

Cure is worth checking before the tool. A part that is 10 percent under-cured has a tear strength about 25 percent lower and a surface that is slightly tacky. The cure can be checked with a hardness reading, a compression set test or a simple tear by hand on a test piece, and it takes minutes.

The release agent itself is a variable that is easy to get wrong. A silicone release agent on a silicone part is a contamination risk; a solvent-based one that has not flashed off is a mould that is wet. The right amount is a thin film that is almost invisible, applied to a clean tool at the start of a run and not every cycle.

Fixing it in the right order

Start with the demould temperature. If the part is coming out at 185 C and the tear rate is 5 percent, dropping to 170 C will take it to about 1.5 percent and the cost is 20 seconds of cycle. That is the cheapest fix and it is usually enough.

Then check the cure. A part that is under-cured tears at a lower force and it feels tacky. A hardness reading against the specification takes a minute and it eliminates a whole class of causes.

Then measure the undercut. A dial indicator around the core at the tear point will find a step that the eye cannot, and a step of 0.1 mm is 12 newtons of force on its own.

Then look at the geometry. A thin section or a sharp corner at the tear point needs a radius or a thickness change, and that is a tool modification rather than a process change.

Table 2 — Reading the tear by where it starts