Silicone Engineering Notes

Too little squeeze leaks; too much makes assembly hard and shortens life. Both figures depend on cord diameter and groove dimensions — calculate before cutting a tool.

Squeeze = (cord diameter − groove depth) ÷ cord diameter. It sets the contact pressure on the sealing face: 15%–25% for static seals, typically 20%–25% for face seals; 10%–15% for reciprocating or rotary dynamic seals, since the rubber needs room to deform during motion; and 25%–30% under vacuum or high pressure to compensate for extrusion.

Fill ratio = O-ring cross-section ÷ groove cross-section. Static seals usually run 75%–85%, never above 90%. Silicone expands roughly 2.5×10⁻⁴ per °C — ten times more than steel — so a section that fills 85% of the groove at 25°C will press against the groove wall at 200°C and the seal fails. For hot service, keep the cold fill ratio at 70%–75%.

Then check compression set. Per ASTM D395 or GB/T 7759 method B, 70 h at 175°C: standard silicone sits at 20%–35%, low-set compounds below 15%. This figure governs long-term sealing reliability — if a customer specifies only hardness, raise the compression-set requirement proactively.

First cure sets the shape; post-cure sets the properties. Skip post-cure and nothing looks wrong at first — compression set, odour and yellowing all surface three to six months later.

First cure happens in the tool and turns compound into a fixed shape: compression moulding at 160–175°C, roughly 1–2 minutes per millimetre of thickness. This only achieves shape — cross-linking is incomplete, and low-molecular volatiles plus peroxide decomposition products remain in the part.

Post-cure runs in an oven: drive out those residues and finish cross-linking. Typical cycle 200°C for 4 h; ramp in stages from 150°C for thick or critical parts to avoid cracking from a thermal gradient. Afterwards compression set drops, resilience improves, odour largely disappears, and residual volatiles are usually held below 0.5%.

The cure system drives much of what follows. Platinum addition cure leaves no by-products — odourless, high clarity, first choice for food-contact and medical parts; but it is poisoned by sulphur, phosphorus and amines (cross-contamination in the mixing room causes under-cure), and both compound and tooling cost more. Peroxide cure is universal, cheaper and slightly better at temperature, but its by-products smell and must be driven off by post-cure.

Yellowing is a common platinum-cure complaint: silicone weathers well, yet platinum-cured parts can still yellow under long light or heat exposure. Use high-purity base gum, add an anti-yellowing agent, and keep sulphur-bearing release agents away. Yellowing in peroxide-cured parts is usually incomplete post-cure — removing the residues normally fixes it without extra additives.

Most defects have three or more possible causes. Decide first whether it is the compound, the tool or the machine settings — that is far faster than trial and error.

For short shots, check three things first: compound flow (too high a Mooney, too much filler), venting (vent depth, vacuum on or off), and injection pressure versus mould temperature. Low mould temperature is the most common and most overlooked cause — the compound scorches or thickens before it reaches the far end of the cavity.

Blisters split into two kinds: gas, from poor venting or air trapped in the compound (mixing without vacuum); and moisture, from wet raw material or humid air. Cut one open — a glossy smooth wall is usually gas, a rough or marked wall usually moisture. Fix the first with venting and vacuum mixing; fix the second with a drying step and humidity control.

Bloom is compounding ingredients migrating to the surface, usually from excess curing agent or accelerator, or from under-cured material leaving unreacted species to migrate. Sticking needs separating into failed release agent, worn tool coating, or over-cure bonding the part to the cavity. Flow marks and weld lines trace back to gate position and venting where flow fronts meet — a tool design issue that parameter changes can only soften, not cure.

"Food-grade silicone" is marketing language, not a standard. When a buyer asks about certification, what actually matters is: which standard, which report, and which test items.

Food contact is governed by food-contact material regulations, not "certification". China uses the GB 4806 series (GB 4806.11 for rubber), the US FDA 21 CFR 177.2600 for repeated-use rubber articles, and at EU level the AP(2004)5 resolution, with Germany adding LFGB. None of these is a one-off lifetime pass — reports are batch-specific and time-limited, and buyers usually want one issued within the past year.

Biocompatibility is a separate framework: the ISO 10993 series, written for medical devices. Common parts are -5 cytotoxicity, -10 sensitisation and irritation, -11 systemic toxicity. The key point: food grade does not imply biocompatibility. Food grade only shows migration stays within limits if ingested; biocompatibility shows no adverse biological response in tissue contact. Medical parts also raise sterilisation and contact-duration questions.

Two boundaries worth stating up front. First, silicone is not oil-resistant — mineral oil, solvents and most fuels swell it; switch to NBR or FKM for those media rather than trying to make silicone work. Second, platinum and peroxide cure behave differently in compliance: platinum leaves no by-products and is the steadier choice for food and medical; peroxide cure with insufficient post-cure leaves decomposition residues that affect odour and migration results.