Silicone Industrial Seals: A Temperature Seal, Not a Chemical One

The medium decides before the temperature

Silicone is sold on temperature and bought for temperature, which is why it ends up in applications where the medium is the actual problem. A silicone O-ring in a hot water flange at 150 Celsius will outlast the pump. The same O-ring in a mineral oil line at 80 Celsius will swell, soften and extrude out of its groove within months.

The mechanism is straightforward. Silicone rubber has a very open, very flexible backbone with weak intermolecular forces, which is exactly why it stays elastic across such a wide temperature range. That same openness lets non-polar hydrocarbons diffuse into the network, where they push the chains apart and the part grows. The growth is called volume swell and it is the single most useful number for choosing a seal compound.

Swell is not automatically a failure. A small amount, two to five percent, can be helpful because it tightens a seal against its groove. The trouble starts above about ten percent, where the swollen material becomes softer, loses tensile strength, and starts to extrude into the clearance gap under pressure. A seal that has extruded has not just failed, it has usually damaged the groove.

So the first question in any seal specification is what the seal will touch, including the fluids it only meets during cleaning, commissioning or a maintenance mistake. The second question is temperature, because temperature accelerates everything: diffusion, oxidation and compression set all roughly double for every 10 to 15 degrees of additional heat.

Swell against temperature

The chart on the left plots volume swell for a standard methyl vinyl silicone compound in four media between 20 and 140 Celsius. Petrol runs from 30 to 50 percent, mineral oil from 18 to 33, glycol coolant from 2 to 8, and water from 1 to 4. The dashed line at ten percent is the usual engineering limit.

The shape of the chart is what makes it useful. Silicone is not marginal in oil and fuel, it is off the scale at the lowest temperature on the axis. Even at 20 Celsius, petrol swells it by 30 percent, three times the limit. No formulation change, no post-cure and no hardness adjustment fixes that, because the interaction is between the hydrocarbon and the polymer backbone itself.

In glycol and water the picture inverts. Swell is a few percent across the whole range, which means the seal stays in its groove and the failure mode becomes compression set rather than extrusion. That is why silicone is the default for engine coolant, hot water, and steam-adjacent sealing, and why it appears in washing and food equipment where the medium is water-based.

Temperature makes a bad case worse faster than it makes a good case worse. Petrol at 140 Celsius is 50 percent swell against 30 at room temperature, a two thirds increase. Water at 140 is 4 percent against 1 at room temperature, a fourfold increase that is still nowhere near the limit. The absolute number matters more than the ratio.

The ten percent rule and what sits behind it

The ten percent swell limit is a rule of thumb, and it is a good one because it marks the point where the material's properties start to fall off a cliff. At ten percent swell a typical silicone has lost perhaps 30 percent of its tensile strength and 20 points of hardness, and its compression set has roughly doubled.

The consequence in a real groove is extrusion. A seal in a pressurised joint is held in a groove with a defined clearance, and the seal resists being pushed into that clearance by its own hardness and strength. Soften it and it flows into the gap, and once a nib of material has entered the gap the pressure acts on it directly and the process runs away. This is why a swollen seal usually fails as a nibbled or chewed ring rather than a clean break.

The limit also explains why hardness selection is not a free choice. A 50 Shore A silicone in a low pressure application can tolerate more swell than a 70 Shore A one at high pressure, because the softer material has more compliance before it extrudes. But a soft material also creeps more, so the same seal cannot be soft and creep-resistant at once. That is a material limit, not a formulation oversight.

Where a small swell is unavoidable, the groove can be designed for it. A groove volume 15 to 20 percent larger than the seal volume leaves room for the material to grow without becoming over-compressed, and a back-up ring on the low pressure side prevents extrusion even at higher swell. Both are cheaper than switching to a fluoroelastomer for a marginal case.

Compression set is the quiet failure

The right hand panel shows why industrial seals fail in service: oil and fuel contact 32 percent, heat ageing 26, compression set 18, chemical attack 14, installation damage 6, other 4. Compression set at 18 percent is the one that gets missed because it produces no visible damage, no leak at first, and no obvious cause.

Compression set is permanent deformation. A seal is squeezed in its groove and held there for years; a good elastomer springs back when the joint is opened, a poor one stays flattened. Once it stops springing back it stops sealing, and the joint weeps. Silicone's compression set is good at room temperature and only moderate at 150 Celsius, where a standard compound might show 25 to 35 percent set after 70 hours against 10 to 15 for a high quality one.

Post-curing is the standard remedy and it is often skipped because it costs time and oven capacity. A silicone part that is post-cured removes the volatile residues from the cure and stabilises the network, which improves compression set noticeably at high temperature. For a seal that will sit at 150 Celsius for a decade, the post-cure is not optional.

Table 1 — How industrial silicone seals actually fail.

Choosing the elastomer for the medium

The table below is the one worth keeping. Silicone owns hot water, steam, glycol and food contact. EPDM owns hot water and steam at lower cost with better tear strength. FKM owns fuels, oils and aggressive chemicals at high temperature. NBR owns mineral oil at moderate temperature and low cost. Nothing owns everything, and the medium is the first column.

The common mistake is to specify one elastomer across a machine for simplicity. A machine with a water circuit, an oil circuit and a fuel line will have three different seal materials if it is designed properly, and the maintenance manual needs to say so, because a fitter with a box of silicone O-rings will use them everywhere.

A second mistake is to accept a material substitution during a shortage. Silicone and FKM are not interchangeable in oil, and the difference is not gradual: one swells by 30 percent and one by 3. A substitution that is approved on the basis of temperature rating alone will fail in the first year.

Table 2 — Elastomer selection by medium, before considering cost or hardness.

Installation and groove design

Six percent of failures are installation damage and that number is understated, because a cut ring that seals for a month before weeping is usually logged as a material failure. A silicone O-ring is soft, has modest tear strength, and is easily damaged by a sharp groove edge, a thread, or a screwdriver used to persuade it into place.

Three habits remove most of it. Round or chamfer every groove edge to at least 0.2 mm. Cover threads and splines with a thin sleeve during assembly. Lubricate with a compatible lubricant, which for silicone in a water application is usually a silicone grease and for silicone in anything else needs checking, because a hydrocarbon grease will swell the ring before it is even installed.

Groove design then decides whether the seal can work at all. The groove fill ratio, the seal cross section divided by the groove cross section, should sit between 70 and 85 percent for a static face seal. Below 70 the seal can move and roll; above 85 there is nowhere for thermal expansion to go and the seal extrudes. A groove designed for a 3 mm cord and used with a 3.5 mm cord is at 100 percent fill before it is even warm.

The last thing to check is the clearance gap. For a 70 Shore A silicone at 10 bar the gap should not exceed about 0.25 mm; at 40 bar it should not exceed 0.15 mm without a back-up ring. These numbers are not conservative, they are the point at which extrusion begins, and a joint that was machined to a generous tolerance for assembly reasons will fail at pressure even with a perfect seal.