Overmolding Silicone on Metal: Primer vs Interlocks

If you're putting metal into a custom RTV silicone mold, a silane primer is chemistry. A dovetail, through-hole, or flared core is geometry. Buy geometry first if that joint has to survive a peel.

This is the RTV analog of “overmolding onto metal”: inner cores and mandrels for hollow chocolates or ice, registration pins in the jacket, threaded hardware that holds a mother mold, a metal master you pour against, or a steel key bushing the rubber sits on. LSR or HCR vulcanizing onto a cookware handle is a different quote for a molded *part*. If the PDF in your inbox is full of heated cavities and a cold-runner, you're not in this article.

A primer film can help the face stay put in shear. It does not replace a rivet of rubber through a hole when the operator peels a 10A–20A skin off a steel core every dump.

What each method is actually doing on an RTV stack

Mechanical interlock. The rubber cannot leave the metal without tearing or shearing a designed feature: through-holes, slots, knurls plus a retaining groove, a T-groove around a core shank, hex flats the jacket wraps, or a flared tube the RTV mushrooms behind. Bond strength becomes tear strength of the cavity and the remaining metal ligament. Hot pectin, dish soap, and steam do not “turn off” a rivet the way they can degrade a primed face.

Primer (typically silane-type coupling). A thin film that bonds to oxides on the metal and presents groups the platinum or tin network can react with as the A/B mix cures. Works in shear on a clean, named alloy. Fails on oil, unexpected passivation, thick scale, or a peel that starts at a sharp free edge. Do not invent a product code. Qualify a system with the RTV supplier and the finish you will actually buy.

Blast / degrease / barrier coat. These are not primers. Degrease until water breaks — machining coolant is the usual poison. Blast raises area on a core shank you will hide. A barrier coat on a metal *master* is there so platinum does not inhibit or stick to the pattern; it is not a lock for a production core you extract every cycle.

Self-bonding RTV grades exist for some metals. Treat them as primed-in-the-rubber. Still clean the insert. Still add a lock if a peel edge exists.

Common mix-up: primer on a polished core, no tooth

A chocolate line wants a 40 mm hollow sphere, two-part platinum RTV, about 800 dumps a month, food-contact.

Right stack: CNC or turned steel core with a pull path, a retaining groove plus through-holes so the *jacket* or the opposite half locates the core, Shore 10A–20A skin, rigid mother mold. Primer on the core shank that stays in the jacket is insurance.

Wrong stack: a polished stainless ball, primer wipe, no groove, and a hope that the RTV will “bond so it doesn't walk.” After a few hot cycles the core clocks, the split steps, and the primer lot gets blamed.

Both quotes can say “overmold metal.” Ask which object ships: a flexible cavity with a removable core, or a vulcanized silicone *part* stuck to steel.

When primer is not enough

Primer-only joints fail in predictable RTV shops:

• A flat washer of rubber on a polished stainless disc (no tooth, easy peel start)
• A food-mold core that sees hot dump, detergent, and thermal shock
• A jacket bolt that flexes the peel edge every open
• Inserts that arrive with machine oil, then get “wiped”
• Aluminum flask hardware with as-machined coolant, or anodize the primer was not qualified on

If you ignore this: first article peels clean with a thumbnail. Production walks after a solvent wipe the kitchen considers normal. The geometry never had a chance.

Use primer as insurance on a locked joint, and as the primary bond only when CAD cannot hide a hole and the load is light, compressive, or well-qualified in peel.

Mechanical features that earn their keep

Ranked by how hard they are to argue with:

1. Through-holes or slots the RTV pours through, with a head or a back-side mushroom so the rubber is a rivet. Use these on jacket plates and on cores that stay in one half.
2. Dovetail or T-groove around a core shank. The peel has to invert a wedge.
3. Knurl or coarse blast in a band, plus a retaining groove. Knurl alone on a shaft can still spiral off under torsion.
4. Wrap-around: the jacket covers a flange or a hex, so extraction is not a straight peel.
5. Keyed bushings in the mother mold, not nubs in 10A rubber. Soft buttons squash and stop locating.

Fill still matters. A 3 mm RTV skin will not pack a 8 mm dovetail. Pour and vacuum so the rubber reaches the lock before it skins. A metal core steals heat from a warm dump; that is a chocolate problem, not a reason to skip the groove.

If you ignore this: a beautiful knurl with no groove, packed with air, and a primer on oil. Three “bonds,” zero joint.

Surface prep that is a process, not a SKU

A usable metal-prep sequence for RTV hardware looks like this. Exact chemicals belong to the qualified primer system.

1. Specify the incoming finish. Bare 304, electropolished 316, anodized Al 6061, zinc-plated flask steel — these are different parts.
2. Degrease until water breaks. Machining coolant is the usual poison.
3. Mechanical profile where the bond lives: blast or knurl. Mask show faces of a metal *master*. Blast grit left in a food-contact bore is a leak and a hygiene fail.
4. Rinse and dry. Trapped blast dust is the issue.
5. Primer in the specified film, flash time, and pot life. Primer on a cold, wet insert is a release agent.
6. Pour the RTV (mix 1:1 or 10:1 by grade, vacuum degas until the foam collapses) inside the open time.
7. Do not treat a post-cure of the mold as a free upgrade for a primer that was never qualified for that oven.

Stainless that must stay unblasted (hygienic ID of a hollow dump) needs the lock on the OD and a primer qualified on that mill finish. Do not blast a food-contact bore to help a jacket bond.

What to ask when you're buying

• Is the metal a removable core, jacket hardware, a master, or a pin that stays in one half?
• Alloy and finish, named — not “stainless.”
• Lock features: holes, dovetail, knurl-plus-groove, wrap?
• Who extracts the core, and what is the pull path so you don't tear the 10A–20A skin?
• Platinum or tin, and is the primer qualified on that chemistry?
• Food-contact path on the *cavity* (21 CFR 177.2600 / LFGB), not a logo on the metal.

Worked example: 40 mm hollow sphere, steel core

A confectioner needed a 40 mm hollow chocolate, two-part platinum RTV, keyed split, about 800 dumps a month. The core was turned 304 with a 2 mm retaining groove and two 4 mm through-holes so the jacket half captured it. Skin Shore 10A–20A, mix 1:1, vacuum degas, rigid jacket. Primer on the captured shank only. We would not have trusted primer on the polished sphere. After the first week, if the core walked, you would see a step at the split — that is a keying and lock problem, not “bad silicone.”

ISO 3302-1 might class a rubber *article* you later cast. It does not class a steel core, and it is not another word for precision.

When this article is the wrong quote

If you actually need vulcanized silicone parts overmolded onto metal (gaskets, handles, sensor boots) in the thousands, this RTV stack is the wrong asset. Quote HCR compression or LSR injection with a cold-runner on a parts OEM. A tiny, shallow tray with no core can skip metal entirely. Geometry can override quantity: one deep hollow still needs a core at 80 dumps.

Related reading: custom RTV silicone molds, keying two-part molds, self-demolding geometry, food-grade molds.

What to send for an RTV + metal quote

Copy this list. A price without these lines is not comparable.

• Metal role: core / jacket hardware / master / pin
• Alloy + finish; lock features (holes, dovetail, knurl+groove)
• Pull path and which half captures the metal
• Cavity Shore, platinum vs tin, mix 1:1 or 10:1
• Casting material and working temperature
• Cavity count, monthly dumps, recast of the RTV skin
• Food-contact path if the cavity dumps food
• Current failure, if any: core walk, peel at the metal, sticky face

Send core drawings (or the physical core), finished cavity dimensions, and the material you will cast. We will review the lock, the jacket, and whether primer is insurance or a wish.

We would not recommend primer-only on a polished food core

We would not recommend raising Shore A first so the rubber “grips the steel.” Add the lock and the jacket. We would not blast a hygienic bore to help a handle bond. We would not cut an LSR steel overmold “so we can scale later” while you still need a cavity to dump chocolate into.

RUUIPON Mold Structure Decision: metal in the stack? name the role. Removable core with undercut? two-part / plug, not a one-piece glove. Soft 10A–20A or wet dump? mother mold so the core does not balloon the skin. Flat back, no metal? open-back, skip this article.

flowchart TD
  A[Metal in the RTV job?] --> B{What is the metal?}
  B -->|removable core / mandrel| C[Pull path + groove or holes]
  B -->|jacket hardware / pins| D[Lock in the rigid shell]
  B -->|metal master| E[Barrier coat; do not bond the pattern]
  C --> F{Peel edge on 10A-20A?}
  F -->|yes| G[Mechanical lock first; primer is insurance]
  F -->|light shear only| H[Qualified primer on named finish]
  B -->|LSR/HCR part on metal| I[Stop: different part RFQ]

Typical shop values

Caption these as typical shop values, not a named lot.

| Item | Typical shop value | | --- | --- | | Cavity Shore | 10A–20A platinum RTV, jacket behind it | | Mix | Platinum 1:1 or 10:1 by grade; vacuum degas until the foam collapses | | Linear shrink | ~0.1–0.3% (platinum); tin keeps moving in storage | | Loaded-tray edge sag | About 0.5–1.5 mm on an unsupported large food tray | | 40 mm hollow, 800 dumps/month | Two-part skin plus captured core is a common first band | | Primer open time | Follow the system; do not primer Monday for a Friday pour unless allowed |

FAQ

Can a silane primer replace knurls or holes when the RTV joint sees peel? No. Primer is a thin coupling film that works in shear on a clean, qualified finish. Peel starts at a free edge and can defeat that film after detergent, steam, or oil the way it cannot defeat rubber poured through a hole. Buy a mechanical lock first and treat primer as insurance on that lock.

Do self-bonding RTV grades let me skip degreasing the insert? They do not. Self-bonding grades behave like primer mixed into the rubber; they still fail on machining coolant, unexpected passivation, or thick scale. Clean to a water-break, name the incoming alloy and finish, and add a lock if a peel edge exists.

Why is a through-hole stronger than knurl alone on a core shank? Rubber that pours through a hole and mushrooms on the back is a rivet; extraction has to tear the silicone. A knurl without a retaining groove can still spiral off under torsion. Combine knurl with a groove, or use holes, dovetails, or a wrap around a flange.

Should I blast a hygienic stainless bore so the jacket bonds better? No. Keep food-contact IDs unblasted. Put the lock and any blast profile on the OD, and qualify primer on the mill finish you will actually buy. Blast grit left in a thread is a leak, not extra bond area.

Why did my quote include a cold-runner and heated platens? That's an LSR or HCR *part* overmold. Ask the supplier to requote a custom RTV mold if what you need is a cavity with a metal core or jacket hardware.