How a Custom RTV Food Mold Is Tested as an Article

If you're qualifying a custom RTV food mold, you test the cavity you'll dump into, not a drum of gum and not a color chip. In the US, rubber articles intended for repeated use are extracted under FDA 21 CFR 177.2600: distilled water at reflux for aqueous food, n-hexane at reflux for fatty food, against milligram-per-square-inch ceilings. On the German/EU path, LFGB is applied to silicone largely through BfR Recommendation XV: composition thinking, volatile organic matter, simulant migration, and an organoleptic check.

The specimen has to be production-real: platinum addition-cure RTV as typical, poured against the production master in the production box, post-cured at the bake you'll run, pigment frozen, jacketed as shipped, cleaned. A spatula plaque doesn't test a chocolate mold. An LSR injection coupon with a cold-runner doesn't test this pour. BPA-free isn't a test method, because silicone isn't polycarbonate.

Common mix-up: a flat plaque, or last year's natural leftover, stands in for the mold

A lot of RFQs treat any silicone coupon as the article. In the shop that usually means the lab extracts a 2 mm plaque, production recasts a 12 mm muffin wall in a holiday red, and the PDF describes something you don't sell.

Labs can't invent your process. They extract or migrate the specimen you hand them. Before the courier, freeze compound and pigment for the RTV A/B; pour and cure as production pours, against the same master and registration keys; post-cure in a ventilated oven, time scaled to thickness; wash as 177.2600 expects a thoroughly cleansed article; and identify SKU, color, cavity, jacket revision, lot logic. Don't send tin-cure as a shortcut. Don't relabel a sister SKU. If the master was sanded “to clean it up” between the lab pour and production recasts, you've also got a new article.

What to ask when you're buying the lab work

• Will the specimen be this production recast (master revision, jacket, color, post-cure), or a plaque / spatula?
• Which flasks: 177.2600 water and/or n-hexane, and/or BfR XV volatiles, simulants, sensory — scoped to the food you'll actually dump?
• What's written on the chain-of-custody bag so the lab doesn't swap geometry?
• Is production post-cure already on the part, or is the lab adding a courtesy bake you can't repeat on recasts?
• Who pays, and is the report SKU-specific with lab, method, and date — no invented cert numbers?
• If color or master changes after the PDF, what's the retest trigger?

US extraction and the German stack

177.2600 isn't a product registration. FDA doesn't “approve” a muffin or praline mold. Aqueous food, distilled water at reflux: first 7 hours not to exceed 20 milligrams per square inch; succeeding 2 hours not to exceed 1 milligram per square inch. Fatty food, n-hexane at reflux: first 7 hours not to exceed 175 milligrams per square inch; succeeding 2 hours not to exceed 4 milligrams per square inch. Surface area as used matters. Dry-food use is framed as suitability under GMP rather than those flasks. Tell the lab the food type or they'll guess wide.

Regulation (EC) 1935/2004 is the EU umbrella. Germany polices that through LFGB. BfR Recommendation XV is what silicone labs methodize — confirm the current text. A typical kitchenware sequence is composition thinking, volatile organic matter (short post-cure fails here), migration against water / 3% acetic acid / 95% ethanol at times and temperatures that represent the claim, and sensory so the article doesn't impart smell or taste. Acetic acid and ethanol aren't n-hexane. That's why an under-post-cured platinum RTV chocolate mold can squeak through 177.2600 and fail Germany on odor.

Buyer add-ons (PAHs, metals) appear when someone writes them. Scope the use or you buy the slowest package.

Worked example: 8-cavity madeleine as the honest specimen

A bakery ordered an 8-cavity madeleine mold, shells about 80 mm long and 18 mm at the thickest RTV wall, platinum RTV in a registered box with a fiberglass jacket, expecting 300–700 oven cycles a month.

The lab bag should hold that production recast: same compound, same pigment, same post-cure, same wash, photographed with the jacket so the lab doesn't swap in a 2 mm plaque because it fits the flask. Food types for batter are usually aqueous and fatty; write them. If Germany is in the market list, send the BfR XV method list in the same bag, not a second chemistry. Dual-market programs still don't nest.

The flask doesn't correct a skipped bake. Scale post-cure to the thickest section you actually pour against the master, then clean, then test. Label the bag. Don't degrease with a solvent the factory will never use. If the courier crushed a thin sheet, send another, because crushed geometry changes surface area. Reject a PDF that only says “food grade” or “FDA approved” with a decorative number. Keep the report with the freeze. If compound, pigment, post-cure, master, or recast recipe changes, the old file describes an article you no longer make.

When this article is the wrong quote

If the mold never touches food, don't buy 177.2600 or BfR XV. If the SKU isn't frozen (weekly color experiments, weekly master recuts), the report will be obsolete before it lands. If you're testing an LSR *part* from a heated cavity and a cold-runner, that's a different specimen and a different OEM. A 2 mm plaque can be representative only when purchasing accepts it *and* it shares compound, pigment, process, and post-cure with a matching thickness — geometry can override that shortcut.

Related reading: FDA vs LFGB, certification filing, what food-grade means, food-grade molds.

What to send with the lab bag

Copy this onto the chain-of-custody. A plaque without these lines isn't this mold.

• SKU, color, platinum RTV, master revision, jacket as shipped
• Post-cure already run in production (time, temperature, thickest wall) — no courtesy lab bake
• Food types: water flask, n-hexane flask, and/or BfR XV simulants + volatiles + sensory
• Photo of the part in the bag; don't crush thin sheets (surface area changes)
• Markets; who pays; retest trigger if pigment or the master changes
• Wash state: cleansed as 177.2600 expects

Send the production recast and the market/food list. We'll write the lab chain so they extract this cavity, not last year's natural leftover.

We would not recommend a 2 mm plaque for a 12 mm pan

We would not recommend a courtesy lab bake you can't repeat on recasts, or last year's natural leftover while this year's red tool is in the box. We would not skip the food-type line and let the lab guess wide.

flowchart TD
  B[Bag the production recast] --> L[Label SKU, color, platinum RTV, post-cure, master rev]
  L --> W{Which flasks?}
  W -->|aqueous| H[Water reflux 177.2600]
  W -->|fatty| X[n-hexane reflux 177.2600]
  W -->|DE-EU| V[BfR XV volatiles + simulants + sensory]
  H --> R[Report names specimen + method + date]
  X --> R
  V --> R

Typical shop values: lab specimen vs production wall

| Item | Typical shop value | | --- | --- | | Honest specimen | Production recast, same jacket, same pigment | | Wall mismatch to avoid | 2 mm plaque standing in for a 12–15 mm muffin rim | | Mix | Platinum 1:1 or 10:1; vacuum degas | | Aqueous limits | 20 mg/in² first 7 h; 1 mg/in² succeeding 2 h | | Fatty limits | 175 mg/in² first 7 h; 4 mg/in² succeeding 2 h | | BfR XV set | Water, 3% acetic acid, 95% ethanol + sensory |

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

FAQ

Can I test a flat plaque instead of the real RTV baking mold? Only if purchasing accepts a representative article and the plaque shares compound, pigment, process, and post-cure. Geometry still changes surface-to-volume. For a food-contact mold, the honest specimen is the cavity poured against the production master.

Do I reflux n-hexane for an ice-sphere mold that never sees fat? 21 CFR 177.2600 assigns n-hexane to fatty food. Aqueous ice is the water flask. Don't skip the section; don't invent fatty conditions unless the use includes them. Write the food type on the lab chain.

Why do labs bake the part again before LFGB volatiles? They shouldn't invent a post-cure you won't run. The specimen must already include production post-cure. If the lab adds a bake you can't repeat on recasts, the report is about a cleaner article than the one in the carton.

Is a COA from the RTV kit supplier the same as 177.2600 testing? The gum letter may support composition. 177.2600 extracts the finished rubber article. You still pour against the master, post-cure, clean, and test that cavity.

How often should an OEM retest a frozen food-contact RTV SKU? Retest when compound, pigment, post-cure, recast recipe, master revision, or claimed use changes, and when a customer quality agreement says so. An unchanged freeze still needs lot discipline. Don't treat a five-year-old sister color as current.