Food-Grade Certification Guide for Custom RTV Molds

If you're collecting “certification” for a custom RTV food mold, what you actually need is a lab report on a defined cavity. It isn't a hologram on a drum, isn't a random number on a carton, and isn't FDA product approval of a baking or chocolate mold. The US rubber rule most buyers mean is 21 CFR 177.2600: composition constraints plus extraction of the finished repeated-use article in water and n-hexane. The German/EU path is LFGB, methodized for silicone largely through BfR Recommendation XV, under Regulation (EC) 1935/2004.

Platinum addition-cure RTV is typical for the articles that pass those files. BPA-free isn't a certification, because silicone isn't polycarbonate. If the PDF describes an LSR injection *part* with a cold-runner, it describes a different product.

Common mix-up: a certificate number, or a DoC with no specimen, is the file

A lot of RFQs treat any PDF with “FDA” or “LFGB” in the filename as certification. In the shop that usually means sales attaches a spatula report, or invents a number to fill a retailer portal field, and the cavity you recast was never in a flask.

A usable document is issued by a laboratory, names a method, names a specimen, and states pass or fail against the method's criteria. Specimen fields that have to exist: this RTV mold (not “silicone products”), color / pigment, process (platinum RTV), thickness of the skin or a note that production post-cure was applied, master revision and whether the jacket shipped with the specimen, date, lab identity, method identity. Don't accept “FDA approved” as the title. Rewrite the claim: finished rubber article tested to 21 CFR 177.2600 for repeated use with named food types.

What to ask when you're buying the certification pack

• Is this a lab report on *this* SKU, this color, this platinum RTV recast — or a marketing PDF / sister file?
• Which method: 21 CFR 177.2600 (water and/or n-hexane), LFGB with BfR XV (volatiles, simulants, sensory), or both?
• Can they trace any certificate number to a lab and a method, or is it decoration?
• For EU, is there a Declaration of Compliance that *points at* the specimen report, not a DoC alone?
• What's the change-control rule when pigment, post-cure, or master revision changes?
• Where do ISO 9001, RoHS, BPA-free, and ISO 10993 live (other folders), so they don't get stapled on as food-contact proof?

United States and German/EU documents, unmixed

177.2600 isn't a registration of your muffin pan. It's extraction of the food-contact surface of the finished article. Aqueous: distilled water at reflux, 20 mg/in² in the first 7 hours, 1 mg/in² in the succeeding 2 hours. Fatty: n-hexane at reflux, 175 mg/in² then 4 mg/in². Dry food is framed under GMP suitability. The article shall be thoroughly cleansed before food contact. A gum supplier letter can support formulation. It doesn't replace extraction of the poured cavity.

CPSIA and California Proposition 65 statements are separate conversations. They aren't 177.2600.

Regulation (EC) 1935/2004 is the EU framework. Regulation (EC) 2023/2006 is GMP for food-contact materials. LFGB is Germany's statute. BfR Recommendation XV is the silicone-specific recommendation labs use — confirm the current text. Typical report contents: composition thinking, volatile organic matter, simulant tests (water, 3% acetic acid, 95% ethanol are common), sensory. Time and temperature have to match the claimed use of this cavity. A DoC should point at those supporting tests. If there's no specimen report, the DoC is a letter.

Throw these out of the certification folder, or file them elsewhere: BPA-free logos, ISO 9001, ISO 10993 or USP Class VI, UL 94, RoHS / REACH without a food-contact method, “platinum-cured” on a website, a sister SKU after you changed pigment, tin-cure craft certificates offered as bakeware, an LSR part-tool report offered as this RTV cavity. LFGB doesn't include 177.2600. Dual-market certification is two reports on the same freeze.

Worked example: filing a 10-cavity cookie-stamp mold

A bakery ordered a 10-cavity cookie stamp, cavities about 50 mm, platinum RTV one-piece with a rigid jacket, 200–500 stamps a month, US retail plus a German specialty account.

Put reports in a folder named by SKU and revision, not in a shared “FDA” heap. The US file is 177.2600 on that stamp (food types as used: often fatty and dry). The German file is BfR XV / LFGB on the same recast recipe, plus a DoC that points at that specimen ID. When sales wants a PDF for a retailer portal, they pull that SKU only. Version the report when the master is recut or the pigment changes. Obsolete files get a watermark or a filename prefix, not deletion, because auditors ask why you retested.

Retail portals that demand a certificate number field get the lab report ID from the real PDF, or they get “not a product registration; see 21 CFR 177.2600 report attached.” They don't get a number you invented to fill a box.

On the RFQ, list markets, food types, use temperatures, process (platinum RTV), color, post-cure, and master ownership. Ask the supplier to attach matching reports or to budget first-article testing.

When this article is the wrong quote

If the mold never touches food, don't build a 177.2600 folder. If you need ISO 10993 or USP Class VI, that's medical, not bakeware. If the retailer only wants a US ice-tray file, don't buy an oven LFGB kitchenware suite. A Declaration of Compliance without a specimen report isn't a skip; a 2 mm plaque isn't the 12 mm pan unless purchasing said representative *and* thickness matches. Geometry can override the coupon.

Related reading: FDA vs LFGB, why “FDA approved” fails, how the cavity is tested, food-grade molds.

What to send so the certification folder matches the SKU

Copy this. A heap named “FDA” isn't comparable.

• SKU, color, platinum RTV, thickness, master revision, jacket as shipped
• Markets and methods: 21 CFR 177.2600 food types; LFGB/BfR XV volatiles, simulants, sensory
• Lab identity, method, date, pass/fail — no invented certificate numbers
• EU: DoC that *points at* that specimen ID, plus 2023/2006 GMP language if asked
• Change control: pigment, post-cure, master recut → retest
• Folder rule: this SKU only; ISO 9001 / RoHS / BPA-free live elsewhere

Send the current PDF (or admit there isn't one) plus the SKU freeze. We'll tell you if it's a real article report, a sister spatula file, or a number invented for a portal.

We would not recommend a portal number you invented

We would not recommend a DoC with no specimen, ISO 9001 stapled as 177.2600, or one report covering every color in the catalog. We would not title the page "FDA approved."

flowchart TD
  D[Document in the folder] --> Q{Names specimen + method + pass/fail?}
  Q -->|no| N[Not a certificate — marketing PDF]
  Q -->|yes| M{Which method?}
  M -->|177.2600| U[US rubber article extraction]
  M -->|BfR XV / LFGB| G[Volatiles, simulants, sensory]
  U --> F[File by SKU + revision]
  G --> F
  F --> C[Retest on pigment, post-cure, or master recut]

Typical shop values: what a usable food-contact file names

| Item | Typical shop value | | --- | --- | | Specimen | This RTV mold, color, thickness, master revision | | US aqueous / fatty | 20/1 mg/in² water; 175/4 mg/in² n-hexane | | BfR XV simulants | Water, 3% acetic acid, 95% ethanol + sensory | | Chemistry | Platinum RTV 1:1 or 10:1; Shore 10A–20A | | Not in this folder | ISO 9001, ISO 10993, BPA-free, invented cert numbers |

Caption these as typical shop values.

FAQ

Is a Declaration of Compliance enough to skip BfR XV testing on this RTV mold? A DoC should point at supporting tests on the article. If there's no specimen report, the DoC is a letter. Ask for both.

Does 21 CFR 177.2600 give me an FDA registration number for the mold? There's no product registration of a baking or chocolate mold under that section. Don't print a fake number. Keep the extraction report.

Can ISO 10993 stand in for food-grade certification? ISO 10993 is medical device biocompatibility. Food-contact rubber in the US is 177.2600; German/EU silicone kitchenware is LFGB/BfR XV. Don't mix the folders.

If I only change PMS color on a recast, do I recertify? Yes if the pigment or masterbatch changes. Color is part of the article. A natural gum pass doesn't cover a red cavity.

Should I publish certificate numbers on the product page? Publish the regulation name, the food types, and that the finished article was tested. Publish a real lab report under NDA if needed. Don't invent numbers. Don't title the page “FDA approved.”