Custom Silicone Food Storage & Snack Bags – Leak‑Proof, BPA‑Free, Reusable

The Leak Rating Is Decided by the Closure, Not by the Silicone

Two silicone snack bags can be moulded from the same 60 Shore A compound and still behave differently in the first week: one carries soup through a dropped bag, the other weeps along the zip rail as soon as the pouch is squeezed. The variable that separates them is not the durometer on the compound sheet. It is the closure geometry, the cure system, and how the film behaves where the bag folds. Those three items are what we quote against, and what this page documents.

1. Cure system before durometer. Addition-cure (platinum) silicone leaves no peroxide decomposition byproducts, so it is the default for fatty foods and for anything that will be chewed or sucked. Peroxide cure is cheaper and more forgiving in a busy workshop, but its residues are a common source of taste transfer. Platinum cure also demands a clean cell: sulfur, amines and tin catalysts poison the catalyst, and a mould that previously ran a peroxide compound will contaminate the batch that follows.
2. Seal on a land, not on a point. A press-to-close rail that meets on a sharp edge holds until the first overfill, takes a permanent set, and then leaks. We place a flat sealing land behind the interlock so closing pressure is distributed along a strip. The interlock keeps the bag shut; the land keeps it leak-proof. Land width, undercut and interference are tooling dimensions, so changing them later means a new insert rather than a new setting.
3. Film behaviour at the gusset. A stand-up pouch stands because the base gusset folds on a controlled line. If the gusset film is thinner than the side wall, the pouch slumps when filled and the corner radius tears first. Wall thickness is therefore quoted per format, and the corner radius is kept larger than the wall thickness so the tear path has somewhere to go.

Where these products actually fail in service

Field returns cluster into five patterns. Each one traces back to a drawing decision, not to a material batch.

• Rail creep. Sealing force drops after a few hundred open-close cycles because the interlock deforms. Fix at design stage: land width, a low compression set compound, and a cycle test on the agreed protocol before tool release.
• Corner tear at the gusset. A sharp internal radius combined with thin film starts a tear that runs along the fold. Fix: radius larger than wall thickness, and a thickness step rather than a long taper.
• Print wear. Screen ink abrades in a dishwasher and fades first at the fold. A moulded-in deboss survives washing; two-pack ink is reserved for artwork that needs more than one colour.
• Hinge whitening. Folding on the same line works the surface and turns it white. Move the fold away from the weld line and keep the notch radius generous.
• Odour retention. Porous surfaces and peroxide residues hold food smells. Platinum cure plus a defined post-cure cycle removes most of it.

How the tool and the batch are controlled

• Mould flow at the gusset and rail. Thin sections fill last, so fill pressure and weld-line position are checked before steel is cut. A rail that sits on a weld line leaks in the field regardless of compound.
• Retention instead of adhesion. Stiffening ribs and any inserted features are moulded in or mechanically retained. Glued joints creep under repeated load and fail at the seam.
• Segregated cure cells. Platinum and peroxide compounds are moulded and post-cured apart. A shared oven or a shared mould is the quickest route to a batch of tacky parts.
• Lot record. Compound lot, cure date, post-cure cycle, hardness per cavity and a dimensional sample are recorded per lot, so a complaint can be traced to a specific cure.

For the sealing side of the same problem, our note on silicone gaskets for food containers covers land width and lid preload in more depth. Buyers comparing formats also review stretchable silicone storage lids, which use a different seal principle again.

What we need to quote, and what comes back

• Three inputs. Fill volume or a physical sample, the intended closure style, and the market whose food-contact rules must be met. A 3D file shortens the loop, but a hand sample is enough to start.
• What comes back. A DFM note covering wall thickness, gusset radius, rail interference and draft, a material shortlist with the cure system named, and a tooling and unit-cost split.
• Sample stage. T1 samples come from production tooling rather than a machined prototype, so the rail and the gusset behave as they will in volume.
• Volume ramp. Cavity count follows annual volume: a 2-cavity tool is often right for a brand test, an 8-cavity tool for a retail programme.