Plenty of plants can produce a sample. Very few can hold it in volume. The difference is not the press — it is how the interface, the temperature and the time are understood.
Of all liquid silicone applications, overmolding is the one that most reliably looks simple and turns out not to be. A customer arrives with a PP or PPSU molding and asks for “a layer of silicone on the outside”. It sounds like one extra process step. It is actually a complete interfacial engineering problem.
This article is about PP only — because PP is the hardest case, and because it is the material we most often inherit after someone else's project stalled.
Bonding silicone to a plastic substrate relies on a combination of mechanical anchorage and chemical interaction at the interface. PP is poorly suited to both:
So whether PP overmolding works has little to do with press tonnage. It comes down to how the interface is modified and how the stress is released.
The most direct symptom is no bond at all between silicone and PP — the layer lifts in one piece under a fingernail or a blade. These cases almost always trace back to the interface: the substrate surface was never treated, was over-treated, or was contaminated after treatment (dust, mold release residue, fingerprints); or the compound was never validated against that specific substrate grade.
Silicone releases trace gases as it cures in the cavity, and heated PP can also evolve volatiles. Inadequate venting, excessive injection speed and uneven mold temperature all produce bubbles or silver streaking at the overmold edge. These defects tend to be intermittent — which makes them the ones that appear at random in a production run.
The worst category. The part looks perfect at demolding, then the interface begins to peel along stress concentrations after assembly, drop testing, sterilization or thermal cycling. The root cause is usually an abrupt wall-thickness transition, poor ejection force distribution, or residual stress left behind when cure timing and PP crystallization are out of step.
A good trial shot means you passed one exam. Volume production is that exam, ten thousand times, with the same pass mark every time.
PP melts around 160–170°C, so a 120°C mold sounds comfortable. But melting point is not service temperature. What actually constrains overmolding is the heat deflection temperature: for homopolymer PP under normal load it typically sits in the 90–105°C band. Boiling water at 100°C is therefore already at the edge of its safe envelope — not comfortably below it.
Platinum-cure LSR, on the other hand, needs mold temperatures above 120°C to crosslink fully. PP overmolding is caught between those two numbers from the very first shot:
The usable window is a narrow slot between those two temperatures. It cannot be widened by running hotter and faster — heat deforms the substrate, and longer cycles kill throughput. It is opened with zoned mold temperature, staged cure and geometry that gives the residual stress somewhere to go.
Which is why PP overmolding failures are often not delamination at all, but it will not seat or it warps after a few sterilizations. Neither of those shows up on a trial-shot report.
Interface treatment is only the entry ticket. What decides whether the part can run at volume is three matching relationships that look identical on a data sheet and are not.
The conclusion: you do not pick a silicone that "sticks". You select it against the substrate grade, the wall thickness and the service temperature together.
Homopolymer, copolymer, talc-filled, colour masterbatch, internal mold release — surface state and crystallization behaviour all differ. Change the grade under the same parameters and adhesion can drop across an entire lot. When we take on a PP program, the first thing we lock is often not the drawing but whether the substrate grade can be fixed. If the customer switches material mid-project, the process is rebuilt from the start.
PP's surface energy is too low, so the surface is usually opened first with a primer (a chlorinated polyolefin type) or with plasma or flame treatment, before silicone can grab. Three common traps:
Primer selection, coat weight and drying conditions are variables to be tuned individually, exactly like cure parameters. It is not something you buy, brush on and forget.
Infant products are sterilized repeatedly. Bottles, training cups, teethers and soft spoons go through dozens of boiling-water or steam cycles in a single service life. The overmold interface therefore does not face a one-off bond-strength test — it faces repeated thermal shock, hydrolysis and fatigue.
Our standard validation is: boil for 5 minutes → remove and cool → boil again, for 70 cycles.
That test applies three stresses at once, and all three matter:
In practice a marginal bond rarely fails on cycle one — it typically starts lifting, whitening and weeping from the edge somewhere after the twentieth or thirtieth cycle. Testing only initial adhesion skips exactly the part of the service life the customer will live through.
Our criterion: after 70 cycles, no delamination, no blistering, no separation, with appearance and critical dimensions still within specification.
“Can it be made” is a question about the trial shot. “Can it be held” is a question about cycle 70.
The previous sections explain why it is hard. This one explains how we made it work — because in PP overmolding, understanding the method is not the same as closing the distance to the floor. That distance is paid in time and failed attempts.
We spent 18 months on this process, in four stages:
What came out is a fixed recipe in which substrate grade, silicone, primer, process window and mold structure are bound to one another. Change any one of them and the whole thing has to be validated again — which is why we tell customers that switching material mid-project means redoing the project.
What those 18 months bought is not the ability to make the part. It is knowing exactly which conditions break it. The first can be luck on a single trial shot; the second only accumulates by rejecting combination after combination.
One boundary worth stating: the specific grades, the primer we settled on and the parameter settings belong to our process file and are not published here. What we do publish is the judgement criteria and the test method — because those are what you actually use when screening a supplier. Recipes and windows belong with the people who earned them.
Across the PP overmolding programs we have held in long-term production, four conditions have to be true at the same time:
They are all called “overmolding”, but the difficulty varies enormously. This is our working judgement from actual projects — qualitative, and always subject to grade and geometry:
| Substrate | Surface polarity | Difficulty | Typical approach | Our status |
|---|---|---|---|---|
| PPSU | Medium | Medium | Surface treatment plus matched process window | Mature at volume |
| PA (nylon) | Relatively high | Low–medium | Control moisture content and mold temperature | Mature at volume |
| PC | Medium | Medium | Prevent stress cracking; control drying | Mature at volume |
| Stainless steel / metal | — | Medium | Surface cleaning plus mechanical anchorage | Mature at volume |
| PP | Low, non-polar | High | Interface modification + dedicated process window + aging validation | Stable volume production |
Misconception one: passing the trial shot means it can be made. A trial run happens on the best machine state, the shortest flow path and the fewest people. Volume production meets shift changes, new material lots, tool wear and delivery pressure.
Misconception two: parameters can be copied. Applying a PPSU window to PP may produce parts in the short term. Long term it shows up as adhesion loss or warpage — the two materials do not even share the same temperature ceiling.
Misconception three: initial adhesion is the metric. Aged data is what the customer will actually live with — especially on infant and medical products that are sterilized repeatedly. A sample that has never been through a thermal cycle has not been validated.
Misconception four: the only risk is that it will not stick. The more common PP failure is it bonded but it moved — dimensions out of tolerance at demolding, parts that will not seat, warpage after sterilization. Those trace back to the temperature window and the geometry, not to more primer.
Overmolding is not the only answer. Where the geometry allows, a snap-fit two-piece assembly, two-shot injection molding or a structured insert may be cheaper and lower risk. We put those alternatives on the table during the DFM stage — rather than committing to overmolding first and explaining later why the yield will not climb.
PP + silicone overmolding remains an advanced process. Plants that can produce a sample are not rare; plants that can hold it in volume are. We treat this process as our core moat, and we are willing to publish the criteria openly — because a customer who genuinely needs it will finish this article knowing exactly what to ask a supplier.
If you are shortlisting suppliers right now, three questions will separate most of them out:
If your project is stuck at this step, send the geometry, the service environment and the validation requirements. We will assess feasibility first, then discuss quoting.
2D / 3D drawings, a sample or clear photographs are all we need to begin. The review is free, and the conclusion will state plainly: can be molded / geometry needs changing / overmolding is not recommended.