Silicone overmoulding either bonds or it does not, and you can tell roughly 70 to 80 percent of the answer before the tool is ever cut. Whether the substrate is an engineering plastic such as PBT, PA66 or PPS, or something like PP or ABS, largely decides the outcome. Whether the primer was applied and then moulded within its pot life decides why one batch from the same tool tears apart while the next peels off with a fingernail.
The direct answer: there is no chemical bond between silicone and plastic. Thermoset silicone will not melt together with thermoplastic plastic. Bonding relies on a coupling agent at the interface plus mechanical interlocking in the geometry. Three layers decide it: the substrate, the interface treatment, and the structure. Miss one layer and you are left with friction alone - edge lift and delamination are just a matter of time.
Part 1: The Substrate - Some Plastics Simply Will Not Bond
Liquid silicone rubber (LSR) injection overmoulding runs a mould cavity temperature of typically 130-170 degrees Celsius, and the silicone cures in tens of seconds to a few minutes. The substrate must not deform, must not bleed additives, and must not release anything that poisons platinum curing.
PBT, PA66, PPS and PET are the conventional choices. PC also works, but a PC part softens around 150 degrees Celsius, and thin-wall parts tend to warp through stress release; we usually increase substrate wall thickness to 1.5 mm or more, or add support in the tool. PEEK works too, at a higher cost.
PP, PE, POM and PTFE should be abandoned outright. Their surface energy is too low, primer adhesion is poor, and there is essentially no production consistency. ABS should also be avoided: its temperature resistance is insufficient, and the additives in it poison the platinum catalyst, leaving the silicone surface tacky and yellowed, cured incompletely. That is a hard defect, not something parameter tuning can rescue.
There is also a category called self-adhesive LSR, whose grade already contains self-adhesive components with an affinity for PA and PBT, allowing overmoulding without primer. It sounds convenient, but its applicable range is narrow - switching to a different substrate grade may mean it no longer bonds. We prefer to treat it as "one process step fewer" rather than "one process step eliminated", because once bonding fails you no longer have an adjustable variable.
Verdict: usable substrates are PBT, PA66, PPS, PET, PC and PEEK. Avoid or treat with caution PP, PE, POM, PTFE, ABS, and any regrind loaded with low-molecular-weight additives. There is also no reliable chemical bond between TPU and silicone - pairing TPU with silicone in a two-shot design is a common material selection error.
Part 2: Three Structural Tasks - Wrap-Around, Through-Holes, Thickness Control
Primer solves adhesion over a surface. Geometry solves the problem at the edge. Silicone parts most often fail not on the flat face but at edges, corners and hole locations. Get those three wrong and the flat-face bond-test numbers will still look good while the part peels within two weeks of use.
Task 1 - the wrap-around must actually wrap. Do not let the silicone layer terminate abruptly on a flat plane; that leaves the entire boundary as a stress concentration. The correct approach is for the plastic substrate to step down or undercut at the termination point, so the silicone wraps around it and forms a mechanical locking edge. We typically leave a step depth of 0.5-1.0 mm and an undercut angle of 15-30 degrees. Too steep and the silicone will not fill; too shallow and it will not lock.
Task 2 - through-hole interlocking is the most reliable of all. Open a row of small holes in the substrate so the injected silicone penetrates through it and forms rivet points. This connection does not depend on primer at all. Hole diameter is conventionally 1.0-2.0 mm, and the hole edge must be rounded, between R0.2 and R0.5 - a square hole edge becomes the tear initiation point under tension. Leave enough wall thickness between hole and edge; do not weaken the substrate until it bends under pressure just to fit more holes. The drawback is that through-holes transmit light and leak water, so they cannot be used on visible cosmetic surfaces - only hidden internally or masked by structure.
Task 3 - do not make the silicone layer too thick. LSR moulding shrinkage is on the order of 1-3 percent, while PC is 0.5-0.7 percent and PBT is typically 0.6-1.4 percent. That shrinkage differential is fixed. The thicker the silicone, the greater the stress it accumulates as it shrinks, and the more likely it is to pull the bond interface open. Thin-wall overmoulding is generally held at 0.8-2.0 mm; beyond 3 mm you should reconsider whether the design really needs overmoulding at all.
A note on the parting line while we are here. Silicone will always migrate into the parting-line gap, and what squeezes out is flash. The conventional measure is to build 0.03-0.05 mm of compression into the parting line, or cut a 0.05-0.1 mm deep, 0.5-1 mm wide overflow groove so the excess material has somewhere to go. Do not leave a parting line on a visible surface - move it to the back or hide it in a decorative groove.
Part 3: Primer and Surface Treatment - Where Mass Production Usually Fails
Once the drawings go out, what actually sends yield off a cliff is usually not the structure but the invisible process steps on the shop floor. The primer is the most fragile link of all.
Primer is essentially a dilute solution containing a silane coupling agent. After application the solvent must evaporate first before the coupling agent reacts with the substrate surface. You cannot inject immediately after coating, and you cannot leave it too long either - there is typically a 4-8 hour pot life, after which reactivity drops and you must recoat. During those hours you must also keep out dust, finger oils and mould-release spray drifting across the shop floor. The most typical case we have seen: primed substrates left overnight in a transfer basket, then moulded the next day, came out with the entire batch delaminated. The trial-moulding parts in front of them had been moulded the same day, straight off the line, and were fine. Same drawings, same material - the only difference was those few hours.
The substrate surface itself must also be clean. Mould-release agent residue on injection-moulded parts is the most common contamination source. If you can avoid release agent, do so; if you must use it, add a cleaning or plasma step before overmoulding. Atmospheric-pressure plasma and flame treatment raise surface energy and are effective on materials such as PA and PBT, but the effect is time-limited - leave the treated part too long and it decays back, so treated parts should enter the tool as soon as possible.
None of this process control can be achieved through drawings. It must be written into the work instruction: surface treatment method, permitted dwell time after treatment, primer grade and dilution ratio, waiting time after coating, cleanliness requirements for transfer containers. The bond quality of an overmoulded part depends half on design and half on whether those few lines were genuinely followed.
How to verify bonding: at the sampling stage, run at least two rounds. The first is dry peel, following GB/T 2791 for a 90-degree peel, examining whether the failure surface occurs in the silicone bulk, at the interface, or in the primer layer - the failure location carries more information than the number. The second is peel after ageing, tested after 85 degrees Celsius / 85 percent RH or 70 degrees Celsius hot-water soak. Many formulations look good dry and peel off in sheets once they see humidity and heat.
Part 4: The Process Boundary
One final note on where the process ends. Overmoulding is a permanent connection - once bonded it cannot be separated, and repair means replacing the whole part. It is also unsuitable for areas that must undergo repeated large-range stretching; interface peel strength is measured in N/cm and will not survive being rolled and torn. If you genuinely need a silicone lid that can be lifted open, that is a different approach entirely, relying on snap-fits and sealing ribs rather than bonding.
About Hezi Industrial Design: Dongguan Hezi Industrial Design Co., Ltd. provides one-stop product appearance design, structural design and CMF design services, with 320+ successfully launched products. Contact: +86 18576718657 | www.hezidesign.com | Dongguan, Guangdong, China.
Top comments (0)