Flow marks are the wavy, faintly dull lines that trail behind the material as it fills a cavity, and weld lines are the seam left where two flow fronts meet and fail to knit. On molded silicone both are almost always a fill problem rather than a compound problem, and most of them respond to a change in gate position, injection speed, or mold temperature. Here is what each mark is telling you, how to separate a cosmetic line from a structural one, and what a molder should change first.
What Flow Marks and Weld Lines Actually Are
A flow mark is a surface artifact. The compound touching the cavity wall cures a fraction earlier than the material behind it, so the skin that forms against the steel never quite heals as the bulk passes over it. You see the boundary as a line, a wave, or a matte band that follows the direction of fill. In liquid silicone rubber the effect is slight because the material cures fast and stays viscous, but it shows up clearly on high-gloss surfaces, on transparent parts, and anywhere the flow front stalls.
A weld line is a different animal. Two flow fronts meet, usually because the cavity has a core pin, an insert, a rib, or more than one gate, and the surfaces of both fronts have already begun to cure. Polymer chains no longer interdiffuse across the interface, so the joint is mechanically weaker than the surrounding wall even when it looks perfect. Push a fingernail across it and you will often feel a notch you cannot see. On a part that flexes, that notch is a crack starter.
The two get used interchangeably on the shop floor, which is why defect reports are often useless. A line that fades out and reappears as the flow path widens is a flow mark. A line that runs from a pin or an insert toward the last point to fill is a weld line. The distinction matters because the cosmetic fix for one is often the structural fix for the other, and doing the cosmetic fix alone can leave a weak knit in a part that will fail in service.
Gate Position and the Fill Pattern
Where the mark starts tells you most of what you need to know. A mark that begins at the gate and fans outward is a gate problem: the gate is undersized, the shear rate through it is high, or the gate is freezing before the cavity packs. Enlarging the gate, or moving it to a non-cosmetic face, removes a surprising number of these lines without touching anything else.
A mark that appears at the far end of the cavity is a delivery problem. The front arrived late, lost heat on the way, and cured before it could merge with the material behind it. The instinct is to raise injection pressure. That usually makes it worse, because higher pressure packs the near end of the cavity first and starves the far end further. The useful move is a fill study: run the mold at partial fill, several shots at increasing shot size, and look at where the front actually stops. Molders call this a short shot series, and it is the cheapest diagnostic in the building.
Weld line position is decided by gate layout, not by pressure. If the knit lands on a cosmetic face, no amount of process tuning will move it, because it is where the last material meets. You either move the gate, add an overflow well at the meeting point and vent it so the front runs past the cosmetic zone, or accept the line and specify it. On multi-cavity tools, imbalance shows up as one cavity filling first and overpacking while the other starves; the starved cavity shows flow marks and short shots together, and the fix is runner balancing rather than a global process change.
Cold Slug, Injection Speed, and Mold Temperature
A cold slug is a plug of partially cured compound that enters the cavity ahead of the main charge and drags a comet tail behind it. It looks like a blob with a streak, not a line, and it is usually found near the gate or wherever the runner turns. In liquid silicone the slug commonly comes from the nozzle, a check valve, or compound that sat too long in a cooled runner. In compression and transfer molding it is a cold preform or a cold pot. Raising barrel or pot temperature to chase it is the wrong reflex; the right question is why the material was sitting there long enough to cure.
Injection speed cuts both ways, which is why the diagnosis has to come before the adjustment. Fill too slowly and the front loses heat, the skin cures ahead of the bulk, and the material folds over itself as it advances. Fill too fast and you overheat the gate, scorch the compound locally, and get a discolored mark that looks like a flow line but is actually thermal damage. If the mark is faint and surface-only, speed up the first stage and move the switchover point earlier. If the mark is dark or has a burnt edge, look at gate size and shear before you touch the profile.
Mold temperature is the variable buyers underestimate. A cold tool makes every flow mark deeper and more visible, because the skin sets earlier and the fronts knit less. A hot tool knits better but scorches more easily and lengthens cycle time. Uneven temperature across the plate is worse than either: one half of the cavity fills differently from the other, and you get a weld line on one side of a symmetric part and not the other. A surface probe across the cavity, taken at several points after the tool has stabilized, settles that argument in ten minutes. Do not trust the controller readout alone; it measures the cartridge, not the steel.
Cosmetic or Structural: How to Tell
The honest test is mechanical, not visual. If the line sits on a sealing lip, a diaphragm, a thin flexing wall, or a thread, assume it is structural until a tear or peel test says otherwise. If it sits on a flat cosmetic face that carries no load, it is a surface note. Visual inspection alone will pass parts that fail in the field, because a properly knitted weld line and a failed one look identical under a lamp.
| What you see | What it usually means | First thing to check |
|---|---|---|
| Line fades under a solvent wipe | Surface-only flow mark | Mold temperature and fill speed |
| Sharp notch felt with a fingernail | Incomplete knit at a weld line | Weld line position versus load path |
| Line only at thin sections | Front froze before the cavity filled | Gate size and first-stage speed |
| Blister or void beneath the line | Trapped gas or volatiles | Venting and vacuum |
| Dark or burnt edge on the line | Shear heating at the gate | Gate size and barrel temperature |
The fingernail test is crude but it is the one most drawing notes are written around, and it correlates well with whether a trim operation will later open the line into a tear. If the mark survives a fingernail drag, it is a surface discontinuity, and a surface discontinuity on a sealing face is a leak path.
What a Molder Changes, and in What Order
Change the cheap things first. Confirm the fill pattern with a short shot series so you are arguing about evidence rather than memory. Then raise mold temperature a few degrees and check uniformity across the plate, because that costs nothing but a cycle time adjustment. Then reshape the fill profile: faster first stage, earlier switchover, vacuum assist if the tool has it.
Only after those do you touch the steel. That means enlarging the gate, moving it to a face that will not be seen, adding an overflow well at the last-fill point and venting it properly, or adjusting a rib or wall transition that is forcing the front to split. Each of these is a tool modification with a lead time, and none of them should be authorized until the short shot series shows the fill pattern you think it shows.
Material is the last suspect and the most over-blamed. Addition-cure silicone does drift: an aged batch, a pot life that has run long, a platinum catalyst that is slightly off ratio, or a compound that has absorbed moisture all change how the front behaves. If two lots of the same compound run differently in the same tool on the same day, that is a material question. If every lot behaves the same way, it is the tool. A wider background on how these defects group together is in our guide to molded rubber defects and their solutions.
One note on clamp tonnage: buyers often ask whether more clamp force will fix a flow line. In liquid silicone injection it will not, because the line forms during fill and clamp force acts after the cavity is full. In compression molding, clamp force affects flash at the parting line, which is a different defect with a different fix.
What to Write on the Drawing and in the RFQ
Vague cosmetic requirements are the most common reason a good part gets rejected. "No flow marks" is not a specification; it is a wish, and it will be interpreted differently by every inspector who reads it. Write down the faces, the lighting, and the viewing distance.
A workable note reads something like this: no visible flow lines or weld lines on surfaces A and B as marked on the view; weld lines permitted on surface C provided there is no notch detectable by fingernail drag and no surface discontinuity visible at ten times magnification. Add the inspection condition, for example visual inspection at arm's length under ordinary shop lighting from a 45 degree angle, and name the surfaces on the drawing rather than describing them in words.
Gate location deserves its own line. Ask for gate position approval before first article, and specify which face the gate vestige may appear on. A molder who is free to put the gate anywhere will put it where the tool is cheapest to cut, which is frequently the cosmetic face. Approving the gate location up front prevents the argument later.
The reject line is simple. A flow mark that can be felt with a fingernail, or that crosses a sealing face or a flexing wall, is a reject. A faint surface line on a non-functional face that disappears under normal handling light is a cosmetic note, and it should be accepted rather than reworked, because every rework pass on silicone adds handling damage and cost without improving the part.
