ISO 9001MSMEMade in India
Back to blog

Shrinkage porosity: why it always shows up at leak test

Illustration of jagged interdendritic shrinkage voids in a die-cast aluminium section

Shrinkage porosity is the defect that passes visual inspection and then fails you at the leak-test bench. The part looks sound, the surface is clean — and under pressure it weeps, and the reject goes back to the melt. To stop it, you have to accept that it isn't a casting flaw so much as a casting certainty that hasn't been controlled.

Where the voids come from

Molten aluminium is denser than solid aluminium. As the casting freezes it contracts by roughly 6–7% in volume. If liquid metal can flow in to feed that contraction while it happens, the casting stays dense. If it can't — because the path to it froze first — the shrinkage has to go somewhere, and it goes into voids.

That is why shrinkage porosity is not random. It appears in the regions that freeze last: thick sections, bosses, the meat behind a core pin, the junction where two walls meet. These hot spots stay liquid after the thin walls around them have solidified and sealed off the feed path. The last liquid to freeze has nothing feeding it, so it tears itself into a network of interdendritic voids.

Why it fails at leak test and not before

Shrinkage voids are jagged and interconnected — the opposite of a round gas pore. Individually they may be microscopic. But because they follow the interdendritic channels of the last-freezing metal, they link into a path. Machining opens one end of that path; pressure at leak test finds the other. A part can carry significant shrinkage and still look perfect until the moment it has to hold pressure across a wall.

The lever that actually moves it: the freeze sequence

You cannot make aluminium stop shrinking. What you can control is the order in which the casting freezes. The goal is directional solidification — the metal farthest from the feed source freezes first, and the freeze front marches back toward a source of liquid that stays open until last. Get the sequence right and the contraction is fed continuously; the void never forms.

Most of the time the hot spot is a place the die simply cannot pull heat out of fast enough. Drilled water lines run where the die can be drilled, not where the metal is thickest, so the thermal balance leaves the heavy section as the last thing to freeze — and unfed.

  • Targeted local cooling. Pulling heat out of the hot spot on a controlled schedule makes it freeze earlier, in sequence, instead of last and starved.
  • Gating and feed paths. The runner and gate have to keep the feed path liquid long enough to reach the section that needs it.
  • Thermal balance of the die. A hot spot is a cooling problem before it is a metallurgy problem — cooling that doesn't reach the mass can't feed it.

What to check on a shrinkage reject

Cut the part through the leak path. If the voids are jagged, branching and sitting in the thickest section or behind a pin, it's shrinkage, not gas. From there the questions are all about heat: where is the last-freezing region, why is it last, and what would it take to cool it on schedule instead of letting it dictate the schedule.

From defect to fix

Shrinkage is a cooling problem before it is a metallurgy problem — the hot spot has to be pulled out of the freeze sequence, on schedule. Explore Jet Cooling

Let’s cut your scrap rate

Tell us the defect.
We’ll bring the technology.

Send whatever you have — a drawing, a photo, a rejection report. An engineer, not a sales script, will come back to you with what we'd actually do.