Sand inclusions are fragments of mold or core sand trapped in the casting, nearly always at or just under the surface, and they get there in one of two ways: the sand was already loose in the cavity when the mold was closed, or the metal tore it off the mold or core during filling. Which of the two happened decides who at the foundry has to change something — the molding floor, the sand lab or the gating designer. So when a batch comes back with sand pockets, the useful question is not whether it was sand but where the sand came from.
What Is a Sand Inclusion in Casting
A sand inclusion is a piece of mold or core sand embedded in the metal, classed in the International Atlas of Casting Defects as G 131 and described there as “generally very close to the surface of the casting”. That last phrase is the part worth keeping: sand does not scatter through a section the way gas porosity does. It sits where the metal froze around it — under cope faces, at the top of bosses, and in the first cut a machinist takes.
The class it gets confused with is G 121, slag, dross and flux inclusions, which come out of the melt and the ladle rather than the mold. The distinction is not academic. Slag is a melting and pouring-practice problem; sand is a molding, sand-control or gating problem, and the corrective action lives in a different department.
A sand inclusion also has a twin on the casting surface, because the sand had to leave a hole somewhere. A wash at the gate, a rough raised patch where a scab lifted off the mold face, a crushed core print: whichever it is points back at the cause, which is why we read the outside of a casting before cutting into the inclusion.
What Do Sand Inclusions Look Like
On the casting, a sand inclusion is a rough, dull, irregular pocket at or just under the surface, often with sand grains still fused into it, and it favours cope faces, the tops of bosses and flanges, and the mold wall downstream of a gate. Sand is lighter than the metal, so whatever the stream picks up floats until it is caught against the cope or in the last metal to freeze. That is why the defect surfaces on the top of the casting as poured, and on the first machining cut of an upward-facing flange.
The surface twins have their own shapes. A wash is excess metal at the gate or under the sprue, filling the hollow the stream scoured out of the mold. A scab is a rough plate of metal lying on the surface with a layer of sand behind it, where a strip of mold face lifted before the metal covered it.
A drop is a rough projection on the cope face where a piece of the mold let go and fell onto the metal; a raised or crushed core shows as a rough projection where the core was. Inside the inclusion itself, a shiny or blue-gray lining is the tell.
On a radiograph, an inclusion is a localized light or dark spot, and ASTM E446 — the reference-radiograph set for steel castings up to 2 in. thick — grades it under Category B, “sand and slag inclusions”, at severity levels 1 through 5. Read that category name literally: radiography grades the inclusion, it does not say whether it is sand or slag. Telling them apart takes the surface twin, or the debris under a microscope or microprobe, so an RT report that calls the inclusion “sand” has made a judgment the film cannot support.
Causes of Sand Inclusions
Sand gets into a casting either because it was already loose in the cavity before the pour or because the metal pulled it off the mold or core during filling, and every row below is one of those two failures.
| Cause | How It Creates Sand Inclusions |
|---|---|
| Low mold strength | Mold surfaces break or erode during pouring |
| Poor core strength | Pieces of the core detach and become trapped |
| High metal velocity | Fast-flowing metal erodes the mold surface |
| Turbulent gating | Turbulence carries loosened sand into the casting |
| Improper ramming | Weak or uneven mold areas break during pouring |
| Excessive pouring temperature | Greater thermal stress damages the mold surface |
| Poor sand preparation | Incorrect moisture, binder, or grain distribution reduces mold stability |
| Mold handling damage | Loose sand remains in the mold cavity before pouring |
| Poor gating design | Metal directly impacts vulnerable mold surfaces |
| Core damage | Broken or cracked cores release particles |
| Inadequate mold cleaning | Loose particles remain inside the cavity |
Read the rows by which side of the plant they point at. Low mold strength, poor core strength, improper ramming, poor sand preparation and core damage are the mold side: the sand could not hold against metal a sound mold would have survived. High metal velocity, turbulent gating, poor gating design and excessive pouring temperature are the metal side: the mold was sound and the metal was let loose on it anyway. Mold handling damage and inadequate cleaning are housekeeping — sand already in the cavity when the cope went on.
On the metal side, the number that matters is how far the metal falls, not how fast the ladle tips. The critical-velocity work behind modern gating design puts the point where a filling front stops being quiet at about 0.5 m/s, and liquid metal reaches that after a free fall of only about 10 mm. Above it the front splashes and folds its own surface film in, and how much mold face it then scours depends on the strength of the sand it lands on. A top-gated pattern with a tall sprue keeps making inclusions however carefully it is poured, because the geometry sets the velocity.
Two rows are not monotonic, and a corrective action that treats them as “less is better” will not hold. Pouring too slowly leaves the mold face exposed to radiant heat before the metal covers it; the sand expands, lifts and scabs, and the scab is a sand inclusion too. Too fast erodes, too slow scabs, and the fix is the right velocity through a gating system that fills quietly. Likewise, raising green strength cures erosion but increases the scabbing tendency — a stiffer sand face cannot relieve its own expansion by yielding locally, so it buckles off as a plate.
How to Prevent Sand Inclusions
You prevent sand inclusions by making the mold and cores able to survive the pour, keeping the metal from attacking them, and keeping loose sand out of the cavity before it is closed — and a corrective action is only credible if it names which of the three failed and changes that one.
- Match the sand system to the mold size: Resin-bonded no-bake sand is far stronger than green sand, and on large iron castings — gearbox housings, counterweights, anything at the top of the weight range — it takes the low-strength and ramming rows out of play.
- Fill from the bottom, through a filter: Bottom gating gives a quiet front and a longer fill; a ceramic filter in the runner takes the turbulence out of the stream and catches sand that has already broken loose. A sprue that drops metal onto the drag face is the gating fault to look for first.
- Control the sand, not just the pour: On green sand the variables that keep proving decisive are green compression strength, moisture content and mold hardness. A foundry that tests them every shift and can show the log has the sand-preparation row covered; one that checks by feel does not.
- Handle cores and molds like finished parts: Cores set in the green state or not fully cured, a core print with no clearance so the core crushes on closing, a cope bumped on the way down — each puts sand in the cavity before any metal arrives. Clearance at the prints, careful fitting and a blown-out cavity before closing are the whole fix.
- Pour inside the window: Pouring temperature stays in the range the grade was qualified at. Hotter metal distorts cores and lifts scabs.
A credible corrective action names the surface twin it found — wash, scab, drop, broken core, loose sand — the side of the plant it points at, and the measurable change made there: a gating redesign with the new ingate velocity, a sand-lab limit tightened, a core-handling step added. “Improved process control” names nothing.
Conclusion
A sand inclusion is a near-surface pocket of mold or core sand, and it always has a source: sand loose in the cavity before the pour, a mold or core too weak for the metal it met, or metal allowed to hit the mold too hard or from too high. Match the inclusion to its twin on the surface and the cause narrows to one department. Judge the supplier’s corrective action by whether it names that department and changes something measurable there.