Casting Defects: Porosity, Shrinkage and Gas Defects in Aluminum and Iron Castings

Casting defects are the most expensive thing in a foundry that nobody budgets for. A scrapped mold costs you the sand, the melt, the labour and the delivery slot. A defect that escapes to the customer costs considerably more.
This guide is a diagnostic reference. It covers what each defect looks like, where it appears on the part, what actually causes it, and what to change to stop it recurring. It deals with defects in the casting itself, not in the mold printing process; if you are looking for defects specific to 3D printed sand molds and cores, our 3D sand printing defects guide covers those separately.

The Direct Answer
Most casting defects trace back to four root causes:
- Filling (gating and pouring): the metal did not enter the cavity cleanly, quietly and completely.
- Solidification (feeding and cooling): the metal did not shrink into a feeder, so it shrank into a void.
- Gas (melting, mold and binder): gas dissolved in the melt or generated by the mold ended up trapped in the part.
- Mold (sand, binder, coating and compaction): the mold itself broke down, eroded or reacted with the metal.
Classify the defect into one of those four families first, then work down to the specific mechanism. Treating a gas problem as a shrinkage problem, or the reverse, is the single most common reason a defect keeps coming back.
How to Diagnose a Defect: Start With Location
Before you look at the defect itself, write down where it is.
| Where it appears | Most likely family |
|---|---|
| Top surface of the casting (cope side), under the skin | Gas, from mold or binder |
| At a thick section, hot spot or isolated heavy boss | Shrinkage |
| Scattered through the whole section, uniform in size | Gas dissolved in the melt |
| At the ingate, or in line with the metal flow path | Filling and erosion |
| On vertical faces or sharp internal corners | Mold breakdown, penetration |
| At a change of section, or where a thin web meets a heavy wall | Hot tearing |
Location narrows the field faster than any laboratory test. A gas pore and a shrinkage pore can look superficially similar in a photograph, but they rarely occur in the same place on a part.

Porosity: Gas vs Shrinkage
Porosity is the umbrella term, and it is the defect that causes the most misdiagnosis because the two main types need opposite corrections. Improving feeding will not fix gas porosity. Reducing gas will not fix shrinkage.
Gas Porosity
Appearance: rounded or spherical pores with smooth, often bright internal walls. Frequently grouped, frequently subsurface, and often uniform in size within a cluster.
Where: concentrated under the cope surface, at the top of a heavy section, or scattered through the section if the gas came from the melt.
Causes:
- Hydrogen dissolved in the melt, which comes out of solution as the metal solidifies. This dominates in aluminium alloys.
- Gas generated by the mold, core or coating: moisture, unburned binder, or decomposition products.
- Turbulence during filling that entrains air and folds it into the metal stream.
What to change:
- Degas the melt and verify with a reduced pressure or density test rather than by eye.
- Dry and vent the mold and cores. Control binder addition and cure.
- Lower the pouring temperature within the alloy’s acceptable range, and slow the fill so the stream stays coherent.
- Redesign the gating so metal enters without free fall, splash or vortexing.
Shrinkage Porosity
Appearance: irregular, jagged, often elongated voids with rough, dendritic internal walls. Sometimes interconnected, sometimes visible only on a section or radiograph.
Where: at thermal centres, that is, the last regions to solidify. Thick sections, isolated bosses, junctions between heavy and light walls, and the region below a riser that stopped feeding too early.
Causes:
- Inadequate feeding: the riser solidified before the casting section it was meant to feed.
- Riser placed where it cannot feed the hot spot, or a feeding path blocked by a thin section.
- Section thickness changes that create an isolated thermal centre with no feed path.
- Pouring temperature too low, so the feeding path froze early.
What to change:
- Design for directional solidification so the last metal to freeze sits in a feeder.
- Move or enlarge the feeder, and check that the feed path stays open.
- Add chills to accelerate cooling at heavy sections and shift the hot spot toward the feeder.
- Where geometry allows, reduce the section change that created the hot spot.
How to Tell Them Apart
| Test | Porosité gazeuse | Porosité de retrait |
|---|---|---|
| Pore shape | Round, spherical | Irregular, angular, branching |
| Internal wall | Smooth, often bright | Rough, dendritic, dull |
| Distribution | Grouped, subsurface, often uniform | Localised at thermal centres |
| Location | Cope side, top of heavy sections | Hot spots, thick-to-thin junctions |
| Response to more feeding | No change | Usually resolves |
| Response to better degassing | Usually resolves | No change |
When a section is ambiguous, a micrograph settles it. Gas pores have smooth walls and a rounded profile; shrinkage voids have a characteristic jagged, inter dendritic shape.
Gas Defects: Blowholes, Pinholes and Subsurface Blowholes
These are gas defects at specific scales and locations, and the fix depends on which one you have.
Blowholes are larger, roughly spherical cavities, typically 1 mm and above, often found just below the casting surface on the cope side. The usual source is mold or core gas: moisture in the sand, insufficient venting, or binder that decomposes too vigorously. Lower the moisture, improve venting and check the binder level.
Pinholes are small gas pores, often below 1 mm, and they frequently break the surface. In aluminium they are commonly hydrogen related. In iron they can come from reaction between the metal and moisture in the coating or sand. Degassing and drying matter more than gating here.
Subsurface blowholes sit entirely under a thin skin of sound metal. They are the ones that pass visual inspection and fail at machining, which makes them the most expensive of the three. Radiography or sectioning is the only reliable way to find them.
A practical sequence for any gas defect:
- Check melt quality first. Degas, and measure rather than assume.
- Check moisture and binder in the sand and cores.
- Check vents. A blocked vent turns a minor gas problem into a scrap rate.
- Check the fill. Turbulence entrains gas that no amount of degassing will remove.
Shrinkage Defects: Cavities, Sink Marks and Hot Tears
Shrinkage cavity is macro shrinkage: a visible internal void, usually below a feeder. It means the feeder ran dry, or the feed path closed. Increase feeder volume, or improve the feed path.
Sink mark is the surface expression of the same problem: a depression on a heavy section where the solidifying skin was drawn inward. It appears when the section is too heavy to feed and too close to the surface to hide the result.
Hot tear, also called hot cracking, is different in kind. It is a crack that opens while the metal is still partly liquid and therefore very weak, usually at a change of section where contraction is restrained. It is not a feeding defect and adding feeders will not cure it. Look at:
- Restraint: does the mold or core geometry prevent the casting from contracting?
- Alloy freezing range: wide-freezing-range alloys are more susceptible.
- Section changes: sharp transitions concentrate strain.
- Pouring temperature and mold hardness: a very hard mold increases restraint.
- Trace elements in some alloy systems, which can dramatically increase susceptibility.
Hot tears usually require a geometry or alloy change rather than a process tweak. If the design cannot change, a more collapsible core or a modified mold coating is often the practical route.
Sand-Related Defects
These come from the mold rather than from the metal, and they are the family most improved by switching from conventional tooling to 3D printed molds, because printed molds eliminate the parting line and the pattern draw.
Sand Inclusion
Loose sand or a lump of mold material ends up inside the casting.
Cause: weak mold or core surfaces, erosion at the ingate, broken core edges, or sand washed off by a turbulent fill.
Fix: improve mold and core strength, reduce pouring velocity, redesign the gating so metal does not impinge on a sand surface, and handle cores more carefully between printing and closing.
Burn-On and Burn-In
Metal penetrates the sand surface, producing a rough, welded layer that is difficult to clean.
Cause: pouring temperature too high for the sand, insufficient or poorly applied refractory coating, sand with low refractoriness, or low compaction.
Fix: lower the pour temperature, improve coating coverage and thickness, and select a sand with better refractoriness for the alloy and section thickness. Chromite and ceramic sands are used where silica alone is marginal.
Scab and Rat Tail
Both are expansion defects. The sand surface expands when the hot metal touches it, and the surface layer buckles.
- Scab: a flake of sand lifts and metal flows behind it, leaving a raised, rough patch.
- Rat tail: a vein-like ridge across the surface where the sand surface buckled without breaking through.
Fix: reduce the sand’s thermal expansion effect by adjusting the sand system (grain shape, additives, and for silica sand, the amount of fines), and avoid excessive compaction. Where the geometry is difficult, a printed mold can be designed without the sharp internal corners that concentrate the expansion stress.
Sand Wash
Metal flow erodes the mold surface, leaving a washed-out depression and usually a sand inclusion downstream.
Fix: reduce fill velocity, move the ingate so the stream does not impinge on a sand face, and increase surface strength at the ingate.

Green Sand Specific Defects and Remedies
Green sand is the workhorse process, and most of its defects trace back to the condition of the sand system rather than to the operator.
| Control | What goes wrong | Defect it produces |
|---|---|---|
| Moisture content | Too wet: gas and blows. Too dry: weak edges and erosion | Blowholes, sand inclusion |
| Compactability | Too high: expansion defects. Too low: soft mold, penetration | Scab, rat tail, burn-on |
| Green strength | Too low: mold walls break, edges crumble | Sand inclusion, mold collapse |
| Permeability | Too low: gas cannot escape | Blowholes, subsurface blowholes |
| Molding sand temperature | Too hot: moisture loss, poor compaction | Erosion, weak mold |
| Fines and grain distribution | Excess fines: low permeability, high expansion | Blowholes, scab |
| Sand to metal ratio and reclamation | High dead clay and dust: everything degrades | Multiple, poorly repeatable |
The important point is that these interact. Increasing compactability to reduce penetration raises the risk of expansion defects, so the sand system has to be tuned as a system, not by adjusting one number.
Practical sequence for a green sand defect:
- Measure the sand. Moisture, compactability, green strength, permeability and loss on ignition, taken at the molding station, not from a sample left in the lab.
- Compare against the recorded values that produced good castings, not against a generic specification.
- Change one variable at a time and keep the melt and the pattern constant.
- Record the result. A green sand defect that comes back is usually a control that drifted, not a new problem.
Iron-Specific Defects in Green Sand
For grey and ductile iron, several defects are specific to the metallurgy.
Slag inclusion. Non-metallic slag enters the mold with the metal. It appears as irregular dark inclusions, often at the cope surface or trapped under overhangs. Fix by improving slag removal: proper slagging, a well-designed pouring basin, a filter in the gating system, and adequate settling time before pouring.
Graphite flotation. In hypereutectic ductile iron, graphite nodules float, producing a layer of excess graphite in the upper part of a heavy section. It shows as areas of reduced strength and poor machined finish. Address it through carbon equivalent and pouring temperature, and by avoiding long solidification times in heavy sections.
Nodularity problems in ductile iron. Low nodule count or degenerate graphite from insufficient magnesium, poor inoculation, or excessive holding time reduces mechanical properties. It is verified by micrograph rather than by visual inspection.
Chill or white iron formation. In thin sections and at the mold face, rapid cooling can produce hard, brittle white iron instead of grey iron. It causes machining problems rather than visible defects. Address through inoculation practice and, in thin sections, by controlling cooling rate.
Aluminum-Specific Defects and Troubleshooting
Aluminium is the alloy family where gas porosity dominates, because hydrogen solubility drops sharply as the metal solidifies. That makes degassing the highest-value habit in an aluminium foundry.
Hydrogen pinholes. Fine, dispersed gas pores, often breaking the surface. They come from hydrogen in the melt, usually from damp charge material, a damp refractory, tools that were not preheated, or prolonged holding in a humid atmosphere. Degas with an inert gas or a rotary degasser, and verify the result. A reduced pressure test is a five-minute check that prevents a lot of scrap.
Oxide films and entrained oxide. Aluminium forms an oxide skin instantly, and turbulence folds that skin into the metal as a planar defect that reduces fatigue strength without showing on a radiograph reliably. Clean, non-turbulent filling matters more in aluminium than in most alloys. So does a properly skimmed melt and a filter in the gating system.
Microshrinkage in aluminium. Aluminium alloys with a wide freezing range tend to form dispersed microshrinkage in thick sections rather than a single pipe. It is difficult to feed out and is usually managed by chilling and by modifying the casting geometry.
Cold shut and misrun. Two streams meet without fusing, or the cavity does not fill, because the metal was too cold or the fill too slow. Raise the pouring temperature within the alloy’s range and increase fill speed, while watching that the higher temperature does not increase gas or penetration.
Modification and grain refinement fade. Sodium modification in Al-Si alloys fades with holding time. If properties drift over a shift, check the time between treatment and pouring before you change anything else.
Mold and Core Defects in 3D Printed Sand Molds
3D printed molds and cores eliminate several conventional defects by removing the causes rather than managing them:
- No parting line and no pattern draw, so no draft-related dimension error and no draw-induced mold damage.
- No ramming, so no compaction variation between one mold and the next.
- Designed-in venting, so gas from the binder and from the metal has a planned path out.
- Free placement of chills and of core prints, because geometry is not constrained by a pattern.
But printing introduces its own failure modes, and they are worth knowing because a printed mold defect reads differently from a green sand defect.
Layer-related surface texture. Printed mold surfaces carry a layer texture. On a casting this can show as a faint stepped pattern. It matters most where surface finish is specified, and it is managed through print orientation, layer thickness and coating.
Veining. Thermal shock cracks the printed mold surface and metal enters the cracks, producing thin fins on the casting. It correlates with binder chemistry, print orientation and pouring temperature. Our 3D sand printing defects guide covers the mold-side mechanisms in detail.
Incomplete cure or residual binder. If the printed mold is not fully cured or not adequately post-processed, it can generate more gas than a conventional mold. Verify the print and cure parameters before blaming the metal.
Sand removal from internal channels. A complex printed core can be difficult to clear. Residual sand in an internal passage becomes a casting defect discovered at the customer.
If you are weighing printed molds against conventional tooling for a specific part, the trade is usually favourable on complex geometry and small batches, and neutral on simple high-volume parts. Our Guide du moulage au sable sans modèle works through that comparison.
Dimensional Defects and Tolerance Capability
Dimensional defects are quieter than porosity and often more expensive, because they are found at inspection rather than at shakeout.
Contraction not compensated. The pattern or the printed mold must be scaled for the alloy’s contraction. Get the factor wrong and every dimension is consistently off in the same direction. Consistent error means a compensation problem; scattered error means a process control problem.
Warpage. Uneven cooling or uneven section thickness distorts the part after it leaves the mold. It is a solidification problem expressed as a dimension.
Sagging or mold movement. Thin sections or soft molds can distort before the metal sets.
On capability: casting tolerances are normally specified using the casting tolerance grade system in ISO 8062, and the grade a foundry can hold depends on the process, the alloy, the part size and the dimension in question. Green sand generally sits in a coarser grade than resin-bonded or printed molds, and printed molds remove the draft and parting-line contribution entirely.
Ask your foundry to state the grade it will commit to for your specific part and dimensions, and to say which dimensions are achievable only by machining. A stated grade is verifiable. A general claim about accuracy is not. For the accuracy side of the printed process itself, see sand 3D printing accuracy and tolerance.

Quick-Reference Defect Table
| Defect | Appearance | Where it appears | Root cause | Immediate action | Long-term prevention |
|---|---|---|---|---|---|
| Porosité gazeuse | Round pores, smooth bright walls | Cope side, top of heavy sections | Gas from mold, core or melt | Improve venting, check moisture | Control sand system and degassing |
| Porosité de retrait | Jagged, branching voids, rough walls | Hot spots, thick sections | Inadequate feeding | Enlarge or reposition the feeder | Design for directional solidification |
| Blowhole | Large rounded cavity, 1 mm and above | Under the cope surface | Mold or core gas, moisture | Vent and dry the mold | Control binder and moisture |
| Pinhole | Very small pores, often at the surface | Surface and subsurface | Hydrogen in the melt | Degas and verify | Charge cleanliness, dry tools, holding control |
| Subsurface blowhole | Hidden cavity under sound skin | Just below the cope skin | Trapped mold gas | Radiograph suspect areas | Venting and lower pour temperature |
| Shrinkage cavity | Visible internal void | Below the feeder | Feeder ran dry or feed path closed | Increase feeder volume | Feed-path design and chills |
| Sink mark | Surface depression | Heavy sections near the surface | Local shrinkage | Add or enlarge the feeder | Reduce section thickness |
| Hot tear | Crack at a section change | Junctions of heavy and thin walls | Restraint during contraction | Improve mold collapsibility | Geometry and alloy review |
| Cold shut | Seam where two streams met | Thin sections, far from the ingate | Metal too cold or fill too slow | Raise pour temperature | Gating and fill-speed design |
| Misrun | Incomplete filling | Extremities of thin sections | Same as cold shut | Raise pour temperature | Gating design, section thickness review |
| Sand inclusion | Loose sand or lump in the casting | Near ingate or from core edges | Erosion or weak mold surface | Reduce fill velocity | Gating redesign, stronger mold |
| Burn-on | Rough welded sand layer | Mold and core surfaces | High pour temperature, poor coating | Improve coating coverage | Sand selection and temperature control |
| Scab | Raised rough patch | Flat mold faces | Sand expansion | Adjust sand system | Control compactability and fines |
| Rat tail | Vein-like ridge | Flat surfaces | Sand expansion buckling | Same as scab | Sand system control |
| Sand wash | Washed depression plus inclusion | Downstream of the ingate | Metal erosion of the mold | Reduce fill velocity | Ingate position and surface strength |
| Metal penetration | Rough, metal-rich surface | Mold faces, coarse sand areas | Coarse sand, low compaction, high temperature | Improve compaction | Sand grading and coating |
| Veining | Thin fins on the casting | Printed mold and core surfaces | Thermal shock cracking of the mold | Review print and cure parameters | Binder and orientation control |
| Slag inclusion | Irregular dark inclusions | Cope surface, under overhangs | Slag carried into the mold | Improve slagging and settling | Filter in gating, pouring practice |
| Flottation du graphite | Graphite-rich layer, poor finish | Upper part of heavy sections | Hypereutectic ductile iron | Review carbon equivalent | Alloy and cooling control |
| Chill, white iron | Hard brittle surface layer | Thin sections, mold face | Rapid cooling | Review inoculation | Inoculation practice |
| Warpage | Distortion after cooling | Long thin sections | Uneven cooling | Check cooling uniformity | Section and mold design |
| Dimensional error, consistent | All dimensions offset one way | Whole part | Wrong contraction factor | Correct the compensation | Pattern or print scaling control |
When the Defect Is Actually a Design Problem
Some defects cannot be fixed in the foundry, and recognising that early saves weeks.
If the part has a heavy isolated boss connected to the rest of the casting by a thin wall, no feeder arrangement will reliably feed the boss, because there is no feed path. If two heavy sections meet at a sharp internal corner, the hot spot is a geometry feature and the restraint is a geometry feature. If a section is specified thinner than the alloy can fill at the required length, the defect is in the drawing.
In those cases the options are limited and should be stated plainly:
- Change the geometry. Often the cheapest fix and the one most resisted.
- Add machining allowance and accept that a region will be cut away.
- Change the process. A printed mold can place a chill exactly where the hot spot is and can eliminate the parting line that contributes to distortion. A different alloy may freeze in a narrower range and tear less.
- Change the acceptance criteria. Where a defect is cosmetic, saying so in the specification is legitimate. Where it is structural, it is not.

If you have a part that keeps producing the same defect through several process changes, send us the drawing and the defect history. Our engineers will tell you whether it is a process problem we can engineer around or a geometry problem that needs a design answer. Get a quote and a free part review.
Foire aux questions
What are the most common casting defects?
The six that account for most scrap are gas porosity, shrinkage porosity, blowholes, sand inclusion, cold shut and hot tearing. Gas and shrinkage porosity together usually dominate, and they are the two most often confused with each other.
What causes porosity in aluminum castings?
Three sources, in order of frequency: hydrogen dissolved in the melt, which comes out of solution as the aluminium solidifies; gas generated by moisture and binder in the mold and cores; and air entrained by a turbulent fill. Degassing, drying and a non-turbulent gating design address all three, and a reduced pressure test tells you which one you are actually dealing with.
How do I tell gas porosity from shrinkage porosity?
Three checks. Shape: gas pores are round, shrinkage voids are irregular and branching. Internal wall: gas pores are smooth and often bright, shrinkage voids are rough and dendritic. Location: gas pores cluster under the cope surface or at the top of heavy sections, shrinkage voids sit at thermal centres. A micrograph resolves anything still ambiguous.
How do you prevent defects in green sand casting?
Control the sand system as a system rather than adjusting one property. Moisture, compactability, green strength, permeability and loss on ignition all interact, and improving one can worsen another. Measure at the molding station, compare against the values that produced good castings, and change one variable at a time.
What tolerance can green sand casting hold?
Tolerances are specified using the casting tolerance grades in ISO 8062, and the grade a foundry can hold depends on the process, the alloy, the part size and the dimension being measured. Green sand sits in a coarser grade than resin-bonded or printed molds. The useful answer for a buyer is not a general figure but the specific grade your foundry will commit to for your part, plus a clear statement of which dimensions will be achieved by machining rather than by casting.
Can 3D printed sand molds reduce casting defects?
Yes, for several defect families, and for structural reasons rather than because the sand is better. Printed molds remove the parting line and the pattern draw, remove ramming variation between molds, allow venting to be designed where it is needed, and allow chills to be placed at the hot spot without pattern constraints. That reduces sand inclusion, draw-related dimensional error, and shrinkage that originates from an unfed hot spot. Printing does introduce its own failure modes, chiefly layer texture and veining, which our 3D sand printing defects guide covers. For a part that is already simple and running well in green sand, printing is unlikely to change the defect rate. For a complex part with a persistent hot spot or a difficult core, it often does.
Conclusion: Classify First, Then Fix
Casting defects are a process problem, not a technology problem, and the fastest route to fixing one is to classify it correctly before touching anything.
Four families, in order of how often they are the answer:
- Gas, if the pores are round, smooth-walled and clustered under the cope surface.
- Shrinkage, if the voids are jagged and sit at a thermal centre.
- Mold, if the defect involves sand, penetration or a surface that broke down.
- Filling, if the defect sits along the flow path or the part did not fill.
Get the family right and the fix is usually straightforward. Get it wrong and you will change three process variables, improve nothing, and lose a week.
If you want a second opinion on a defect that keeps returning, send us the part, the alloy and the defect history. We print molds and cores for castings that conventional tooling struggles with, and we will tell you honestly if the defect is a geometry problem rather than a process one.
Articles connexes
- 3D Sand Printing Defects: Causes, Prevention & Fixes for Molds and Cores
- Sand 3D Printing Accuracy and Tolerance: What the Process Can Hold
- What Can You Cast with 3D Printed Sand Molds?
- Sand 3D Printing Materials: Silica, Ceramic and Chromite Sand
- Patternless Sand Casting: Process, Cost and Lead Time
- Traditional vs Patternless Sand Casting: 5 Key Differences
- Sand Mold and Sand Core 3D Printing Service




