Porosity Welding Causes: 22 Reasons and How to Fix Them

A fabricator welds a 6 mm fillet on primed mild steel, and the bead looks fine until the inspector puts a radiograph on the screen: a bright herring-bone trail of voids right through the fillet. Nothing in the procedure changed, the same welder ran the same machine — the only difference was a fan pointed across the bay. That is the uncomfortable truth about porosity: it is rarely one single mistake, and it is almost never random. It is gas, trapped in metal that froze too fast to let it out.

Porosity in Welding: What It Actually Is and Why It Forms

That herring-bone trail of voids on the radiograph had a cause, and it was a fan. The direct answer to the main question: porosity welding causes all trace back to one thing, gas that entered the molten pool and could not escape before the metal froze. That gas arrives by four routes: the shielding system fails, the joint is contaminated, hydrogen comes in with moisture, or the welding parameters and joint geometry trap the bubbles in place.

The definition, in plain terms

The Training and Research Organisation (TWI) defines porosity as cavities in the weld metal caused by the freezing-in of gas released from the pool as it solidifies. That sentence holds the whole mechanism: a gas source, a liquid pool that absorbs it, and a solidification front that races to trap it.

A single void is a pore. Porosity is the condition of a weld containing pores, however many and whatever shape they take. For inspectors working to the international system, porosity sits in group 2 (cavities) of the ISO 6520-1 classification, the same family that also covers cracks, slag inclusions and other voids.

Molten metal absorbs gas, usually hydrogen, nitrogen or oxygen pulled from the air, and sometimes shielding gas itself. The solidification front advances inward from the pool edges while bubbles formed in the liquid try to rise and escape. If the pool is large, hot and slow enough, they make it. If the front freezes over first, the bubbles lock into solid metal and you get voids. It is a race between buoyancy and freezing speed, and that race is the whole story of the defect.

The four forms you will meet

TWI groups porosity into four forms, and the shape tells you a lot about where the gas came from.

• Distributed porosity — fine pores scattered through the bead, usually a shielding or cleanliness problem.
• Surface-breaking pores — cavities that open at the surface, usually with a larger population of fine porosity beneath them.
• Wormholes — elongated, often linked cavities that show up on a radiograph as a herring-bone trail, a sign that a large volume of gas formed and was trapped.
• Crater pipes — a pipe-like depression at the weld stop, where the pool froze hard as the arc was extinguished.

Why it matters beyond appearance

Porosity is not a cosmetic issue. According to ESAB’s guide to welding defects, it degrades tensile strength and impact toughness, and it can cause leaks in pressure-retaining joints. In stainless steels and aluminium it may also initiate corrosion, because voids create sheltered, oxygen-depleted pockets where corrosion products set up and spread.

Damage depends on pore size, number, location, and what the joint will do afterwards. A few small pores in a statically loaded structural joint rarely matter. In a cyclically loaded or corrosion-exposed joint the same pores are a very different proposition, which is why acceptance limits in welding codes are not one-size-fits-all.

Aluminium and stainless work deserves particular care, because both punish contamination more than carbon steel does. Preparation rules are worth reading before a job starts, and our guide to cleaning aluminium correctly before welding covers the process in detail.

Porosity is not cracking, and it is not slag

Porosity versus cracks. A pore is rounded and volumetric, like a tiny gas bubble frozen in place. A crack is sharp and planar, a fracture through the metal with almost no opening. Inspectors judge cracks far more severely: a crack can propagate, while a pore is a closed cavity that does nothing unless conditions change. Sharp, planar and running into the base metal means it is a crack, and the repair strategy is completely different.

Porosity versus slag inclusions. Slag is non-metallic flux residue trapped inside the weld. ESAB notes it is most common in flux-based processes (MMA, FCAW and SAW) and that inclusions interrupt metallic continuity, reducing ductility and fatigue resistance. Slag comes from the consumable and the process; porosity comes from gas. The advantages of stick welding are well documented, but no process is immune to the slag-cleanliness rules that keep inclusions out.

Porosity is a symptom, not a disease. The gas came from somewhere, and every case has a source you can find: a leaking hose, a painted joint, a damp electrode, a travel speed set too high. Finding that source is a solvable diagnostic exercise, and the fastest way to start is by reading the shape of the defect.

Shielding gas choice, cylinder supply and delivery hardware all matter here. Our explanation of how air gas welding and shielding gas delivery actually work is a useful primer before we dig into flow rates, leaks and drafts.

The Four Root Cause Categories: Shielding, Contamination, Hydrogen, Parameters

Porosity is not random. When a welder stands in front of a hole-filled bead and asks why, there is a specific answer waiting to be found, and it always falls into one of four families. Build that map in your head and troubleshooting stops being guesswork.

The mechanics are simple once you see them. As The Fabricator frames it, porosity is weld metal contamination in the form of a trapped gas: gases are absorbed into the molten metal and released as solidification takes place. The gas can enter from the arc, from the joint, from the consumables, or from the way you moved the torch. Where it came from determines everything about how you fix it.

1. Shielding problems

Air entrainment, the wrong flow rate, leaks, drafts across the arc, and turbulence at the nozzle. Whenever the protective gas is diluted by room air or simply fails to cover the pool, the weld absorbs atmospheric nitrogen and oxygen. This is the most common category in a shop with extraction fans or an open overhead door. Setting up cylinders, regulators and flowmeters correctly is covered in our guide to airgas welding supplies.

2. Contamination

Paint, primer, rust, oil, grease, zinc from galvanised steel, shop grime, antispatter gels, and dirty consumables. Anything sitting on the joint that burns, boils or smears into the pool will generate gas in the arc. Aluminium is unforgiving here, which is why prep matters so much, as we explain in how to clean aluminium for welding properly.

3. Hydrogen and moisture

Damp electrodes, wet wire or flux, moisture on the base metal, rust, humid shop air, and water trapped in faying surfaces. Hydrogen is the most dangerous of the four because, under the right conditions, it does not stop at porosity and moves on to cracking. Rod storage and handling is covered in our all-position welding rod guide.

4. Process parameters and joint geometry

Arc length, travel speed, voltage, weave width, crater filling, crevices and open roots. Travel too fast and the pool freezes before the bubbles can rise, so the geometry of the joint decides how much time gas gets to escape. TWI and ESAB both list long arc length as a direct cause for exactly this reason.

Which gas attacks which metal

TWI’s material susceptibility table is the single most useful reference for narrowing a suspect down, because it pairs the likely gas with the cleaning method that removes it.

Base metal	Gas most likely to cause porosity	Cleaning method before welding
C-Mn steel	Hydrogen, nitrogen and oxygen	Grind to remove scale coatings
Stainless steel	Hydrogen	Degrease, wire brush, degrease again
Aluminium and alloys	Hydrogen	Chemical clean, wire brush, degrease, scrape
Copper and alloys	Hydrogen and nitrogen	Degrease, wire brush, degrease
Nickel and alloys	Nitrogen	Degrease, wire brush, degrease

Note the pattern. Stainless and aluminium are hydrogen problems, so cleaning and dryness decide the outcome. Nickel and copper are nitrogen problems, so shielding coverage decides the outcome. Mild steel is exposed to all three, which is why it is the hardest to diagnose by reasoning alone.

Nitrogen and oxygen versus hydrogen

The Welder draws the distinction cleanly. Nitrogen and oxygen porosity comes mainly from inadequate shielding gas coverage. Hydrogen porosity typically comes from contaminants at the joint or in the electrode or filler material.

That difference shows up in the weld itself. A shielding failure tends to produce fine, evenly distributed pores spread through the bead, because atmospheric gas was available everywhere the arc touched the pool. A hydrogen failure is more localised and more severe, because the moisture was concentrated in one place: one dirty patch, one wet consumable, one rusty area. Larger voids, wormholes and even cracks cluster around the source.

The diagnostic table

Cause category	Mechanism	Typical appearance	First thing to check
Shielding	Room air dilutes the gas; nitrogen and oxygen are absorbed by the pool	Fine, evenly distributed pores, sometimes breaking the surface as coarse pits	Measure flow at the nozzle with the torch raised, then leak-test the line
Contamination	Coatings, oils or zinc burn off and release gas in the arc	Coarse pores clustered near the coated area, heavy spatter, discolouration	Inspect the joint and the area around it for paint, primer, rust, grease and antispatter gel
Hydrogen and moisture	Moisture breaks down and releases hydrogen into the pool	Localised large voids, wormholes, crater cracking and fish-eyes	Verify electrode and flux dryness, and look for water in crevices and faying surfaces
Parameters and geometry	The pool freezes faster than gas can escape; crevices trap gas	Elongated voids at the root, a pipe at the weld stop	Compare travel speed and arc length against the WPS, and look for crevices or an open root

Gas Porosity: How Shielding Flow, Leaks, and Drafts Let the Air In

Shielding gas failures are the quietest cause of porosity in a shop, because nothing about the setup looks wrong. The cylinder is full, the valve is open, the arc sounds normal. The real issue is the small percentage of room air mixed into the gas stream.

According to TWI, as little as 1% air entrainment in the shielding gas causes distributed porosity, and more than 1.5% produces gross surface-breaking pores. That makes 1% the number to remember. A bead with fine craters or a faintly foamy look does not need a new procedure qualification. It needs a gas system that is actually delivering shielding gas.

TWI also names the usual suspects: leaks in the gas line, too high a flow rate, draughts, and excessive turbulence in the weld pool.

Too Little Gas, Too Much Gas

Too little gas means the envelope around the arc is not full, so the pool breathes room air. The signature is fine, evenly scattered pores, often worse mid-bead than at the edges.

If the flowmeter at the regulator reads correctly but the flow at the nozzle comes up low, the regulator is not the problem, the delivery system is. As The Welder puts it, the blame falls on the sealed connections, the sealants and the O-rings, the parts that get crushed, aged out or swapped by mistake.

Too much gas is just as damaging: gas leaving the nozzle too fast creates a turbulent, swirling envelope instead of a smooth blanket, and that turbulence pulls room air back into the zone you were protecting. The Fabricator recommends 50 to 60 CFH at the GMAW nozzle and 20 to 30 CFH at the GTAW torch, except shielding gas with more than 50% helium, which wants a higher rate. Leaving the valve wide open wastes gas and causes defects.

Process	Flow at the nozzle
GMAW (MIG)	50 to 60 CFH
GTAW (TIG)	20 to 30 CFH
FCAW-G and SMAW	Per WPS and consumable data; SMAW is self-shielded, FCAW-G uses external gas

The Leak Checklist

Before you touch the procedure, walk this list. Each has ended a fabrication job somewhere:

• An empty or nearly empty cylinder. The regulator still shows pressure, just not the pressure you set.
• A pinched or crushed hose. Over 20 ft, a hose kinks easily, and a kink cuts flow without looking like one.
• A cut or burnt hose anywhere from the regulator flowmeter to the feeder connection.
• Damaged O-rings on the GMAW gun whip or the GTAW torch cap.
• A defective gas solenoid at the wire feeder or on the GTAW set.
• A borrowed hose once used for something else. The Fabricator tells of a hose with a wasp nest inside it and an air hose full of tool oil. Neither looked wrong from outside.

Drafts and Moving Air

Drafts get blamed on everything else. The usual suspects are open overhead doors, extraction fans, cooling fans and any cross-draught in the bay. The Fabricator points out that overhead or floor fans 25 feet away can disrupt gas delivery, and that drafts of more than 4 to 5 mph can affect SMAW and FCAW. That matters commercially, since the advantages of stick welding disappear fast if the extractor pulls the envelope apart.

Codes get involved as well. The Welder notes that some codes limit wind speed at the weld area when a process uses an external shielding gas system, covering GTAW, GMAW and FCAW-G. A weld that meets the WPS at the bench can fail inspection on second shift when the extraction unit switches on.

Two Start-of-Weld Classics

1. The gas-free start. After a break or lunch, the line has been purged and the gas has not arrived when the arc strikes. The first few centimetres of bead come out full of voids while the rest is clean. The gas is there, it just has not got there yet.
2. The corner that never purges. On a tight corner or in a T-joint, the first burst of shielding gas sweeps over the pocket instead of flooding it, so trapped air sits where the arc is. Pull the trigger for a second or two, cut the wire, then start in the cleared pocket. Regulated TIG work such as aerospace TIG welding treats it as mandatory.

Gas Purity and Dew Point

Gas quality matters too. Ask your supplier to certify the gas, including moisture content and dew point, and check the regulator and flowmeter, since a damaged float or dirty gauge gives a number you cannot trust. When replacing cylinders or regulators, our guide to air gas welding supply, cylinders and safety covers what to look for.

I’ll rewrite this section, tightening it to land within the 560–840 word range while keeping every fact and link.

Hydrogen Porosity and Hydrogen Cracking: Following the Moisture Trail

If shielding gas is the most common source of trapped bubbles, hydrogen is the most dangerous one. Nitrogen and oxygen porosity looks like fine, even fizz spread across the bead. Hydrogen porosity behaves differently. It is more localised, the individual voids are larger, and it does not stay a cosmetic defect. Trapped during solidification, hydrogen can set up a delayed crack that appears hours or days after the weld looked fine. That is why this second gas family deserves its own trail.

According to TWI, hydrogen can originate from moisture from inadequately dried electrodes, fluxes or the workpiece surface, and grease and oil on the workpiece or filler wire are also common sources. The distinction that matters for diagnosis: hydrogen porosity is a cleanliness and moisture failure, not a shielding failure. Your gas can be pure, your cylinder full and your flow perfect, and you can still fill a weld with voids, because the hydrogen arrived with the metal, the wire or the rod rather than with the air.

Where the Water Hides

Most welders check the top face of the joint. The moisture usually is not there.

• Faying surfaces. Water hides between the plates of a lap joint, or between a butt joint and its backing bar. It cannot evaporate, it cannot easily be seen, and it sits exactly where gas evolves.
• Cutting oils. Oils used to saw tubing to length or drill holes in an I-beam are common contaminants near weld joints. Films survive handling, wiping and a change in the weather.
• Condensation and morning dew. Overnight temperature drops leave moisture on heavy plate and lap joints. The Fabricator notes this becomes a genuine problem below about 50°F, and the practical fix is to preheat the metal to 200–220°F so the moisture evaporates before the arc strikes.
• Rust. As The Welder explains, rust occurs when iron, water and oxygen combine, and it mixes with weld metal chemistry to generate hydrogen bubbles that may not escape before solidification.
• The consumable itself. Electrodes, wire and filler metal pick up moisture in a humid shop even when the material looks and feels fine.

From Voids to Cracks

Here is where hydrogen stops being a cosmetic headache. In high-strength low-alloy steels such as A514, A588 and A709, hydrogen trapped during solidification can lead to hydrogen-induced cracking. The crack may not exist when the weld cools. It develops over time, and it can become catastrophic once fatigue loading is added on top of it. A joint that passed visual inspection at the end of the shift can be cracking apart months later under cyclic service loads.

Treat hydrogen-related porosity as a process failure rather than a bad-luck weld. If you regularly run A514 or A588, review consumable handling, storage ovens and preheat requirements as a system. Our guide to alloy steel welding electrodes covers matching filler selection to the base metal, the other half of that control.

Storing Consumables So They Do Not Drink the Shop Air

Storage is the cheapest fix here, and the one most often skipped. SMAW electrodes, FCAW electrodes and SAW flux absorb moisture in an unprotected environment, and SAW flux in particular behaves like a sponge. Codes and manufacturers are clear about the response:

• Use ovens and dryers, holding consumables at the manufacturer’s specified temperature for the specified time.
• Store opened packages according to the manufacturer’s directions, not according to what is convenient on the rack.
• Keep wire in its original vacuum packaging until use, and reseal or clip the run-off rather than leaving a spool open all week.
• Rotate stock first-in, first-out so the oldest consumable is used first and forgotten material does not sit for months.
• Never use a rod that has been left out in a humid shop. If it has been out, treat it as suspect.

For aluminium work in particular, getting this sequence right matters more than most operators expect. Our walkthrough of how to prep aluminum right for welding covers the chemical clean step in detail, because that is where the hydrogen trouble usually starts in that process family.

Put the pieces together and the moisture trail is straightforward. Hydrogen porosity is nearly always a storage problem, a cleaning problem, a joint-design problem or a preheat problem, and in every case the fix is cheaper than the rework. The full list of porosity causes is worth bookmarking next to the consumable storage area, because the same wet rod that causes voids today is the same wet rod that causes a crack rejection next quarter.

Contamination Causes: Paint, Primer, Rust, Zinc and Shop Grime

Contamination is the porosity cause that hides in plain sight. The plate looks clean, the settings are right, and yet the radiograph comes back peppered with voids. It is also the most common cause in real shops because it is rarely deliberate: a primer sprayed too thick, a galvanised coating nobody mentioned at handover, or a gun liner quietly collecting shop grease for two years.

Paint and Primer: The Biggest Single Culprit

TWI notes that surface coatings such as primer paints, and treatments like zinc coatings, may generate copious amounts of fume during welding. As the coating breaks down under the arc it releases gases, and those gases have to go somewhere. When they cannot escape the pool, they freeze in as voids. TWI also flags a point that helps in diagnosis: the risk of trapping evolved gas is greater in T joints than in butt joints, especially when you fillet weld both sides, because the second side reheats and traps gas from the first, making the double-sided T joint the classic setting for a porosity complaint. See the full TWI guidance on porosity causes.

The Weldable Primer Question and the 20 Micron Rule

Shop folklore says strip primer back to bright metal. TWI’s position is more measured: it is not necessary to remove weldable (low zinc) primers before welding, but if primer thickness exceeds the manufacturer’s recommendation, porosity is likely, particularly with processes other than MMA. Stick welding gets more latitude because the flux covering lays down its own protective layer over the coating, and MIG, TIG and flux core are not so forgiving.

The number worth memorising: for T joints in primed plate, coating thickness on the edge of the vertical member should not exceed the manufacturer’s maximum, typically around 20 microns, and over-spraying is the usual way that limit gets broken. One extra coat of primer on that edge is often the difference between a sound weld and a void-ridden one. When the coating gets very thick the failure mode changes: TWI notes that wormholes are indicative of a large amount of gas being formed and trapped, from gross surface contamination or very thick paint or primer coatings.

Galvanised Steel and Zinc

Zinc melts at approximately 420 degrees F. At welding temperatures far in excess of 2,000 degrees F it turns from solid to gas in a fraction of a second, and zinc dust is a byproduct of welding galvanised metal. Galvanised work therefore belongs to a trained welder: procedures exist to weld it successfully, but they depend on grinding or burning the coating back, controlling fume and managing zinc exposure, and that takes practice, not guesswork.

Rust and Mill Scale

When mill scale and rust are welded over, decomposition gases are formed and oxidation begins, with a strong possibility of cold lapping and lack of fusion at the weld toe. Mill scale is also hard, and it makes the wire roll and spit spatter, feeding more contamination back into the joint. Grinding or wire brushing back to clean metal is not optional.

Everyday Grime and the Antispatter Trap

The Fabricator’s list of offenders that release large volumes of gas at arc temperatures is worth reading twice: paint, grease, oil, glue and sweat, plus tape, water, cutting fluid, chalk, soapstone and residue left by NDT operations such as dye penetrant and magnetic particle testing.

Contaminated Consumables

Contamination does not have to arrive on the base metal. Dirty gloves during GTAW carry sweat and grease onto the tungsten and filler rod. GMAW gun liners collect shop grease, oil, dust and dirt, then wipe it onto the wire as it feeds through. Filler wire that has been dropped, dragged across a dirty bench or stored open carries paint, grease, oil, tape and glue into the puddle. Seeing the same defect on multiple joints in one shift? Suspect the consumables before the procedure.

How to Clean a Joint Properly

Method	What it removes	Notes
Solvent degrease	Oil, grease, cutting fluid, drawing compound	Wipe a generous margin around the joint
Mechanical cleaning	Rust, mill scale, paint, primer, zinc, old spatter	Grinding, wire brushing, scraping or shot blasting to bright metal
Chemical cleaning	Aluminium oxide and surface contaminants	Essential for aluminium; see our guide to cleaning aluminium correctly for welding
Blast cleaning	Everything, back to bare metal	Most reliable when contamination is widespread

Cleaning aluminium is a separate discipline: the oxide layer reforms in minutes, so aluminium work needs a chemical clean and a dedicated wire, much like the prep in our aluminium MIG welding guide.

I’ll rewrite this section to fit the 560–840 word range while keeping the heading, facts, and all links intact.

Process and Parameter Causes: Arc Length, Travel Speed, Voltage and Joint Design

This is the category welders tend to skip. The gas system is set, the plate is clean, the electrode is dry, and the bead still comes up full of holes. Nothing looks wrong with the setup, so the shop blames the gas, the wire or the weather. Often the real culprit is a parameter or a joint detail nobody thinks to question.

The reason parameters matter is simple: they decide how much time the bubbles get. Gas trapped in a molten pool wants to rise and escape. Whether it can depends on how long the pool stays hot and fluid. Push the travel speed up, or let the pool get cold and sluggish, and you have shortened the escape window. Freeze the pool with bubbles still in it and you have a void.

Arc length and voltage

ESAB’s welding defects guide lists long arc length as a direct cause of porosity, alongside gas leaks, incorrect gas type or flow, damp electrodes, wires or flux, and inadequate back-purge on stainless roots. That list is short enough to memorise, and four of those five causes are covered elsewhere in this guide. Arc length is the one that sneaks through.

Why does a long arc create porosity? Two reasons. A long arc spreads out and weakens the shielding gas column, so less of the pool is properly protected and more air gets in. It also raises the arc voltage, and a higher voltage tends to a wider, less fluid pool that traps gas rather than releasing it. A short, stable arc keeps the heat concentrated where the shielding is strongest.

Travel speed, weave and puddle size

Travel speed is the parameter most likely to be quietly wrong, because the welder is often trying to keep up with production. The bead looks fine, the fill passes, and the porosity shows up on the radiograph. The same goes for excessive weave and wide puddles. A wide, cold pool with the wire whipping from one edge to the other gives bubbles a long, sluggish path to the surface. ESAB makes the same argument for slag: tight groove geometries and high deposition rates increase the risk of inclusions because they impede slag float-out. The logic for gas is the same in reverse. A wide puddle traps what should have floated out.

Joint geometry and crevices

TWI’s porosity guidance notes that wormhole entrapment is more likely in crevices, such as the gap beneath the vertical member of a horizontal-vertical T joint fillet welded on both sides. That gap is a pocket. Molten metal runs into it, gas collects in it, and the solidifying fronts close over the top and trap it. Prime wormhole territory, and it looks like a joint problem because it is one.

An open root is the other end of the same problem. As The Fabricator points out, if the joint is open at the root it will suck in air from the back side, and unprotected liquid metal absorbs air easily. You cannot shield a root that is breathing straight into the pool.

Then there is the corner problem. Welding a GMAW gun over the edge of an outside corner joint means the nozzle physically cannot cover the joint. Gas spills off the edge, turbulence pulls room air back in, and the bead fills with fine pores. The same applies to a gun laid at an angle across the joint, which spreads the gas flow and sucks atmosphere in from the back side. A gun angle of 5 to 15 degrees, perpendicular to the joint, is acceptable.

Root condition and back-purge

For stainless and aluminium, an unpurged root bead is a guaranteed problem. The root face sits in air, so the first metal to freeze contains oxygen and nitrogen rather than argon. For stainless that means a discoloured, sometimes oxidised root layer that can hide a lack of fusion; for aluminium, an oxide layer that the MIG process cannot weld through reliably.

Back-purge fixes it by flooding the root with argon from behind, usually through a purge dam built from cardboard, tape or a purpose-made dam, with a hose and flowmeter set to vent the displaced air at the far end. Purge until gas flow is established, then start the root weld. Some shops purge through the torch for the first pass; either way, the root needs clean argon behind it. Note that flux-cored aluminium work still needs root protection, since the slag does not shield the root from the atmosphere.

Crater pipe and crater cracks

Crater pipe is the classic endpoint of a bad stop. TWI describes it as forming during the final solidification of the weld pool and resulting from shrinkage on solidification, with the arc being extinguished causing rapid solidification of the pool. In TIG, autogenous techniques or stopping the welding wire entering the weld pool before extinguishing the arc will effect crater formation and may promote the pipe imperfection.

The remedies TWI gives are worth borrowing:

• Run-off tag: let the pool run off the end of the joint onto a sacrificial tab, where the crater does no harm.
• Grinding out the crater stop if a tag is not practical.
• Slope-down or crater-fill current functions on the machine, which taper the current down rather than cutting it dead.
• Adding filler in TIG during the stop so the pool does not freeze empty and deep.

Never extinguish a welding wire before the arc. Whether it is a GMAW gun or TIG rod, stopping the feed and then killing the current leaves a deep, fast-freezing hole that shrinks and pulls open into a pipe.

Why one spot?

Porosity often appears in one location on an otherwise sound weld, and the reason is usually local: a change in travel speed over a thin-to-thick transition, a gap that opens up, a nozzle running low on gas just at that point, a mill scale patch, a change in weld position, or the arc entering a section with different chemistry. The defect marks the spot where something changed, so treat the location as the clue rather than the whole weld.

Types of Porosity: Distributed Pores, Surface Pores, Wormholes and Crater Pipes

Once you know that gas got trapped in your weld, the next useful question is a shape question: what does the defect actually look like? Porosity is not one thing. It arrives in four recognisable forms, and each one narrows the list of suspects fast. A fine haze of bubbles spread through the bead points at shielding gas or cleanliness. A long channel blown through the root points at something else entirely, usually a wet joint or a cavity in the fit-up. Reading the shape turns guesswork into diagnosis.

One piece of vocabulary first, because it matters when you read a report. A single void is a pore. Porosity is the condition of a weld that contains them. That is not pedantry: a code can reject a weld for a single oversized pore even when the rest of the bead is sound, so inspectors track pore size and count separately.

Distributed porosity: fine pores through the bead

This is the most common form and the least dramatic to look at. According to TWI’s guidance on porosity, distributed porosity is normally found as fine pores throughout the weld bead. On a cut section you see a scattering of small round voids like bubbles in fizzy drink. On a radiograph it reads as a mottled, speckled density rather than a single hard-edged void.

Cause-wise it is the classic symptom of a shielding or cleanliness problem. Air pulled in at the nozzle, a gas flow reading low because a hose is crushed, or a joint carrying paint and shop grime will all scatter fine pores through the whole deposit. It is also what you get over wet or rusty steel, where hydrogen comes in diffusely and bubbles out across the pool.

Because it is a whole-bead problem, the cause is systemic rather than local. If every weld on the floor has it, stop blaming the welder and check the cylinder, the flowmeter and the primer.

Surface-breaking pores: the visible evidence of what is underneath

Surface-breaking pores are round holes that open on the weld surface, sometimes with a small raised lip or a trail of escaping gas. They look dramatic, and they are a clue: TWI notes that surface-breaking pores usually indicate a large amount of distributed porosity beneath the surface. The hole you can see is the exit point for a much bigger population of trapped gas.

TWI’s own illustration is a fillet weld made on primed plate, and that image tells the story. The primer did not leave a few specks on the surface. It boiled at arc temperature, generated a large volume of gas, and had nowhere to go but into the pool.

So when you find a burst pore on the face of a bead, do not simply grind it out and move on. Ask what was feeding the gas: check primer thickness, joint cleanliness immediately before the arc, and whether flow was where your WPS says it should be.

Wormholes: the herring-bone signature

Wormholes are elongated pores, and TWI describes them as producing a distinctive herring-bone appearance on the radiograph. That pattern is one of the most recognisable images in weld inspection.

A wormhole forms when a large volume of gas forms in the pool and is trapped as the solidification front advances across it. The channel is long rather than round, which tells you the gas was generated in quantity, not trace amounts. Typical sources are gross contamination such as heavy paint or thick primer, moisture in the joint, and cavity-forming joint geometry where the gas has an easy escape route that then seals itself over.

Wormholes also have a strong location bias. They favour the root of a fillet weld, especially a horizontal-vertical T joint filleted on both sides, where the gap under the vertical member acts as a pocket that fills with gas and then traps it. If you find them at a consistent depth, look at the fit-up gap before the gas bottle.

Crater pipe: the defect at the weld stop

A crater pipe is a pipe-like depression right where the arc went out. It forms because the pool shrinks as it solidifies and the last metal freezes fast when the heat source is removed. Some also carry trapped gas, which is why gas porosity and crater fill problems so often appear together at the same end of a weld.

Classic causes are stopping the welding wire before extinguishing the arc, TIG autogenous techniques where no filler is added at the end, and simply extinguishing the arc without any slope-down. The remedies are well established: run-off tags, crater-fill or slope-down functions on the machine, grinding the crater stop out, and adding filler at the end of a TIG bead.

Porosity against the other common weld defects

Most rejection disputes come down to confusing one defect for another. Here is how the main five separate on appearance, cause, detection and severity.

Defect	Appearance	Typical cause	Detection	How it is judged
Porosity	Rounded volumetric voids: fine and distributed, elongated channels (wormholes), or a depression at the weld stop (crater pipe)	Air entrainment, contamination, hydrogen from moisture, wrong parameters, poor joint geometry	PT or MT if surface-breaking; RT or UT if subsurface	Judged by size, distribution and governing code limits
Cracking	Sharp, planar, linear, sometimes branched	Thermal stress, restraint, hydrogen-assisted cracking, design problem	PT or MT for surface; RT and UT for subsurface	Judged far more severely than porosity, often rejected outright
Slag inclusion	Non-metallic material trapped in the weld, often elongated or stringy along the bead	Flux residue, poor slag float-out, contaminated or oxidised pool	RT is primary; UT with good technique can also find them	Serious because it breaks the load path and can initiate cracking
Undercut	Groove along weld toe or side, sharp and continuous	Excessive travel speed, wrong arc length or work angle, excess current, unstable arc	Visual, sometimes aided by profile gauges	Judged on depth and length against code limits
Lack of fusion	Cold, unfused boundary where weld metal failed to wet or melt the parent metal	Travel speed too high, heat input too low, dirty or oxidised surface, wrong bead placement	RT and UT usual; visual for obvious cases	Judged strictly, often requires removing the entire weld

The distinction that catches people out is the first one on that list. TherNess puts it plainly: a pore is a rounded, volumetric cavity, while a crack is sharp and planar, and a crack is judged far more severely. On a radiograph a pore shows as a rounded dark or bright void with a soft edge, a crack as a thin, jagged line. Same bead, different outcome. On aluminium work, where cleanliness and filler handling dominate the defect list, our guide to prepping aluminium for welding covers the surface preparation that keeps porosity out of the pool in the first place.

How porosity gets found

For surface-breaking defects, TWI’s guidance is straightforward: liquid penetrant inspection or magnetic particle inspection will show them. PT finds pores open to the surface on non-porous materials like clean steel and stainless. MT works on ferromagnetic material such as carbon steel and is often more sensitive to fine surface-breaking indications, though it only finds indications that break the surface.

For subsurface porosity the options are radiography or ultrasonic inspection. TWI notes that radiography is normally more effective at detecting and characterising porosity, because the film gives a two-dimensional picture of the voids and lets you measure them against code limits. The limitation is real: small pores are hard to detect, especially in thick sections where film contrast and geometry both work against you. UT finds porosity reliably in a trained operator’s hands, but characterising the defect is harder, and it is less certain about a very small pore near the surface.

On the shop floor, two crude checks still earn their keep. A water or soapy-water test tells you whether a non-pressure item is leak-tight, turning a porosity question into a yes-or-no answer in seconds. The hammer and dye check, where a light tap leaves a bleed mark at a leak path, is rough screening only. Treat both as indicators, never as a substitute for proper NDT. For pressure-retaining items, the rules in OSHA 1910.252 and the applicable code apply, and a leak test does not prove the weld meets specification.

Does It Matter? Code Limits, NDT Rejection and Real Consequences

Short answer: yes, but it depends on where the weld is and what the service does to it. Fine pores in a static structural fillet are a nuisance. The same pores in a cyclically loaded riser or a stainless pressure vessel root are a rejection waiting to happen.

Start with the published limits, which are more generous than most fabricators expect. According to The Welder (April 2022), AWS D1.1 allows porosity not to exceed 3/8 in. in any linear inch of weld and not to exceed 3/4 in. in any 12 in. length. ASME B31.1 sets a limit on pore dimensions of no greater than 3/16 in. API 1104, the pipeline welding code, is tighter still: porosity cannot exceed 1/8 in. diameter.

Read D1.1 carefully. It is an aggregate allowance, not a zero-porosity rule, so a weld can pass with visible porosity in it. Most shops still treat any pore as a defect to gouge out on sight, which is a very different conversation.

The International Framework: ISO 5817 and ISO 6520-1

For work outside North America, or for European and international clients, the reference framework is ISO. ESAB’s welding defects guide explains that standards including ISO 5817, AWS D1.1 and ASME Section IX define terminology and acceptance levels, but that critical industries often require defects to be eliminated entirely rather than merely brought inside a limit.

ISO 5817 is the acceptance standard, setting quality levels B, C and D. The more demanding the level, the smaller the permitted pore and the less area the pores may take within the weld section. ISO 6520-1 handles classification: it places cavities in the group 2 family, where a single void is a pore and porosity is the condition of a weld containing them. Keep both in mind, because a procedure qualification and an acceptance criterion are different documents.

Why a Code Pass Is Not a Client Pass

Here is the nuance that catches people out. The same The Welder article points out that the acceptance criteria suggest the codes are considerably more forgiving on porosity than the quality assurance departments enforcing them. A weld that satisfies D1.1 without argument can still be rejected by the client, the owner’s representative or an inspection agency, because their criteria may be tighter or the specification simply says no porosity.

Contract language wins over code defaults. Get the project specification and the client’s inspection and test plan before the job starts, not after the radiograph comes back, because that document is the one the inspector holds in his hand.

What Porosity Actually Does to the Metal

Porosity is not merely cosmetic. TherNess sets out the mechanics: porosity above the acceptance limit reduces the effective load-bearing section of the joint, acts as an initiation site under fatigue loading, and is among the most frequent causes of NDT rejection and rework.

• Reduced effective section. A pore is a void, so less material carries the load and stress concentrates at its edges.
• Fatigue initiation. Sharp pore edges act as a stress raiser and a crack starter, so a bead that passes a static check can fail under vibration.
• Leakage. In a pressure-retaining joint, a connected path of pores is a leak path.
• Corrosion initiation. In stainless steels and aluminium, pores trap moisture and chloride and become sites for pitting and crevice corrosion.

That last point matters most for aluminium and stainless. Aluminium MIG welding and stainless root beads both depend on the weld root staying gas-tight, and both punish porosity harder than a carbon steel fillet. Shielding and root purge advice for aluminium is in our aluminium cleaning guide.

What Rejection Actually Costs

Rejection is rarely just the weld failed. It is a chain of costs that lands weeks after the arc went cold, and TWI notes that remedial action normally means removal by localised gouging or grinding, and that if the porosity is widespread the entire weld should be removed and re-welded to the agreed welding procedure. That is the expensive part: you are back to bare metal, re-fitting, re-purging and re-running a qualified procedure.

Add it up:

• NDT on the original weld, then on the repair, then again if the repair is rejected.
• Rework hours for gouging, grinding and re-welding, usually the biggest cost since it is skilled labour on a structure that may no longer be accessible.
• Schedule slippage and the knock-on effect on the erection crew and the client’s own milestones.
• Consumables used and scrapped, plus the wire or electrode lot to trace if the cause is a defective consumable.

In a pressure or offshore context there may be regulatory reporting obligations on top. That is why porosity showing up on a radiograph costs so much more than the hours spent welding it.

The Practical Takeaway

TherNess makes a reassuring point: a few small pores rarely compromise a statically loaded joint. A bracket under dead load will not fail because of three small voids, while the same voids in a fatigue-loaded bracket or a submerged riser is a different proposition. Porosity matters in proportion to three things: how much there is, where it sits relative to the load, and what the service environment will do to it.

A few small pores in a static, non-corrosive joint is a blemish and a paperwork item. The same porosity in a fatigue-loaded or corrosion-exposed joint ends up in the repair log. So do not chase zero pores on every weld, which is expensive and often impossible. Know your acceptance criteria before the first arc strikes, build the inspection plan around the joint’s real duty cycle, and treat the code limit as a ceiling on the argument, never as the target.

Prevention and Troubleshooting: A Step-by-Step Shop-Floor Playbook

Most porosity defects never reach a report. They get ground out, re-welded and shipped, and the cause walks back to the bay to repeat the cycle. The fix is a routine, not a better wire.

The playbook follows the order of work: before the arc, the first minutes of the shift, during the weld, and after a defect appears. TWI’s guidance on porosity frames prevention as a checklist rather than a mystery: clean joint edges before welding, seal air leaks, cut excessive gas flow, avoid draughts, dry electrode and flux, degrease the workpiece.

Phase 1: Before the Arc

• Fit-up and joint preparation. Close the gap, align the members, clear anything that could trap gas. Crevices are wormhole territory.
• Clean within a few inches of the joint, not just the weld line. The heat-affected area wicks contamination toward the pool.
• Remove oil, grease, rust, mill scale and NDT residue. Penetrant dyes, chalk and soapstone must come off first. On aluminium, use a dedicated chemical cleaner and prep the oxide separately. Our guide to prepping aluminium for welding covers the solvents that work.
• Measure primer thickness. Weldable primer over the manufacturer’s maximum, typically around 20 µm on the edge of a vertical member in a T joint, is a reliable troublemaker.
• Verify consumable storage and bake status against the datasheet: oven log, drying temperature, hold time. A rod left open overnight in a humid shop is not usable.
• Check the gas system: cylinder contents, regulator, flowmeter, hose, O-rings, gun liner.
• Set flow to the WPS range and confirm it at the nozzle, not the regulator. Workpiece and consumables must be dry and above the dew point.

ESAB’s welding defects guide covers the same ground: degreasing, baked consumables, gas purity, leak-free hoses, laminar flow, short arc and back-purging.

Phase 2: The Fifteen-Minute Leak Test

This is the cheapest diagnostic in the trade, and it belongs before the first weld, not after the third rejection. The Fabricator describes it plainly: open the cylinder handwheel, pressurise the system for 15 to 20 seconds, close the cylinder and check that pressure holds.

Phase 3: While Welding

• Set laminar flow, not a blast. Too little gas gives poor coverage; too much stirs room air into the pool.
• Keep a short, stable arc. A long arc raises voltage and weakens shielding.
• Match travel speed to the pool. Slow down so bubbles escape before the solidification front traps them. Cutting wire feed with travel speed fixes elongated porosity.
• Hold the nozzle close and the gun 5 to 15° perpendicular to the joint for best coverage with least turbulence.
• Clear spatter from the nozzle regularly. A blocked nozzle is a leaky nozzle.
• Use a run-off tag and crater-fill or slope-down functions to avoid a crater pipe at the weld stop.
• Purge the gas line after a break. A restart on a purged line begins gas-free and voids the bead head.
• Purge corners before starting. Per The Fabricator, the first burst of shielding gas in a tight corner seldom purges the pocket. Trigger a second or two, cut the wire, then go.
• Back-purge stainless and aluminium roots with argon and a purge dam where required.

For gun work and travel control, see our guides to aluminium MIG welding and aluminium flux core welding.

Phase 4: After the Weld

Inspect rather than assume. Surface porosity shows under liquid penetrant or magnetic particle inspection; internal porosity needs radiography or ultrasonic testing. Gouge out what you find, re-prepare, re-weld to the qualified procedure and verify by NDT.

Diagnostic Table: Read the Symptom

Symptom	Probable cause	First checks
Fine, even pores	Air entrainment	Nozzle flow, hose leaks, drafts
Large void at bead start	Gas-free start	Purge line, delay start, solenoid
Wormholes at T-joint root	Crevice geometry plus primer	Gap under member, primer, fit-up
Fine pores at bead start only	Contamination or moisture	Clean edges, primer, oven log
Coarse pores in heavy section	Excessive travel speed or arc length	Slow down, short arc, less wire feed
Pipe at the weld stop	Crater fill	Run-off tag, crater fill, TIG filler
Root porosity on stainless	Weak back-purge	Purge dam, flow, timing before arc

Safety: Rework Is Hot Work

Grinding a porous weld means grinding filler metal that may carry chromium-bearing alloying elements, inside the fume plume. OSHA 1910.252 sets the general requirements for welding, cutting and brazing, including ventilation and protection from hazards; OSHA’s welding hazards and solutions guidance covers the specifics. OSHA’s hexavalent chromium overview sets a permissible exposure limit for chromium (VI) of 5 µg/m³, and the fact sheet on controlling hazardous fume and gases explains the controls.

• Move the fume away: local exhaust and general ventilation first, hot-work awareness second.
• Know your material and match the respirator to it: stainless, galvanised and aluminium fume each need their own handling. Add eye and skin protection, and gloves.

Once the cause is fixed, write it down: defect shape and root cause in the log turn a bad month into a better year.

Conclusion: A Systematic Approach Beats Guesswork

Porosity is not bad luck. It is gas trapped in metal that froze too quickly to let it escape, and every case has a traceable source: the shielding supply, the air in the bay, moisture on the joint, a dirty consumable or a crater that froze shut. The pool absorbs gas, and bubbles that cannot escape before the solidification front advances stay locked in the bead. Find the source and the defect goes with it.

That is why experienced shops treat porosity as a diagnostic problem, not a guessing game. TWI’s guidance on prevention is a list of gas-source removals: sealed air leaks, no pool turbulence, filler with adequate deoxidants, sensible gas flow, no draughts, dry electrode and flux, clean and degreased workpieces. ESAB’s welding defects guide adds skilled operators, correct parameters, sound joint design, clean consumables and rigorous quality assurance. As The Fabricator points out, a competent welder can turn reject parts into weldments acceptable under most welding codes once the causes are identified.

The Four Buckets, One Last Time

Every case traced back to one of four places: shielding, contamination, hydrogen, and parameters plus joint geometry. Shielding problems pull air in through leaks, drafts, wrong flow or turbulence. Contamination puts paint, primer, rust, grease, oil, zinc and shop grime into the pool. Hydrogen comes from moisture in the joint, the consumables or the atmosphere. Parameters and geometry decide how long bubbles have to escape. One bad weld can inherit causes from several buckets at once.

The numbers worth memorising:

• 1% air entrainment in the shielding gas causes distributed porosity; more than 1.5% gives gross surface-breaking pores.
• 50 to 60 CFH at the GMAW nozzle and 20 to 30 CFH at the GTAW torch is a sane baseline, adjusted for the gas and the joint.
• 200 to 220°F preheat drives off moisture when ambient temperature drops below about 50°F.
• About 20 µm maximum primer thickness on the edge of a T-joint vertical member, unless the manufacturer allows more.
• 15 to 20 seconds of pressurised gas to leak-test the system before the first arc.

A Five-Step Checklist That Actually Works

1. Identify the gas family from the appearance. Fine pores, surface-breaking pores, wormholes and crater pipes each point somewhere different. Decide what you are looking at before you change anything.
2. Test the gas system. Open the cylinder handwheel, pressurise for 15 to 20 seconds, shut it off and watch whether the pressure holds. Check flow at the nozzle, not the regulator, and look for pinched hose, cut lines and damaged O-rings.
3. Clean the joint and consumables. Degrease, remove rust, mill scale and paint or primer beyond the manufacturer’s limit, and confirm the electrode, wire and flux are dry and properly stored.
4. Check parameters against the WPS. Arc length, travel speed, voltage, weave, nozzle angle and crater fill. More time in the pool means more time for gas to leave.
5. Record the root cause: which bucket, which job, which consumable lot, which shift. The step most shops skip, and the one that stops you paying for the same defect twice.

Fix the Process While You Can

This is the part that matters beyond any single job. A shop that logs porosity causes and acts on them improves measurably: fewer radiograph rejects, less grinding and re-welding, fewer NDT calls on hold, and welds that pass the first inspection instead of the third. A qualification test or client audit is a good moment to fix a process that has been quietly producing voids.

Rework carries its own cost and hazards. Grinding out a weld puts the operator back in front of the fume, and on stainless that means hexavalent chromium exposure and local exhaust. OSHA’s welding hazards and solutions guidance and 29 CFR 1910.252 cover the requirements, and OSHA’s fact sheet on controlling hazardous fume and gases covers ventilation and respiratory protection. In a confined space, remember that a porous weld can leak: a leaking joint in a poorly ventilated area is an asphyxiation risk as well as a quality failure.

Where to Go Next

Our guides to aluminium MIG welding, aluminium flux core welding, cleaning aluminium before welding and the all-position welding rod guide cover the next steps. Aerospace TIG welding shows the discipline zero-defect work demands, and the aerospace welding guide covers gas handling for stainless and aluminium shops.

Porosity will always show up somewhere. But it will not keep appearing in the same corner of the same joint because nobody asked where the gas came from. The welder who would rather hear it from a radiograph than a customer asked first.

Frequently Asked Questions

What are the most common causes of porosity in welding?

By a wide margin it is shielding gas problems and contamination. Air getting into the shielding gas — a leak, an empty cylinder, a draft across the pool, or flow set too low or too high — is the most frequent trigger, followed by oil, grease, paint, primer, rust and zinc on the joint or the consumables. Damp electrodes, wire and flux are the leading cause of hydrogen porosity.

Can porosity be fixed without re-welding the whole joint?

Often, yes. If the porosity is localised, TWI advises removing it by localised gouging or grinding, re-preparing the joint and re-welding to the agreed procedure. If the porosity is widespread, however, the entire weld should be removed and re-welded. Cosmetic porosity on a non-critical weld may need nothing more than surface dressing.

Is porosity a serious defect or just cosmetic?

It depends entirely on the joint and the loading. A few small pores in a statically loaded structural joint rarely cause failure, but the same pores reduce the effective cross-section, act as fatigue crack initiation sites, can leak in pressure-retaining applications, and can initiate corrosion in stainless steel and aluminium. Codes set limits, but client quality departments are often stricter than the code.

Why do I get porosity only at the start of a weld?

That is almost always a gas-free start. After a break, lunch or any pause, the gas line may have been purged or the cylinder may have emptied, so the initial burst of shielding gas is not enough to displace the air in the torch or the corner pocket. The fix is to flow gas for a second or two before striking the arc, and to purge tight corners first.

Does more shielding gas always prevent porosity?

No — and this trips up a lot of shops. Flow that is too low gives poor coverage, but flow that is too high creates turbulence at the nozzle that draws room air back into the pool. As a working baseline, 50–60 CFH at a GMAW nozzle and 20–30 CFH at a GTAW torch is usually plenty, adjusted per the WPS and the gas mix.

Why is porosity a bigger problem in aluminium and stainless steel than in mild steel?

Different gases attack different metals. Aluminium and stainless steel are primarily vulnerable to hydrogen, so moisture and cleanliness dominate, and TWI recommends a chemical clean plus wire brush and degrease for aluminium. Mild steel can pick up nitrogen and oxygen as well, so shielding quality matters more, and rust and mill scale must be ground away.

How much porosity is allowed under AWS D1.1 and ASME?

Published figures cited by The Welder are: AWS D1.1 — porosity not to exceed 3/8 in. in any linear inch of weld and not to exceed 3/4 in. in any 12-in. length; ASME B31.1 — no dimension greater than 3/16 in.; API 1104 — not to exceed 1/8 in. diameter. Always confirm against the current edition of the code and the contract that governs your work.

Can porosity turn into cracking?

Not directly — but hydrogen trapped in a pore can contribute to hydrogen-induced cracking in susceptible steels such as A514, A588 and A709, and those cracks develop over time and can become catastrophic once fatigue loading is added. Porosity also creates stress concentration points, so it makes crack initiation more likely in its own right.

How do I find porosity that is hidden inside the weld?

Radiography is normally the most effective technique for detecting and characterising porosity, though small pores are hard to find in thick sections. Ultrasonic testing is used for subsurface indications, while liquid penetrant or magnetic particle inspection catches surface-breaking pores and cracks only.