Porosity in Welding: Definition, Causes, Fixes & Inspection

A fabricator once told us the story of a 40-foot pressure vessel seam that passed every visual check, only to be rejected by a radiograph riddled with pinholes the size of pepper flakes. Porosity is the defect welders often cannot see until an inspector, a leak test, or a fatigue failure puts it on the table.

What Is Porosity in Welding? A Clear Definition

Picture this: a structural joint looks textbook-perfect on the cover pass, the bead has that consistent ripple pattern inspectors like to see. Then a radiograph comes back, and the weld is riddled with tiny voids. This scenario plays out in structural shops, pressure vessel fabricators, and pipeline crews every week, because porosity is one of the most frequently cited discontinuities in inspection reports industry-wide. Before we get into causes, codes, and fixes, let’s pin down exactly what porosity is.

In the simplest terms, porosity is trapped gas. As the molten weld pool solidifies, gas that was dissolved in the liquid metal can no longer stay in solution. It forms bubbles, and if those bubbles do not escape to the surface before the metal freezes, they are locked in as voids. In aluminum work especially, this is the defect welders fight most often, and it shows up everywhere from an aluminum MIG welding project on thick plate to fine aerospace assemblies.

The AWS Definition: Cavity-Type Discontinuities

The formal wording matters when you are arguing a rejection or writing a repair procedure. AWS D1.1, the structural welding code for carbon and low-alloy steels, defines porosity as cavity-type discontinuities formed by gas entrapment during solidification or in a thermal spray deposit. In plain language: any void created because gas got caught as the metal turned from liquid to solid falls under the porosity umbrella.

A related AWS term worth knowing is piping porosity, defined as elongated porosity whose major dimension is roughly normal, meaning perpendicular, to the weld surface. When a pore like that breaks through the face of the weld, welders commonly call it a pinhole. Other code books and inspection reports use overlapping terms, which is why the vocabulary can be confusing.

The Vocabulary of Porosity: Pores, Blowholes, Pinholes and Wormholes

You will run into a handful of names in shop talk and inspection reports. They are not all interchangeable, so here is the working glossary:

– Gas pore: A single, usually rounded void caused by gas entrapment. The most basic form.
– Blowhole: A large, often spherical or irregular cavity formed by a substantial volume of gas.
– Pinhole: A small pore that breaks the weld surface. Pinholes often show up as tiny needle-like dots or dark specks in the finished bead.
– Wormhole (or worm track): An elongated, tunnel-like pore trailing in the direction of travel, left behind by a bubble that tried to rise but froze in place.
– Clustered porosity: A group of pores concentrated in one zone rather than spread evenly.
– Piping porosity: The AWS term described above, elongated pores running from inside the weld toward the surface.
– Shrinkage porosity: A related but metallurgically different void caused by metal contraction rather than gas, which can change the repair approach.

Discontinuity Versus Defect: The Distinction That Matters

This is where many welders lose arguments they should win. Codes define porosity as a discontinuity, which is simply an interruption in the normal physical structure or configuration of a weld. A discontinuity only becomes a defect when it exceeds the acceptance criteria of the applicable code or when it renders the part unfit for service.

In other words, finding porosity on a radiograph does not automatically mean the weld fails. It means the weld needs to be evaluated against limits. Section 7 of this article breaks down those limits in detail, including AWS D1.1 spacing rules and ISO 5817 acceptance levels. For now, hold onto this principle: porosity in welding is a discontinuity by definition and a defect only by judgment against a standard.

Why Porosity Matters

Three reasons put porosity near the top of every inspection checklist:

– Strength: Voids reduce the load-bearing cross-section of the weld and act as stress raisers, particularly under cyclic loading.
– Tightness: In pressure vessels, tanks, and pipe welds, surface-breaking porosity is a direct leak path, even when the weld is structurally sound.
– Corrosion and fatigue: Pores concentrate stress and can trap moisture or corrosive media, accelerating degradation far beyond what their size suggests.

The root mechanism is the same across nearly all gas porosity: molten metal holds far more gas than solid metal, so as the pool freezes it rejects the excess and the bubbles have nowhere to go. The variables that decide whether those bubbles escape, travel speed, heat input, shielding coverage, and joint cleanliness, are what the following sections unpack, process by process. We will also cover why aluminum prep matters so much and how to read pore shape and pattern as a diagnostic tool rather than just a rejection reason.

How Porosity Forms: Gas Solubility and Solidification

Porosity is not random. Every pore an inspector catches on a radiograph, and every pinhole you grind out of a bead, is the final chapter of a short and predictable story: gas dissolves into liquid metal, the metal freezes, and the gas suddenly has nowhere to go. Once you understand that sequence, you can usually predict where pores will show up before the arc is even struck.

Step 1: The Molten Pool Picks Up Gas

A welding arc breaks atmospheric molecules apart. Oxygen, nitrogen, and hydrogen from the air, from moisture, from oil, rust, or a wet electrode coating all become available to the molten puddle, which is an unusually willing host. In steel, dissolved oxygen can react with carbon to form carbon monoxide bubbles, while nitrogen taken from a compromised shield can later come out of solution as the bead cools.

Hydrogen is the usual culprit because it is everywhere. Water vapor splits on contact with molten metal, freeing atomic hydrogen that dissolves straight into the pool. The effect follows Sievert’s law: the amount of dissolved diatomic gas scales with the square root of its partial pressure in the surrounding atmosphere. That is why a humid day, a damp gas line, or a wet storage bucket can push porosity over the edge even when everything else looks unchanged.

Aluminum behaves much worse than steel here. Molten aluminium is effectively thirsty for hydrogen, absorbing it eagerly from the atmosphere, from a hydrated oxide layer, and from any solvent residue or condensation left on the joint. Miller’s aluminium-versus-steel welding guidance makes the same point from the shop floor: cleaning discipline matters far more on aluminium than on mild steel, because the pickup happens so quickly. A wiped joint that still carries a film of cleaning agent or shop dust can weld up porous on the very first pass.

Step 2: Solidification Makes Gas Unwelcome

Liquid metal can hold dissolved gas fairly comfortably. Solid metal cannot. As the pool begins to freeze, gas solubility drops, sometimes dramatically, and the excess has to leave solution the same way carbon dioxide leaves a opened soda can: it forms bubbles.

The numbers are stark for aluminium. Hydrogen solubility falls by roughly 95% during solidification, meaning almost all the gas the liquid was carrying is rejected in the narrow window where the metal turns from liquid to solid. That single figure explains why aluminium welds are so famously prone to trapped gas pores, and why a puddle that looked clean and quiet can reveal a cluster of pinholes the moment it is sectioned. Work published in Welding in the World has examined exactly this solubility-driven pore formation, and review literature on porosity in welded joints traces how pickup, rejection, and entanglement combine into the defects codes actually measure.

Step 3: Bubbles Need Something to Nucleate On

Free-floating bubbles in a pristine melt are rare. Gas pores nucleate heterogeneously, meaning they preferentially grow on existing surfaces inside the puddle: oxide particles, broken-up anodized layers, slag inclusions, and other suspended solids give the gas a head start. This is one more reason aluminium is vulnerable, since its tough oxide film can be torn into the pool by inadequate cleaning or a contaminated arc. Good preparation dissolves the problem before it starts, which is what a proper aluminum cleaner for welding routine is really doing: removing the nucleation sites along with the dirt.

Step 4: Can the Bubble Escape?

Forming a bubble is not the same as trapping one. A bubble that rises while the pool is still fluid reaches the surface, bursts, and leaves nothing behind. A bubble frozen in place becomes a spherical internal pore, or, if it clings to the fusion boundary while freezing progresses around it, an elongated wormhole or piping pore.

Two shop variables control that race:

• Travel speed. A slower travel speed keeps the pool fluid longer and gives bubbles time to float out. A fast, cold, fast-freezing puddle traps them exactly where they were forming.
• Heat input. Adequate heat keeps viscosity low and convection active, letting buoyancy do its work. Too little heat, too long an arc, or an incorrect gun angle produces a shallow, sluggish pool that cannot release gas.

That distinction matters when you troubleshoot. Gas porosity tells you to look at atmosphere, moisture, and cleanliness. Shrinkage porosity tells you to look at solidification geometry. The welding defects reference guide lays out both patterns side by side, and learning to tell them apart at a glance is the first real skill in porosity diagnosis. The next section turns that skill into a full catalogue of shapes and what each one signals.

Types and Shapes of Porosity You Will Encounter

Porosity is not one defect but a family of them, and the pattern matters as much as the quantity. Two welds can both contain “gas pores” and tell completely different stories: one is a minor, isolated nuisance, the other is a shielding failure that will show up again on the next joint. Learning to read the pattern on a bead crown is one of the fastest diagnostic skills a welder can build. Below are the recognized forms, what they look like, and what each one usually signals.

Isolated or Spherical Pores

Individual, rounded cavities scattered across the bead surface or buried just under it. Isolated spherical pores are the mildest form. A single small pore near the toe of a weld is a discontinuity, not necessarily a defect, and many codes tolerate a limited number of them. Diagnostically, scattered isolated pores point to marginal conditions rather than a hard failure: slightly dirty base metal, a borderline gas flow setting, or a travel speed a touch too fast for the heat input.

Surface Porosity: Craters, Pinholes and Grainy Beads

Surface porosity appears as craters, pinholes, or a “grainy,” bumpy bead crown, and it is usually visible without any nondestructive testing. You can see it, and often feel it with a gloved thumb. Welders sometimes describe a badly shielded bead as “peppery” or “cheesy.” Because it breaks the surface, this form is also the easiest to remedy: grind it out and re-weld. Pinholes sit at the smaller end of the spectrum, tiny blowholes that vent gas right at the solidifying surface.

Clustered Porosity

Grouped pores concentrated in one zone rather than spread evenly. Clustered porosity typically points to a localized contamination source: a spot of oil, paint, or moisture on the joint, a smear of grinding residue, or a shielding gas lapse at one specific point in the weld. If the cluster lines up with where you paused, restarted, or changed gas cylinders, that location is your evidence.

Linear or Aligned Porosity

Pores arranged in a row, usually along the weld centerline or along the fusion line of a pass. Aligned porosity suggests a systematic cause tied to the arc itself rather than random dirt: consistent shielding interruption, a wire that picked up surface moisture along an entire length, or deoxidation failing pass after pass. Linear patterns concern inspectors more than scattered ones because they concentrate weakness along a line, which is exactly where cracks like to start.

Subsurface Porosity

Pores fully embedded beneath a sound-looking surface. Subsurface porosity requires radiography or ultrasonic testing to detect, and it is the type most often caught during code inspection. The bead looks clean, then the inspector’s radiograph comes back with a string of dark round indications. This is why many fabricators routinely shoot film or use phased-array UT on critical work even when the surface passes visual inspection.

Piping Porosity

Elongated vertical voids trailing upward from a subsurface point and opening toward the bead surface, often visible on a radiograph as comet-shaped tails. Piping porosity deserves special attention because AWS D1.1 treats it with explicit limits: the code sets maximum lengths for piping porosity depending on material thickness, and its acceptance differs from ordinary rounded pores. A trail of it usually means gas was still evolving as the top of the pool froze.

Wormholes and Trail Porosity

Wormholes are long, elongated, often snaking cavities. They indicate directional gas evolution, usually tied to arc blow, contamination on one side of the joint, or a shielding gas jet disturbed by a draft. In flux-based processes, wormholes frequently trace back to flux conditioning problems or, in out-of-position work, to travel speed that outran the gas escape path.

Blowholes

Blowholes are large gas cavities, common in castings and heavy multi-pass deposits, and they can compromise the full cross-section of a weld. A cavity big enough to eat through the throat of a fillet is a rejection every time, and on pressure-boundary welds it is an automatic removal.

Shrinkage Porosity

Spongy, irregular micro-void networks caused by solidification shrinkage rather than gas pickup. They cluster in the last liquid to freeze, typically the center of a heavy bead or the top of a crater. The distinction matters for repair: grinding out and re-welding fixes a gas pore, but if the root cause is an improperly filled crater or an undersized final pass, the same pattern returns on the next weld.

For a broader catalog of how these patterns appear on production welds, the welding defects guide from Xiris is a useful visual reference, and for laser and high-speed processes where porosity forms differently, see our Alpha Laser Welding: Systems, Uses, and Safety Guide.

Why Porosity Happens: Causes by Material and Filler

Every pore starts with gas trying to escape a molten pool that will not let it. The next question is where that gas came from, and the answer changes with the metal under your arc: carbon steel, aluminum, stainless, and nickel alloys each have a signature porosity mechanism. That is why a fix that works on mild steel does nothing for an aluminum hull.

Carbon and Low-Alloy Steel: Carbon Monoxide, Nitrogen, and Deoxidation

On carbon steel the usual suspects are carbon monoxide and nitrogen. Carbon monoxide forms when carbon in the melt reacts with oxygen that entered the arc zone from mill scale, rust, oil, cutting fluid, or a shielding gas stream that broke down. As the weld solidifies, CO is no longer soluble, so it nucleates bubbles, often at the root or along the bead centerline.

Nitrogen porosity is different: it comes from air sweeping into an arc that lost its shielding. The nitrogen dissolves in the liquid metal and comes out of solution during freezing, typically producing fine, evenly distributed pinholes rather than large blowholes.

The defense is deoxidation. Fillers carrying silicon, manganese, titanium, and aluminum scavenge dissolved oxygen before it can reach carbon, forming stable oxides instead of gas. That is why a killed steel paired with a properly deoxidized filler is far more forgiving than a rimmed or poorly killed heat. Our guide to alloy steel welding electrodes walks through those filler selections.

Aluminum: The Hydrated Oxide Layer and Hydrogen

Aluminum is hydrogen’s playground. Its oxide layer is porous and chemically bonded to water, so a freshly cleaned plate left in humid air rehydrates within hours. That hydrated oxide is the dominant hydrogen source, and hydrogen is exactly the gas aluminum weld metal rejects violently as it freezes. Mechanical or chemical cleaning removes the layer and its water together, which is the purpose of proper aluminum prep before welding.

Historical research in the Welding Journal found that atmospheric contamination of the shielding gas, meaning air leaking into or across the gas envelope, was the greatest single source of porosity in aluminum alloy welds. Shop practice still reflects it: on aluminum, chase the gas system and surface moisture before blaming technique. If you run aluminum wire, our aluminum MIG welding guide covers the gas coverage, drive rolls, and liner care that control this.

Stainless Steel: Ferrite Loss and Limited Deoxidation Backup

Austenitic stainless with little or no ferrite is more prone to porosity than welds holding roughly 4 to 10 percent ferrite. Ferrite raises fluidity and opens the pool structure so gas can escape; fully austenitic compositions, especially those pushed to very low ferrite for corrosion or high-temperature service, hold bubbles until they freeze. Austenitic fillers also carry far fewer strong deoxidizers than carbon steel fillers, so there is less chemical margin when oxygen or nitrogen intrudes. The practical consequence: stainless needs tighter gas coverage and cleaner joints, not different amperage.

Nickel Alloys and Other Specialty Metals

Nickel alloys trap gas easily because their weld pools are viscous and freeze quickly, giving bubbles little time to rise. Sulfur, lead, and other low-melting contaminants make it worse, so joint cleanliness is the controlling variable. Procedures for values like Alloy 20 are written around strict pre-weld cleaning; see our Alloy 20 welding procedures.

Consumables, Coatings, and the Contaminant List

Across every base metal, moisture in coatings and electrodes raises diffusible hydrogen, which is why AWS low-hydrogen classifications exist. An electrode such as E7018-H4R is qualified to limit diffusible hydrogen to 4 mL per 100 g of deposited weld metal, a threshold defined in the AWS D1.1 structural welding code (AWS D1.1/D1.1M). Beyond moisture, documented porosity triggers include:

• Oil, grease, and cutting fluids trapped in the joint
• Paint, primer, and marking crayon near the fusion zone
• Mill scale, rust, and mill lubricants on hot-rolled plate
• Lint from wiping cloths, and oil from leather gloves
• Wet or damaged gas hoses, leaking regulators, and contaminated gas cylinders

Current research keeps landing on the same conclusion: contamination, not arc physics, is the usual root cause. Reviews of fusion welding defects trace most gas porosity back to surface and consumable contamination rather than inherent process behavior (Springer, 2024).

Source	Gas introduced	Most affected metals
Hydrated oxide / surface moisture	Hydrogen	Aluminum, magnesium
Oil, paint, cutting fluid	Hydrogen, carbon, oxygen	Carbon steel, stainless
Mill scale and rust	Oxygen, carbon monoxide	Carbon steel
Moist electrode coatings	Hydrogen	Carbon and low-alloy steel
Air entrainment in the arc	Nitrogen, oxygen	All metals, critical on stainless

The pattern to remember: identify the gas, then find its source. Hydrogen points to moisture and coatings. Oxygen points to shielding and scale. Nitrogen points to air leakage. Know which gas you are fighting, and troubleshooting in the next section becomes a short list instead of a guessing game.

Porosity Causes by Welding Process: SMAW, GMAW, GTAW, FCAW, SAW

Porosity rarely shows up for one universal reason. It shows up for process-specific reasons, and once you know which process you are running, the suspect list gets short fast. A welder who lays down clean stringers all week on MIG and suddenly gets pinholes on Monday is almost never dealing with a mystery metal problem. Usually something changed in the gas line, the wind, the electrode storage, or the pass schedule.

Process	Most common porosity source	Quick field check
GMAW / GTAW	Gas flow too high or too low, drafts, leaky hoses, empty cylinders	Soap-test hose joints; watch the gas envelope at the nozzle in a breeze
SMAW	Moisture or damage to low-hydrogen electrode coating	Check electrode oven and how long rods have been exposed
FCAW	Trapped slag between passes; wind on self-shielded wire	Clean every pass; confirm gas is flowing on gas-shielded wire
SAW	Damp flux, rust or oil under the flux bed	Check flux bake date; vacuum and dry the joint before welding

GMAW and GTAW: It Is Almost Always the Gas Envelope

Gas-shielded processes are the biggest offenders, and the counterintuitive part is that shielding gas problems work in both directions. Too little gas leaves the arc under-protected. But excessive or turbulent flow can aspirate ambient air right into the arc zone and cause porosity just as reliably as too little gas. Flow rates are not a “more is safer” setting. Follow the machine and wire manufacturer’s recommendation, usually somewhere in the 20 to 35 CFH range for GMAW, and increase it modestly for out-of-position or torch-angled work. Our aluminum MIG welding guide covers gas setup in depth, because aluminum, which is fussier than steel about hydrogen pickup, is where sloppy gas discipline gets punished fastest.

After flow rate, the next suspects are physical: leaking hose fittings, a cracked gun liner, a missing or stuck check valve, a damaged regulator, or a cylinder running so low that pressure drops and air can be drawn back through the regulator. A partially empty bottle left out in the rain can also carry moisture into the mix. This is one reason reputable gas suppliers and shop procedures push cylinder exchange before a bottle drops below roughly a quarter full.

Outdoors, drafts dominate. Wind speeds that disrupt the gas envelope appear at surprisingly low velocities and are one of the most common field causes of sudden porosity in open-air GMAW and GTAW work. If the shop cannot move the work inside, use wind screens, shorten the stick-out, or switch to a process less sensitive to air movement. For TIG technique fundamentals and how arc length changes everything, see our aerospace TIG welding guide.

SMAW: The Coating Is the Gas Shield

Stick electrodes shield the arc through the burning coating, which means damaged coatings become porosity generators. Low-hydrogen electrodes (E7018 and friends) are especially sensitive: the coating absorbs moisture from shop air, and that moisture becomes hydrogen in the weld pool. AWS D1.1 requires low-hydrogen electrodes to be stored and re-dried according to the manufacturer’s recommendations (AWS D1.1), and exposure limits are measured in hours, not days. Rods left out overnight on a machine deck are a routine, preventable source of hydrogen pickup. For the broader stick-electrode context, our advantages of stick welding and alloy steel welding electrode guides cover selection and storage practices.

Technique matters too. An arc length that is too long on a thin-coated rod, dragging the electrode at too shallow an angle, or pausing and letting the puddle freeze over gas that has not escaped all produce the same honeycomb of small pores.

FCAW and SAW: Slag and Flux Conditioning

In flux-cored and multi-pass stick work, inadequate cleaning between passes allows trapped slag to read as porosity-like indications on radiography. Those rounded dark spots on a film are not always gas pores, and misreading them wastes repair time. Chip, wire-brush, and grind the crown of each pass before the next one goes over it. Self-shielded flux core has an extra vulnerability: no external gas to fall back on, so it is highly sensitive to wind, while gas-shielded flux core shares all the GMAW gas problems above. Our aluminum flux core welding piece explains where flux-cored processes genuinely work well and where they struggle.

Submerged arc welding adds flux conditioning to the list. Damp or contaminated flux, oil or rust under the flux bed, and un-baked recovered flux all reintroduce moisture and contaminants into the pool. Flux should be baked and stored dry, and joints cleaned and dried before the blanket goes down.

How Porosity Is Detected: Visual, Penetrant, Magnetic Particle, RT and UT

No inspection method sees everything. The available tools fall into two families: surface techniques that only find pores breaking the weld face or toe, and volumetric techniques that look through the full thickness. A seam can pass every surface check and still hold a cloud of internal voids, which is why the pressure vessel in our opening story passed visual inspection and was rejected by its radiograph.

Visual Inspection (VT) and Dye Penetrant Testing (PT)

Visual inspection is always step one. With decent light, a cleaned surface and a bit of magnification, an experienced eye catches surface-breaking pinholes, blowholes, crater pipes and the telltale rough texture of entrapped gas. It is fast, cheap and required by virtually every code, but its hard limit is simple: light cannot travel through steel. Anything sealed beneath the weld surface is invisible to VT.

Dye penetrant testing (PT) extends that reach by a hair. A liquid penetrant soaks into any opening, is wiped away, then drawn back out by a developer as a visible bleed line. Fluorescent versions glow under ultraviolet light and reveal pinholes far too fine for the naked eye. The critical requirement is cleanliness: oil, oxide, paint and shop grime block the penetrant from entering the pore and produce a false negative. That is why a proper aluminum cleaning procedure matters as much for inspection as it does for the weld itself. PT finds only surface-breaking discontinuities and struggles on rough or absorbent surfaces. In aerospace welding work, where a single pinhole can ground an aircraft component, PT is a routine complement to visual checks. See the aerospace welding guide for how inspection requirements tighten in that sector.

Magnetic Particle Testing (MT)

Magnetic particle testing goes one layer deeper. The part is magnetized, and fine magnetic particles dusted over the surface collect where a defect leaks the magnetic field. MT finds surface porosity and, importantly, near-surface voids lying just beneath the face, typically a millimetre or two down, which no pure surface method can see. The catch is material: it only works on ferromagnetic metals. Austenitic stainless steel, aluminum and most nickel alloys cannot be magnetized, so PT or RT has to do that job instead.

Radiographic Testing (RT): The Classic Answer

Radiographic testing remains the standard way to see inside a weld. X-rays or gamma rays pass through the joint onto film or a digital detector, and pores show up as dark rounded spots because less material absorbs the beam there. RT reports exactly what a code reader needs: pore size, distribution, spacing and alignment through the thickness, whether the porosity is isolated, clustered, linear or piping. That is why RT underpins most code acceptance readings of porosity. Its limitations are practical rather than technical: you need access to both sides, the source must be positioned correctly for the geometry, and radiation safety zones must be controlled. Orientation also matters, since flaws running parallel to the beam can be under-revealed, though roughly spherical pores are about the easiest defect there is to catch radiographically.

Ultrasonic Testing (UT) and AWS D1.1 Acceptance

Ultrasonic testing bounces high-frequency sound waves through the weld and reads the echoes. It needs access to only one side, carries no radiation hazard, and is sensitive to internal reflectors. The challenge is interpretation: a rounded pore cluster and a slag inclusion can produce similar A-scan responses, and separating them honestly takes an experienced technician. AWS D1.1 includes specific UT acceptance and rejection criteria, with different allowable provisions for statically loaded and cyclically loaded connections, reflecting the fact that vibration and fatigue make the same porosity far less tolerable. A general welding defects reference is useful for seeing how porosity fits among the other discontinuities inspectors are trained to separate.

Phased Array, X-ray CT and Research Methods

Newer techniques are closing the gaps. Phased-array ultrasonic testing (PAUT) steers and focuses the beam electronically, producing sectorial and layered scans that characterize pore clusters in ways conventional UT cannot. At the research end, X-ray micro-computed tomography reconstructs a weld in three dimensions, exposing pore shape, connectivity and proximity to the surface that no flat radiograph can show. Studies combining micro-CT with LECO hydrogen analysis are being used to correlate a weld’s hydrogen content with its three-dimensional pore networks (see this materials study and this related research). These tools excel at process development, but scan time, equipment cost and specimen-size limits keep them out of routine production inspection.

Method	Finds	Main limitation
VT (visual)	Surface-breaking pores, blowholes, rough bead texture	Nothing subsurface
PT (penetrant)	Fine surface-breaking pinholes and channels	Surface only; needs a spotless part
MT (magnetic particle)	Surface and shallow near-surface voids	Ferromagnetic materials only
RT (radiography)	Internal pore size, count, distribution, alignment	Two-sided access and radiation control
UT / PAUT	Internal reflectors, clustered and piping porosity	Skilled interpretation required
Micro-CT (research)	Full 3D pore networks and connectivity	Slow, costly, not a production tool

The practical takeaway is that the detection method shapes what you are allowed to conclude. A weld that “looks clean” has only proved it has no visible surface pores, and knowing that distinction is what stops a visual pass being mistaken for a final acceptance.

Acceptance Criteria: When Porosity Becomes a Weld Defect

Here is the distinction that settles arguments on the shop floor: a porosity indication is a discontinuity by default, and it only becomes a defect when it exceeds the acceptance criteria of the governing code. The welder who sees three pinholes and the inspector who stamps the joint read the same bead through different rulebooks. Knowing what those rulebooks say turns a subjective debate into a measurable one.

AWS D1.1: The Numbers That Matter

In the United States, AWS D1.1/D1.1M, Structural Welding Code – Steel governs most structural work, and its porosity rules live in the visual inspection and radiographic acceptance provisions. The limits are not one number; they depend on joint type, loading, and inspection method.

Situation	AWS D1.1 limit
Complete joint penetration groove weld in a butt joint transverse to computed tensile stress	No visible piping porosity permitted
Other groove and fillet welds	Sum of visible piping porosity 1/32 in. (1 mm) or larger must not exceed 3/8 in. (10 mm) in any line of weld
Fillet welds, porosity frequency	No more than one instance of piping porosity per 4 in. (100 mm) of weld length
Maximum piping pore diameter	3/32 in. (2.5 mm)

Two of these deserve a closer look. The blanket ban on visible piping porosity in transverse-tension CJP butt joints exists because that weld carries the full calculated load with no redundancy to hide behind. And the “any line of weld” wording matters: porosity is measured as a running total along a line, so small pores in a row add up faster than the same pores scattered randomly.

D1.1 also imposes a spacing rule on larger cavities and voids. Unless the drawing says otherwise, voids 1/16 in. (1.6 mm) or larger must be separated by at least the maximum dimension of the larger void. The logic is straightforward: two nearby pores behave like one elongated flaw, and coalescence is exactly how a cavity becomes crack-like.

Static Versus Cyclic Loading

Porosity tolerance shrinks as service loading becomes less forgiving. Under static tension, a few rounded, well-dispersed pores have limited effect on load capacity, because surrounding metal simply carries around them. Under cyclic loading, each pore is a stress concentrator and a potential fatigue crack initiation site, so the tightest porosity and fusion requirements land on connections in cyclic tension. This is the code-level echo of decades of fatigue research: a single pore at a fillet weld toe can cut fatigue life dramatically, while the same total porosity buried mid-weld where stresses are lower may barely register.

That split drives inspection planning. A column splice designed for static load and a crane runway girder detail seeing millions of cycles do not get the same scrutiny, even when the beads look identical. Shops serving critical industries already work this way. A firm such as Alamo Welding & Boiler Works, handling pressure and boiler work, builds its acceptance discipline around service consequences rather than visual convenience.

ISO 5817: A Graded International Framework

Outside the United States, and increasingly on international projects, ISO 5817 is the common yardstick for arc-welded joints in steel, aluminum, and nickel alloys. Rather than a single pass/fail line, it defines three quality levels for each imperfection type, including porosity:

• Level B — highest quality, with the tightest limits on pore size, count, and distribution. Typical for cyclically loaded structures, pressure-retaining components, and aerospace work, where organizations follow the standards covered in our guide to aerospace welding processes, standards, and careers.
• Level C — medium quality, the everyday default for general structural fabrication.
• Level D — the lowest permitted level, suitable for statically loaded, non-critical joints where leak-tightness and fatigue are not design drivers.

Each level specifies maximum individual pore diameter, the number of pores or total porosity percentage per unit length, and rules for clustered or linear distributions. The practical value of the scale is that a drawing can simply state “quality level C to ISO 5817” and remove the ambiguity that fuels disputes.

ASME Section VIII and B31.3

Pressure equipment and process piping follow a different philosophy again. Under ASME Boiler and Pressure Vessel Code Section VIII, Division 1, radiographic acceptance of porosity is tied to the joint’s construction category: joints built to full radiography meet the strictest limits, while spot- or non-radiographed joints get correspondingly more lenient treatment based on weld thickness. ASME B31.3 for process piping goes further, linking inspection extent and acceptance to fluid-service category, so a hazardous-service line is held to a much tighter standard than a utility water line in the same plant.

Reading the Criteria Before You Strike the Arc

The cheapest rework is the weld that was never at risk. Before the first pass, identify the governing code, the loading regime, the inspection method that will be applied, and the acceptance table that goes with it. If the criteria call for radiography of a CJP butt joint in tension, plan for zero visible piping porosity and weld accordingly: clean joint, dry consumables, verified shielding. Acceptance criteria are not paperwork. They are the specification for how the weld has to be made.

Effects of Porosity on Weld Strength, Fatigue, Leaks and Corrosion

Finding pores in a bead does not automatically mean the weld is scrap. What matters is how much metal the pores remove, where they sit, and what the joint has to do in service. A cosmetic frame bracket and a cryogenic tank seam can hold the same number of bubbles and get completely different verdicts. Here is what porosity actually does to a welded joint.

Tensile Strength and Ductility: What NASA Measured

The cleanest public data on porosity versus mechanical properties comes from NASA research on the aluminum alloys 2014-T6 and 2219-T87, both used in aerospace structures. Across the tensile properties they measured, ultimate tensile strength showed the strongest linear relationship with porosity level of any property tested. In plain terms: as the volume fraction of pores went up, the maximum stress the specimen could carry came down in a predictable, straight-line fashion.

Two other findings matter just as much. At high porosity concentrations, both yield strength and elongation decreased markedly, which tells you that ductility loss arrives later than strength loss. A weld can give up a noticeable share of its ultimate strength while still bending and stretching normally, then start behaving in a brittle way once pores become severe. That is why a weld that “looks fine and bends fine” in a shop check can still be under-strength. Practical aluminum work, from Aluminum MIG Welding: The Complete Practical Guide to structural fabrication, leans on this: shielding and cleaning are strength concerns, not cosmetic ones.

Fatigue: The Pore’s Distance From the Surface Matters Most

Fatigue is where porosity earns its reputation. Cyclic loading concentrates stress at notches, and an internal void behaves exactly like one. Fatigue testing consistently shows that pores located close to the specimen surface are the most detrimental to fatigue life in welded joints, because the crack initiates where tensile stress is highest and can grow immediately into sound metal.

A study published in Welding in the World quantified this: pores sitting closer to the surface than about 13 percent of the sheet thickness can cut the applicable design fatigue curve by up to 4 FAT classes. FAT classes come from the IIW nominal-stress approach, where each class number is the stress range in megapascals the detail survives at two million cycles. Losing four classes means the joint must be designed for a much lower stress range, which usually translates into more material, more weld, or a shorter service life. For a fatigue-critical detail in Aerospace Welding: Processes, Standards, and Careers Guide territory, that is the difference between a qualified joint and a redesign.

Leak Paths: Why Pressure and Pipe Welds Fail on Tightness, Not Strength

Pipelines, boilers, and pressure vessels rarely reject a porous weld because it collapsed. They reject it because gas or fluid found a route through it. Porosity in the root pass of a pressure or pipe weld creates a connected or near-connected path from the inside surface to the outside, while the static tensile strength of that weld stays perfectly acceptable. Leak-before-break logic does not help if the leak itself is the failure.

This is exactly why AWS D1.1 applies stricter radiographic acceptance rules to piping porosity than to scattered isolated pores, and why radiographic or ultrasonic testing is mandatory rather than optional in many pressure code welds. Fabricators working on boilers and vessels, such as those profiled in Alamo Welding & Boiler Works: Services & Boiler Expertise, know that the root pass is the one place where a visually perfect cap hides the only defect that matters.

Corrosion: Trapped Contaminants and Crevice Attack

Surface-breaking pores give corrosive media somewhere to sit. Salt water, condensation, or process chemicals collect in a pore and stay there, turning a small void into a miniature crevice. In stainless steels and nickel alloys this can trigger localized crevice or pitting attack long before the surrounding weldment shows any sign of trouble. Subsurface pores are less of an issue unless they sit near enough to break through later, but in immersed structures like hulls they can eventually open and start feeding a leak-corrosion cycle.

Where Porosity Is Tolerable and Where It Is Not

Application	Typical tolerance	Why
Static, non-critical bracket or cosmetic weld	Some scattered porosity usually acceptable	Loads are low; no cyclic stress; no pressure
Cyclically loaded structural member	Tight, often near-zero allowance	Surface pores cut fatigue design curves sharply
Pressure vessel, boiler, piping root pass	Very strict; RT/UT required	Leak path risk even at acceptable strength
Cryogenic or vacuum service	Effectively none	Pores provide leak paths and stress concentrators at temperature extremes
Corrosive or marine environment	Surface porosity not tolerated	Trapped media drives crevice and pitting attack

The takeaway is straightforward: porosity reduces strength roughly in proportion to the cross-section it removes, but it punishes fatigue, leak tightness, and corrosion far harder than a simple strength calculation would suggest. Strength tests forgive porosity longer than service does.

Preventing Porosity: Practical Fixes and Common Mistakes

Porosity is one of the few welding defects that is almost always preventable at the source. It comes from a chain: a dirty plate, a damp rod, a pinhole in a gas hose, a gust from an open door. Work that chain in reverse and you stop gas pores before the inspector sees them.

Start With the Three-Question Diagnostic

Before you change anything, answer three questions: which gas, which spot, and when did it start? Those answers narrow the cause faster than any equipment swap.

Step 1: Clean the Joint the Way It Actually Needs It

Remove everything that can boil in the arc: oil, grease, paint, primer, mill scale, cutting fluids, marker ink, and moisture. A shop rag can undo that work, because lint and oil-tanned leather have been documented as hydrogen sources in welding operations. Use lint-free wipes and cotton gloves, plus a dedicated aluminum cleaner for welding on aluminum, where the hydrated oxide layer is the main hydrogen supplier.

Let solvents flash off completely before striking an arc; trapped solvent under a hot bead produces gas where you can least afford it.

Step 2: Control Moisture and Store Consumables Properly

Hydrogen enters the puddle through three doors: atmosphere, base metal, and filler. You control the third completely.

• Store low-hydrogen SMAW electrodes in sealed holdovers or a rod oven, and re-bake per the manufacturer’s schedule.
• Keep FCAW and SMAW rods in a dry room; one night on a damp floor pushes moisture into the flux.
• Condition SAW flux before use and never top off a batch with unchecked leftovers.
• Wipe condensation from gas hoses and regulators when the shop temperature swings.

Step 3: Verify the Gas System Before You Blame the Arc

Check the whole circuit: correct gas, regulator settings, flow at the gun rather than the gauge, hose condition, liner wear, and tip-to-work distance. Typical culprits: a cracked hose, a leaking fitting, a blocked nozzle, a clogged tip. Change cylinders with the arc off, then re-check flow at the gun; a nearly empty bottle reads fine on the gauge and delivers nothing at the puddle. A reputable gas supplier can confirm your mixture and purity when a batch behaves badly.

Stop the wind: open bay doors, fans, grinders, and crane traffic all displace shielding gas. Enclose the work, drop a wind screen, or shift position. The Miller welding knowledge base recommends flow rates low enough to avoid turbulence but high enough to survive minor drafts, the balance most shops miss by cranking flow higher.

Step 4: Set Heat High Enough for the Gas to Escape

A puddle that stays fluid longer gives hydrogen time to rise before solidification locks it in. Shorten the arc, slow the travel speed, and avoid running cool and fast on thin material. On aluminum this matters more: hydrogen’s solubility drops drastically at freezing, so helium-blend shielding gases are recommended where deeper penetration and faster hydrogen escape are needed.

For carbon steel, the most direct control against carbon monoxide porosity is filler selection: choose a wire with adequate deoxidizer content when the plate still carries mill scale. A heavily deoxidized ER70S-6 wire absorbs minor surface contamination that an S-2 wire exposes immediately.

Step 5: Lock In Technique

Arc length, work angle, and steady travel speed prevent more porosity than any machine knob. Long arcs pull in air, fast travel traps gas, and excessive weaving cools the leading edge of the puddle so bubbles hide under the ripples. Keep the arc at the leading edge so the gas escapes behind you.

The Most Common Mistakes Shops Make

Mistake	Why it causes porosity	Fix
Welding over fresh paint or cutting oil	Organics vaporize into the puddle	Grind and degrease 1 in. beyond the joint
Reusing a dirty wipe or glove	Lint and tanned leather release hydrogen	Use lint-free wipes and cotton gloves
Cranking gas flow up to fix it	Turbulence pulls air into the arc	Return to spec; fix the leak
Running a near-empty or wrong cylinder	Wrong mixture, no arc coverage	Log changes; verify the mix
Wet or out-of-date electrodes	Coating moisture breaks down in the arc	Re-bake or replace; store in holdovers
Traveling too fast	Bubbles freeze in place	Slow down; widen the puddle

The pattern on the finished bead is your guide: isolated pores at the start mean clean-up issues, a porous cluster in one spot means a local contaminant, and uniform porosity across every part means the process is off. Fix the cause and run a test coupon before production. Prevention is a habit: check the joint, moisture, gas, parameters, environment, and technique in that order, every time.

Repairing Porosity, When to Reweld, and Final Takeaways

The radiograph, penetrant test, or leak test flagged porosity in a bead that looked clean to the eye. Next comes a strict sequence: remove, verify, re-weld, re-inspect. Skip a step and a routine repair becomes a second repair, with a bigger groove and a harder conversation with the inspector.

Step One: Remove the Metal Back to Sound Material

You cannot weld over porosity and hope the next pass seals it. Any cavity left behind stays open to the new weld pool, and its gas migrates into the fresh bead. The affected metal must come out by grinding, arc-air gouging, or machining until the groove is free of visible and test-confirmed voids, with feathered edges for a clean tie-in.

Arc-air gouging is the fastest way to chase a long line of linear or piping porosity, but it leaves a carburized layer that must be ground off before rewelding. If the torch is unfamiliar, our walkthrough of air arc welding and carbon arc gouging covers depth control and cleanup.

Once the metal is out, run magnetic particle or dye penetrant over the prepared groove to confirm the cavities are gone. Repairing a groove that still holds pores is the most common reason a weld fails its post-repair film.

Step Two: Know Whether Repair Is Allowed at All

Repair is not an automatic right. Under AWS D1.1, repair work must meet every requirement applied to original welds, and when the nonconformance results from the contractor’s own work, the repair is done at the contractor’s expense, gouging, rewelding, and reinspection included.

Some specifications are stricter. Pressure vessel, boiler, and certain piping work may prohibit porosity repair in dynamic or cyclically loaded joints outright, or require written approval from the engineer of record before an arc is struck. On a critical joint, attach that approval to the welder’s repair record before you start.

Radiography is the standard post-repair confirmation, since the replacement weld is where an inspector looks hardest. If the second film fails, the removal zone grows, so the first repair pass is the one that must be right.

Step Three: Reweld Without Reintroducing the Problem

Reweld to the original procedure: same filler metal, shielding gas, and heat input range, with preheat and interpass temperatures re-established rather than assumed from surrounding metal. Clean the groove once more right before welding, since oil, moisture, and grinding dust settle fast in an open preparation. On aluminum, remove the fresh oxide layer with a stainless wire brush and a proper aluminum cleaner for welding, since hydrogen from a hydrated surface reproduces the original defect.

Technique matters more in a repair than in production. Use stringer beads over wide weaves, keep the arc short, hold a consistent gun angle, and slow down enough for trapped gas to rise out of the puddle. Check gas flow and hose integrity before the arc starts, not after the first pore appears.

Pre-Weld Checklist: Run It Before Every Arc

1. Joint cleanliness: bright metal; no oil, paint, mill scale, or moisture.
2. Gas system: correct gas and flow, no kinks or leaks, draft shields up.
3. Consumables: low-hydrogen rods and flux in ovens or sealed containers.
4. Environment: fans and doors routed away; wind screens outdoors.
5. Parameters: voltage, wire speed, travel speed, stick-out, and preheat verified against the WPS.
6. Technique: gun angle, short arc, steady travel, no over-wide weaving.
7. Documentation: welder ID, procedure, heat number, and inspection plan recorded up front.

Most porosity traces back to one of those seven lines being skipped. When a pore appears, ask three questions: which gas (or its absence), which spot, and when it started.

Conclusion: The Definition Worth Remembering

Porosity in welding is a cavity-type discontinuity: a void left when gas dissolved in the molten pool is trapped as the metal solidifies. It becomes a defect only when it exceeds the acceptance criteria of the governing code, which is why knowing the limits matters as much as knowing the causes. Four priorities cover nearly all of it: clean, dry joints; intact shielding; properly stored consumables; and parameters that let trapped gas escape before the puddle freezes.

Get those four right and you will rarely see porosity on your film. Catch it early and you fix it with a grinder. Catch it late and it costs a radiograph, a gouge, and a repair that has to pass inspection all over again.

Frequently Asked Questions

What is porosity in welding?

Porosity is a cavity-type weld discontinuity formed when gas becomes trapped in the molten weld pool and cannot escape before the metal solidifies. AWS defines it as cavity-type discontinuities formed by gas entrapment during solidification. It becomes a defect only when it exceeds the applicable code’s acceptance criteria.

What causes porosity in a weld?

The main causes are hydrogen from moisture and contaminants, oxygen and nitrogen reactions in the arc, shielding gas disruption from drafts or incorrect flow, dirty base metal or filler, and consumables that are wet or improperly stored. The specific gas depends on the material: hydrogen dominates in aluminum, while carbon monoxide and nitrogen dominate in carbon steel.

What is the difference between surface porosity and subsurface porosity?

Surface porosity breaks the bead surface as pinholes, craters, or a grainy crown and can be caught by visual inspection. Subsurface porosity is hidden inside the weld and requires radiography or ultrasonic testing to detect, which is why it commonly appears only during formal code inspection.

How much porosity is allowed under AWS D1.1?

It depends on the joint and loading. For example, complete penetration groove welds transverse to tensile stress allow no visible piping porosity, while other groove and fillet welds limit the sum of piping porosity 1/32 in. or larger to 3/8 in. in any line, and fillet welds are limited to one instance per 4 in. of weld length with a maximum diameter of 3/32 in.

Does porosity weaken a weld?

Yes, but the effect depends on amount, size, and location. NASA testing showed ultimate tensile strength correlates strongly with porosity level, and elongation drops sharply at high porosity concentrations. Pores near the surface are especially damaging because they act as fatigue crack initiation sites.

Can a weld with porosity be repaired?

Yes, if the code and engineer allow it. The porous metal must be fully removed by grinding, gouging, or machining back to sound material, then re-welded using the original qualified procedure. Repaired welds are typically re-inspected radiographically to confirm the porosity is gone.

How do you prevent porosity when welding aluminum?

Remove the hydrated oxide layer before welding, keep the surface and filler free of oil and moisture, verify shielding gas purity and flow, avoid drafts, and use enough heat to keep the puddle fluid long enough for hydrogen to escape. A helium-argon blend can also help reduce porosity.

What is the difference between porosity and shrinkage?

Porosity is gas trapped as cavities during solidification, while shrinkage porosity forms when liquid metal contracts as it freezes and no liquid metal is available to fill the void. Both appear as cavities on radiography, but their causes and remedies differ.