Porosity in Welding: Ways to Prevent & Detect It

Porosity in welding is gas (hydrogen, nitrogen or oxygen) trapped as cavities when the weld pool freezes. Whether pores are rejectable depends on the code. AWS D1.1 lets a statically loaded fillet weld show visible pores totaling 3/8 in. in any inch, but a cyclically loaded one only one pore, at most 3/32 in., per 4 in.

That difference is why porosity has to be handled at two levels. The welding process decides whether pores form at all, and the acceptance code decides whether the pores that did form can stay. This guide covers both: the cause chain and how to stop it, the forms porosity takes, the US acceptance limits in AWS D1.1, ASME Section VIII, ASME B31.3 and API 1104, and why radiography usually finds rounded pores better than ultrasonic testing.

Contents: What causes it · Types · Prevention · Detection methods · RT vs UT · US acceptance limits · Worked examples · Repair · Limits · FAQ

About the numbers: AWS D1.1, ASME and API 1104 are paywalled. The limits below come from the publishers' public pages and from secondary sources that agree with one another. Check every value against the edition named in your contract before accepting or rejecting a weld.

What causes porosity in welding?

Porosity forms when more gas enters the weld pool than can escape before it solidifies. TWI describes the cavities as:

"caused by the freezing in of gas released from the weld pool as it solidifies." (TWI Job Knowledge 42)

The three gases that matter are hydrogen, nitrogen and oxygen. They arrive by two main routes:

• Nitrogen and oxygen usually come from the surrounding air when gas shielding breaks down. TWI reports that about 1% air entrained in the shielding gas causes distributed porosity, and above 1.5% it produces gross surface-breaking pores.
• Hydrogen comes from moisture and hydrocarbons: damp electrodes and fluxes, condensation on the joint, and oil or grease on the workpiece or filler wire.

The sequence is the same in every case. A gas source reaches the pool, the molten metal absorbs it, the gas's solubility drops as the metal cools, and bubbles form. Bubbles that rise out of the pool before it freezes do no harm. Bubbles caught by the solidification front become pores.

That means porosity is a process problem first and an inspection finding second. Repairing pores without finding the gas source reproduces them in the next weld.

Key takeaway: find the gas source, because each source leaves a recognizable pattern.

Why is porosity worse in aluminum welds?

Aluminum is especially prone to hydrogen porosity because hydrogen's solubility falls sharply on freezing. TWI states that solubility "is much lower in solid aluminium (approximately 20x lower)" than in the liquid, so the excess gas is forced out as the weld solidifies (TWI: avoiding porosity when welding aluminium). The hydrogen comes from hydrocarbons and moisture on the plate and wire and from water vapor in the shielding gas.

TWI calls cleanliness the single most important control. The practical steps are to degrease before mechanical cleaning, remove the oxide shortly before welding, keep gas runs short and leak-free, and purge an idle gas system before use.

What are the types of porosity in welding?

Porosity in welding takes five common forms: uniformly distributed, clustered, linear (aligned), elongated (wormhole or piping) and surface-breaking. A sixth cavity, the crater pipe, looks like porosity but is caused by shrinkage, not gas.

ISO 6520-1 places all of these in its cavity group of weld imperfections, and AWS A3.0 defines porosity as cavity-type discontinuities formed by gas entrapment during solidification. US codes use slightly different names (D1.1 says "piping porosity" for elongated pores open to the surface), but the forms are the same.

Distribution matters as much as pore size. One small pore and a dense cluster of the same size pores are different findings, and every code below judges size, spacing and total quantity separately.

Is crater porosity the same as a crater pipe?

No. A crater pipe is a shrinkage cavity, so it needs a different fix from gas porosity. TWI explains that it "results from shrinkage on weld pool solidification", typically when the current is switched off abruptly and a large pool freezes quickly (TWI FAQ: crater pipes). The cures are crater filling, a gradual current slope-down, run-off tabs, or grinding out the stop before the next pass. AWS D1.1's visual table treats unfilled craters as their own item, separate from porosity.

How do you prevent porosity in welding?

Prevent porosity by controlling every gas source before, during and after welding, and by holding the parameters in the qualified welding procedure specification (WPS). No single control is enough, because porosity usually appears when two or three small controls slip at once.

Before the arc starts

1. Clean the joint. Remove oil, grease, cutting fluid, paint, primer, rust and mill scale from the joint faces and the adjacent surface. Degrease before grinding or brushing, or the abrasive spreads the contamination.
2. Remove coatings from the weld zone. Galvanizing and many primers must come off unless the WPS was qualified with them in place.
3. Store consumables dry. Keep low-hydrogen electrodes and fluxes in the manufacturer's recommended storage and reconditioning conditions, and follow the WPS and code requirements for exposure time.
4. Dry the work. Cold steel picks up condensation; preheat drives it off where the WPS calls for it.
5. Check the gas system. Inspect cylinders, regulators, hoses, fittings and the torch for leaks, and confirm the right gas is connected.

During welding

• Keep the shield stable. Set the flow within the WPS range. Too little lets air in; too much causes turbulence that pulls air into the shield.
• Block the wind. Use screens or enclosures outdoors. A flowmeter reading can look correct while a draft strips the gas away at the arc.
• Keep the nozzle clean of spatter, and hold contact-tip-to-work distance and torch angle within the qualified range.
• Hold the qualified parameters. Current, voltage, arc length and travel speed all decide whether gas has time to escape. Excessive travel speed and long arcs are common causes.
• Clean between passes. Remove slag and surface contamination so later passes don't trap it. Our guide to slag inclusion in welding covers the related defect.
• Finish the stop properly. Fill the crater or slope the current down to avoid crater pipes.

After welding

• Inspect visually right away for surface pores, pinholes and unfilled craters.
• Record the conditions for welds that show porosity: welder, consumable lot, gas, flow, weather and position. Patterns show up quickly.
• Treat repeated porosity as a process nonconformance, not a series of local repairs. Fix the cause, then resume.

Process monitoring systems that track arc signals or use weld cameras can flag disturbances in real time. They supplement a qualified procedure and inspection; they don't replace them. Procedure qualification itself is covered in our guide to welding procedure specifications.

Key takeaway: porosity prevention lives in the WPS and in shop discipline, not in the NDT report.

Which NDT methods detect weld porosity?

Visual testing finds surface pores, MT and PT find surface-breaking pores, and radiography or ultrasonic testing is needed for buried porosity. TWI notes that subsurface porosity is detected by radiography or ultrasonic inspection, and that radiography is normally the more effective of the two.

Which method the job uses is usually set by the code or the contract, not chosen by the inspector. A wider comparison is in our guide to non-destructive testing for welding inspection.

Why does radiography usually beat UT for porosity?

Radiography usually beats ultrasonic testing for porosity because a rounded pore shows clearly on a radiograph but scatters ultrasound instead of reflecting it back to the probe. The two methods respond to different properties of the flaw.

RT measures density. A pore is empty space, so less radiation is absorbed along that path and the pore shows as a dark, rounded spot. The image shows the size, shape and spacing of every pore in the projection, which is exactly what porosity acceptance charts judge. TWI's one caveat is that small pores are hard to detect, especially in thick sections.

UT measures reflected sound. A planar flaw such as lack of fusion acts like a mirror: when the beam hits it at the right angle, much of the energy comes straight back. A spherical pore has no flat face. It scatters the sound in all directions, so only a small fraction returns to the probe, and the echo is low even for a pore of meaningful size. Scattered porosity produces several weak, shifting signals that can sit near the noise from the weld's grain structure.

That leads to three practical effects:

• Low UT amplitude doesn't mean little porosity. Amplitude-based UT acceptance (AWS D1.1 indication rating, ASME Section VIII Appendix 12, B31.3 paragraph 344.6.2) can accept porosity that a radiographic rounded-indication chart would reject, simply because the echoes stay under the reference level.
• Porosity can mask other flaws on UT. A cluster can scatter the beam enough to reduce the signal from a flaw behind it.
• UT is still the better tool for planar flaws. Lack of fusion and cracks are the opposite case, as explained in our guide to lack of fusion in welds. That's why many specifications allow either method and some critical welds get both.

How a pore behaves on UT compared with the other common weld flaws:

More forms are compared in our guide to weld defect types.

Key takeaway: if porosity is the main concern on a weld, ask for RT; if planar flaws are, ask for UT.

What porosity is acceptable under US welding codes?

Acceptance depends entirely on the governing code: AWS D1.1 for structural steel, ASME Section VIII for pressure vessels, ASME B31.3 for process piping and API 1104 for pipelines. ISO 5817, often cited in porosity articles, is the European equivalent for quality levels; US contracts rarely call it up.

AWS D1.1 porosity limits: static vs cyclic

AWS D1.1 visual acceptance (Table 8.1) sets different porosity limits for statically and cyclically loaded connections, and the often-quoted 3/32 in. limit applies only to the cyclic case. The nontubular limits as summarized in CWI study material and AWS forum quotations of the code are:

Under D1.1, NDT beyond visual inspection applies only when the contract documents call for it. When RT is specified, Clause 8 has its own radiographic criteria for static, cyclic and tubular connections, which limit discontinuity size against weld size and spacing; check those against your edition. Full D1.1 acceptance, including undercut and UT indication rating, is in our guide to AWS D1.1 weld inspection criteria.

ASME Section VIII rounded indications

Section VIII, Division 1 judges porosity on radiographs as rounded indications under Mandatory Appendix 4, with size limits based on the thickness t. Forum quotations of the appendix give these limits:

A rounded indication is generally one whose length is no more than three times its width. Two qualifiers matter. Appendix 4 applies to welds that are fully radiographed under UW-51. Under spot radiography (UW-52), secondary sources agree that rounded indications are not a factor in acceptance. Section V, the examination code, sets how the radiograph is made but contains no acceptance criteria.

ASME B31.3 porosity

ASME B31.3 Table 341.3.2 refers radiographic porosity to the Section VIII Appendix 4 size and distribution limits, with different criteria for normal and severe cyclic fluid service. Practitioner readings disagree on whether surface porosity is evaluated the same way or is not permitted at all, so read Table 341.3.2 in your edition before accepting surface pores on process piping. For UT, B31.3 paragraph 344.6.2 rejects a linear-type indication only when its amplitude exceeds the reference level and its length exceeds the thickness-based limit, which is why low-amplitude porosity rarely triggers a UT reject. Code-by-code UT limits are in our guide to ultrasonic testing acceptance criteria.

API 1104 porosity

API 1104 rejects an individual pore larger than 1/8 in. (3 mm) or 25% of the specified wall thickness, whichever is less, and limits overall scattered porosity with distribution charts. Secondary summaries of the standard also give these limits:

For regulated gas pipelines, 49 CFR 192.7 incorporates the 21st edition of API 1104, and 49 CFR 192.241 requires welds to meet its acceptability standards. The 22nd edition (2021) is current, but it isn't the regulatory edition for PHMSA gas lines.

Key takeaway: the same pore can pass one code and fail another, so quote the code, the edition and the loading or service class with every porosity call.

Worked examples: is this porosity acceptable?

The examples below use the limits as summarized above. They show the method; confirm the values against your edition.

Example 1: AWS D1.1 fillet weld, visual. An inspector finds four visible pores within one inch of a fillet weld, measuring 1/8, 1/8, 3/32 and 1/16 in.

• Sum: 4/32 + 4/32 + 3/32 + 2/32 = 13/32 in.
• Static limit: 3/8 in. = 12/32 in. in any linear inch. Reject, by 1/32 in.
• Cyclic limit: one pore per 4 in., max 3/32 in. diameter. Reject: there are four pores in one inch, and two exceed 3/32 in.

Remove only the 1/8 in. pores and the remainder (5/32 in.) passes the static rule, but still fails the cyclic frequency rule. The loading case changes the answer.

Example 2: ASME Section VIII, Appendix 4, t = 1/2 in. A radiograph of a fully radiographed seam shows a 9/64 in. (0.141 in.) rounded indication.

• General limit: 1/4 × 0.5 = 0.125 in., versus 5/32 = 0.156 in.; the smaller is 0.125 in. The pore exceeds it.
• Isolated limit (if nothing else lies within 1 in.): 1/3 × 0.5 = 0.167 in., versus 0.25 in.; the smaller is 0.167 in. The pore is within it.
• Result: the pore is acceptable only if it is isolated, and the total porosity in the length must still fall within the distribution chart for that thickness. If the seam were spot radiographed under UW-52, rounded indications wouldn't be graded at all.

Example 3: API 1104 girth weld, two wall thicknesses.

On thin-wall pipe, the wall-thickness rule is stricter than the 1/8 in. cap, so a pore that would pass on heavy-wall pipe can fail on thinner pipe.

How should rejectable porosity be repaired?

Remove the porous metal by grinding or gouging, confirm it's gone, reweld to a qualified procedure, and re-inspect with the same method that found it. The repair sequence is:

1. Evaluate against the governing code; only porosity that exceeds the limits is a defect.
2. Find the cause before repairing: shielding, moisture, contamination, coating or technique.
3. Excavate the affected length. Gouging a porous area often opens more pores than the radiograph showed, so dig until the cavity is clean.
4. Verify the excavation visually, and with MT or PT where the code or procedure requires it.
5. Reweld to the WPS with the cause corrected.
6. Re-inspect with the original method, and record the repair.

Widespread porosity suggests a systemic problem. It may mean cutting out the whole weld and repreparing the joint, and it should trigger a check of other welds made under the same conditions.

What inspection cannot tell you about porosity

• RT doesn't show depth on a single exposure, and very small pores can be missed in thick sections. It also needs a radiation exclusion zone and usually access to both sides.
• UT can underreport porosity. Amplitude depends on pore size, shape and grouping, so a low signal can represent significant porosity. UT also can't reliably tell a pore cluster from slag without supporting evidence such as scanning from several directions.
• Surface methods see only the surface. VT, MT and PT won't find buried porosity in a groove weld.
• Passing a code doesn't prove the process is under control. Acceptable porosity on every weld can still be a sign of drifting shielding or damp consumables that will produce a reject next week.
• No inspection result tells you the cause. The pattern (distributed, clustered, linear, wormhole) narrows it down, but the cause is found in the shop records and the WPS.
• This article doesn't replace the code. Clause, table and figure numbers change between editions; the acceptance decision belongs to the CWI or Level III working from the licensed text.

How EMA Industries applies this

EMA Industries performs AWS D1.1 structural weld UT, MT and PT, and provides NDT Level III and WPS/PQR services from Southern California, mobilizing nationwide. On a porosity problem, the useful work is usually twofold: confirming whether indications meet the governing code, and reviewing the WPS, consumable handling and shielding setup so the porosity stops. Where porosity is the main concern and the specification calls for radiography, the inspection plan should say so. See our weld ultrasonic examination services or contact EMA Industries to scope an inspection or a procedure review.

Related reading: AWS D1.1 weld inspection criteria · Ultrasonic testing acceptance criteria standards · Lack of fusion in welds · Slag inclusion in welding · Weld defect types · ASME weld ultrasonic inspection

Frequently asked questions

What are the types of porosity in welding?

There are five common forms: uniformly distributed (scattered), clustered, linear or aligned, elongated (wormhole or piping) and surface-breaking. Crater pipe looks similar but is a shrinkage cavity at the weld stop, not gas porosity, so it is prevented by crater filling rather than better shielding.

What is the most common cause of weld porosity?

Loss of gas shielding and hydrogen from moisture or contamination cause most porosity. Drafts, wrong flow rates and leaking gas lines let air into the pool, while damp consumables, condensation, oil and coatings supply hydrogen. The porosity pattern usually points to which one it is.

Does porosity always mean the weld must be repaired?

Only when it exceeds the governing code's limits. AWS D1.1, ASME Section VIII, B31.3 and API 1104 each allow some porosity, judged by pore size, spacing and total amount, with limits that depend on loading, service class and wall thickness. Cracks are rejected at any size, but porosity is not.

Is ultrasonic testing good for detecting porosity?

UT detects larger pores and clusters and gives their depth, but it is weaker than radiography for porosity. Rounded pores scatter sound rather than reflecting it back to the probe, so small scattered pores return low signals that can fall under the reference level. RT is generally preferred when porosity is the main concern.

What causes porosity in aluminum welds?

Hydrogen is the cause. Molten aluminum dissolves hydrogen from moisture, oil and hydrated oxide, and solid aluminum holds roughly 20 times less, so the excess forms pores as the weld freezes. Thorough degreasing, fresh oxide removal, dry filler wire and a leak-free, purged gas system are the main controls.

Sources

• Defects/imperfections in welds: porosity (Job Knowledge 42), TWI (definition, gas sources, 1% and 1.5% air entrainment, RT vs UT)
• Avoiding porosity when welding aluminium, TWI (hydrogen solubility about 20x lower in solid aluminum; cleanliness)
• What causes crater pipes and what will prevent their formation?, TWI FAQ
• AWS announces release of D1.1/D1.1M:2025, American Welding Society, April 17, 2025
• 49 CFR 192.7, Incorporation by reference, eCFR (text read on Cornell LII): API 1104 21st edition
• 49 CFR 192.241, Inspection and test of welds, eCFR (text read on Cornell LII)
• ASME BPV Committee on Nondestructive Examination (BPV V), committee scope (Section V excludes acceptance criteria)
• AWS D1.1/D1.1M:2025 Clause 8 and Table 8.1; ASME BPVC Section VIII, Division 1, UW-51, UW-52 and Mandatory Appendix 4; ASME B31.3 Table 341.3.2 and para. 344.6.2; API Std 1104, 21st and 22nd editions, Section 9; ISO 6520-1:2007; AWS A3.0 (all paywalled; values above from agreeing secondary sources, to be confirmed against the licensed text)