5 Ways Gases Form Porosity in Welds
Weld porosity forms when gas bubbles remain in molten metal as it solidifies. Air entering the shielding zone, hydrogen-bearing contamination, coating vapor, chemical reactions, and unstable pool or keyhole behavior can all contribute. To find the cause, trace both the gas source and what prevented escape, then compare equivalent welds using the required inspection method.
How gas becomes a pore
Porosity is a gas cavity in solidified weld metal. It can be internal, open to the surface, scattered, clustered or elongated. A gas source alone does not guarantee a pore: a bubble that reaches a free surface can escape before the metal freezes around it.
Gas dissolves in the liquid, develops through a reaction, or becomes isolated inside the pool.
Liquid flow and buoyancy affect its path toward a free surface.
The advancing solidification front closes around gas that has not escaped.
The five routes below overlap. For example, a shielding leak can introduce moisture, that moisture can supply hydrogen, and pool motion can determine whether the resulting bubble stays inside the weld. Treat them as investigation paths rather than five mutually exclusive diagnoses.
First confirm the discontinuity. A dark spot on a radiograph or a hole at the weld end can also involve an inclusion, lack of fusion or shrinkage. In particular, a crater pipe can form by solidification shrinkage and may coexist with gas pores.
1. Air enters an incomplete gas shield
A leak, draft or disrupted gas stream can let the atmosphere reach hot weld metal. Nitrogen, oxygen and moisture then become part of the investigation. A flow reading at the cylinder does not prove that the molten pool is covered.
TWI identifies gas-line leaks, drafts and excessive flow as common porosity contributors. Too little gas leaves inadequate coverage; too much can create turbulence that draws air into the stream. No single flow rate suits every gas, nozzle, torch position and process.
Trace the delivery path from the identified gas supply through the regulator, hoses, connections, valve and torch. Inspect nozzle or cup condition, blockage, torch distance and angle, and airflow at the actual joint. Verify delivery using the equipment manufacturer’s method and the welding procedure specification (WPS).
For TIG welding, torch assembly and gas timing also matter. Miller explains how long gas lines can produce a start-up surge that disturbs coverage. A good steady-state reading may miss a problem limited to the first part of a weld.
A useful comparison: if porosity began after moving the workstation or changing a nozzle, restore that verified setup and repeat the same coupon before changing heat or travel speed.
1 Travel direction · 2 Contact tube · 3 Electrode wire · 4 Shielding gas · 5 Molten weld metal · 6 Solidified weld metal · 7 Workpiece
Match the check to the process. TIG and MIG/MAG depend on external shielding gas. Stick welding and self-shielded flux-cored welding rely on their consumable shielding systems; a cylinder-flow fix does not apply to them. Gas-shielded flux-cored welding needs both the correct wire conditions and external gas supply. Laser gas choice is also material-dependent; see protective gas selection for laser welding.
2. Moisture and hydrocarbons supply hydrogen
Water, oil, grease and other hydrogen-bearing contamination can feed hydrogen into a weld pool. Possible sources include the workpiece, filler, electrodes or flux, handling, and moisture in the gas system. The key question is where the contamination entered.
Aluminum is particularly sensitive because hydrogen is much more soluble in liquid aluminum than in the solid metal. As the pool solidifies, hydrogen can leave solution and form bubbles. TWI’s aluminum guidance identifies hydrated surface oxide, filler contamination, atmospheric moisture and gas lines as important sources. Cast or sintered material can also contain hydrogen from its manufacture, so repeated surface cleaning may not resolve every case.
Review the sequence of preparation and storage. Degreasing, oxide removal where needed, clean dedicated tools, dry consumables and protection from recontamination work together. Clean before assembling a joint that would trap cleaner, and allow the approved cleaning process to finish before welding.
A useful comparison: prepare matched coupons with documented clean material and correctly stored filler. Keep the joint, shielding and machine program the same. If the pores persist, investigate the gas system and material history rather than repeatedly polishing the weld face.
In susceptible steels, hydrogen also raises cracking concerns. Hydrogen-assisted cracking depends on the material, thermal cycle and restraint. Absence of visible pores does not establish adequate hydrogen control. Follow the exact consumable storage and WPS requirements; there is no universal electrode redrying or preheat temperature.
3. Residues and coatings release gas or vapor
Paint, primer, oil, sealant and material trapped between joint faces can release gas when heated. Volatile metallic coatings introduce another mechanism: zinc can vaporize before the underlying steel melts.
In a tightly clamped galvanized lap joint, zinc at the hidden interface needs an escape path. Vapor forced through the weld pool can disturb the liquid, produce spatter or leave cavities. TWI’s work on zinc-coated automotive steel explains why coating position, joint arrangement and vapor escape matter. A clean outer face does not establish that the overlap is clean or adequately vented.
Identify the coating and its thickness, examine where it intersects the weld, and record the fit-up. A solution may involve permitted coating removal, controlled joint spacing, a vent feature or a process developed for that coating. Each changes more than gas release: gap and preparation also affect fusion and corrosion protection. Use the galvanized steel laser welding guide for the wider joint-design considerations.
A useful comparison: compare the original joint with one approved preparation or venting change, preserving the other conditions. Check penetration and internal pores as well as surface appearance. A fixed gap or grinding width from another job is not a validated remedy for this one.
Choose the cleaning and fume controls together. Identify the coating and cleaner from their safety data sheets. Keep chlorinated solvent residues and vapors away from welding heat and arc radiation, which can generate toxic decomposition products. Coating removal and welding also require suitable fume controls; CCOHS explains the relevant fume and gas hazards.
4. Weld-pool reactions generate gas
Gas can form through a chemical reaction inside the pool. In steel welding, carbon can react with oxygen to produce carbon monoxide (CO). If the gas cannot escape before solidification, it can leave pores.
[C] + [O] → CO(g)
A simplified reaction: the brackets represent carbon and oxygen dissolved in the metal, and “g” denotes gas. It shows a mechanism, not a formula for predicting pore volume.
Deoxidizers such as manganese and silicon can combine with oxygen, reducing the oxygen available for this reaction. ESAB’s steel MIG-wire guidance explains this role. The relevant combination includes the base metal, filler composition, surface condition and shielding gas.
This does not mean that CO₂-containing shielding gas is always wrong. Such gas can be part of an appropriate steel MAG procedure with a compatible filler. Eagar’s discussion of joining thermochemistry shows why shielding and deoxidizer chemistry must be considered together.
A useful comparison: if the problem follows a material heat, wire lot or gas change, retain the certificates and samples. Verify the approved combination and compare with the last acceptable lot before making broad parameter changes. Do not substitute a more heavily deoxidized wire without checking the required weld properties.
5. Pool motion or keyhole collapse traps bubbles
A bubble can remain even when its original source has been reduced. Liquid flow, unstable droplet transfer and the rate of solidification affect whether gas reaches a free surface. Changing heat or speed can alter both gas uptake and escape, so the direction of a useful adjustment is not universal.
For wire-fed arc welding, check whether feeding and metal transfer are steady. Record disturbances, torch distance, joint fit and the location of pores. Restoring a known stable setup makes a more useful first comparison than changing several settings at once.
Deep-penetration laser welding adds a keyhole: a vapor-filled cavity maintained under the beam. A local narrowing or collapse can separate a bubble from that cavity. The bubble may escape or rejoin the keyhole; it becomes a pore only if solidifying metal traps it.
Laser diagnosis therefore includes actual penetration, power, travel speed, focus, beam shape or wobble, fit-up and shielding. A clean surface cannot by itself establish keyhole stability. Record these conditions alongside the pore depth and position.
What an internal-weld study actually shows
Pang and colleagues studied fiber-laser welding of 3 mm 304 stainless steel. Their Figure 2 compares radiographs and longitudinal sections at different ultrasonic assistance powers. The welding conditions were 2,000 W laser power, 1.2 m/min travel speed, +3 mm defocus and 20 L/min nitrogen shielding.
View larger figure
Scroll across and down to inspect the pores and scale bars.
In this comparison, 750 W of ultrasonic assistance produced fewer pores than 1,000 W. The lesson is that more input did not produce a steadily better result, and internal examination revealed the difference. These settings belong to that experimental arrangement; they are not a recipe for a different laser, handheld head or joint.
The study’s nitrogen choice is also material- and process-specific. See nitrogen versus argon for laser welding before treating a gas change as a general porosity fix.
Use the pore pattern to choose the next check
Shape and location help narrow an investigation, but they do not identify the gas by themselves. Combine the indication with what changed in production and evidence from the affected part.
| Observed evidence | What it can suggest | Check before concluding |
|---|---|---|
| Scattered or surface-opening pores after a setup change | A gas coverage or contamination change. | Gas delivery, nozzle condition, drafts, preparation and the first affected weld. |
| Fine internal pores in aluminum | A dissolved-hydrogen route is plausible. | Hydrated oxide, filler, condensation, gas-line moisture and material history; shape alone cannot locate the source. |
| Elongated pores or a cluster at an overlap | Continuing gas evolution or a local trapping condition. | Hidden coatings, residues and joint geometry; confirm the cavity relative to the interface. |
| Laser pores near the root or centerline | Keyhole or pool behavior may be involved. | Penetration, fit-up and process stability alongside gas and coating checks. |
| A cavity at the weld termination | Gas entrapment, shrinkage, or more than one imperfection. | Section or examine the termination as needed; review filling and shutdown behavior before naming it porosity. |
Swipe sideways to read all columns on small screens. These are investigation prompts, not acceptance limits or a ranked diagnosis.
Build a controlled comparison
- Keep a baseline. Retain the failed part or coupon and record material, thickness, joint, coating, filler, gas and the actual machine settings. Mark where the pores occur and when the issue began.
- Choose the best-supported change. Use the evidence above to select one source or process condition. Restore a known acceptable setup where possible, and hold the remaining conditions constant.
- Compare equivalent evidence. Inspect a comparable weld length, location and section using the same method. Check fusion, penetration and other required properties along with porosity. Repeat sufficiently to establish consistency under the production quality plan.
For example, if pores began with one filler lot, a matched comparison using the previous acceptable lot tests that observation. Changing the filler, gas and travel speed together would leave the source of any improvement unresolved. This is a proposed troubleshooting comparison, not a reported production result.
Confirm the weld and the repair against the right requirements
A smooth face does not rule out internal pores. Visual inspection finds surface evidence; radiography often helps reveal volumetric porosity. Small pores, thickness and geometry affect detectability. Ultrasonic testing, computed tomography or sectioning may suit particular examination or development tasks, but they are not interchangeable methods.
Separate the name of an imperfection from its acceptance. ISO 6520-1 provides a classification framework. ISO 5817:2023 covers quality levels for specified fusion-welded materials and excludes beam welding. ISO 13919-1:2019 addresses electron- and laser-beam welded steel, nickel, titanium and their alloys. The applicable product specification, selected quality level and examination requirements govern the actual job.
Welding over a pore is not proof of repair. Where porosity is rejectable, an agreed repair normally requires removal of the affected weld metal, correction of the cause, re-preparation, rewelding and re-examination. TWI’s repair guidance distinguishes local removal from widespread porosity that may require removing the whole weld.
If the indication could be a crack or lack of fusion, or the acceptance requirement is unknown, resolve that question before releasing the part. A trial coupon with fewer visible pinholes does not establish that the production weld meets its requirements.
Discuss a laser-welding porosity problem with Oceanplayer Laser. Share the base grade and thickness, joint and coating, filler, gas and nozzle, power–speed–focus settings, when the problem started, and photographs plus available internal inspection results. Include the required weld criteria so a sample comparison addresses the actual part.
Discuss your welding application