Porosity is gas that was dissolved in the molten pool and did not escape before the metal froze. It appears as rounded voids, either on the surface or just beneath it, and their shape tells you they were bubbles rather than tears or inclusions. The metal was liquid, gas was trapped in it, and the pool solidified faster than the bubbles could rise.
Two things have to be true for a pore to form: gas has to get into the pool, and the pool has to freeze before the gas leaves. That second condition is why porosity is often blamed on gas supply when it is really a heat or travel speed problem. A pool that stays fluid longer gives bubbles time to surface.
Separating those two causes is the whole job of diagnosing porosity. The location, size and distribution of the pores usually say which one you are dealing with before you touch the torch.
Gas reaches the pool by a small number of routes, and almost every porosity investigation ends at one of them. It comes from the atmosphere when shielding fails. It comes from the surface of the base metal as oil, paint, moisture or oxide. It comes from the filler as moisture or contamination. It comes from the torch itself, through a leaking hose or a porous body.
There is also a metallurgical route that is easy to overlook: hydrogen picked up earlier in the process, in the base metal or from a previous pass, which comes out of solution as the weld solidifies. That version is harder to fix because there is nothing obviously wrong at the bench.
If porosity is scattered evenly along a bead, in small pores with a regular spacing, the cause is usually at the torch. Either the flow is too low, the cup is too far from the work, or the stream is turbulent and pulling air in. On stainless this often comes with discolouration; on titanium the colour is a stronger signal than the pores.
The tell is that the problem follows the position rather than the part. If the same joint welded from a different angle comes out clean, the issue is coverage on that angle, not the material.
Interior purge is the other half of shielding and the one most often skipped. A closed tube welded without purge produces a root that oxidises from the inside, and the porosity appears on the backside where nobody looked. Our stainless steel exhaust system work treats purge as a step with a check, not as a courtesy.
Oil, cutting fluid, paint, marker, adhesive residue and mill scale all release gas when the arc reaches them. The quantity needed to ruin a weld is small enough that a wiped surface can still be contaminated. Solvent that has not fully evaporated is itself a source of hydrogen.
Oxide is the more subtle case. On aluminium the oxide layer has to be removed before welding and reforms quickly afterwards. On stainless, heavy oxide from a previous thermal operation can trap moisture and release it into the pool as the arc passes over it.
The reliable practice is mechanical cleaning followed by a solvent wipe, done close enough to the weld that contamination cannot return. Cleaning a joint the day before and leaving it open on a bench is closer to doing nothing than most shops realise.
Filler rod picks up moisture from the air, and the rate depends on how it is stored. Rods left in an open tube on a bench, or in a rack beside a wash tank, will absorb enough moisture to cause porosity in a joint that is otherwise perfect.
Contamination on the rod surface is a separate problem. Grinding dust, spatter and handling oils all sit on the outside of the rod and go straight into the pool. A rod that has been dragged across a dirty bench should be cleaned or cut back before use.
A torch that has been dropped, dragged or left with a kinked hose can draw air into the gas stream. The regulator still reads correctly because the leak is downstream of the gauge, so the flow meter tells you nothing about the mixture arriving at the cup.
Testing is simple and rarely done. Holding the cup against a clean gloved hand and listening, or checking for a change in the arc when the hose is flexed, will find most leaks in a few seconds. A torch that behaves differently at different angles is a torch with a problem.
Porosity often concentrates at the beginning and end of a bead rather than being spread evenly. At the start, gas coverage has not stabilised and the pool is small and cooling fast. At the end, the arc stops, the gas stops, and the crater sits exposed while it freezes.
The crater is the worst case because it is the last metal to solidify and the first to lose protection. A welder who lifts off quickly leaves a small, unprotected pool that can absorb air for as long as it stays liquid.
Managing it is a matter of technique: establish gas flow before striking, taper the current at the end rather than breaking the arc, and keep the torch over the crater while it freezes. Those habits remove a whole class of defects that no amount of regulator adjustment will fix.
The distribution of pores is diagnostic. Fine, evenly scattered pores along the whole bead point to gas supply or moisture. Large, isolated pores point to a specific event, such as a dip of the tungsten or a bit of contamination at one point. Pores in a line along one edge point to the joint surface or to a gap that was not clean.
Pores on the root only point at purge. Pores on the face only point at the torch. Pores at the stop point at termination technique. Each pattern narrows the search to one or two variables, which is far faster than changing everything and hoping.
Photographing the defective area before it is ground out is worth the effort. A shop that keeps images of its porosity finds that the same pattern keeps coming back for the same reason, and the fix becomes procedural instead of reactive. You can see how clean roots should look on our exhaust pipe and stainless steel exhaust mufflers pages.
Porosity is a leak, a weak point and, on a polished part, a visible flaw. Buyers cannot inspect for subsurface pores, so the questions have to go to the process that prevents them.
Tigindustry has been welding stainless and titanium in Shanghai since 2009, and shielding discipline is one reason parts arrive leak-free rather than leak-tested. You can see the range on our titanium exhaust system pages and TIG stainless steel welding, or browse by alloy under shop by material.
