TIG welding stainless steel: heat, color, and corrosion
Heat decides a stainless weld: sensitization, heat tint, filler choice, purging and duplex — what to control, and why the numbers beat technique.
Stainless steel is not difficult to weld. It is difficult to weld and keep stainless — and almost every fault that matters comes down to the same variable: how much heat went in, and how long the metal spent getting rid of it.
What makes stainless stainless
The corrosion resistance does not come from the bulk of the metal. It comes from a passive chromium oxide film, a few atoms thick, that forms on the surface and repairs itself when scratched — as long as there is enough free chromium in the metal underneath to rebuild it.
Everything on this page is about protecting that film, or about restoring it afterwards.
Sensitization: the fault you cannot see
Hold austenitic stainless between roughly 425–815 °C and chromium starts combining with carbon into chromium carbides, which precipitate at the grain boundaries. The carbide itself is not the problem. The problem is what it takes with it: the metal immediately beside the boundary is left depleted of chromium, and depleted metal cannot rebuild the passive film.
The part still looks like stainless. In service it corrodes along the grain boundaries — intergranular corrosion, or weld decay — and it does so out of sight, from the inside.
Two defenses, and good practice uses both:
- Low-carbon grades. 304L and 316L cap carbon at 0.03%, leaving too little to form meaningful carbide. Stabilised grades 321 and 347 do it differently, by adding titanium or niobium that grabs the carbon first.
- Low heat input. Sensitization is a function of temperature and time. The less time the heat-affected zone spends in that band, the less carbide forms — which is why travel speed and heat input matter more here than technique.
Heat tint: the fault you can see
The colors that appear alongside a stainless weld are the oxide film thickening. They read as a temperature record:
| Color | Roughly | Meaning |
|---|---|---|
| Silver, pale straw | lowest | Film essentially intact |
| Straw to gold | moderate | Thickened film, chromium still adequate |
| Bronze to purple | higher | Film altered, chromium depleted beneath |
| Blue to gray, black scale | highest | Heavy oxide, significant depletion |
Tinted metal corrodes faster than the metal next to it. On decorative work that is cosmetic. On a pharmaceutical line, a heat exchanger or an offshore manifold it is a defect, and the specification will say so.
Two ways to deal with it. Prevent it with a proper purge and low heat input, or remove it afterwards — pickling paste, electropolishing, or mechanical cleaning with a brush that has never touched carbon steel. Removal is slower, uses aggressive chemicals, and on a long weld it costs more than getting the heat right in the first place.
Purging the root
A full-penetration weld exposes the root to whatever is behind it. Unpurged, it oxidizes into a rough, dark, granular scale — sugaring — which no amount of face-side technique fixes, cannot be reached for cleaning inside a closed pipe, and becomes a corrosion site for the life of the part.
Purge with argon, and displace the air properly before striking. The usual target for stainless is below about 50 ppm residual oxygen in the purge volume; an oxygen meter beats guessing. Keep the purged volume small with dams rather than filling an entire pipe run. The shielding gas guide covers the detail.
Distortion: two numbers that catch people out
Austenitic stainless behaves differently from carbon steel in two ways that both push the same direction:
- Thermal expansion is about 50% higher. The same heat moves the part further.
- Thermal conductivity is about a third. The heat does not spread away; it concentrates near the joint.
More expansion plus less dissipation means noticeably more distortion for the same weld. Tack more often than carbon steel would need, keep the interpass temperature down — typically below 150 °C unless the procedure says otherwise — and use backing bars or chill bars on thin sheet.
Filler selection
| Base metal | Filler | Note |
|---|---|---|
| 304, 304L | 308L | The standard austenitic pairing |
| 316, 316L | 316L | Keeps the molybdenum for pitting resistance |
| Stainless to carbon steel | 309L | Tolerates dilution without going brittle |
| Duplex 2205 | 2209 | Over-alloyed in nickel to balance the weld ferrite |
| 321, 347 | 347 | Keeps the stabilising element |
The L matters. Standard 308 and 316 carry enough carbon to sensitise; the low-carbon versions are what most modern procedures specify, and the strength difference is negligible for the joint.
Duplex is a different problem
Duplex stainless — 2205 and the super-duplex grades — is roughly half austenite and half ferrite, and that balance is what gives it both strength and chloride resistance. Welding disturbs it in both directions:
- Too little heat, and the weld cools too fast for austenite to re-form. The joint ends up ferrite-heavy, hard, and low in toughness.
- Too much heat, and intermetallic phases — sigma among them — precipitate and destroy the corrosion resistance the alloy was bought for.
So duplex is not a “keep it as cold as possible” material. It has a heat input window, commonly quoted around 0.5–2.5 kJ/mm, and the qualified procedure will name it. Backing gas usually carries a small nitrogen addition to keep nitrogen in the root, and interpass temperature is limited tightly.
If there is one material where the WPS is not a formality, this is it.
Housekeeping
- Dedicated brushes and grinding media. Anything that has touched carbon steel embeds iron particles, which rust and seed pitting. This is the most common cause of “stainless that rusted” in a fabrication shop.
- Separate storage. Do not lay stainless on the same bench as carbon steel cutting.
- Clean before, not after. Degrease first; brushing an oily surface drives the oil into the metal.
Why heat input is the whole argument
Read back over the faults above and the pattern is hard to miss. Sensitization is time at temperature. Heat tint is peak temperature. Distortion is total energy into the part. Duplex is a heat input window. Not one of them is a hand-skill problem.
That is why the process matters on stainless more than on most materials, and it is where hot wire TIG makes its case. Feeding the filler mechanically and pre-heated means the arc does not have to melt cold wire, so the same joint is filled at a substantially higher travel speed — and less time over any one spot is less heat into the part. TIP TIG delivers the lowest heat input of any wire-fed process, which on stainless translates directly into less time in the sensitization band, less tint to remove and less distortion to straighten.
The practical evidence is on inspected work: on ID Inconel-clad subsea pipe for CNOOC, where conventional TIG was too slow and pulsed MIG could not hold the quality, the requirement was 100% X-ray with zero rework — the case study has the detail. The metals and alloys guide covers where else the same argument applies.
In short
Stainless welds easily and loses its corrosion resistance easily, and both come down to heat. Use low-carbon filler that matches the base metal, purge any full-penetration root, keep interpass temperature down, and treat heat tint as a defect rather than a finish. Duplex needs its heat input inside a window rather than simply low. And keep the carbon-steel brushes on the other side of the shop.