The film is thin, and welding removes it

Stainless steel resists corrosion because of a chromium-rich oxide film a few nanometres thick that reforms itself in air. That is the whole mechanism. A weld bead can look perfectly acceptable — uniform ripples, no undercut, no spatter — and still be the place where the fabrication starts rusting, because the heat has replaced that self-healing film with a thick oxide scale and stripped chromium out of the metal immediately beneath it. The weld is not finished when the arc stops.

The visible evidence is heat tint. The straw, blue and purple bands beside a bead are an interference-colour scale: thicker oxide reads darker, and a dark band means a thicker oxide, which means more chromium pulled out of the surface layer and less left to reform the film. Grinding the colour off is not the same as fixing it. What matters is what is left underneath.

Why heat lingers in a stainless joint

Austenitic stainless holds heat where carbon steel sheds it. The 904L data sheet puts the figure plainly: austenitic steels have only about 30 per cent of the thermal conductivity of non-alloyed steel, a lower fusion point, and a substantially higher coefficient of thermal expansion. The practical consequences listed there are worth taking into a procedure: weld with lower heat input, use higher travel speed on thin section, expect more distortion, prefer a double-V over a single-V above 12 mm with a 60–70 degree included angle, avoid accumulating weld seams, and space tack welds closer than you would on carbon steel.

Expansion makes it worse. Type 304 runs at 17.2 × 10^-6 /K against roughly 11.0 to 12.4 for the ferritic 409 grade, so an austenitic joint both distorts more and spreads a wider heat-affected zone for the same arc energy. And the high-temperature grades sit in the same trap: Alloy 310S runs continuously to 2000 °F and resists scaling precisely because it forms oxide readily — which is another way of saying it tints readily, and its own sheet refers corrosion questions to NACE.

Free iron: the second problem, and the quieter one

Heat tint is visible. Embedded iron is not. A flap disc that finished a carbon steel job, a wire brush shared between materials, slings and stillages used for both, or grit from a blast cabinet that has not been changed — all of these leave free iron pressed into the stainless surface, and every particle becomes a rust spot and a pit initiation site. The 316 and 316L sheet carries an unusually direct warning about this: make sure copper and zinc do not become contaminants at the site of welds, because they can cause cracking.

It goes the other way too. Type 416HT is a free-machining martensitic grade, and its own data sheet states that high-sulfur free-machining grades are unsuitable for marine or other chloride exposure. Putting a 416HT component into a 316L assembly that is going to be pickled asks for a stained, pitted result no amount of cleaning will fix. 302, a higher-carbon 304 under AMS 5516, will sensitise faster than 304L if the tinted zone is left alone.

Cleaning, pickling and passivation are three different jobs

Abrasive cleaning removes oxide mechanically. It is fast, it is what most shops reach for, and it has two drawbacks: it smears rather than dissolves, and a disc that is not dedicated to stainless drives free iron into the surface it just cleaned. Pickling is chemical — typically a nitric-hydrofluoric bath or paste — and it dissolves both the oxide and the chromium-depleted metal beneath it, letting the surface re-establish a uniform film. Passivation is the mildest of the three: an oxidising treatment that removes free iron and thickens the existing oxide. It will not remove heat tint and it will not restore a depleted layer, which is the single most common misunderstanding on a fabrication drawing.

The finish designations make the distinction visible. On the 304/304L coil range, No.1 is the surface finished by heat treatment and pickling after hot rolling; 2B is cold rolled, heat treated, pickled or equivalently treated, then given a final light cold roll; BA is bright heat treated after cold rolling. Our S32205 duplex pipe range offers annealed and pickled, bright annealed, polished and mirror finishes, and the 2205 seamless pipe listing includes 2B, pickling, polished, brushed, sandblast, BA and electropolished. Pickling is a named finish, not an afterthought.

What changes with the grade

Low carbon and stabilisation are the two levers, and they determine how much post-weld work is strictly necessary. 304L at 0.03 carbon against 304 at 0.08 exists for heavy-gauge weldability; 316L is preferred in corrosive environments and its Mo content sits at 2.0–3.0 per cent. 317LMN adds nitrogen and keeps carbon at 0.035, which gives resistance to sensitisation during welding on top of higher strength. 321 is titanium stabilised at 5×(C+N), capped at 0.70 or 0.75; 347 is stabilised with columbium at 10×C to 1.10, and its data sheet recommends it specifically for welding under conditions that prevent a post-weld anneal — that is the clearest statement of intent you will find on any grade page. 347H then carries 515 MPa tensile, 205 MPa yield and 35 per cent elongation.

The high-temperature and high-alloy families each have their own rule. 310S is the low-carbon variant of 310, chosen to minimise carbide precipitation during welding. 904L guarantees intergranular corrosion resistance in the welded condition on the strength of its 0.02 carbon. Duplex 2205 yields roughly twice what austenitic grades do and is suited to the −50 to +600 °F band with restrictions on welded structures outside it; 2507 should be kept below 316 °C; S32760 / F55 holds PREN 40 and ships with 3.1 certification. 254 SMO at 6.0–6.5 molybdenum and 0.18–0.20 nitrogen is the usual answer when chlorides rule out everything below it. On the ferritic side, 430 is titanium stabilised and described as having good mechanical properties at the weld, and 409 is stabilised with niobium, titanium or both specifically for weldability.

Specifying it, and checking it was done

Put four lines on the drawing or the enquiry. State the finish required after fabrication, using the real designations — No.1, 2B, BA, AP, pickled — rather than “clean”. State whether heat tint must be removed and how far back from the toe of the weld. Prohibit contact with carbon steel tooling, slings and blast media. And ask for the cleaning sequence to be recorded, because the difference between a ground bead and a pickled bead is invisible in a photograph and obvious in six months.

Verification is cheap. A free-iron test swab darkens where iron is present; a water-break test shows where the surface is still carrying oil. Then protect what you have achieved, because pickling and passivation leave a clean surface that will pick up iron again from the first carbon steel forklift that touches it. Our surface treatment and plate welding services cover this sequence, bespoke fabrication runs through custom processing, and the full material list sits under stainless steel.