Duplex stainless is not a single metal so much as a controlled mixture of two. Its microstructure carries both austenite and ferrite, and the corrosion resistance and the strength both depend on that mixture staying roughly where the mill put it. Lose the balance in the weld and you keep the chemistry on the certificate while losing the properties you paid for.
Shops like it for two reasons. Duplex resists chloride attack far better than the common austenitic grades, and it is much stronger, so a wall can be thinner for the same duty. Shops fear it for one: the same microstructure that gives those properties is sensitive to how much heat it sees and how fast it cools afterwards.
Every weld has a working range of heat input. On plain austenitic stainless that range is wide enough that an experienced hand can stay inside it by feel. On duplex it is narrow, and it is bounded on both sides by different failures.
Too much heat and the ferrite in the heat-affected zone turns into austenite, or into coarse grains that leave the region weaker and less corrosion resistant than the parent. Too little heat and the ferrite stays excessive, which makes the joint harder and more crack-sensitive, and leaves the balance wrong in the other direction. There is only a band, and it moves with wall thickness and joint type.
Practically, this is why duplex work needs a written procedure and a welder who follows it rather than one who works by ear. The feedback you would normally use, bead appearance and puddle behaviour, tells you about fusion but not about phase balance.
The two phases do not expand at the same rate and do not transform at the same temperature. That mismatch is the root of most duplex difficulty. Within a single bead, one constituent is transforming while the other is only growing, and the result is internal stress that a homogeneous metal would not develop.
It also means the cooling rate is doing metallurgical work, not just thermal work. Slow cooling gives the austenite time to re-form in the right proportion. Fast cooling freezes in more ferrite than you want. So the cooling curve is part of the specification, and anything that changes it, such as a chill bar, a heavy fixture or a cold shop, is a process variable rather than a convenience.
Duplex fillers are usually over-alloyed compared with the parent, and the reason is the arc itself. Welding preferentially burns out nitrogen and shifts the balance towards ferrite, so the filler carries extra austenite-forming elements to push the weld metal back towards the intended mixture. Choosing a filler that matches the parent chemistry looks neat on paper and produces a weld with the wrong structure.
Mixing is the other half of the same problem. If the joint is a dissimilar one, or if the root is welded with a different filler from the cap, the final composition is an average rather than a design. On duplex that average is rarely where you want it. Single-filler procedures, applied consistently through the thickness, are easier to control than mixed ones.
Duplex moves more than austenitic stainless under the same heat, and it moves in a way that is harder to correct afterwards, because the strength that makes it attractive also makes it resist straightening. A fabricated section that racks out of tolerance is not going to come back with a few careful hammer blows.
The response is to plan the sequence around movement rather than chase it. Balance passes on opposite sides of the neutral axis, leave the assembly free to move where movement does not matter, and constrain only where the geometry demands it. Pre-setting the part slightly out of square, so welding pulls it into tolerance, is standard on duplex for exactly this reason.
Tack placement deserves more thought than it usually gets. Fewer, longer tacks on duplex reduce the number of times the part is heated and cooled before the real weld starts, and every one of those cycles contributes to both distortion and phase imbalance.
The root of a duplex weld has to be shielded like any other stainless root, and the same discipline applies on the inside as on the outside. An oxidised root loses corrosion resistance exactly where a tube is thinnest and most likely to be attacked, and on duplex that loss is compounded by the metallurgical disturbance the root pass already causes.
Where the root is welded without filler, the shrinkage is greater and the phase balance is more difficult, because there is no filler to restore what the arc removed. Where the root is welded with filler, the pass has to be kept small enough that the heat-affected zone does not grow beyond the procedure. Either way, the root is the pass that decides whether the part lasts.
On thick sections, the interpass temperature is the variable that quietly destroys duplex properties. The first pass is done to procedure, the second starts on metal that is still hot, and by the fourth the part is being welded in conditions nobody planned. The result is a joint with good-looking caps and a heat-affected zone that has been held at temperature far longer than intended.
Control means measuring rather than estimating. A contact thermometer on the joint face, checked before each pass and recorded, turns an assumption into a number the shop can defend. Waiting is the cheapest tool in the box and the hardest one to sell to a production planner, but on duplex it is not optional.
Duplex earns its cost in wet, chloride-bearing service and in structures where weight and strength matter at the same time. In dry, hot exhaust work it is often specified by habit rather than need, and a well-made austenitic part would do the same job with less process risk.
The honest split is this: where corrosion resistance in a wet environment drives the design, duplex is the right answer and the discipline is justified. Where the driver is only that duplex sounds more advanced, the specification is adding cost and risk without adding life. That is a conversation worth having at the drawing stage, before the material is ordered.
Duplex parts are bought on paper as though they were ordinary stainless, and that is where most of the trouble starts.
Most of the duplex we build sits in exhaust sections and bracketry where the material was chosen for a reason. The austenitic side of the shop, which follows the same documentation habits, is on our TIG stainless steel welding pages, and finished systems are listed under stainless steel exhaust systems. Pipe and tube sections for these builds are under exhaust pipe work, fabricated sub-assemblies under fabricated components, and titanium parts, where the duty allows it, under titanium TIG welding parts. Aluminium fabrications are shown separately under aluminium TIG welding.
