Titanium is reactive in a way stainless steel is not. At room temperature it carries a thin, stable oxide skin that protects it and gives it the colour people expect. Heat that skin and it stops being a barrier: the metal underneath takes oxygen, nitrogen and hydrogen out of any atmosphere that contains them, and does so quickly, while the weld is still hot.
The practical consequence is that a titanium weld is not judged only on fusion. It is judged on what the metal was breathing while it cooled. Two welds with the same bead profile, the same penetration and the same fit can be one acceptable part and one scrap, purely because one of them had gas on it until it was cool and the other did not.
Oxygen is the obvious contaminant and the one that shows up first as colour. Nitrogen is slower to show and more damaging inside the joint, because it does not announce itself with a bright surface. It stiffens the metal, and the stiffness surfaces later as cracking, sometimes after the part has been through service heat cycles. Hydrogen is the quietest of the three and the one that leaves the least visible trace on the outside of a bead.
Because the damage is cumulative and mostly invisible, purge discipline is a process question rather than an inspection question. You cannot cut a section and decide whether the shielding was adequate. You have to be able to say, from the setup and from the colour, that the part never had a chance to absorb anything in the first place.
A torch cup throws a cone of gas that covers the arc and a short distance around it. That is enough to protect the molten pool. It is not enough to protect the hot solid metal immediately behind the pool, and it is certainly not enough to protect the inside of a tube, where the root of the weld is cooling in whatever atmosphere the tube happens to contain.
Titanium keeps absorbing contaminants well below the point where it stops glowing. The metal is still reactive while it looks dull grey and, to the eye, finished. So the requirement is not "gas while welding". It is "gas until the part is cool enough that the surface has stopped being hungry", and that window extends further behind the arc than most people estimate.
The three tools in the title are not alternatives. The torch covers the pool, a trailing shield covers the cooling bead, and an internal purge covers the root. Each fails in a place the other two cannot reach.
A trailing shield is a second gas outlet that follows the torch and blankets the bead behind it: usually a flat or curved shoe with a diffusing screen, fed from its own line so the torch flow is not disturbed. The separate line is about stability, not generosity. Tap the torch supply and the flow drops exactly when you need it, and the arc wanders. The shoe must sit close to the work without touching, follow the joint without twisting, and stay over the bead as the torch changes angle, because on a curved section a shoe that lifts at the change of direction stops protecting the place that needs it most.
For small components, a chamber is often the cheaper answer. A sealed box with an inlet, an outlet and a glove or a window is filled with argon until the atmosphere inside is effectively inert, and then the whole part is welded inside it. Everything is protected at once: root, bead, back face and the areas you never thought about because they are hidden behind a flange.
With a chamber the question is whether the atmosphere is actually inert when the arc starts, not whether gas is flowing. Fill it, let it settle, monitor the outlet instead of assuming. A chamber that is "probably fine" is the most expensive way to scrap a small titanium part.
On tube and pipe the root is the problem, and no chamber is big enough for a full exhaust system, so the purge has to be local. The tube is dammed at both ends of the section being welded, gas is fed into the sealed volume, and the joint is welded while the inside is full of inert gas. Water-soluble dams and heat-resistant tape close the run because they can be removed without cutting the tube open.
Two failures matter. A dam that leaks means the purge volume never reaches a clean state and the root goes grey however long you wait. A dam set too close to the joint stops gas flowing past the weld, so the root starves while the outside of the bead looks perfect.
Time is the variable shops get wrong most often on tube. Purging a long section is not instant, because gas has to displace air in a confined volume with limited outlets. Purge until the outlet reads clean, hold a while longer, and keep the gas on after the arc stops until the joint is cool. Turning the purge off when the arc stops is the commonest cause of a grey root on a good weld.
Colour on a titanium weld is a map of peak temperature and atmosphere. Light straw or pale gold is what a properly shielded bead looks like; bronze and dark blue mean the metal got hotter than the shielding could cope with; a dull grey or a white powdery deposit means the surface reacted. Where the colour sits matters as much as the shade. A grey band at the end of a run says the gas was cut early, grey at the toe says the shield was too narrow, and the back face has to be read with a mirror or a borescope, because the root is where the purge worked or did not.
Ends of tubes are the hard part. The dam has to be close enough to seal the volume and far enough away to let gas circulate, and the last joint on a section is often welded when the surrounding metal is already warm from previous passes. Warm metal absorbs faster, so the same shield that worked on the first joint is marginal on the last one.
The fix is procedural rather than clever. Weld the awkward end first, while the section is cold and the purge volume is fresh. Keep the trailing shield on past the end of the bead. Leave the internal gas on until the joint can be handled with a bare hand, which is a crude test but a reliable one. Then inspect the root before the part leaves the bench, not after it is packed.
A titanium part that is priced on geometry alone is priced on the wrong basis, because the shielding regime is where most of the real cost sits. Buyers who understand that ask better questions, and they get fewer surprises on arrival.
The same logic runs through everything we build. The titanium parts we produce are on our titanium TIG welding parts pages, the systems they go into under titanium exhaust systems, and the fabricated sub-assemblies under titanium fabrication components. The stainless side of the shop works to the same discipline, which you can compare on our stainless steel exhaust systems and TIG stainless steel welding pages, and pipe and tube sections are listed under exhaust pipe work if you want to see how the runs are assembled.
