Amine cracking: what it is, where it hides, and how to look for it

Amine units strip acid gas out of hydrocarbon streams using an aqueous alkanolamine solution which is then regenerated and recirculated. They corrode in two quite separate ways, and the first thing to get straight is which one you are dealing with.

• Amine corrosion is thinning. It is driven by amine type and strength, temperature, velocity, acid gas loading and — often dominantly — the concentration of heat-stable amine salts that build up in a solution that is not being properly reclaimed. It attacks the hot lean and rich lines, the regenerator, the reboiler and anywhere the solution flashes. You find it with thickness readings.
• Amine cracking is alkaline stress corrosion cracking. It attacks the weld and the heat-affected zone of carbon steel, it removes essentially no metal, and no thickness reading will ever detect it.

The rest of this is about the second one.

What it needs to run

Three ingredients, and the third is the one you control:

1. The environment

   An aqueous amine solution. Lean amine is generally more aggressive for cracking than rich, which is counter-intuitive to people who expect the acid-gas-loaded stream to be the dangerous one — cracking here is an alkaline mechanism, not an acidic one.

2. Temperature

   It gets worse with heat. The hot lean amine circuit — regenerator bottoms, lean amine to the exchangers, the reboiler and its piping — is where cracking concentrates.

3. Tensile stress, and specifically residual stress

   The applied pressure stress is rarely enough on its own. What drives the cracking is the residual stress left in and around a weld, which in an as-welded carbon steel joint can approach the yield strength of the material.

What it looks like

Cracking is typically intergranular, initiating on the surface wetted by the amine — the inside — and running parallel to the weld in the heat-affected zone, or in the weld metal itself. It can be extensive and fine rather than a single obvious feature, and it may be under a layer of scale or deposit that hides it completely from a casual internal visual.

Because it starts on the process side and grows outward, it spends most of its life invisible from where inspectors usually stand.

Distinguishing it from wet H2S damage

Both crack carbon steel welds in units that handle H₂S, both are controlled largely by PWHT and weld hardness, and the two get conflated constantly. The distinction that matters practically:

They also coexist. An amine unit handles H₂S by definition, so a vessel can be susceptible to both, in different places, for different reasons. That is an argument for controlling residual stress and hardness properly rather than for deciding which mechanism to inspect for.

How to actually look for it

Everything follows from the fact that it is surface-breaking on the inside, at welds.

• From the inside, if you can get in. Wet fluorescent magnetic particle inspection of the internal weld seams and nozzle welds is the reference technique, and it needs the surface properly cleaned first — grit blasting, not a wire brush. Scale hides fine intergranular cracking completely.
• ACFM or eddy current where the surface cannot be brought to bare metal, since both tolerate coatings better than MT does.
• From the outside, angle beam ultrasonics, phased array or TOFD aimed at the weld — this is how piping and any vessel that cannot be entered gets inspected. Detection is achievable; sizing fine, branched, multiple-initiation cracking is considerably harder than sizing a single planar flaw.
• Hardness testing on welds, particularly repairs, as a proxy for whether stress relief happened and whether the weld procedure produced a hard HAZ.
• Nothing volumetric for thickness helps at all. A full corrosion map of a cracked vessel will report it as sound, because it is sound — right up until it is not.

Where to spend the inspection

Rank by residual stress and temperature together. Non-PWHT'd welds in hot lean amine service come first, then repairs and modifications of unknown heat treatment history anywhere in the unit, then nozzle and attachment welds where restraint is highest, then PWHT'd main seams. A unit that was fully heat treated and has stayed that way through every modification is a genuinely low-risk unit for this mechanism — and knowing that with confidence is worth more than any amount of scanning.