Sulfidation is the reason F5 and F9 exist in refinery flange service. It is a steady, predictable wall thinning mechanism, and chromium is the element that slows it. Unlike hydrogen attack, there is nothing sudden or hidden about it, which makes it one of the more manageable high temperature problems, provided the alloy was chosen correctly at the start.
What happens
Hot hydrocarbon streams carry sulfur compounds. Above roughly 500 degrees F those compounds react with the steel surface to form iron sulfide scale. The scale is not protective the way a passive oxide film is, so the reaction continues and the wall thins steadily.
Rate rises sharply with temperature and with sulfur content. It is a general thinning mechanism rather than a localised one, so it is amenable to thickness monitoring and to a corrosion allowance, which is how most units manage it.
Why chromium works
Chromium changes the nature of the scale. Chromium-bearing sulfide scales are denser, more adherent and considerably more protective than iron sulfide, so they slow the reaction instead of merely recording it.
The improvement is substantial and it scales with chromium content, which is why the refinery alloy ladder is essentially a chromium ladder:
| Material | Relative sulfidation resistance |
|---|---|
| Carbon steel | Lowest, and sensitive to silicon content |
| 1.25Cr-0.5Mo, F11 | Modest improvement |
| 2.25Cr-1Mo, F22 | Better |
| 5Cr-0.5Mo, F5 | Substantially better. The common refinery step |
| 9Cr-1Mo, F9 | Better again |
| 12Cr and austenitic stainless | Best, and a different cost bracket |
Note that F11 and F22 are chosen mainly for hydrogen and creep duty. When sulfidation is the dominant mechanism the jump to 5Cr is what actually buys you something.
Estimating the rate
Two established curve sets, and which one applies depends on hydrogen.
- Modified McConomy curves, for high temperature sulfidic corrosion without hydrogen present. Crude and vacuum unit hot sections, hot oil service.
- Couper-Gorman curves, where hydrogen is present, as in hydroprocessing. Hydrogen changes the chemistry of the attack, so the no-hydrogen curves do not transfer.
Both give corrosion rate against temperature and sulfur content, by alloy. Both are estimation tools with real scatter, so they inform a corrosion allowance and an inspection interval rather than giving a guaranteed life.
Two mechanisms that change the answer
Naphthenic acid corrosion
Some crudes carry naphthenic acids, and they attack in a way that chromium alone does not solve. Severity depends on the acid content of the crude, the temperature, and critically on velocity and turbulence, so it concentrates at elbows, pump internals, control valves and anywhere flow is disturbed.
The usual answer is not more chromium but an austenitic stainless with sufficient molybdenum. A 5Cr flange chosen for sulfidation can be attacked rapidly in naphthenic acid service, so the two mechanisms have to be assessed separately. If your crude slate has an acid content worth mentioning, mention it.
Silicon content in carbon steel
Not a chrome-moly issue, but it appears in the same units and it has caused real failures. Carbon steel components with low silicon content can sulfidize considerably faster than normal-silicon components in the same service, and low silicon carbon steel sitting alongside ordinary carbon steel has produced unexpected local thinning.
The industry response has been to pay attention to silicon on carbon steel components in sulfidation service. If a specification calls out a minimum silicon on carbon steel, this is why, and it is not a typographical oddity.
Practical implications for flanges
- Match the flange alloy to the piping alloy. A carbon steel flange in a 5Cr line becomes the thin spot, and a flange is a worse place to have one than a straight pipe run.
- Remember the bore. The wetted surface on a weld neck includes the bore through the hub, and that is where the fluid is fastest.
- Corrosion allowance is on thickness, not on class. Going up a pressure class adds flange thickness but does not change the bore wall, so it is not a substitute for the right alloy.
- Sulfidation does not care about PWHT, but the grade you pick for sulfidation still brings its own welding and heat treatment requirements. 5Cr and 9Cr are strongly air hardening. See preheat and PWHT.
What to tell us
- Design and operating temperature
- Sulfur content or the sulfidation rate basis the unit works to
- Whether hydrogen is present, and its partial pressure
- Whether naphthenic acid is a factor in the crude slate
- The alloy of the connecting pipe
- Corrosion allowance required
With that we can confirm the grade suits the service. Sulfidation is one of the mechanisms where getting the alloy right at purchase is genuinely cheaper than managing the consequences later, because the alternative is a unit turnaround.
See high temperature and hydrogen service for the other mechanisms and chrome-moly grades for the material reference.