Amine and Caustic Stress Corrosion Cracking: Where to Look and How to Find It

Short answer: amine and caustic stress corrosion cracking are both alkaline cracking mechanisms that attack carbon steel welds that were never post-weld heat treated (PWHT). Look first at weld heat-affected zones, nozzles, attachment welds, cold-formed bends and anywhere steam tracing or steam-out heats the metal. The cracks are tight and often oxide-filled, so liquid penetrant can miss them; wet fluorescent magnetic particle (WFMT), ACFM, eddy current and shear-wave or phased-array UT are the usual detection methods.

This guide is written for the people who have to plan the inspection: mechanical integrity engineers, API 510 and API 570 inspectors, turnaround planners and QA/QC leads at refineries, gas plants and chemical sites in the USA and Canada. It explains what the industry guidance says, where the cracks actually show up, how each NDE method performs on tight alkaline cracks, and what records your inspector of record will expect. It does not reproduce any standard. For the governing requirements, read the current editions of API RP 945, API RP 571 and AMPP SP0403 alongside your own inspection code.

Two mechanisms, one family: what amine SCC and caustic SCC have in common

Amine SCC and caustic SCC belong to the group often called alkaline stress corrosion cracking. Both need three things at the same time: a susceptible material (most commonly carbon steel and low-alloy steel), a tensile stress (usually residual stress from welding or cold work rather than operating pressure stress), and an alkaline environment at a temperature high enough to drive cracking. Remove any one of the three and the cracking does not develop. That is why the industry's main control for both mechanisms is the same: PWHT of carbon steel welds and cold-worked areas in susceptible service, which relieves the residual stress that the cracks need.

The two mechanisms differ in the environment. Amine SCC happens in aqueous alkanolamine solutions used to remove hydrogen sulfide and carbon dioxide from gas and liquid hydrocarbon streams in amine treating units, sulfur recovery tail-gas units and gas plants. Caustic SCC, historically called caustic embrittlement, happens in sodium hydroxide (and potassium hydroxide) service: caustic treating, neutralisation, caustic injection, boiler water treatment and spent-caustic systems.

API RP 571 describes both mechanisms in its damage-mechanism catalogue, and each has a dedicated industry document. API RP 945, Avoiding Environmental Cracking in Amine Units, is now in its fourth edition (September 2022). It addresses carbon steel equipment in amine units and the four cracking mechanisms found there: sulfide stress cracking, hydrogen-induced cracking, stress-oriented hydrogen-induced cracking and amine SCC. Cracking of stainless steels in amine units is outside its scope. For caustic service, NACE (now AMPP) SP0403, Avoiding Caustic Stress Corrosion Cracking of Refinery Equipment and Piping, contains the long-standing Caustic Soda Service Chart that plots caustic concentration against temperature to show where carbon steel can be used as-welded, where PWHT is needed, and where an alloy is needed. Use the chart from the current edition of SP0403 rather than a copy from an old handbook.

Amine stress corrosion cracking: where to look

The highest-probability locations for amine SCC are non-PWHT carbon steel welds that see lean amine, especially at elevated temperature. Experience reported in the industry literature shows cracking is most frequent in MEA (monoethanolamine) and DEA (diethanolamine) systems, but it has also been found in MDEA, DIPA and other amine services, so a "safe amine" assumption is not a substitute for checking the PWHT history.

• Weld heat-affected zones and the residual-stress zone next to them. Amine cracks are typically found in the HAZ or in the band of base metal just outside it where welding residual stress is highest. They usually run parallel to the weld.
• Nozzles and attachment welds. Cracks radiate from set-on nozzle welds and follow the HAZ of set-through nozzles. Internal attachment welds such as tray-support rings and clips, and external attachments welded directly to the shell, can produce cracks on the opposite (process) side of the wall.
• Longitudinal and circumferential seams of contactors, absorbers, regenerators (strippers), flash drums and reflux drums that were never stress relieved.
• Lean amine piping and exchangers: lean/rich exchangers, lean amine coolers, reboiler piping and lean amine lines downstream of the regenerator, where temperature and lean solution combine.
• Repair welds. A vessel that was PWHT at fabrication can be put back at risk by a later repair or hot-tap that was not heat treated. Repair records are as important as the original data report.
• Cold-formed components such as bends and formed heads where cold work left residual stress.

Rich amine sections are not exempt; they are dominated by wet hydrogen sulfide damage (blistering, HIC and SOHIC) rather than classic amine SCC, which is why RP 945 treats all four mechanisms together. If you are planning an amine-unit turnaround, read the companion guide on sour cracking, HIC and SOHIC so the inspection scope covers both families.

Caustic stress corrosion cracking: where to look

Caustic SCC risk depends mainly on caustic concentration, metal temperature and residual stress. The point inspectors most often miss is that the metal temperature can be far higher than the process temperature, so the most important locations are often the ones heated by something other than the process.

• Steam-traced and electrically heat-traced caustic lines. Tracing that is too hot, or tracing installed directly against the pipe without spacers, can push the local metal temperature into the cracking zone. Industry experience includes failures in improperly heat-traced piping and equipment.
• Equipment and piping that is steamed out while caustic residue is still present. Steam-out of non-PWHT carbon steel that has seen caustic is a recognised trigger.
• Heating coils, internal heaters and their connections in caustic tanks, where the metal near the heater runs much hotter than the bulk liquid.
• Nozzles, welds and areas next to welded seams, and other locations with residual or concentrated stress.
• Caustic injection points where concentrated caustic enters a hot hydrocarbon or water stream before it is fully mixed, and evaporation points where caustic can concentrate (drips, leaks under insulation, boiler crevices).
• Spent and contaminated caustic. Caustic contaminated with sulfides has caused cracking under conditions that the service chart would otherwise place in the as-welded carbon steel zone, so treat spent-caustic systems conservatively.

Carbon steel, low-alloy steel and 300-series stainless steel are all susceptible to caustic SCC. In carbon steel the cracks are usually intergranular, branched and oxide-filled, often in a network of many fine cracks; in austenitic stainless steel they are more often transgranular and can be hard to tell apart from chloride stress corrosion cracking, which is one reason metallography is used to confirm the mechanism after a first finding.

Why liquid penetrant underperforms on alkaline cracks

Both mechanisms produce surface-breaking cracks, so it is tempting to scope penetrant testing (PT) because it is quick and cheap to mobilise. The problem is the crack itself. Amine and caustic cracks are typically very tight and filled with oxide or scale. Penetrant has to enter the crack and then bleed back out to make an indication; a crack packed with corrosion product gives the penetrant nowhere to go. Industry references on amine cracking specifically note that PT may not detect these cracks, and the same physics applies to oxide-filled caustic cracks.

Wet fluorescent magnetic particle testing works differently. It relies on magnetic flux leaking from the crack faces, which does not require the crack to be open, and fluorescent particles under UV-A light give much higher contrast than visible dry powder. That is why WFMT has long been the standard surface method on the inside of carbon steel amine and caustic vessels. The catch is surface preparation: WFMT needs a clean surface, usually achieved by abrasive blasting or grinding to remove scale and deposits, and the quality of that preparation sets the detection limit. A rushed blast job turns a good method into a poor one.

NDE methods for amine and caustic SCC: what each finds and where it struggles

No single method suits every location. The table below summarises how the common methods behave on tight alkaline cracks. It is a planning aid, not a procedure: the examination procedure, reference standards and acceptance criteria come from your inspection code, the owner's specifications and the procedure approved by the NDT Level III.

For process-side cracking found from the outside, phased-array UT has largely replaced manual angle-beam scanning on accessible welds because it records the full scan and gives better coverage of the HAZ band. When a crack is confirmed, TOFD or PAUT tip-diffraction techniques give the through-wall dimension that engineers need. For a broader cross-reference of methods against mechanisms, see NDE method selection by damage mechanism.

How the inspection is applied in practice: a worked example

Consider, qualitatively, a carbon steel amine regenerator and its lean amine circuit at a Gulf Coast refinery. The fabrication records show the tower shell was PWHT, but the reboiler return nozzle was replaced in a later repair with no record of heat treatment, and the lean amine cooler inlet piping was never stress relieved. The unit runs MDEA, and the integrity team had treated it as low-susceptibility.

The inspection plan written by the owner's API 510/570 inspector would typically do the following. First, it would review the damage-mechanism assessment and flag every non-PWHT weld in lean amine service, including the repaired nozzle, because the repair broke the PWHT assumption. Second, during the turnaround entry, internal welds at the repaired nozzle and selected shell seams would be abrasive blasted and examined by WFMT, with the extent set by the inspector. Third, the lean amine piping welds, which cannot be examined internally, would be screened externally with a qualified shear-wave or PAUT procedure. Fourth, any indication would be sized (ACFM for surface length and depth estimate, PAUT or TOFD for through-wall height), photographed and mapped to the weld map.

If cracks are found, the decision on what happens next belongs to the owner. Options include grinding out shallow cracks and verifying removal by WFMT, a weld repair followed by PWHT where the code and the owner's engineers require it, replacement of a spool, or an engineering fitness-for-service assessment commissioned by the owner under API 579-1/ASME FFS-1. The NDE contractor's job is to deliver accurate locations, dimensions and morphology so those decisions are made on good data.

Prevention, mitigation and what the inspection plan should verify

Inspection finds cracks; it does not prevent them. A credible plan also checks that the mitigation measures the risk assessment relies on are actually in place.

• PWHT records. Confirm PWHT for every weld credited as stress relieved, including repair welds, hot taps and attachments added after fabrication. API RP 945 and AMPP SP0403 both give guidance on when PWHT is recommended; use the current editions for any temperature or concentration criteria.
• Material choice. Austenitic stainless steels and nickel alloys such as Alloy 400 are used where carbon steel is not suitable in amine service, while caustic service at high concentration and temperature may require nickel alloys. Remember that 300-series stainless steels are themselves susceptible to caustic SCC.
• Heat tracing design. Verify tracing type, spacing and temperature control on caustic lines against the owner's specification.
• Steam-out procedures. Check that procedures require caustic to be flushed out before non-PWHT equipment is steamed.
• Process control. Caustic concentration and injection mixing, amine strength and lean amine temperature are integrity operating windows that should be monitored, with excursions triggering inspection review.

Documentation the inspector of record expects

The owner's API-certified inspector remains the inspector of record under API 510 or API 570, and the inspection records must stand up to audit. For alkaline SCC work the NDE deliverables normally include:

• The NDE procedure reference and revision, the Level III approval, technician certification level and method, and equipment calibration records.
• Surface preparation method and the condition of the surface at examination.
• A weld map or drawing showing every examined weld and the percentage examined, so the coverage is reproducible at the next inspection.
• For each indication: location referenced to datum points, length, depth or through-wall estimate where measured, orientation, photographs under UV-A for WFMT, and the method used to size it.
• A clear statement of areas not examined and why (no access, insulation, internals in place).

Those records feed the owner's damage-mechanism review and inspection interval decisions under the applicable code. Atlantis does not set those intervals; the owner and the inspector do.

Common mistakes on amine and caustic cracking inspections

• Using PT on carbon steel in place of WFMT because it is quicker, then reporting "no relevant indications" on a crack-prone weld.
• Trusting the fabrication PWHT record and forgetting later repairs, hot taps and added attachments.
• Ignoring metal temperature. Scoping caustic lines by process temperature and missing the heat-traced sections that run much hotter.
• Too little surface preparation. WFMT on a poorly blasted surface with scale still in place.
• No crack sizing. Reporting crack presence without length and depth leaves engineers unable to assess the finding.
• Treating MDEA or other amines as immune. Cracking is less common in some amines but has been reported across most of them.
• Not confirming the mechanism. Caustic SCC in stainless steel can look like chloride SCC; the fix and future inspection plan differ, so metallography is worth the time.

Regulatory and jurisdiction overlay (USA and Canada)

In the United States, amine and caustic units at refineries and chemical plants are usually covered by OSHA Process Safety Management (29 CFR 1910.119). Its mechanical integrity element requires inspection and testing to follow recognised and generally accepted good engineering practices (RAGAGEP). API 510, API 570, API RP 571, API RP 945 and AMPP SP0403 are among the documents owners commonly cite as RAGAGEP for this equipment. The risk of these units is not theoretical: the 1984 Union Oil refinery disaster near Lemont (Romeoville), Illinois, which killed seventeen people, began with the rupture of an amine absorber pressure vessel. That failure was attributed to hydrogen-related cracking next to a repair weld rather than to classic amine SCC, but it drove industry surveys of cracking in amine service, an OSHA standard interpretation on potentially hazardous amine absorber vessels in 1986, and the guidance that RP 945 now consolidates. Pressure vessels may also fall under state or provincial boiler and pressure vessel law, which can impose its own inspection and repair rules.

In Canada, provincial regulators administer pressure equipment safety. In Alberta, ABSA requires owners operating under an integrity management system to follow their approved programme, and in Ontario the TSSA administers the boilers and pressure vessels regulation. Confirm the specific requirements with your jurisdiction, as repair, PWHT and reporting rules vary.

How Atlantis supports amine and caustic cracking inspection

Atlantis NDT performs the NDE that the owner's inspection plan calls for: wet fluorescent magnetic particle examination of internal welds and nozzles, ACFM and eddy current surface examination, shear-wave UT and phased-array UT from the external surface, and TOFD or PAUT crack sizing, all by ASNT-certified technicians under ASNT Level III oversight. Results, weld maps and indication data go to the owner's API-certified inspector, who remains inspector of record and decides repairs and intervals. Atlantis does not perform fitness-for-service or risk-based inspection assessments. See our pressure vessel inspection services, magnetic particle testing for refining and ACFM pages, or request a quote for amine or caustic cracking inspection. Quotes are returned within 24 hours.

Frequently asked questions

What is amine stress corrosion cracking?

It is cracking of carbon steel and low-alloy steel in aqueous amine solutions used to remove hydrogen sulfide and carbon dioxide. It needs tensile stress, usually residual welding stress, so it concentrates at non-PWHT welds in lean amine service. Cracks are typically intergranular, oxide-filled and run parallel to welds.

Which amines cause amine cracking?

Cracking has been reported most often in MEA and DEA systems, but it has also been found in MDEA, DIPA and other amines. The safer approach is to judge susceptibility by PWHT status, temperature and amine loading, not by amine type alone.

What is the best inspection method for amine cracking?

Wet fluorescent magnetic particle testing on a properly blasted surface is the standard internal method. ACFM and eddy current are used for screening and sizing, and shear-wave UT or PAUT is used from the outside when internal access is not available.

Can liquid penetrant find amine or caustic cracks?

Not reliably. These cracks are usually tight and filled with oxide, so penetrant may not enter them. Industry references specifically warn that PT may miss amine cracks, which is why WFMT is preferred on carbon steel.

What does API RP 945 cover?

API RP 945, fourth edition (2022), covers environmental cracking of carbon steel equipment in amine units: amine SCC, sulfide stress cracking, HIC and SOHIC, with guidance on materials, fabrication, inspection, monitoring and repair. Cracking of stainless steels in amine units is outside its scope.

Where does caustic stress corrosion cracking occur?

Most often at non-PWHT carbon steel welds and cold-worked areas in caustic service, and especially where metal runs hotter than the process: steam- or heat-traced lines, heating coils, and equipment steamed out with caustic still present.

What is the caustic soda service chart?

It is a chart in NACE/AMPP SP0403 that plots caustic concentration against temperature and shows where carbon steel can be used as-welded, where PWHT is recommended and where alloys are needed. Use the current edition of SP0403 for the actual boundaries.

Does PWHT prevent amine and caustic cracking?

PWHT is the main mitigation for both, because it relieves the residual stress the cracks need. It has to cover repair welds and later attachments too; one untreated repair can reintroduce the risk.

Is caustic embrittlement the same as caustic SCC?

Yes. Caustic embrittlement is the older name, from boiler practice, for caustic stress corrosion cracking. The mechanism and the inspection approach are the same.

Can stainless steel suffer caustic cracking?

Yes. 300-series stainless steels are susceptible to caustic SCC, and the cracks can resemble chloride SCC. Metallography is often used to confirm which mechanism is active.

Who decides what happens when amine cracks are found?

The owner and its API-certified inspector, with engineering support where needed. Options include grind-out, weld repair with PWHT, replacement, or an owner-commissioned fitness-for-service assessment. The NDE contractor supplies the location and sizing data.

Planning a turnaround on an amine or caustic unit? Talk to our NDE team about WFMT and PAUT scope, browse all inspection services, or ask for an amine unit turnaround NDE quote.