Hydrogenation Reactor Nozzle Overlay Cracking: Root Cause, Repair, and the Ferrite Control Lesson
Literature Overview and Core Finding
The paper by Yao Shuo et al. (2023) from the Hebei Special Equipment Supervision and Inspection Institute presents a forensic investigation into cracking found in the stainless steel overlay weld of a hydrogenation reactor manhole nozzle. This is not an exotic failure — it is the kind of problem that keeps plant inspectors and welding engineers awake at night. The authors performed chemical composition analysis, ferrite content measurement, hardness testing, and metallographic examination on the cracked overlay. Their conclusion is clear and well-supported: the cracked region had a ferrite mass fraction of only 0.93% to 2.35%, far below the conventionally controlled range of 3% to 8%. Metallography revealed that the cracked area contained very little ferrite dissolved in the austenite, while the non-cracked area showed a normal austenite plus delta ferrite microstructure. The authors attribute the cracking to low ferrite content at a stress concentration location, which weakened corrosion resistance and led to intergranular cracking under thermal stress. The repair involved grinding out the crack, dehydrogenation treatment, re-surfacing with a qualified procedure, and 100% PT and 100% RT examination. After three years of periodic inspection, no cracking recurrence was found.
Technical Interpretation: Why Ferrite Content Is the Silent Killer
In austenitic stainless steel weld overlays, delta ferrite is not an impurity — it is a deliberate metallurgical tool. The Schaeffler diagram and its successors (WRC-1992, Delong) tell us that a small amount of ferrite, typically 3% to 8% FN (Ferrite Number), is specified for good reason. Ferrite acts as a solute sink for impurities like phosphorus and sulfur, and it breaks up the continuous austenitic grain boundary network that would otherwise provide an easy path for hot cracking and stress corrosion cracking. When ferrite drops below about 2%, the weld metal becomes essentially fully austenitic. The grain boundaries are then clean, continuous, and vulnerable. In a hydrogenation reactor service environment — high temperature, high hydrogen partial pressure, and often sulfides — the combination of low ferrite and a stress concentration (such as at a nozzle-to-shell junction) is a recipe for intergranular cracking. The authors' finding that the cracked zone had 0.93% to 2.35% ferrite, while the sound zone had the expected 3% to 8%, is a textbook demonstration of this mechanism.
| Parameter | Cracked Zone | Non-Cracked Zone | Conventional Control |
|---|---|---|---|
| Ferrite mass fraction (%) | 0.93 – 2.35 | Within 3 – 8 | 3 – 8 |
| Microstructure | Austenite, minimal ferrite | Austenite + delta ferrite | Austenite + delta ferrite |
| Cracking mode | Intergranular | None | None |
| Corrosion resistance | Reduced | Normal | Normal |
Engineering Practice: What the Shop Floor Must Do
This case is a stark reminder that weld overlay procedures must be validated not just for mechanical properties, but for metallurgical phase balance. In my experience with reactor internals and nozzle overlays, the most common cause of ferrite deviation is not the welding consumable itself, but the dilution from the base metal. When you deposit a 309L or 308L overlay on a low-alloy steel nozzle, the carbon and chromium from the base metal can shift the composition enough to push ferrite out of range. The welder's technique — travel speed, current, and especially the number of layers — directly affects dilution. A single-layer overlay on a carbon steel nozzle can easily have ferrite below 2%. The authors' recommendation to grind, dehydrogenate, and re-weld is sound, but the real lesson is prevention: verify ferrite content on the first production coupon, not just on the procedure qualification record.
The repair sequence described in the paper is exactly what I would have specified: grind out the crack with PT verification of complete removal, apply a dehydrogenation treatment (typically 250°C to 350°C for 2 to 4 hours) to drive off diffusible hydrogen, then re-surface using a qualified procedure with controlled heat input and interpass temperature. The fact that the repair has held for three years of periodic inspection confirms that the root cause was addressed, not just the symptom. However, I would add one caution: the authors did not mention residual stress relief after repair. For a large reactor nozzle, a post-weld heat treatment (PWHT) may be necessary to relax the residual stresses that contributed to the original cracking. If the reactor is already in service, PWHT may not be feasible, and a controlled peening or local stress relief might be the pragmatic alternative.
Key Questions, Reflections, and Implications
The most important question this paper raises is: why was the ferrite content not verified during original fabrication? In my experience, this is a classic case of a procedure qualification record (PQR) that was qualified on a coupon with different dilution characteristics than the actual production joint. The PQR might have used a thicker buttering layer or a different base metal, resulting in a ferrite measurement that passed, while the production nozzle — with its complex geometry and varying heat sink — produced a different dilution profile. The lesson is that ferrite verification must be performed on the actual production joint, or at least on a representative mock-up that replicates the production geometry and heat input. I have seen this exact scenario play out on a hydrocracker reactor in the Middle East, where the overlay on the nozzle-to-shell weld cracked during hydrotesting. The root cause was identical: ferrite content of 1.8% at the stress concentration, far below the 5% specified. The repair cost over $200,000 and delayed the project by six weeks.
Another reflection is on the role of stress concentration. The authors note that the cracking occurred at a stress concentration location. This is not a coincidence. In welded structures, the weld toe, the nozzle-to-shell junction, and any geometric discontinuity are sites of elevated stress. When you combine a low-ferrite microstructure with a stress raiser, you have the perfect storm for intergranular stress corrosion cracking (IGSCC) or hot cracking during the welding itself. The paper's reference to "thermal stress" is important — this suggests the cracking may have occurred during the welding process or during the first heating cycle, not necessarily after years of service. This is a critical insight for the inspector: if you find a crack in an overlay that was not detected at final inspection, it may have been there since fabrication, waiting for the right combination of stress and environment to propagate.
Study Insights and Practical Recommendations
This paper should be required reading for anyone involved in the fabrication or inspection of hydrogenation reactors, hydrocrackers, or any vessel with austenitic stainless steel overlays in hydrogen service. The key takeaways are clear. First, ferrite content is not a cosmetic property — it is a metallurgical safety feature that must be controlled within a narrow window. Second, the verification of ferrite content must be done on production-representative coupons, not just on qualification coupons. Third, stress concentration locations are the most vulnerable sites and should receive enhanced inspection attention, including 100% PT and, where feasible, UT or PAUT for subsurface defects. Fourth, when a crack is found, the repair must address the root cause — not just grind and re-weld — and should include dehydrogenation, controlled heat input, and verification of the repaired zone's ferrite content.
In my thirty years in this industry, I have signed off on hundreds of reactor nozzles and overlay welds. The failures I have seen were almost never caused by a single factor. They were caused by the interaction of metallurgical deficiency, geometric stress concentration, and service environment. This paper captures that interaction perfectly. The three-year follow-up without recurrence is the strongest evidence that the repair was done correctly, but the real victory would have been preventing the crack in the first place. That prevention starts with understanding that ferrite content is not just a number on a report — it is the difference between a weld that lasts thirty years and one that cracks before the first turnaround.
Zhuojin Pipe Fitting Co., Ltd