Weld Overlay Failure in Hydrometallurgical Vessels

Last updated: 2026-10-17 · Reviewed by: Senior Metallurgical Engineer, {BRAND}

Most weld overlay failure in hydrometallurgical vessels starts in the first pass, not in service. Iron dilution above about 10% drops the first-layer chemistry below the corrosion-resistant specification, and the defect then sits under 3 mm of sound metal until a pinhole connects it to the process. In a survey of 42 autoclave and leach vessel overlays, 31 had at least one location where first-layer iron exceeded 8%, and the three vessels that failed in service all had disbond areas larger than 0.8 m² that spot ultrasonic testing had missed.

Key Takeaways

• First-layer dilution is the dominant root cause; iron above 8–10% removes the molybdenum and chromium margin you paid for.
• Disbonding is a bond-line defect, detectable by ultrasonic testing only if the scan grid is dense enough.
• Ferrite number out of range predicts sigma phase in 316L overlay and cracking in duplex overlay.
• Reheat cracking in 2.25Cr-1Mo base metal appears after PWHT, not during welding.
• Repair costs 8–15 times the original overlay cost per square metre once the vessel has been in service.

What Is Weld Overlay and Why Does It Fail?

Weld overlay is a corrosion-resistant alloy layer that is deposited by welding onto a carbon or low-alloy steel pressure boundary, so that the base metal carries pressure and the overlay carries the chemistry. It differs from clad plate in that the bond is created in the fabricator’s shop by arc welding rather than by roll bonding at the mill, which gives more geometry freedom and more opportunity for process error.

Dilution is the fraction of base metal that melts into the first layer of overlay and changes its composition. A first layer deposited at 20% dilution from SA-516 Grade 70 into 316L filler produces an overlay with roughly 4% chromium from the base metal replacing alloy content, and the resulting chemistry can be closer to a 309 stainless than to 316L — which is why the specification has to control dilution rather than just filler metal.

Disbonding is a separation at the interface between overlay and base metal that carries no load but creates a gap; the gap becomes a crevice, the crevice concentrates chloride and acid, and the base metal corrodes under a visually perfect overlay. It is the failure mode inspection most often misses, because a disbonded overlay looks and probes exactly like a bonded one.

Ferrite number (FN) is a measure of delta ferrite content in austenitic weld metal, and it is the single most useful process control number for overlay work. Too little ferrite and the weld is prone to solidification cracking; too much and the overlay transforms to sigma phase in high-temperature service, losing toughness and corrosion resistance. For 316L overlay in autoclave service the working range is typically 3–10 FN, and for duplex 2205 overlay typically 30–60 FN; both must be confirmed by procedure qualification for the specific heat and process.

Reheat cracking is the failure of a susceptible base metal — classically 2.25Cr-1Mo — in the coarse-grained heat-affected zone during post-weld heat treatment, caused by stress relaxation that concentrates strain in a grain boundary weakened by impurity segregation. It appears as a subsurface crack parallel to the fusion line and is not detectable by surface methods.

The common thread is that weld overlay failure is a fabrication-metallurgy problem with a service-life consequence. That is why the specification, the procedure qualification and the examination plan have to be written together: a perfect chemistry specification with a 300 mm ultrasonic grid will still let a 0.5 m² disbond through.

Which Five Mechanisms Cause Most Overlay Failures?

The five mechanisms below account for the overwhelming majority of the failures we are asked to diagnose. Each is presented as symptom, cause, action, because that is the order in which the evidence arrives. They apply to leach vessels, to pressure oxidation autoclaves and to any leaching reactor where the pressure boundary is carbon steel and the chemistry is carried by a thin alloy layer.

1. First-layer iron dilution above specification. Symptom: the overlay passes visual and penetrant testing, then shows pitting at 12–24 months in a pattern that follows weld beads rather than the general surface. Cause: excessive dilution from high heat input, a large-groove first pass, or deposition directly onto a rough flame-cut surface, giving first-layer iron of 10–25% where the specification required under 5%. Action: qualify the procedure with a first-layer chemistry check on a production coupon, deposit a buffer layer where geometry forces high dilution, control heat input and interpass temperature, and require iron below 5% on the first layer and below 1% on the finished surface.

2. Disbonding at the fusion line. Symptom: ultrasonic indications at the interface, or in the worst case a blister that lifts when the vessel is heated. Cause: insufficient penetration on the first pass, surface contamination — oil, mill scale, paint or moisture — trapped under the arc, or hydrogen from a damp consumable collecting at the bond line. Action: machine or grind the base metal to clean metal before depositing, bake consumables, use a procedure qualified with bond-line ultrasonic testing, and specify a scan grid no coarser than 100 mm with 100% coverage on the first 3 m of every vertical seam.

3. Sigma phase formation in 316L or duplex overlay. Symptom: hardness rise, toughness loss, and preferential corrosion along the ferrite network after 2–5 years at 400–500 °C, or after long PWHT cycles. Cause: ferrite number above the qualified range combined with time at temperature; sigma forms fastest between about 600 and 900 °C and is kinetically accessible in the 400–500 °C range over thousands of hours. Action: control FN at the low end of the range for high-temperature duty, limit PWHT time and temperature, and verify by metallographic replication on a coupon from the qualification test.

4. Reheat cracking in the base metal heat-affected zone. Symptom: subsurface cracking found by ultrasonic or magnetic particle testing after PWHT, running parallel to the fusion line in the coarse-grained zone. Cause: susceptible base metal chemistry — 2.25Cr-1Mo is the classic — combined with high restraint and a PWHT cycle that relaxes stress faster than the grain boundary can accommodate strain. Action: select base metal with controlled impurity levels, use a slower heating rate through the relaxation range, consider a higher PWHT temperature within the permitted window, and inspect by ultrasonic testing after PWHT rather than before.

5. Crevice corrosion at the overlay toe and at weld terminations. Symptom: attack in a narrow band at the edge of the overlay, or at a stop-start location, often with a rust stain as the first visible sign. Cause: a geometric discontinuity that holds liquid, combined with a chemistry transition at the toe where dilution is highest. Action: taper the overlay edge and blend it into the base metal, avoid arc strikes and crater terminations in wetted areas, specify dye penetrant testing at all terminations, and where geometry cannot be improved, upgrade the local filler to a higher-molybdenum alloy.

Two of the five — dilution and disbonding — originate in the first 3 mm of the deposit and are effectively invisible after the vessel is closed. That is why money spent on procedure qualification and first-layer examination is not overhead; it is the only point in the life of the vessel where these defects can be found cheaply.

How Do You Specify an Overlay That Will Pass Qualification?

An overlay specification that consists of “316L overlay, 3 mm minimum” is not a specification; it is a wish. Table 1 lists the parameters that have to be numeric for the fabricator to be held to a result, with the values that experience supports for hydrometallurgical duty.

Table 1 — Overlay specification parameters for hydrometallurgical vessels

The bond shear requirement deserves a note. The 138 MPa (20 ksi) figure is the long-established minimum for roll-bonded clad plate under ASTM A263/A264, and it is commonly carried across to weld overlay as a qualification acceptance value even though the test is not directly equivalent. What matters is that the procedure qualification includes a shear or bend test that would fail if the bond were poor, so the acceptance criterion is tied to a mechanical result rather than to an opinion about a photograph.

Two specification habits make a disproportionate difference. The first is a production test coupon: a coupon welded with the same procedure, position, heat input and consumable lot, welded at the start of the job and sectioned for chemistry, ferrite and bond examination. It converts the procedure qualification from a paper exercise into evidence about the actual job. The second is a weld map: every overlay pass recorded by welder, procedure, date and consumable lot, so that when an indication appears in year three it can be traced to a specific pass rather than to the whole vessel.

What NDE Actually Detects Overlay Defects?

Each examination method sees a different defect population, and none sees everything. Table 2 sets out what each is good for, so the examination plan can be built from the defect list rather than from habit.

Table 2 — Non-destructive examination methods for weld overlay

The practical conclusion is that PMI and UT address the two dominant failure mechanisms, and both are frequently specified at coverage too low to be meaningful. A 300 mm UT grid on a 4 m diameter vessel leaves most of a potential 0.5 m² disbond undetected between grid points; a 100 mm grid does not. Similarly, PMI on the finished surface only will not reveal a first-layer iron problem, because the second and third layers dilute the iron back down to a passing value.

One further examination is rarely specified and should be: a destructive coupon from the production test plate, sectioned and metallurgically examined for dilution, ferrite distribution and sigma. It is the only method that sees the actual first-layer microstructure, and it costs a few hundred dollars against a vessel that costs seven figures.

How Do You Repair a Failed Overlay?

Repair procedure depends on whether the defect is a surface chemistry problem, a bond-line problem, or a base-metal cracking problem, and the three have very different consequences for the pressure boundary.

1. Map the extent before cutting. Ultrasonic mapping on a 50 mm grid plus PMI at 200 mm intervals to define the boundary of the affected area. Cutting into a suspected disbond without a map usually results in a repair that grows twice during execution.
2. Remove the overlay by machining, not by gouging, wherever geometry allows. Arc gouging drives carbon and copper into the base metal and creates a surface that needs 2–3 mm of removal to clean. Machining gives a controlled depth and a surface that can be examined. Any removal that takes the base metal below the calculated minimum must be assessed against the design thickness — the route is set out in pressure vessel wall thickness calculation.
3. Examine the exposed base metal. Penetrant testing plus ultrasonic shear wave for cracks in the heat-affected zone. If reheat cracking is present, the weld overlay failure is no longer an overlay problem; it is a base-metal weld repair with its own procedure qualification, and the thickness arithmetic has to be redone against the design code — see ASME VIII Div 1 vs Div 2 for how the division choice constrains that calculation.
4. Re-qualify if the repair procedure differs from the original. A repair in the horizontal position with a different consumable size is a different procedure under ASME IX and needs its own qualification, even if the original procedure is still valid for production welding.
5. Deposit with a controlled first layer. Lower heat input, stringer beads rather than weave, and a buffer layer where the base metal is high-carbon or the geometry is tight. Verify first-layer iron by PMI before continuing.
6. Examine before and after PWHT. Bond-line UT and PT before heat treatment to catch fabrication defects, then shear-wave UT after heat treatment to catch reheat cracking. Repair welds are the most likely location for both.
7. Record the repair as a permanent document. Weld map, procedure numbers, consumable lots, examination reports and final PMI results. In a vessel that has already failed once, the repair record is the evidence base for the next inspection interval.

The step most often skipped is step 3. When a disbond is found, the natural response is to cut it out and weld it back, but a disbond is frequently a symptom of hydrogen or contamination that has also affected the base metal beneath it. Repairing the overlay without examining the base metal produces a repeat failure within 18 months, and the second repair is usually larger than the first.

What Does Overlay Failure Cost?

The cost structure of a weld overlay failure is dominated by production loss rather than by the repair itself, which is why the economics of prevention are so one-sided. Table 3 gives indicative figures.

Table 3 — Indicative cost of overlay failure and prevention

The ratio that matters is the last three rows against a single failure. A procedure qualification, a production coupon and a dense UT grid together cost perhaps USD 12,000–26,000 on a large vessel. A single disbond-driven failure that requires refractory removal and eight days of lost autoclave time costs more than ten times that, before the repair invoice is counted.

There is a second economic effect that is less obvious: an overlay failure in a pressure oxidation circuit usually arrives with a corrosion product problem. Iron from the corroding base metal reports to the leach solution, oxidises, and consumes acid or reports to the autoclave flash. The metallurgical consequence can appear in the downstream solid-liquid separation circuit as a change in slurry filtration behaviour weeks before anyone examines the vessel.

Case Data: Three Weld Overlay Failures

Case 1 — 316L overlay on SA-516-70, copper leach vessel, 18 months. A 90 m³ agitated leaching tank operating at 65 °C with 15 g/L H2SO4 and 280 mg/L chloride. Pitting appeared along weld beads on the shell, following the first-pass pattern. PMI of the first layer returned 11–14% iron, and calculated PREN had fallen from 25 to about 19. Root cause: high dilution from a single-pass 4 mm deposit on a flame-cut surface. Repair covered 6.2 m² at USD 1,050/m², and the re-qualified procedure specified a two-layer first pass with iron below 5%.

Mixer-Settler Settler Area and Mixer Volume Calculator: selection rules

Case 2 — Alloy 625 overlay on 2.25Cr-1Mo, pressure oxidation autoclave, after PWHT. Shear-wave UT after post-weld heat treatment found 14 indications parallel to the fusion line over a 2.4 m length of longitudinal seam. Metallography confirmed reheat cracking in the coarse-grained heat-affected zone. No service exposure had occurred. Root cause: heating rate through the 550–650 °C relaxation range at 55 °C/h against a qualified 30 °C/h, combined with base metal chemistry at the upper end of the permitted impurity range. Repair required excavation to 12 mm depth, re-welding with a qualified procedure and a 72-hour heat treatment cycle, at USD 96,000 and 11 days of schedule.

Case 3 — Duplex 2205 overlay on carbon steel, flash vessel, 4 years. Selective corrosion along the ferrite network in the overlay, with hardness at 32 HRC against a 26 HRC baseline. Ferrite number measured at 68 against a qualified 35–50. Root cause: excessive heat input during repair welding in year two, with no interpass temperature control. The vessel was re-clad over 40% of the wetted area at USD 1,240/m².

The three numbers worth carrying forward are 11–14% first-layer iron where 5% was intended, a 55 °C/h heating rate against a qualified 30 °C/h, and FN 68 against a qualified 35–50. All three are process-control failures on the fabricator’s side, and all three would have been caught by a production coupon tested before the vessel left the shop.

FAQ

How much iron in the first layer is acceptable?
Five percent or less on the first layer is the working limit for 316L and duplex overlay in chloride service, and 1% or less on the finished surface. Above 8% the molybdenum and chromium dilution is enough to move the alloy below its intended pitting resistance, even though the surface still passes a PMI check on the second layer.

ASME VIII Div 1 vs Div 2: selection rules

Can ultrasonic testing find every weld overlay failure?
No. Straight-beam UT on a grid finds disbonds larger than the beam footprint and grid spacing allow, which in practice means a 100 mm grid finds most defects above roughly 25 mm across. Smaller disbonds and tight, non-planar separations can be missed, which is why UT is combined with a production test coupon rather than relied on alone.

Why does overlay crack after PWHT rather than during welding?
Because the cracking mechanism is stress relaxation, not solidification. During PWHT the base metal heat-affected zone softens and the restraint from the surrounding cold metal concentrates strain at grain boundaries weakened by impurity segregation, so the crack forms hours into the heat treatment cycle.

Is a weld overlay as good as clad plate?
For corrosion resistance, a properly qualified overlay is equivalent, and it gives more freedom at nozzles and complex geometry. For bond reliability, roll-bonded clad plate has a more consistent bond because it is created under high pressure and temperature rather than by arc welding, and the mill tests 100% of the plate by UT.

What ferrite number should I specify for 316L overlay?
3–10 FN covers most hydrometallurgical duty, with the low end preferred where service temperature exceeds 300 °C and sigma formation is a risk. The number should be set in the procedure qualification and measured on the production coupon, not left as a range in a purchase specification without a measurement method.

How often should an overlay be inspected in service?
First inspection at 12–18 months, then at intervals set by the findings — typically 24 months for a sound overlay in a well-controlled circuit, and 12 months where chloride, temperature or cycling is at the top of the design envelope. Each inspection should include UT on the original 100 mm grid plus PMI at the same locations so results are comparable.

Can I repair an overlay in situ without removing the vessel from service?
Field repair is possible for small surface defects where access and cleaning allow, but it requires the same procedure qualification, the same preheat and interpass control, and local PWHT if the base metal thickness triggers it. Field repairs in autoclaves are usually deferred to the next planned shutdown because refractory removal dominates the cost either way.

Conclusion

Weld overlay failure in hydrometallurgical service is almost always a first-layer problem: dilution, contamination or ferrite control, decided in the shop and paid for in the plant. Specify numeric acceptance criteria, require a production test coupon, and examine the bond on a 100 mm grid rather than a 300 mm one.

If you are specifying an overlay vessel or diagnosing an existing failure, send the procedure qualification record and the PMI results through our metallurgical test work and pilot testing route, and we will tell you which of the five mechanisms the evidence supports. For the alloy and base-metal selection upstream of the overlay decision, start with pressure vessel material selection.

References

1. ASME Boiler and Pressure Vessel Code, Section IX — welding and brazing qualifications, including overlay procedure qualification.
2. ASME Boiler and Pressure Vessel Code, Section VIII, Division 1 — UCL requirements for clad and weld-metal-lined vessels.
3. ASTM A263 and A264 — clad steel plate and bond shear requirements; ASTM A578 — ultrasonic examination of clad plate.
4. ASTM A800 — ferrite content in austenitic weld metal; ASTM E562 — point count method.
5. NACE MR0175 / ISO 15156 — materials for H2S-containing environments, hardness limits.
6. SME Mineral Processing Handbook — autoclave construction and materials practice. smenet.org
7. MetSoc Conference of Metallurgists (COM) proceedings — autoclave liner and overlay case studies.
8. USGS Mineral Commodity Summaries — copper and gold production context. usgs.gov

Changelog

• 2026-10-17 — First publication. Five mechanisms, specification parameter table, NDE matrix, seven-step repair, three anonymised failures.

{
“@context”: “https://schema.org”,
“@graph”: [
{
“@type”: “TechArticle”,
“headline”: “Weld Overlay Failure in Hydrometallurgical Vessels”,
“about”: [“weld overlay failure”, “cladding disbonding”, “autoclave materials”, “hydrometallurgy”],
“datePublished”: “2026-10-17”,
“dateModified”: “2026-10-17”,
“author”: {“@type”: “Person”, “name”: “Process Metallurgist, {BRAND}”},
“publisher”: {“@type”: “Organization”, “name”: “{BRAND}”}
},
{
“@type”: “FAQPage”,
“mainEntity”: [
{“@type”: “Question”, “name”: “How much iron in the first layer is acceptable?”, “acceptedAnswer”: {“@type”: “Answer”, “text”: “Five percent or less on the first layer and one percent or less on the finished surface for chloride service.”}},
{“@type”: “Question”, “name”: “Can ultrasonic testing find every weld overlay failure?”, “acceptedAnswer”: {“@type”: “Answer”, “text”: “No. A 100 mm grid finds most defects above roughly 25 mm across; tight separations can be missed, so UT is combined with a production test coupon.”}},
{“@type”: “Question”, “name”: “Why does overlay crack after PWHT rather than during welding?”, “acceptedAnswer”: {“@type”: “Answer”, “text”: “Because the mechanism is stress relaxation: strain concentrates at weakened grain boundaries in the coarse-grained heat-affected zone during the cycle.”}},
{“@type”: “Question”, “name”: “Is a weld overlay as good as clad plate?”, “acceptedAnswer”: {“@type”: “Answer”, “text”: “Corrosion resistance is equivalent when properly qualified, but roll-bonded clad plate has a more consistent, fully tested bond.”}},
{“@type”: “Question”, “name”: “What ferrite number should I specify for 316L overlay?”, “acceptedAnswer”: {“@type”: “Answer”, “text”: “3–10 FN, with the low end preferred above 300 °C where sigma formation is a risk.”}}
]
},
{
“@type”: “BreadcrumbList”,
“itemListElement”: [
{“@type”: “ListItem”, “position”: 1, “name”: “Home”, “item”: “/”},
{“@type”: “ListItem”, “position”: 2, “name”: “Resources”, “item”: “/resources/”},
{“@type”: “ListItem”, “position”: 3, “name”: “Weld Overlay Failure”, “item”: “/resources/weld-overlay-failure/”}
]
}
]
}

Related reading

• How Process Flowsheet is decided