Welding Duplex Stainless Steel How to Maintain the 50/50 Austenite-Ferrite Balance

● 50/50 balance is NOT automatic - it requires precise control of heat input, interpass temperature, cooling rate, and filler metal chemistry. One wrong parameter destroys both corrosion resistance and toughness.

● Heat input window for standard duplex (2205): 0.5–2.5 kJ/mm. Below 0.5 = excessive ferrite (brittle). Above 2.5 = sigma phase (embrittled + corroded).

● Interpass temperature: ≤150°C for 2205, ≤100°C for super duplex (2507). Exceeding these limits triggers sigma phase precipitation within minutes.

● Filler metal MUST be over-alloyed - ER2209 for 2205 (Ni ~9% vs. base 5%), ER2594 for 2507. Never use 308L/316L filler.

● Standard PWHT (600–720°C) is FORBIDDEN - it creates sigma phase. Only solution annealing at 1040–1100°C + water quench is acceptable.

All Duplex Properties

Duplex stainless steels derive their name - and their engineering value - from a roughly equal mixture of austenite (γ) and ferrite (α) phases. This 50/50 microstructure is not cosmetic; it is the very mechanism that delivers the alloy's defining characteristic: the combination of austenitic corrosion resistance with ferritic strength, at roughly double the yield strength of standard 304L/316L stainless steels.

When duplex stainless steel solidifies from the molten weld pool, it does so AS 100% FERRITE. Austenite forms only during subsequent cooling, via a solid-state phase transformation in the temperature range of roughly 1350°C down to 800°C. How much austenite eventually forms - and whether harmful third phases (sigma, chi, nitrides) also appear - is entirely determined by three welding variables:

① Heat input (kJ/mm) - the total thermal energy deposited per unit length of weld.

② Cooling rate (°C/s) - governed by heat input, plate thickness, interpass temperature, and preheat.

③ Chemical composition - especially nickel and nitrogen, which stabilize austenite.

If austenite content falls below 25% (ferrite >75%), the joint becomes brittle with poor corrosion resistance. If the cooling rate is too slow, chromium-rich sigma phase (σ) precipitates at grain boundaries between 700°C and 950°C, consuming chromium and molybdenum from the adjacent matrix. The result: catastrophic loss of toughness AND pitting resistance - in a single thermal cycle.

The Core Metallurgical Principle

• Duplex weld solidification: Liquid → 100% Ferrite → Partial transformation to Austenite upon cooling.
• The goal: control cooling so that 30–70% austenite forms BEFORE the temperature drops below ~800°C.
• The enemy: time spent between 700–950°C - where sigma phase nucleates and grows within minutes.
• The solution: stay in the "Goldilocks zone" of heat input (not too hot, not too cold) and enforce strict interpass temperature limits.

Heat Input of Standard Duplex Steel

Heat input is the single most important welding variable for duplex stainless steel. It is calculated as:

Heat Input (kJ/mm) = (Voltage × Current × 60) / (Travel Speed × 1000)

The following table defines the acceptable heat input range for the three most common duplex grades. These numbers are not suggestions - they are metallurgical boundaries. Exceeding them results in immediate and irreversible damage to the welded joint.

Source: Industeel (ArcelorMittal), Duplex Stainless Steels Welding Guidelines, Rev. 2019; Sandvik Welding Handbook; AWS D10.18 Guide for Welding Duplex Stainless Steels.

The difference between 2205 and 2507 is significant: super duplex (2507, PREN ≥41) has a HIGHER alloy content (Mo ~4%, Ni ~7%), which makes it MORE sensitive to sigma phase formation. This is why its allowable heat input is capped at 1.5 kJ/mm - 40% lower than 2205's upper limit. Lean duplex (2304, PREN ~26) has less alloying, so its window is tighter on both ends.

Practical Implications for Welders

● Low heat input (<0.5 kJ/mm) = fast travel, low current → rapid cooling → ferrite stays, austenite doesn't form → Brittle joint.
● High heat input (>2.5 kJ/mm) = slow travel, high current → slow cooling → sigma phase precipitates → Embrittled + corroded.
● The Goldilocks Zone (0.5–2.5 for 2205) provides just enough cooling time for 30–70% austenite to form, without crossing into sigma territory.

Interpass Temperature of Duplex 2205

Managing interpass temperature is arguably MORE critical than heat input control, for one simple reason: sigma phase precipitation kinetics accelerate exponentially with temperature. At 850°C, sigma phase can form in under 2 minutes. At 700°C, it may take 15–30 minutes. A welder who continues depositing passes without allowing the joint to cool below the maximum interpass limit is effectively heat-treating the HAZ inside the sigma phase danger zone.

The interpass temperature limits defined by industry standards are as follows:

Source: AWS D10.18/D10.18M; NORSOK M-601:2016; API RP 582:2021; Industeel Welding Guidelines (ArcelorMittal, 2019).

For thin-wall sections (<6 mm), some specifications recommend even tighter limits (70–100°C) because the heat from subsequent passes accumulates more rapidly in thin material. The welder must measure temperature using a contact pyrometer or temperature-indicating crayon at the weld toe IMMEDIATELY before starting the next pass. Forced air cooling (but NEVER water quenching) may be used to accelerate interpass cooling.

The Consequences of Exceeding Interpass Limits

When interpass temperature exceeds 150°C for 2205 (or 100°C for 2507):

● Sigma phase (Cr-rich intermetallic) nucleates at ferrite / austenite boundaries within 2–5 minutes at 800–900°C.
● The adjacent matrix is depleted of chromium → PREN drops below threshold → pitting corrosion initiates at depleted zones.
● Charpy impact toughness drops from >100 J to <20 J at room temperature.
● The damage is cumulative: each hotter pass adds more sigma, more chromium depletion, and more embrittlement.

Filler Metal of Duplex Steel

Duplex filler metals are deliberately formulated with 2–4% MORE nickel than the base metal. This is not a manufacturing error - it is a metallurgical compensation strategy that addresses the unavoidable nitrogen loss that occurs during welding.

Source: AWS A5.9/A5.9M:2021 - Specification for Bare Stainless Steel Welding Electrodes and Rods; ASTM A240/A240M-22 for base metal chemistry.

Why the over-alloying? During welding, nitrogen - a potent austenite stabilizer - escapes from the molten weld pool. If the filler metal has the same nickel content as the base metal, the nitrogen loss will shift the phase balance toward excessive ferrite (>75%). By over-alloying nickel by 2–4%, the filler metal ensures that sufficient austenite forms during cooling, keeping the deposit in the 30–70 FN target range.

What You Must NEVER Do

● NEVER use ER308L, ER316L, or ER309L filler on duplex stainless steel. These grades contain ~10% Ni and ~20% Cr - designed for fully austenitic deposits. They will produce a ferritic (>80%) weld metal with PREN far below the base metal requirement.
● NEVER perform autogenous (no filler) GTAW on duplex stainless steel without adding nitrogen to the shielding gas. Without filler, the nitrogen loss is catastrophic - ferrite exceeds 85%.
● ALWAYS verify that the MTR (Material Test Report) for the filler metal lists the correct AWS classification - ER2209 for 2205, ER2594 for 2507.

Phase Balance Must Be Measured, Not Guessed

After welding, the ferrite content must be verified. The following standards define the acceptable range and the methods for measurement:

Source: ASTM A923-22; ASTM E562-19; ISO 17781:2017; WRC-1992 Constitution Diagram for Stainless Steel Weld Metal (Kotecki & Siewert).

Key Measurement Point

Measure ferrite on EACH pass - root, fill, and cap - not just the cap. The root pass typically has the highest ferrite content because it cools fastest (thinnest section, highest heat sink to the base metal). If the root pass exceeds 75 FN, the joint will fail Charpy impact testing regardless of acceptable cap pass readings.

Post-Weld Heat Treatment (PWHT) of Duplex Stainless Steel

This is the most counterintuitive rule for welders familiar with carbon steel and austenitic stainless steel fabrication. With carbon steel, PWHT at 600–720°C relieves residual stress and tempers the HAZ. With duplex stainless steel, the same temperature range precipitates sigma phase and destroys the joint.

Source: AWS D10.18/D10.18M:2020 - Guide for Welding Ferritic/Austenitic Duplex Stainless Steel Piping and Tubing; API RP 582:2021 - Welding Guidelines for the Chemical, Oil, and Gas Industries.

This Rule Has No Exceptions

If a code-mandated PWHT is required for the assembly (for example, due to carbon steel components in a dissimilar metal joint), the duplex stainless steel section MUST be protected from the PWHT cycle - either by designing it as a removable spool piece or by conducting the PWHT BEFORE the duplex component is installed. There is no welding procedure that makes duplex stainless steel immune to sigma phase at 600–720°C.

Grade-by-Grade Comparison

The following table provides a comprehensive one-page reference for welding engineers. Use it as your go/no-go checklist before striking an arc on any duplex grade.

Source: AWS D10.18/D10.18M:2020; ASTM A240/A240M-22; Industeel Duplex Welding Guidelines; Sandvik Welding Handbook; NORSOK M-601.

Frequently Asked Questions

304L and 316L are fully austenitic - they remain austenite from solidification to room temperature. Their welding requirements are straightforward: control heat input to avoid hot cracking, interpass ≤175°C to prevent sensitization, and use matching or 308L/316L filler. Duplex stainless steel is fundamentally different: it solidifies as 100% ferrite and MUST be cooled through a specific temperature window (1350°C → 800°C) to form the correct 30–70% austenite. Using 304L-level heat input on duplex will produce a fully ferritic, brittle weld with no corrosion resistance.

Q2: What happens if my ferrite reading is 85 FN instead of 30–70?

A reading of 85 FN means your weld metal is approximately 80–85% ferrite by volume - far above the acceptable limit. The consequences: (1) Charpy impact toughness drops below 27 J at room temperature - the joint is brittle. (2) The absence of austenite means there are no sites for nitrogen solubility, so chromium nitrides precipitate within the ferrite grains, destroying pitting resistance. (3) The weld is susceptible to hydrogen-induced cracking. The root cause is almost always one of: heat input too low, travel speed too fast, pure argon shielding gas (no N₂), or no filler metal added.

Q3: Can I weld duplex to carbon steel?

Yes - but with strict procedures. Use ER2209 filler metal (NOT ER309L, which is standard for carbon-to-stainless welds). The higher nickel in ER2209 compensates for carbon steel dilution and ensures sufficient austenite in the diluted zone. Control heat input per duplex limits (not carbon steel limits). The most critical consideration: if any post-weld heat treatment is required for the carbon steel portion, the duplex side WILL be destroyed by sigma phase unless the duplex component is a removable spool piece that can be installed AFTER PWHT.

Q4: How do I know if sigma phase is already present in a weld?

Sigma phase is invisible to the naked eye. Detection requires: (1) Metallographic examination per ASTM A923 Method A - NaOH electrolytic etch reveals sigma as a darkened phase at ferrite/austenite boundaries at 400–500× magnification. (2) Charpy impact testing per ASTM A923 Method B - if Charpy values are below 54 J, sigma is likely present. (3) Ferric chloride corrosion testing per ASTM A923 Method C - if pitting occurs within 24 hours of immersion, chromium depletion from sigma phase is confirmed. A portable ferritescope CANNOT detect sigma phase - it only reads ferrite content.

Q5: Is solution annealing a practical option for completed weldments?

Rarely. Solution annealing requires heating the ENTIRE component to 1040–1100°C and then rapidly quenching in water. For large fabrications (pressure vessels, FGD absorbers, pipe spools over 2 meters), this is physically impossible without a furnace and quench tank sized for the component. Distortion during quenching is severe. For this reason, the industry standard is to design the welding procedure to PRODUCE acceptable properties as-welded - no PWHT. Solution annealing is reserved for mill products (plate, pipe) or small components that can be furnace-treated.