Heat Input and Interpass Temperature in Duplex Welding: A Practical Field Guide

Duplex welding failures rarely come from bad welders. They come from parameters that were correct on the WPS and wrong at the arc. The welder followed the procedure sheet. The heat input number was within range. The interpass temperature was checked. And the weld still pitted in service, or shattered under a hammer tap at the next turnaround.

The reason is that duplex welding is governed by a thermal window, not a single parameter. Heat input sets the window. Interpass temperature determines how many times you pass through it. Both are measured with the same instruments you already have, and both are routinely recorded incorrectly.

This guide covers the practical numbers, the calculation, the measurement methods, and the record-keeping that separates a duplex WPS that passes audit from one that produces a weld that survives its service environment.

Why These Two Parameters Control Everything

Duplex stainless steel gets its properties from a roughly 50:50 ferrite-austenite microstructure. Welding destroys that balance locally. The weld pool solidifies as nearly 100% ferrite, and austenite only re-forms as the metal cools.

The cooling rate is set by heat input and interpass temperature:

• Cool too fast (low heat input, cold interpass): austenite has insufficient time to re-form. Ferrite stays high, chromium nitrides precipitate, toughness and corrosion resistance drop.

• Cool too slow (high heat input, hot interpass): the weld sits in the 700–950°C range long enough for sigma phase, chi phase, and secondary austenite to form. Sigma phase is brittle and chromium-depleted; it destroys both toughness and pitting resistance.

There is no “safe side” to err on. Both edges of the window are failure modes. A weld at 0.4 kJ/mm and a weld at 3.0 kJ/mm can both fail, for opposite reasons.

Heat Input: The Working Range

Heat input is the energy delivered per unit length of weld. It determines the cooling rate through the critical temperature range of 1200–800°C.

Recommended Ranges by Grade

For 2205, the broad range of 0.5–2.5 kJ/mm is widely cited. Some welding investigators recommend a narrower practical band of 0.5–1.5 kJ/mm for super duplex, with the root pass and small beads toward the low end of the range.

For 2507, the limits tighten considerably. Sandvik recommends 0.2–1.5 kJ/mm. Other sources recommend not exceeding 1.5 kJ/mm, with 1.6 kJ/mm as an absolute ceiling. The reason is the higher alloy content: more chromium and molybdenum means faster sigma-phase precipitation kinetics. What is safe for 2205 at 2.0 kJ/mm is already dangerous for 2507.

The Thickness Adjustment

The heat input is not a fixed number for a grade. It scales with wall thickness:

• Thin-wall tube (t = 1.5 mm): <0.5 kJ/mm is optimum. Fast cooling is less of a concern because the mass is small and the weld solidifies quickly regardless.

• Heavy wall (t ≥ 25 mm): Heat input closer to the upper end of the range. The concern shifts from excessive ferrite to insufficient fusion and lack of penetration.

For 2507 wall thicknesses ≥25 mm, Sandvik explicitly flags that “attention must be paid” — meaning the window is narrowest at the thick end because the thermal mass makes interpass temperature control hardest.

Interpass Temperature: The Limits That Actually Apply

Interpass temperature is the temperature of the weld zone measured immediately before depositing the next pass. It is a maximum limit, not a target. For duplex, preheat is generally unnecessary and often harmful. The only temperature control that matters is keeping the interpass low enough to avoid cumulative heat buildup.

The Numbers

NORSOK M-601 specifies 150°C as the mandatory maximum for 2205 duplex, and 100–150°C for 2507 super duplex, with some project specifications limiting to 100°C. The IMOA shop sheet states the maximum interpass temperature is limited to 150°C for lean and standard duplex, and to 100°C for super duplex.

Some technical standards apply an “extremely conservative” limit of 100°C for 2507 across all cases. The difference between 100°C and 150°C is not academic. At 150°C, the risk of sigma-phase formation in the HAZ during multi-pass welding is significantly higher than at 100°C, particularly in thick sections where the surface reading understates the internal temperature.

The Critical Measurement Trap

This is the most important practical point in this entire guide: interpass temperature is measured on the surface, but the failure occurs inside the weld metal.

Sandvik’s welding guidance states it directly: “As the interpass temperature is measured on the surface of the weld or on the base metal close to the weld, the actual temperature will be higher deeper inside the weld metal. This can cause embrittlement and low impact values in the root region”.

In a 25 mm thick 2507 spool welded with a 100°C surface interpass limit, the internal temperature near the root can be 130–150°C. If the WPS specifies 150°C maximum and the welder waits for the surface to drop to 149°C, the root may already be well beyond the safe sigma-phase range.

Field rule: For 2507 in thickness ≥20 mm, use 100°C as the surface maximum, not 150°C. The 50°C margin accounts for the internal-to-surface gradient. For 2205, 150°C surface is generally acceptable, but the same gradient logic applies in thick sections.

How to Calculate Heat Input Correctly

The standard formula for arc energy is:

Heat input (kJ/mm) = (V × A × 60) / (1000 × v)

Where V is arc voltage (volts), A is welding current (amperes), and v is travel speed (mm/min).

For thermal heat input (accounting for arc efficiency), multiply by η:

Thermal heat input = η × (V × A × 60) / (1000 × v)

Arc efficiency factors vary by process:

The distinction matters for qualification. ASME IX and most project specifications reference arc energy (η = 1.0) unless explicitly stated otherwise. NORSOK M-601 allows the qualified heat input to vary within ±15% of the average value used during qualification. If your procedure qualification was performed at 1.0 kJ/mm arc energy, the production range is 0.85–1.15 kJ/mm — not the broader published range for the grade.

A Worked Example

A GTAW root pass on 2205 tube: 120 A, 12 V, travel speed 100 mm/min.

Arc energy = (12 × 120 × 60) / (1000 × 100) = 0.864 kJ/mm

Thermal heat input (η = 0.7) = 0.605 kJ/mm

Both values fall within the 0.5–2.5 kJ/mm range for 2205. But note how the process selection shifts the effective number. A SMAW fill pass on the same joint at 140 A, 24 V, and 150 mm/min gives:

Arc energy = (24 × 140 × 60) / (1000 × 150) = 1.344 kJ/mm

The arc energy is higher despite the faster travel speed because the voltage and current are higher. The point is not the absolute number; it is that the calculation must use the actual parameters for each pass, not a single assumed value for the whole weld.

Measurement: What to Use and Where to Point It

Four instruments are in common use for interpass temperature:

The contact pyrometer is the workhorse. EN ISO 13916:2017 specifies pressing the probe against the metal surface adjacent to the weld zone, typically at a distance of 4× the material thickness from the weld edge. Measuring directly on the hot weld bead reads high; measuring too far away reads low.

The crayon is useful for a quick “is it still too hot” check but should not be the primary record value. Infrared readings on freshly welded duplex are unreliable because the oxidized surface has variable emissivity, and the reading is surface-only — which is exactly the wrong place given the internal-to-surface gradient.

For critical welds — 2507 in thickness ≥25 mm, or any weld in sour service — weld-attached thermocouples with data logging are the preferred method. They provide a continuous record that shows not just whether the maximum was exceeded, but how long the joint sat in the sigma-phase range.

Five Field Errors That Cause Failures

1. Using 2205 Limits on 2507

A fabricator welds 2507 with a WPS that lists 150°C interpass and 1.5 kJ/mm heat input — both within the 2205 range, both at or beyond the edge of the 2507 range. The weld passes visual and radiographic inspection. It fails impact testing, or pits in service at the weld line.

Fix: Maintain separate WPS ranges for each grade. 2507 interpass should be specified as 100°C maximum, not 150°C.

2. Measuring Interpass on the Cap Instead of the Root

The welder waits for the cap pass to cool to the specified limit, then starts the next pass. The root pass, which was deposited earlier and has been insulated by the fill passes, is still well above the limit. The root region develops sigma phase while the surface reads compliant.

Fix: For multi-pass welds in thickness ≥20 mm, require interpass measurement at the root or use embedded thermocouples. Surface measurement alone is insufficient.

3. Ignoring the ±15% Qualification Window

The WPS lists “heat input: 0.5–2.5 kJ/mm.” The welder runs the root at 2.3 kJ/mm because it is within the published range. But the procedure qualification was performed at an average of 1.2 kJ/mm, and NORSOK M-601 limits production variation to ±15% of that average — 1.02–1.38 kJ/mm. The 2.3 kJ/mm pass is outside the qualified window even though it is within the grade range.

Fix: Write the qualified range on the WPS, not the published grade range. The welder needs the number that was actually tested.

4. Cooling with Water or Compressed Air

A welder hits the interpass limit and cools the joint with a water spray or compressed air blast to speed up the next pass. The rapid quench produces a locally ferritic microstructure with chromium nitrides in the HAZ. The weld passes inspection and fails in service.

Fix: Cool in still air. If the schedule is tight, rotate between two joints so one cools while the other is welded. Never quench duplex.

5. Recording a Single Temperature for the Whole Weld

The weld map shows “interpass: 95°C” for all 12 passes. The number was measured once, after the root pass, and copied. The fill passes ran at 140°C or higher. The audit trail is false.

Fix: Record interpass temperature before every pass for critical welds, or at a defined frequency (e.g., every third pass) with a written note of the measurement points. Auditors under EN ISO 3834-2 cross-reference WPS values against production records; a copied number is a finding.

Verification: What to Test After the Weld

Heat input and interpass temperature are process controls. Their effect is verified by three tests:

Ferrite content: The weld should land at 30–70% ferrite as-welded, with 35–60% typical for the parent metal. A Feritscope reading outside this range indicates a heat input or interpass problem. But note the blind spot: sigma phase is non-magnetic, so a Feritscope reads a perfect number on a weld that is already embrittled.

Impact toughness: Charpy V-notch at the design minimum temperature. For 2507 in NORSOK service, this is typically -46°C with a 45 J average minimum. A sigma-affected weld will fail this test dramatically.

Corrosion testing: ASTM A923 Method C (ferric chloride) is the acceptance test for detrimental intermetallic phases in duplex welds. It is designed to demonstrate absence of sigma phase in a cost-effective way, with acceptance typically expressed as a maximum corrosion rate. If the weld passes A923 Method C, the heat input and interpass controls were effective.

The One-Page Field Summary

The WPS number is not the published grade range. It is the tested range, narrowed by the qualification, and further narrowed by the project specification. Write that number on the procedure, measure it at the root as well as the surface, and cool in still air. The rest is metallurgy that happens whether you account for it or not.