Welding Austenitic Stainless Steel: Sensitisation, Hot Cracking, and Filler Selection

Austenitic stainless steels — the 3xx series — dominate corrosion-critical fabrication in chemical processing, food production, pharmaceutical manufacturing, and nuclear plant because of their excellent corrosion resistance, ductility, and toughness down to cryogenic temperatures. Yet they impose metallurgical penalties that are absent in carbon steel welding: a narrow window of acceptable delta ferrite content, a real risk of sensitisation in the heat-affected zone (HAZ), and a solidification morphology that controls susceptibility to hot cracking. This article provides a rigorous technical treatment of all three mechanisms — and the filler metal selection logic that manages them — so you can make defensible engineering decisions for 304L, 316L, 321, 347, and related grades.

1. Austenitic Stainless Steel — Grades and Metallurgical Basis

Austenitic stainless steels derive their microstructure from sufficient nickel (and manganese, nitrogen) additions to stabilise the face-centred cubic (FCC) austenite phase at room temperature. The basic composition is 16–26% Cr, 6–22% Ni, with C typically below 0.08 wt%. Chromium provides the passive film, nickel stabilises austenite, molybdenum (in 316/317 series) enhances pitting resistance, and nitrogen (in 316N, 304N) increases strength without compromising toughness.

The most widely welded grades — 304L and 316L — dominate fabrication because their ultra-low carbon content (C ≤ 0.03%) eliminates sensitisation risk in most single-pass weld geometries. Understanding why requires examining the chromium carbide precipitation kinetics in detail.

1.1 Common Grades and Nominal Compositions

2. Sensitisation — Mechanism, Kinetics, and Prevention

Sensitisation is perhaps the most practically critical metallurgical phenomenon in austenitic stainless steel welding. The term describes the precipitation of chromium-rich carbides — primarily Cr₂₃C₆ — at austenite grain boundaries when the material is exposed to temperatures in the 425–850°C range. Because carbon diffuses two to three orders of magnitude faster than chromium in austenite at these temperatures, the carbide precipitate draws carbon from the adjacent matrix far faster than chromium can diffuse to replenish it. A chromium-depleted zone forms on both sides of the grain boundary, with local Cr content falling below approximately 12 wt% — the minimum for passive film stability.

2.1 The C-Curve Kinetics of Sensitisation

The time-temperature-sensitisation (TTS) diagram for a given grade has a characteristic C-curve shape analogous to a TTT diagram. The “nose” of the C-curve for standard 304 (C ~0.06%) falls around 650–675°C, where sensitisation can begin within minutes. At lower temperatures, diffusion rates drop and sensitisation requires hours to days. Above 850°C, the carbide solvus is exceeded and any precipitates dissolve.

The welding HAZ traverses this temperature range rapidly on heating and more slowly on cooling. The critical zone is the region that dwells longest in the 500–750°C window during interpass cooling. For multi-pass welds, successive thermal cycles accumulate sensitisation exposure incrementally — this is why interpass temperature control below 150°C is specified in codes such as AWS D1.6 and EN 1011-3.

2.2 Effect of Carbon Content

Carbon content is the dominant material variable controlling sensitisation kinetics. Lowering C from 0.08% (standard 304) to 0.03% (304L) shifts the TTS nose dramatically to longer times — a factor of 10–100 at equivalent temperature. For most single-pass welds and even many multi-pass welds on 304L and 316L, the total thermal exposure below 850°C is insufficient to cause measurable sensitisation, provided interpass temperature is controlled.

2.3 Stabilised Grades — 321 and 347

The stabilised grades use titanium (321) or niobium (347) additions that preferentially form carbides (TiC, NbC) at higher temperatures, leaving insufficient free carbon available to form Cr₂₃C₆. The minimum addition is Ti ≥ 5×C and Nb ≥ 8×C (by weight). These grades are specified for service above 400°C where the low-carbon grades would eventually re-sensitise through long-term diffusion.

2.4 Prevention and Remediation

• Material selection: Specify 304L, 316L, or stabilised grades (321, 347) as appropriate for the service temperature range.
• Heat input control: Lower heat input reduces HAZ width and time spent in the sensitisation range. Use GTAW or short-circuit GMAW on thin sections.
• Interpass temperature: Maintain below 150°C (maximum) or 100°C for critical corrosion service per EN 1011-3.
• Post-weld solution annealing: 1050–1100°C, hold 1 h/25 mm thickness, water quench — dissolves Cr₂₃C₆ and re-homogenises Cr. Only practical for smaller components or where distortion is acceptable.
• Stabilising anneal for 321/347: 870–900°C, 2 h, air cool — re-precipitates TiC/NbC without sensitising.

3. Hot Cracking — Solidification and Liquation Mechanisms

Hot cracking is unique to the weld metal and immediately adjacent HAZ regions and occurs while the metal is still partially liquid during solidification or on reheating of a previously deposited bead. It is the second major metallurgical risk in austenitic stainless steel welding, and its control is intimately linked to the delta ferrite content of the weld deposit.

3.1 Solidification Cracking

Austenitic stainless steel weld metal solidifies over a temperature range, producing a mushy zone of dendrites and interdendritic liquid. The final solidifying liquid is enriched in sulphur, phosphorus, and silicon — elements with large solid-liquid partition coefficients in austenite. This liquid concentrates between dendrite boundaries and grain boundaries as a thin, continuous film. When weld metal contraction stresses are applied before this film solidifies, the film tears open — solidification cracking (also called hot cracking or centerline cracking).

The crack-susceptibility index (CSI) defined by Matsuda relates to the combined S + P content and the liquid film area. Modern austenitic grades are specified with low S (≤ 0.03%) and P (≤ 0.045%) partly for this reason, but the key metallurgical lever is solidification mode.

3.2 Solidification Modes and the Role of Delta Ferrite

Four solidification modes are recognised for austenitic-ferritic weld metals, depending on the Cr_eq/Ni_eq ratio (defined by the WRC-1992 or Schaeffler approach):

In the FA solidification mode, primary delta ferrite solidifies first, with its much higher solubility for S and P. As austenite nucleates and grows at the expense of ferrite on cooling, the harmful impurities partition preferentially into the residual ferrite and ferrite-austenite interfaces rather than concentrating in a continuous liquid film. The film is thus disrupted both compositionally and morphologically, dramatically reducing hot cracking susceptibility.

3.3 Target Ferrite Number Range

The Ferrite Number (FN) — measured magnetically by instruments calibrated to AWS A4.2 or ISO 8249 — is the accepted practical measure of delta ferrite content in weld metal. The standard engineering target for austenitic stainless steel weld deposits is FN 3–8:

• Below FN 3: Insufficient ferrite to disrupt the hot-cracking film; solidification cracking risk increases sharply.
• FN 3–8: Optimum range — crack-resistant, adequate corrosion resistance, negligible embrittlement risk at ambient service temperatures.
• Above FN 8: Increasing risk of sigma-phase formation (>475°C service), 475°C embrittlement, and magnetic response that can affect inspection or applications requiring non-magnetic welds.
• FN 10–15: Specified for some high-dilution overlay applications; accept higher sigma risk in exchange for hot-cracking protection.

3.4 Liquation Cracking in the HAZ

Liquation cracking occurs in the HAZ immediately adjacent to the fusion line, where peak temperatures are high enough to melt grain boundary films enriched in low-melting-point phases — NbC eutectic, phosphide/sulphide films, or sigma + austenite eutectic in over-aged material. Unlike solidification cracking, it is a base-metal phenomenon not preventable by filler metal selection alone. Keeping sulphur and phosphorus at the lower end of specification, controlling heat input, and using grades with clean grain boundaries mitigates liquation cracking risk.

4. The Schaeffler and WRC-1992 Diagrams

Weld metal microstructure prediction uses constitutional diagrams based on compositional equivalency factors — summing the austenite-stabilising and ferrite-stabilising elements into nickel-equivalent and chromium-equivalent axes respectively.

4.1 Schaeffler Diagram (1949)

Cr_eq (Schaeffler) = %Cr + %Mo + 1.5×%Si + 0.5×%Nb
Ni_eq (Schaeffler) = %Ni + 30×%C + 0.5×%Mn

Microstructure regions (approximate):
  Cr_eq/Ni_eq < 1.25   → Austenite (A) — fully austenitic
  Cr_eq/Ni_eq 1.25–1.48 → Austenite + Ferrite (AF mode)
  Cr_eq/Ni_eq 1.48–1.95 → Ferrite + Austenite (FA mode, preferred)
  Cr_eq/Ni_eq > 1.95   → Ferrite (F) dominated

The Schaeffler diagram was developed from experimental data on weld deposits using manual electrodes available in 1949. It does not include nitrogen as a variable (a significant austenite stabiliser), and its carbon coefficient (30) overestimates carbon’s stabilising effect at low carbon levels. Its predictions for modern L-grade and N-bearing steels can be significantly in error.

4.2 WRC-1992 Diagram

Cr_eq (WRC-1992) = %Cr + %Mo + 0.7×%Nb
Ni_eq (WRC-1992) = %Ni + 35×%C + 20×%N + 0.25×%Cu

Advantages over Schaeffler:
  - Nitrogen explicitly included (coefficient 20) — critical for duplex and N-bearing austenitics
  - Copper included (coefficient 0.25) — relevant for 904L and Cu-bearing grades
  - Ferrite Number (FN) isolines directly readable — no conversion needed
  - More accurate for low-C, N-bearing modern grades

The WRC-1992 diagram is the preferred tool for filler metal qualification and welding procedure specification in chemical plant and process piping. ASME Section II Part C, AWS A5.4, and AWS A5.9 classification testing all use FN measured by the WRC-1992 method. For most 304L and 316L welding with matching or near-matching fillers, the target FN 3–8 is reliably achievable.

Dilution Effects

When welding austenitic stainless steel to carbon steel or low-alloy steel — or when depositing the first overlay pass — base metal dilution can shift the deposit composition significantly toward the martensite field on the Schaeffler diagram. The dissimilar filler selection must account for this dilution (typically 20–40% for SMAW root passes) to ensure the diluted deposit remains in the austenitic or FA region. ER309L and ER309LMo are standard choices for this reason.

5. Filler Metal Selection for Austenitic Stainless Steel

Filler metal selection for austenitic stainless steel follows four principles: (1) match corrosion resistance to the base metal service environment; (2) achieve the target FN 3–8 in the as-deposited weld; (3) maintain low carbon where sensitisation is a risk; and (4) use an overalloyed transition buffer when dissimilar welding.

5.1 Matching Fillers

5.2 Dissimilar Welding — Austenitic SS to Carbon/Low-Alloy Steel

Dissimilar joints between austenitic stainless steel and carbon or low-alloy steel require the filler to accommodate compositional dilution from the carbon steel side without producing a martensite band at the fusion boundary — which would be hard, brittle, and prone to hydrogen-assisted cracking from post-weld conditions. ER309L achieves this because its composition (23% Cr, 13% Ni) on the Schaeffler diagram lies in the FA or AF region even when diluted 30–40% with carbon steel. The resulting deposit remains austenitic or at most austenite + ferrite.

Example: ER309L buttering on P265GH carbon steel (30% dilution by SMAW)

  ER309L undiluted:  Cr_eq ≈ 24,  Ni_eq ≈ 13.5
  P265GH dilutant:   Cr_eq ≈ 0.2, Ni_eq ≈ 0.4

  Mixed at 30% dilution:
    Cr_eq_mix = 0.70 × 24 + 0.30 × 0.2 = 16.86
    Ni_eq_mix = 0.70 × 13.5 + 0.30 × 0.4 = 9.57

  Cr_eq/Ni_eq = 16.86 / 9.57 = 1.76  → FA mode, FN ~8 (Schaeffler)
  ✓ No martensite; appropriate hot-cracking resistance

5.3 High-Alloy and Special Situations

• Fully austenitic deposits (FN 0): Specified for cryogenic service, non-magnetic applications, or some nuclear components. Use ERNiCr-3 (Alloy 82) or ER310, accepting higher hot cracking risk managed by process control (low heat input, stringent cleanliness).
• High-temperature 310S/330 applications: Fully austenitic fillers necessary. Control S + P at base metal level, use GTAW with precise heat input.
• Corrosion-resistant overlays: ER309L or ER312 for first layer (accounts for dilution), ER316L for second layer achieving specified corrosion resistance.
• 316L in pharmaceutical/semiconductor service: Specify low-sulphur ER316L (S ≤ 0.005%) and GTAW with 100% Ar shielding and backing gas to achieve mirror-finish internal bores and maximum corrosion resistance per ASME BPE standard.

6. Welding Process and Procedure Considerations

6.1 Process Selection

GTAW (TIG) is the reference process for root passes, thin-wall tubing, and corrosion-critical components because it offers the lowest heat input per unit volume deposited, no flux contamination, and precise control of shielding. For austenitic stainless, backing gas (100% Ar or 95%Ar/5%N₂) is mandatory on root passes to prevent oxide formation (sugaring) on the back surface, which would compromise corrosion resistance. Learn more in our guide to HAZ microstructure formation.

GMAW in pulse mode is commonly used for heavier-section pipe and vessel fabrication, providing higher deposition rate than GTAW with acceptable heat input control. Short-circuit transfer is appropriate for out-of-position work. SAW is used for automatic girth welding of large vessels. SMAW is used in field conditions and for repair welds. For all processes, shielding gas should be Ar or Ar+2%CO₂ — do not use CO₂-rich mixtures that can carburise the weld deposit and shift it toward the martensite field on the Schaeffler diagram.

6.2 Heat Input

Austenite has approximately 50% lower thermal conductivity than carbon steel, concentrating heat and producing wider HAZs for equivalent heat input. The heat input formula applies identically:

Where η is the thermal efficiency factor (GTAW: 0.60; SMAW: 0.80; GMAW/SAW: 0.85). For sensitisation control, lower heat input reduces the width and duration of the HAZ exposure in the 425–850°C range. EN 1011-3 recommends maximum heat input of 1.5 kJ/mm for most austenitic grades in corrosion service, with stricter limits for thin gauges.

6.3 Distortion Control

Austenitic stainless steel has a coefficient of thermal expansion of 16–17 × 10⁻⁶ /°C — approximately 50% higher than carbon steel — combined with lower thermal conductivity. These properties combine to produce significantly higher weld distortion than equivalent carbon steel joints. Balanced welding sequences, strong-back fixtures, and controlled back-step welding sequences are essential on fabrications where dimensional tolerance is critical.

6.4 Cleanliness and Contamination Prevention

Iron contamination of the stainless steel surface from carbon-steel wire brushes, grinding wheels, or fixturing causes local corrosion (surface rusting) and can shift local composition on the Schaeffler diagram toward the martensite field. Dedicated stainless-specific tools, stainless brushes, and segregated storage are mandatory good practice on any quality stainless steel fabrication shop. Chloride contamination (from perspiration, cutting fluids, marking inks) must also be eliminated — chlorides attack the passive film and nucleate pitting and stress-corrosion cracking (SCC) in service.

7. Post-Weld Treatment and Inspection

7.1 Pickling and Passivation

Post-weld pickling (typically HNO₃ + HF acid mixture) removes the heat tint (chromium-depleted oxide layer) formed during welding. Passivation (20–25% HNO₃, or citric acid per ASTM A967) re-establishes the passive chromium oxide film. Both treatments are specified in ASTM A380 and are mandatory for pharmaceutical, food-contact, and aggressive chemical environments. Electropolishing further enhances the passive film quality and is standard in ASME BPE applications.

7.2 NDT Considerations

Austenitic welds present specific challenges for non-destructive testing. The coarse, columnar austenite grain structure causes strong anisotropic ultrasonic scattering, making conventional UT unreliable for volumetric inspection. Phased array UT with low-frequency transducers (1–2 MHz) and TOFD techniques, combined with radiographic testing (RT), are the preferred methods. Penetrant testing (PT) is effective for surface-breaking defects. Ferrite content verification by calibrated magnetic instruments per AWS A4.2 / ISO 8249 is performed on procedure qualification coupons to verify FN target. Shubham’s readers interested in inspection qualification may also find our article on hydrogen-induced cracking in low-alloy steel welds useful for contrast.

8. Industrial Applications and Standards

Austenitic stainless steel welding is governed by a matrix of material, welding, and inspection standards depending on industry sector:

For deeper understanding of corrosion mechanisms that drive material selection in these sectors, see our articles on corrosion mechanisms and pitting corrosion in stainless steel. The iron-carbon phase diagram provides foundational context for understanding how austenite stability is established by composition. Our coverage of grain boundary segregation is directly relevant to understanding why sensitisation initiates at boundaries rather than within grains. For broader weld testing knowledge, see Charpy impact testing and hardness testing methods.

Recommended References