Liquation Cracking — Mechanism, Susceptible Alloys, and Prevention

Liquation cracking is a form of hot cracking that originates in the heat-affected zone (HAZ) and in reheated weld metal during welding. Unlike solidification cracking — which forms within the weld pool as it freezes — liquation cracking occurs in material that was never fully melted: localised grain boundary films liquefy at temperatures well below the bulk solidus, and those films tear open under weld shrinkage strains before they can resolidify. The result is a family of short, intergranular, oxidised cracks that sit immediately outside the fusion boundary, often invisible to surface inspection and capable of acting as fatigue crack initiation sites in service.

For fabricators working with nickel-base superalloys, austenitic stainless steels, aluminium alloys, and certain precipitation-hardening steels, liquation cracking is a persistent engineering challenge. Understanding the thermodynamic and kinetic conditions that create the liquid films — and the mechanical conditions that open them into cracks — is essential for writing effective welding procedure specifications (WPS) and selecting appropriate base material conditions, filler metals, and heat treatment sequences. This guide covers the full picture: mechanism, susceptible alloy systems, design considerations, and the engineering controls that reliably reduce cracking risk.

Liquation cracking is often discussed alongside reheat cracking in Cr-Mo steels, but the two are mechanistically distinct and require different prevention strategies. Where reheat cracking is driven by precipitation hardening during PWHT, liquation cracking is driven by constitutional or segregation-induced melting during the weld thermal cycle itself — a difference that matters enormously when writing the correct remedial action.

Hot Cracking in Context — Where Liquation Cracking Fits

Hot cracking is the collective term for weld discontinuities that form at elevated temperatures while metal is either liquid, partially liquid, or in a very-low-ductility solid state. Three distinct mechanisms fall under this umbrella, each with a different location and driving force:

Recognising which type of hot cracking is present in a given failure is the critical first step in root cause analysis. Liquation cracking is identified by its location (PMZ, within 0.5–3 mm of the fusion line), its intergranular morphology, and the presence of resolidified material or oxidation within the crack. The companion guide on the WeldFabWorld joint types and weld joint geometry reference is useful context for understanding how joint configuration influences the location and severity of the tensile strains that open liquation cracks.

The Liquation Mechanism — How Grain Boundary Films Form

Two distinct routes produce the intergranular liquid films responsible for liquation cracking. Both can operate simultaneously in the same weld, and in susceptible alloys such as Inconel 718 they frequently do.

Route 1 — Constitutional Liquation

Constitutional liquation is the formation of a localised liquid phase at a temperature below the bulk alloy solidus, caused by a second-phase particle reacting with its adjacent matrix before equilibrium dissolution can occur. The mechanism was first described systematically by Pepe and Savage in 1967 for austenitic stainless steels containing NbC and TiC particles, and it remains the dominant route in nickel superalloys.

The sequence is as follows:

1. A coarse second-phase particle (NbC, TiC, M₂₃C₆, boride, or intermetallic) exists in the base material before welding.
2. During rapid heating of the HAZ, the particle begins to dissolve, but the heating rate is too fast to allow equilibrium dissolution — the particle and the adjacent matrix form a localised binary or pseudo-binary system.
3. At the particle-matrix interface, the local composition enters a two-phase (solid + liquid) or eutectic field at a temperature significantly below the bulk alloy solidus.
4. A thin liquid film nucleates at the interface and spreads outward along the adjacent grain boundary because the liquid-boundary surface energy is lower than the solid-boundary energy (i.e., the liquid wets the boundary).
5. On subsequent cooling, the thin film may resolidify as a eutectic microstructure — but if tensile strains from weld shrinkage exceed the liquid film’s capacity before resolidification, a crack opens.

Route 2 — Segregation Liquation

Segregation liquation does not require pre-existing particles. Instead, it results from the equilibrium (or near-equilibrium) segregation of solute elements to grain boundaries during prior thermal processing — casting, hot working, or previous heat treatment. Elements with strong grain boundary partitioning (P, S, B, Si in steels; B, Zr, S in nickel alloys; Mg, Si in aluminium alloys) concentrate at boundaries to levels far above their bulk concentration. When the weld HAZ is heated into the partially melted zone, these boundary-enriched regions reach their local liquidus temperature before the grain interior does, producing liquid films at the boundaries without requiring any second-phase particle to initiate the process.

The Role of Grain Boundary Wetting

Whether the liquid produced by either route forms a continuous, cracking-susceptible film depends on the dihedral angle. When the liquid at a grain boundary groove has a low dihedral angle (approaching zero degrees), it forms a continuous thin film that separates the two adjoining grains completely. This is the most dangerous configuration. When the dihedral angle is large (greater than 60 degrees), the liquid forms isolated lenses at triple junctions and does not compromise grain boundary cohesion along its length.

The dihedral angle is controlled by the ratio of grain boundary energy to solid-liquid interfacial energy. For nickel alloys, boron additions dramatically lower the dihedral angle of Ni-B eutectic liquids at austenite boundaries, which is why even trace boron concentrations (above ~100 ppm in base metal or filler) dramatically increase liquation cracking susceptibility. For austenitic stainless steels, P, S, and Si segregation at boundaries reduces the dihedral angle of the Fe-rich liquid.

The Partially Melted Zone (PMZ)

The partially melted zone is the thin band of base metal immediately adjacent to the fusion boundary that is heated above the local liquidus of specific grain boundary regions but below the bulk liquidus. It is the anatomical location of nearly all liquation cracking. Understanding the PMZ’s microstructural character helps both in NDE strategy and in root cause analysis.

PMZ Formation and Extent

The PMZ exists because the weld thermal cycle imposes a steep temperature gradient across the fusion boundary. Within a distance of typically 0.5–3 mm of the fusion line, the peak temperature passes through the regime where:

• Bulk liquidus temperature is not reached (so the grains themselves do not fully melt)
• Local boundary or particle-induced liquidus temperature is exceeded (so liquid films form at boundaries)

The width of the PMZ increases with heat input — higher heat input creates a broader temperature gradient and a wider band where the temperature lies in this partial-melting range. Grain boundary orientation relative to the isotherms also matters: boundaries approximately parallel to the isotherms are wetted over their full length, while boundaries perpendicular to the isotherms may only be wetted over a narrow portion.

PMZ Microstructural Features

On metallographic cross-section, the PMZ in nickel superalloys is characterised by:

• A zone of coarse, irregularly shaped grains immediately inside the fusion boundary
• Intergranular films of resolidified eutectic (often appearing as bright phases in backscattered electron SEM imaging)
• Cavities or cracks at triple junctions where liquid concentrated
• A compositional gradient measurable by EDS or EPMA, showing enrichment of Nb, Ti, Si near the boundaries
• In severe cases, open cracks following the grain boundary network, often with oxidised crack surfaces (from exposure to the weld atmosphere during the thermal cycle)

Susceptible Alloy Systems

Liquation cracking susceptibility is not evenly distributed across alloy families. The key variables that determine susceptibility are: the presence and type of grain boundary segregants or second-phase particles, the temperature range between the local boundary solidus and the bulk solidus (the “liquation window”), and the alloy’s resistance to tensile strain in the partially liquid state (its “ductility” in the PMZ, often characterised by the Nil Ductility Temperature range).

Nickel-Base Superalloys — Highest Susceptibility

Nickel-base superalloys are the most susceptible alloy class, and welding them demands the most rigorous controls. The susceptibility arises from the combination of a wide solidification range, multiple low-melting second phases, and high restraint in aerospace and power turbine applications.

For IN718 specifically, the niobium content (typically 4.75–5.50 wt%) is the primary driver. NbC particles, particularly those with diameters above 5 microns, are potent constitutional liquation sites. The delta phase (Ni₃Nb), which forms in aged or long-service material at grain boundaries, is also susceptible to liquation. This is why the base metal condition at the time of welding — solution-annealed versus peak-aged — is one of the most influential variables in IN718 weldability. The GTAW process with its precise, low heat input is strongly preferred for IN718 and similar superalloy root passes.

Austenitic Stainless Steels

Fully austenitic stainless steels (310, 330, high-alloy nitrogen grades) are more susceptible than duplex or ferritic grades because austenite solidification concentrates impurities at grain boundaries. The primary liquation agents in austenitic stainless are P, S, and Si segregation at boundaries, which produce low-melting Fe-P, Fe-S, or Fe-Si-Ni liquid films. Delta ferrite, when present in dual-phase solidification, tends to interrupt these liquid films and improve resistance — which is why the delta ferrite content in stainless weld metal is carefully controlled. See the dedicated guide to delta ferrite importance in austenitic stainless welds for more on this point.

Grade 347 stainless (Nb-stabilised) also shows constitutional liquation of NbC particles similar to the mechanism in IN718, particularly in heavy section fabrications where particles are coarser. Grade 316 and 304, which solidify with a ferritic-austenitic (FA) mode, are much more resistant because the delta ferrite provides a skeletal network that interrupts boundary liquid films. Understanding sensitisation and weld decay in stainless steels provides complementary context on the role of grain boundary carbides in stainless weld metallurgy.

Aluminium Alloys

Aluminium alloys of the 6xxx (Al-Mg-Si) and 7xxx (Al-Zn-Mg-Cu) series exhibit liquation cracking through both constitutional liquation of Mg₂Si, MgZn₂, and CuAl₂ particles, and through boundary segregation of low-melting elements. The solidification range is large (50–150°C for many 7xxx alloys), and the liquid films have very low surface tension, promoting effective grain boundary wetting. In aluminium, heat input control is especially critical because the high thermal conductivity of Al reduces the natural self-quenching that limits PMZ extent in steels.

Low-Alloy and Carbon Steels — Generally Resistant

Plain carbon and low-alloy steels are generally resistant to liquation cracking because they contain few second-phase particles stable at near-solidus temperatures, they solidify over a narrow range, and the ferritic solidification mode provides some resistance to boundary wetting. Exceptions include steels with significant residual S or P (above 0.03 wt%) and rephosphorised free-machining steels (12L14, etc.) where Fe-P or Fe-FeS eutectic can produce grain boundary liquation. For high-strength low-alloy (HSLA) steels with boron additions, boride-induced constitutional liquation has been documented at very high boron concentrations. The carbon equivalent calculation is the primary preheat and cracking risk screening tool for these steels.

Multi-Pass Weld Liquation Cracking

An important and sometimes overlooked form of liquation cracking occurs not in the base metal HAZ but in the previously deposited weld metal that is reheated by subsequent passes. This is called weld metal liquation cracking or inter-run liquation cracking, and it is the predominant form in multi-pass heavy-section nickel alloy fabrications.

Mechanism in Reheated Weld Metal

During solidification of a weld pass, solute elements (Nb, Ti, Si in nickel alloys) segregate to the interdendritic regions and solidify last as low-melting eutectic phases. When the next weld pass deposits, the previously deposited weld metal is reheated and these interdendritic eutectics — which have a lower melting point than the dendrite cores — reliquefy. This creates a situation analogous to constitutional liquation in the base metal HAZ: liquid films form in the reheated weld metal, and shrinkage strains from the new pass open cracks along the prior solidification structure.

Bead Sequence and Overlap

The geometric arrangement of successive beads significantly influences multi-pass liquation cracking risk. Wide, flat bead profiles create large areas of reheated weld metal where interdendritic eutectics can reliquefy. Narrow, convex bead profiles produced by GTAW or orbital welding concentrate reheating on the weld bead centreline rather than the bead edges, reducing the area of reheated inter-run material. For the Submerged Arc Welding (SAW) process, which typically produces wide, flat bead profiles and high heat input, multi-pass liquation cracking in nickel alloy cladding applications is a particular concern that must be addressed in WPS development.

Liquation Cracking vs Solidification Cracking — Practical Distinctions

In failure analysis, distinguishing liquation cracking from solidification cracking is essential because the root cause and corrective action differ substantially. The following comparison covers the key diagnostic criteria:

Prevention Strategies

Effective prevention of liquation cracking requires addressing the three components of the mechanism: the source of the liquid (the low-melting constituent), the grain boundary wetting (the dihedral angle), and the tensile strain (from restraint and weld shrinkage). The most robust prevention strategies target all three simultaneously.

1. Pre-Weld Homogenisation Heat Treatment

For nickel superalloys where constitutional liquation of coarse NbC or other carbides is the dominant mechanism, a homogenisation anneal before welding dissolves these particles and redistributes the solute elements into uniform solid solution. This eliminates the localised eutectic-forming compositions that drive constitutional liquation. Standard homogenisation treatments for IN718 involve heating to 1065–1080°C for 1–2 hours, followed by water quench or rapid air cool to prevent reprecipitation on cooling.

2. Heat Input Control

Reducing heat input narrows the PMZ, reduces the time at high temperature (limiting the extent of grain boundary wetting), and reduces the total strain energy available to open liquid films. For IN718 GTAW, heat inputs are typically limited to 0.8–1.5 kJ/mm. For MIG/GMAW on austenitic stainless in thick sections, a maximum heat input of 2.5 kJ/mm is commonly specified. Use the MIG welding settings calculator and TIG welding settings calculator on WeldFabWorld to verify heat input from your WPS parameters.

3. Filler Metal Selection — Chemistry Controls

Filler metal chemistry directly controls the weld metal’s susceptibility to inter-run liquation cracking and influences the PMZ wetting angle through the weld pool chemistry gradient. Key controls include:

• Boron: Restrict to below 50 ppm in filler metal for nickel alloy applications. Boron dramatically lowers the dihedral angle of liquid at austenite grain boundaries. Some specifications for critical nickel alloy welds restrict B to below 20 ppm in both base metal and filler.
• Sulphur and Phosphorus: Minimise both in filler and base metal for austenitic stainless. Many specifications limit S + P ≤ 0.010 wt% in filler for critical applications.
• Silicon: Si lowers the weld metal solidus and widens the solidification range; restrict to below 0.2 wt% where possible for nickel alloy weld metal.
• Niobium and Titanium: Essential for IN718 strength but drive liquation susceptibility. Use fillers where Nb is at the lower end of the specification range.

4. Welding in the Correct Base Metal Condition

As noted for IN718, always weld in the solution-annealed condition, not the aged condition. For austenitic stainless steels, welding in the solution-annealed (not cold-worked) condition reduces residual stresses and grain boundary segregation. For precipitation-hardening steels (15-5PH, 17-4PH), welding should be performed in the annealed condition with aging applied post-weld, because the aged microstructure contains fine coherent precipitates that reduce ductility in the PMZ.

5. Restraint Reduction

High joint restraint increases the tensile strain acting across liquid boundary films during weld cooling. Design measures to reduce restraint include:

• Welding free to shrink before tack-welding to rigid structures
• Allowing unrestricted root bead thermal contraction before applying subsequent passes
• Avoiding full joint restraint (both ends clamped) when possible
• Using pre-bending or pre-springing to allow the joint to move during welding in a direction that partially cancels thermal shrinkage strains
• In highly restrained assemblies, applying slight compressive pre-stress in the joining direction (stress-assisted welding)

The welding joint types guide on WeldFabWorld provides useful reference on how joint geometry and access influence residual stress and restraint.

6. Preheat and Interpass Temperature Management

Unlike for cold cracking, preheat for liquation cracking prevention is not about reducing hydrogen diffusion — it is about controlling the thermal gradient and hence the tensile strain rate. Moderate preheat (100–150°C for nickel alloys; the minimum specified for austenitic stainless is usually ambient, but 50–80°C in thick sections reduces thermal shock) reduces the severity of thermal gradients and hence the tensile strains that open liquid films. For the SMAW process, interpass temperature must be controlled both from below (minimum) and above (maximum) for superalloy applications.

7. Post-Weld Heat Treatment Sequencing

For IN718 and similar alloys, the PWHT sequence can either help or worsen the situation. If aging is applied immediately after welding without a solution anneal, the precipitates formed in the PMZ are fine and coherent, limiting ductility in a region that may already contain micro-crack nuclei. The preferred PWHT for IN718 critical welds is: solution anneal (985–1010°C / 1 hour / rapid cool) followed by a two-stage aging (720°C / 8 h + 620°C / 8 h). This sequence heals minor micro-cracks, dissolves any Laves phase formed in the PMZ, and produces the optimum precipitation state. The solution anneal temperature must not exceed 1020°C or grain growth in the PMZ can negate the benefit.

NDE and Detection of Liquation Cracks

Liquation crack detection is challenging because of the cracks’ small size, tight opening, and location in the acoustically noisy near-surface region adjacent to the fusion boundary. No single NDE method is universally sufficient. A combined approach is standard for high-integrity applications.

Code and Standard References

Several codes and standards address hot cracking susceptibility and its prevention, either directly or through the preheat, heat input, and filler metal controls that underpin prevention:

For code-specific weld procedure qualification requirements relevant to Cr-Mo and nickel alloy pressure equipment, the ASME Section IX qualification quiz and the P-Number, F-Number, and A-Number guide on WeldFabWorld provide practical reference material. The mechanical testing guide covers the qualification tests that demonstrate weld quality in hot-cracking-susceptible alloys.

Hot Cracking Test Methods

When developing welding procedures for susceptible alloys, quantitative hot cracking testing provides the data needed to rank materials and procedures before committing to full production WPS qualification. The three most widely used methods for liquation cracking susceptibility testing are:

Varestraint Test (and Trans-Varestraint)

The Varestraint test applies a controlled strain to a weld specimen during the weld thermal cycle by bending it over a die block of defined radius. The total crack length or maximum crack length on the HAZ surface is measured as a function of applied strain. The Transvarestraint variant applies the strain transversely to the weld direction, making it more sensitive to HAZ liquation cracks in the PMZ. Results are expressed as the threshold strain below which no cracking occurs (CST — Critical Strain for Cracking), which provides a direct comparison of susceptibility between material-procedure combinations. ISO 17641-3 covers the Trans-Varestraint method.

PVR Test (Programmierter Verformungs-Risstest)

The PVR (Programmed Deformation Cracking) test applies a continuously increasing tensile strain rate during welding and measures the critical deformation rate at which cracking initiates. It is particularly useful for ranking the relative susceptibility of filler metals and base metal conditions rather than for quantitative cracking threshold determination. The test is referenced in the German DVS standards and in EN ISO 17641-1.

Sigmajig Test

The Sigmajig test applies a fixed transverse stress to a specimen during autogenous welding and determines the threshold stress above which hot cracking occurs. It is simpler to execute than the Varestraint test and provides a direct stress-based susceptibility measure. It is less widely used for liquation cracking specifically but can differentiate solidification from liquation susceptibility when the weld pool geometry is carefully controlled.

Related WeldFabWorld Resources