Weld Solidification Cracking: Causes, Mechanisms and Prevention

Weld solidification cracking, also called hot cracking or centerline cracking, forms while the weld pool is still in its terminal stage of freezing, when shrinkage strain ruptures a thin film of segregated liquid trapped between growing dendrites. It is the most frequently encountered hot-cracking mode in fusion welding and accounts for a disproportionate share of weld rejections in austenitic stainless steel, nickel-base, and high-restraint carbon steel fabrication. This article works through the dendritic solidification mechanism, the compositional and process variables that control it, the standardized tests used to rank susceptibility, and the measures that prevent it in production welding.

What Is Weld Solidification Cracking?

Weld solidification cracking is a hot-cracking mechanism that develops in the fusion zone while the weld metal is completing solidification, typically at a temperature at or just below the equilibrium solidus. It is fundamentally different from cracking that occurs after the joint has fully cooled: the metal is still partly liquid when the crack initiates, and the crack surface itself often shows a dendritic, interdendritic morphology under the microscope rather than the faceted, transgranular fracture typical of cold cracking.

Macroscopically, solidification cracking most often appears as a straight crack running along the weld centerline, following the last-to-freeze plane where dendrites growing from opposite sides of the pool meet. A second common form is the crater crack, a star-shaped, radiating crack that forms at the point where the arc is extinguished and the weld pool solidifies inward from several directions at once. Both forms share the same underlying cause: a thin, mechanically weak liquid film that cannot accommodate the strain imposed by weld metal shrinkage.

Solidification cracking is one of several mechanisms grouped under the broader term hot cracking, alongside liquation cracking in the heat-affected zone and ductility-dip cracking in the solid state. These mechanisms are frequently confused with each other, and with the unrelated, sub-solidus mechanism of hydrogen-induced cold cracking covered in our guide to hydrogen-induced cracking; the comparison table later in this article distinguishes them by location, temperature regime, and driving mechanism.

Metallurgical Mechanism

Non-Equilibrium Solidification and Segregation

Fusion welds solidify far from equilibrium. As columnar dendrites advance into the weld pool, solute elements with an equilibrium partition coefficient below 1, principally sulfur, phosphorus, boron, and in some alloys niobium and silicon, are rejected ahead of the solid-liquid interface rather than incorporated into it. Because solidification in a weld pool happens in seconds rather than the hours available in a casting, there is no time for back-diffusion to homogenize this rejected solute. It accumulates instead in the shrinking pool of residual liquid between dendrite arms, exactly as described by Scheil-type non-equilibrium solidification behavior covered conceptually in our guide to the iron-carbon phase diagram.

By the time solidification is nearly complete, this terminal liquid can be enriched in impurity elements by an order of magnitude or more relative to the bulk composition. In carbon and low-alloy steel, the relevant reaction product is an iron-iron sulfide eutectic with a solidus near 988°C (1810°F), far below the bulk steel solidus. This film persists as a continuous, essentially strengthless layer along solidification boundaries long after the surrounding dendrites have interlocked into a coherent solid skeleton, which is exactly the condition needed for cracking.

Borland’s Generalized Theory of Solidification Cracking

The most widely cited framework for the mechanism is the generalized theory proposed by J.C. Borland in 1960, which divides the terminal stage of solidification into a sequence of structural states rather than treating solidification as a single event. In the earliest stage, primary dendrites form within a continuous liquid, and the liquid is still free to move and can heal any incipient separation. As solidification advances, the dendrites interlock into a coherent network and liquid mobility becomes restricted to narrow interdendritic channels, reducing the structure’s ability to feed shrinkage. In the most crack-susceptible stage, the residual liquid exists only as thin, largely isolated films decorating the boundaries between adjacent grains, with essentially no capacity to flow and backfill an opening separation. Once solidification is complete and solid bridges have formed across these boundaries, ductility recovers and the crack-susceptible window closes.

This sequence explains why solidification cracking is fundamentally a competition between two rates: the rate at which the structure transitions through this brittle window toward full solid bridging, and the rate at which shrinkage strain accumulates within it. Anything that prolongs the brittle window, widens it, or speeds up strain accumulation increases susceptibility; anything that shortens the window or slows strain accumulation reduces it.

The Brittle Temperature Range and Strain Theory

The temperature interval over which this near-zero ductility condition exists is called the brittle temperature range (BTR), also referred to as the solidification temperature range or, in N.N. Prokhorov’s earlier strain-based formulation, the high-temperature brittleness range. Prokhorov framed the problem quantitatively as a plasticity reserve, the product of the width of the BTR and a critical strain rate the structure can tolerate within it; cracking occurs once the strain rate actually imposed by weld shrinkage exceeds that reserve. This is why susceptibility correlates more closely with the size and ductility characteristics of the BTR than with the total freezing range of the alloy: two alloys can have similar liquidus-to-solidus spans yet very different cracking behavior if their BTR ductility recovers at different rates.

A = CST × BTR

A   = plasticity reserve within the brittle range
CST = critical strain rate the partially solid structure can tolerate (%/°C)
BTR = width of the brittle temperature range (°C)

Cracking initiates when the imposed shrinkage strain rate
exceeds the reserve A within the BTR.

Metallurgical and Compositional Factors

Sulfur and Phosphorus

Sulfur is the single most influential impurity element in carbon and low-alloy steel weld metal. Its solubility in solid iron is extremely low and its partition coefficient during solidification is far below 1, so it concentrates almost entirely into the terminal liquid. There it forms the iron-iron sulfide eutectic described above, with a solidus around 988°C, well below the temperature at which the surrounding steel dendrites have already interlocked into a load-bearing skeleton. Manganese counters this by reacting preferentially with sulfur to precipitate manganese sulfide instead of iron sulfide; manganese sulfide melts near 1600°C and, critically, tends to form as discrete, more rounded particles rather than a continuous grain-boundary film, so it does little to compromise hot ductility. This is the metallurgical basis for the manganese-to-sulfur ratio appearing as a control variable in nearly every solidification-cracking guideline.

Phosphorus behaves similarly to sulfur in that it segregates heavily and forms low-melting iron phosphide eutectics, and on a per-unit-weight basis it is generally considered at least as potent a crack promoter as sulfur. Unlike sulfur, however, phosphorus is not effectively neutralized by manganese, which is why most specifications hold phosphorus to a strict independent limit rather than relying on a compositional ratio.

Carbon and Niobium

In carbon and low-alloy steel weld metal, carbon increases susceptibility largely by widening the effective freezing range and increasing the volume fraction of terminal liquid available to segregate impurities into; this is reflected directly in the heavy weighting carbon receives in the quantitative susceptibility formulas discussed below. Niobium acts differently and more aggressively: even in small quantities it forms niobium carbide and, in nickel-base alloys, niobium-rich Laves-phase eutectic liquids that persist to unusually low temperatures, widening the brittle temperature range. This is a well-documented weldability concern in niobium-stabilized austenitic grades such as 347 stainless steel and in nickel-base superalloys such as Inconel 718, where niobium segregation is a primary driver of both solidification cracking and the related liquation cracking discussed in our guide to heat-affected zone microstructure.

Solidification Mode in Austenitic Stainless Steel

Austenitic stainless steel weld metal can solidify either with austenite forming directly from the liquid (primary austenite, or AP mode) or with ferrite forming first and austenite developing afterward by solid-state transformation (primary ferrite, or FA mode). The distinction matters enormously for cracking resistance. Austenite has very limited solid solubility for sulfur and phosphorus, so in primary austenite solidification the segregated liquid film persists as a continuous network right up to the point of complete solidification. When ferrite forms first instead, sulfur and phosphorus partition preferentially into it, and the resulting two-phase ferrite-austenite boundary structure breaks up the continuity of the liquid film. A small amount of retained delta ferrite, typically specified as a ferrite number in roughly the 3 to 10 FN range, is therefore one of the most effective and widely used tools for controlling solidification cracking in austenitic stainless steel weld metal. Whether a given composition solidifies in the FA or AP mode is predicted from the chromium and nickel equivalents on a constitution diagram such as the Schaeffler or WRC-1992 diagram, with elements such as chromium, molybdenum, and silicon promoting ferrite and elements such as nickel, manganese, carbon, and nitrogen promoting austenite.

Susceptible Alloy Systems

Solidification cracking susceptibility varies enormously across alloy families, and the dominant driver differs from one family to the next. The table below summarizes the principal mechanism and mitigation approach for the systems most commonly encountered in fabrication.

Welding Process and Joint Design Factors

Weld Pool Shape and Solidification Pattern

The plan-view shape of the weld pool strongly influences how dendrites converge at the centerline. A pool elongated into a teardrop shape, typical of fast travel speeds, forces dendrites growing from the two fusion boundaries to meet at a shallow, almost head-on angle directly along the centerline, concentrating the terminal liquid film into a single continuous plane. A more elliptical pool, typical of slower travel and higher heat input per unit length, allows dendrites to meet at a steeper angle across a less continuous boundary, which is generally more resistant to cracking at equivalent composition. This is one reason that simply slowing travel speed, without changing composition at all, is a common and effective field fix for intermittent centerline cracking.

Depth-to-Width Ratio and Restraint

Narrow, deep weld beads concentrate solidification and the accompanying shrinkage strain into a thin centerline plane, while wide, shallow beads spread the same shrinkage over a larger volume and a less sharply defined boundary. Production guidance commonly targets a bead width-to-depth ratio in the range of roughly 1.1 to 1.4 to minimize this effect; ratios below this range, often produced by high-current, low-travel-speed submerged arc or keyhole-mode laser and electron beam welds, are recognized risk factors for centerline cracking even in compositions that would otherwise screen as acceptable. Joint restraint compounds the effect: thick sections, short weld lengths between strong points, and tight fit-up all increase the strain imposed on the solidifying weld pool independent of bead geometry.

Process Comparison

Gas tungsten arc and gas metal arc welding generally allow good control over both heat input and bead shape, making them comparatively forgiving processes for crack-sensitive alloys when parameters are set deliberately. Submerged arc welding, valued for its high deposition rate and deep penetration, is more prone to producing the narrow, deep bead profile associated with higher risk unless travel speed and electrical parameters are specifically tuned to widen the pool. Laser and electron beam welding, with their characteristically narrow, deep keyhole-mode penetration profiles and very high solidification rates, are recognized as some of the more demanding processes for solidification-crack-sensitive stainless steel and nickel alloys, which is why hot-cracking susceptibility testing is routinely performed when qualifying these processes for such materials.

Identification and Testing

Visual and Non-Destructive Detection

Solidification cracks are usually straight, run parallel to the weld axis along the centerline, and may be surface-breaking or entirely subsurface. Crater cracks appear at arc-stop locations as a short, radiating, star-shaped pattern. Liquid penetrant and magnetic particle testing reliably find surface-breaking cracks; radiography and ultrasonic testing are needed for subsurface cracking, and radiographic interpretation of a fine centerline crack can be difficult when the crack is tight and oriented favorably to the beam. Weld procedure qualification under codes such as ASME Section IX, AWS D1.1, and API 1104 requires crack-free macro-etch sections and radiography, so a centerline solidification crack found during qualification or production inspection is cause for rejection regardless of the original joint design intent.

Standardized Susceptibility Tests

Several self-restraint and externally loaded test methods have been developed specifically to rank alloy and filler metal combinations for solidification cracking susceptibility before they are committed to production.

Varestraint and Trans-Varestraint Tests

The Varestraint (variable restraint) test bends a welded specimen over a die block of known radius at a controlled point during welding, imposing an augmented strain directly onto the still-solidifying weld pool. The total length of cracking produced, often reported as the maximum crack distance, is used to rank susceptibility across compositions under a known, reproducible strain. The Trans-Varestraint variant applies the augmenting strain transverse to the weld axis rather than longitudinally, which more closely represents the strain direction relevant to centerline cracking.

Sigmajig Test

The Sigmajig test pre-loads a thin test plate to a known transverse tensile stress before a single autogenous weld bead is run down its centerline. The threshold stress at which cracking first appears is reported as the susceptibility metric, making the test useful for ranking materials by a stress-based criterion rather than a strain-based one.

Houldcroft (Fishbone) Test

The Houldcroft test cuts a series of slots of progressively increasing length into one edge of a test plate, then runs a single weld bead along the plate from the slotted edge toward the solid end. Each slot progressively reduces local restraint, so the point at which cracking stops as the slots lengthen, or the total cracked length summed across all slots, gives a simple, low-cost relative ranking of susceptibility that does not require specialized loading equipment.

Solidification Cracking Compared with Other Weld Cracking Mechanisms

Several distinct cracking mechanisms are routinely grouped together informally as “weld cracking,” but they occur in different locations, at different temperatures, and respond to different countermeasures. Conflating them leads directly to ineffective fixes, such as adding preheat to address a problem that is actually compositional.

Prevention Strategies

Composition Control

For carbon and low-alloy steel, the most reliable lever is keeping sulfur and phosphorus as low as the specification allows and maintaining an adequate manganese-to-sulfur ratio, screened quantitatively with the units-of-crack-susceptibility (UCS) formula covered in the calculator above. Free-machining or resulfurized grades, selected elsewhere in a project for their machinability, should never be welded as though they were standard structural grades; if a connection to such material is unavoidable, it needs its own qualified procedure and careful dilution control. For austenitic stainless steel, the equivalent lever is selecting filler metal and managing dilution to land within a target ferrite number, generally in the 3 to 10 FN range, verified by magnetic ferrite measurement or predicted in advance from the chromium and nickel equivalents on a WRC-1992 diagram. For niobium-stabilized grades and nickel-base alloys, controlling niobium and silicon content in the filler is the more relevant lever, since ferrite control is often not available or not desirable in these systems.

Joint Design and Welding Technique

Reducing restraint wherever the design permits, through joint sequencing, backstep or block welding techniques on long seams, and avoiding unnecessarily rigid fit-up, directly reduces the strain imposed on the solidifying pool. Favoring a wider, shallower bead profile over a narrow, deep one, by moderating travel speed and current rather than maximizing deposition rate, addresses the geometric driver discussed earlier. At arc stops, a controlled current decay (crater fill) or a brief reversal of travel direction widens and shallows the crater before it freezes, which is the standard production fix for crater cracking.

Industrial Significance

Solidification cracking is consequential well beyond the immediate cost of a failed radiograph. In pressure vessel, piping, and structural steel fabrication, a centerline crack discovered during weld procedure qualification forces requalification with revised parameters or filler metal, delaying project schedules and consuming consumable test budgets. In austenitic stainless steel and nickel-base alloy fabrication for chemical process, power generation, and aerospace applications, the same mechanism governs whether a procedure qualifies at all, since these alloy families sit much closer to the susceptible end of the spectrum than ordinary carbon steel. In nuclear and offshore service, where weld integrity is verified by extensive radiography and where in-service repair access can be severely limited, a solidification crack that escapes initial inspection represents a long-term integrity risk rather than a one-time rework cost. For all of these reasons, susceptibility screening at the procedure development stage, using the quantitative tools and tests described in this article, is consistently cheaper than discovering the problem in production.

Frequently Asked Questions

Recommended Reference Reading

Disclosure: MetallurgyZone participates in the Amazon Associates programme. If you purchase through these links, we may earn a small commission at no extra cost to you. This helps support free technical content on this site.

Further Reading