A discontinuity in welding is any interruption in the normal structure of a weld, whether it’s a tiny gas bubble, a small crack, or a spot where the metal didn’t fuse completely. Every weld has them. The critical distinction is that a discontinuity only becomes a “defect” when it exceeds the tolerance limits set by the applicable welding code. Until it crosses that threshold, it’s simply an imperfection that falls within acceptable range.
Understanding the types of discontinuities, what causes them, and when they cross the line into rejection territory is essential for welders, inspectors, and anyone working with welded structures.
Discontinuity vs. Defect
These two terms get used interchangeably in casual conversation, but they mean very different things in welding codes. A discontinuity is any deviation from the ideal weld: a pore, a bit of trapped slag, slight undercut along the toe. Every weld has slight imperfections like these. A defect is a discontinuity (or group of discontinuities) that exceeds the tolerance specified by the governing code, such as AWS D1.1 for structural steel or ASME Section IX for pressure vessels.
The code sets the limits. For example, under AWS D1.1, undercut up to 1/32 inch deep is permitted for the entire length of a weld on thinner materials. But undercut deeper than 1/16 inch on material one inch thick or greater is unacceptable at any length. A single discontinuity might be perfectly fine on one project and grounds for rejection on another, depending on which code applies and whether the joint carries cyclic or static loads.
Porosity
Porosity is one of the most common internal discontinuities. It consists of small cavities formed when gas gets trapped inside the weld metal as it solidifies. The American Welding Society defines it as a “cavity type discontinuity formed by gas entrapment during solidification.” These voids can appear as scattered individual pores, clustered groups, or elongated channels sometimes called worm tracks.
The causes are straightforward: contamination on the base metal (oil, moisture, rust), inadequate shielding gas coverage, or incorrect welding parameters. With flux-cored arc welding, too much voltage is a particularly common culprit. As voltage increases toward the threshold of creating porosity, you’ll often see worm tracks forming on the weld cap just under the slag, a sign that the puddle is cooling and solidifying before gases can escape to the atmosphere. Drafts that blow shielding gas away from the arc, damp electrodes, and excessive travel speed all increase the risk.
Slag Inclusions
Slag inclusions are pockets of non-metallic material trapped within the weld. They occur most often in processes that produce slag, like shielded metal arc welding (stick) and flux-cored arc welding. When slag from one pass isn’t fully removed before the next pass is deposited, it gets buried in the joint.
These inclusions can sit between weld passes or along the fusion boundary where the weld meets the base metal. If porosity contains slag, it’s classified as a slag inclusion by definition. Poor joint preparation, insufficient cleaning between passes, and incorrect electrode angle all contribute. Slag trapped at critical locations like the toe of a fillet weld is especially concerning because it can act as a stress concentrator.
Incomplete Fusion and Penetration
These are two related but distinct problems. Incomplete fusion means the weld metal didn’t fully bond to the base metal or to a previous weld pass, leaving an unfused gap at one side of the joint. Incomplete penetration means the weld didn’t reach all the way through the root of the joint, leaving both sides of the root region unfused.
Several setup and technique errors cause these issues:
- Root face too thick: An excessively thick root face in a butt weld leaves too much metal for the arc to penetrate through.
- Root gap too small: Without enough space at the root, the weld pool can’t reach the bottom of the joint.
- Heat input too low: Insufficient current for the size of the root face means the arc doesn’t have the energy to melt through.
- Heat input too high: Counterintuitively, too much current can also cause problems. The welder moves faster to compensate, and the weld pool bridges over the root without actually penetrating it.
- Bevel angle too small or electrode too large: Both restrict arc access to the root of the joint.
These discontinuities are particularly dangerous because they directly reduce the load-carrying cross-section of the weld. They’re more common in consumable electrode processes like MIG, stick, and flux-cored welding, where the welder has limited ability to control penetration independently of depositing filler metal.
Cracking
Cracks are the most serious category of weld discontinuity. Most codes treat any crack as a rejectable defect regardless of size because cracks can propagate under load. They fall into two broad categories based on when they form.
Hot Cracking
Hot cracking, also called solidification cracking, happens while the weld is still cooling from liquid to solid. Weld metals are alloys with a range of freezing temperatures rather than a single melting point. As the metal solidifies, the growing crystal structures push low-melting-point constituents to the boundaries between grains. This leaves a thin film of liquid persisting at temperatures well below where the bulk of the weld has already solidified. If shrinkage stresses pull the joint apart before that last bit of liquid freezes, a crack forms right down the centerline or along grain boundaries. Certain elements in the base or filler metal widen this freezing range and increase the risk significantly.
Cold Cracking
Cold cracking, often called hydrogen-induced cracking, develops hours or even days after the weld cools to room temperature. It requires three conditions acting together: hydrogen dissolved in the weld from moisture or contamination, a hard and brittle microstructure in the heat-affected zone, and residual tensile stress from the welding process. This is why preheat and post-weld heat treatment matter so much on higher-carbon and alloy steels. Removing any one of those three factors largely eliminates the risk.
Undercut and Overlap
Undercut is a groove melted into the base metal along the toe or edge of the weld that isn’t filled by weld metal. It creates a notch, which is a stress concentration point that can initiate fatigue cracking under cyclic loading. High arc voltage, excessive travel speed, and improper electrode angle are the primary causes. In laser hybrid arc welding, undercut severity increases sharply with welding speed, as the weld pool dynamics change at higher travel rates. A high arc current can depress the weld pool surface, leaving only a thin film of molten metal in contact with the base metal at the edges, which solidifies before it can fill the groove.
Under AWS D1.1, the acceptable depth depends on material thickness and loading conditions. For cyclically loaded connections, undercut deeper than 1/32 inch is generally unacceptable. For statically loaded joints on thicker material, undercut up to 1/16 inch deep may be allowed for a limited length. Many experienced welders simply aim for zero undercut to avoid the conversation entirely.
Overlap is the opposite problem: weld metal that flows over the base metal surface without fusing to it. It creates a mechanical notch similar to undercut. Overlap results from too slow a travel speed, too low a voltage, or incorrect work angle that lets gravity pull the molten pool over unfused base metal.
Lamellar Tearing
Lamellar tearing is a base metal discontinuity rather than a weld metal one. It occurs beneath the weld in rolled steel plate that has poor ductility in the through-thickness direction (perpendicular to the plate surface). During rolling, sulfide and oxide inclusions in the steel get flattened and elongated parallel to the plate surface. When welding stresses pull in the through-thickness direction, these inclusion layers can separate in a stepped, staircase-like pattern beneath the weld.
No single grade of steel is immune, but plates with a high concentration of elongated inclusions are most susceptible. The risk increases with highly restrained joints, thick plates, and weld designs that concentrate shrinkage stress through the plate thickness. Lamellar tearing only occurs in rolled plate, never in forgings or castings, because the rolling process is what creates the layered inclusion structure in the first place. Joint redesign to reduce through-thickness stress, using buttering layers, or specifying steel with guaranteed through-thickness properties (often called Z-grade plate) are the standard preventive measures.
How Discontinuities Are Detected
Surface discontinuities like cracks, undercut, overlap, and porosity open to the surface can be found through visual inspection or enhanced with liquid penetrant testing, where a colored or fluorescent dye seeps into surface-breaking flaws. Magnetic particle testing works similarly on ferromagnetic steels, using magnetic fields and iron particles to reveal surface and near-surface flaws.
Internal discontinuities require more involved methods. Radiographic testing (X-ray or gamma ray) produces an image of the weld’s interior, making porosity, slag inclusions, and some types of incomplete fusion visible on film or a digital detector. Ultrasonic testing sends sound waves through the weld and interprets the reflections, making it especially effective at finding planar discontinuities like cracks and lack of fusion that radiography can miss. The choice of inspection method depends on the code requirements, the type of discontinuity expected, and the joint geometry.