The toe of a weld is the point where the surface of the weld bead meets the base metal. It forms a visible line running along each side of a weld, marking the exact boundary between added weld material and the original workpiece. This small geometric detail matters far more than its size suggests: the weld toe is the most common site where fatigue cracks begin in welded structures, and its shape directly determines how long a welded joint will last under repeated loading.
Why the Weld Toe Creates a Weak Point
Every weld creates a change in geometry where the raised weld bead transitions back down to the flat base metal. That transition happens at the toe, and it acts as a stress concentrator. When a load passes through the joint, stress doesn’t flow evenly across the surface. Instead, it piles up at the toe the same way water speeds up when a river narrows. The sharper and more abrupt that transition, the higher the stress concentration becomes.
Engineers quantify this effect using stress concentration factors, and getting them right is critical. Even a few percent error in estimating the peak stress at a weld toe can translate into errors of tens or even several hundred percent when predicting how long a joint will survive under cyclic loading. In the worst case, if the weld toe radius approaches zero (essentially forming a sharp corner), the theoretical stress at that point becomes infinite. That’s why a smooth, gradual transition at the toe is so important in practice.
How Fatigue Cracks Start at the Toe
The weld toe isn’t just a stress concentration because of its overall shape. Up close, the surface is rarely perfectly smooth. Tiny secondary notches, minor undercuts, slag inclusions, and even spatter that landed on cooler metal nearby all create additional microscopic stress risers right at the toe. These small defects serve as initiation points where fatigue cracks begin.
In manually welded joints, researchers have identified minor undercuts, trapped slag, and surface spatter at the toe as the most common crack initiation sites. In automated welds, undercut-type defects dominate. Once a small crack forms at one of these imperfections, the already-high stress at the toe drives it deeper into the base metal with each loading cycle. This is why fatigue failures in welded structures almost always trace back to the weld toe rather than the middle of the weld bead or the base metal itself.
Common Defects at the Weld Toe
The two defects inspectors most often look for at the weld toe are undercut and overlap. Undercut appears as a groove or recessed channel running along the toe where the arc melted away the base metal but the filler didn’t fill it back in. Under proper lighting, it shows up as a linear notch right at the transition line. Undercut is one of the single most damaging defects for fatigue life because it sharpens the stress concentration exactly where stresses are already highest.
The most common causes of undercut are straightforward welding parameter problems. Moving the torch too fast doesn’t give filler metal enough time to deposit along the edges. Running too much voltage or too long an arc widens the arc cone and erodes the sidewalls instead of fusing them properly. An unbalanced wire feed rate, especially in mechanized welding, throws off the ratio of filler to heat and increases the chances of undercut forming.
What Codes Require at the Weld Toe
Welding codes set strict limits on how much undercut is acceptable at the toe, and those limits depend on the loading direction and the type of structure. The 2025 edition of the American Welding Society’s D1.5 Bridge Welding Code uses three tiers. For welds running across a primary tension stress (the most critical orientation), undercut depth is limited to just 0.01 inches. For most other locations, the limit is 1/16 inch. At stiffener and connection plate corners, it relaxes to 1/8 inch. If undercut exceeds 1/32 inch deep, grinding alone won’t fix it, and the defect must be repaired by welding.
The international standard ISO 5817 takes a similar approach, grading weld quality into three levels: B (strictest), C, and D. At Level B, continuous undercut at the toe isn’t permitted at all, and intermittent undercut can be no deeper than 5% of the material thickness, up to a maximum of 0.5 mm. The standard also addresses the weld toe angle itself. For the highest quality butt welds (Level B), the angle between the weld face and the base metal must be 150 degrees or greater, ensuring a very gradual, smooth transition. Level C requires at least 110 degrees, and Level D allows angles down to 90 degrees.
Improving the Toe After Welding
Because the weld toe is so critical to fatigue life, several post-weld treatments exist specifically to improve its geometry and stress state. The most common are grinding, TIG dressing, and mechanical impact treatment.
Grinding the weld toe with a burr or disc smooths out the transition and removes surface defects like small undercuts and slag. The goal is to create a gradual radius rather than an abrupt angle change. Codes typically require that grinding not reduce the cross-sectional area of the base metal by more than 2%.
TIG dressing involves remelting the weld toe with a TIG torch and no filler metal. The heat creates a new, smoother solidification profile at the transition. In practice, this transforms the toe radius from near zero in the as-welded condition to roughly 6 to 7 mm after treatment. Beyond the geometric improvement, the remelting also redistributes residual stresses at the toe from tension to compression, which resists crack initiation. Both effects contribute to measurably better fatigue performance.
High-frequency mechanical impact (HFMI) treatment, sometimes called ultrasonic peening, uses a rapidly vibrating tool to physically indent the weld toe. This does several things at once: it increases the toe radius, introduces compressive residual stresses that can exceed the material’s yield strength to a depth greater than 1.5 mm, hardens the local material, and reorients any existing micro-cracks away from the most damaging direction. HFMI is increasingly used to extend the service life of existing welded structures without having to re-weld them.
Identifying the Toe During Inspection
For anyone inspecting a weld, the toe is the first place to look. Run your eye (or a magnifying lens) along each side of the weld bead where it meets the parent metal. You’re checking for a smooth, gradual transition with no visible grooves, sharp notches, or places where the weld metal has rolled over onto the base metal without fusing. Proper lighting at a low angle helps reveal undercut that might not be visible from directly above.
The toe angle, the toe radius, and the presence or absence of surface defects at the toe collectively tell you more about a weld’s likely fatigue performance than almost any other visual feature. A weld can look perfectly acceptable across its face and still have a toe profile that will crack under cyclic loading. That’s what makes understanding this small geometric detail so important for anyone working with welded structures.