Welding Is a Metallurgical Process, Not Just a Joining Process
It is tempting to think of welding as simply melting two pieces of metal together. Metallurgically, it is far more consequential: welding subjects a narrow band of metal to one of the most extreme, rapid thermal cycles it will ever experience — heating from room temperature to well above its melting point in seconds, then cooling back down almost as fast — and that thermal cycle rewrites the local microstructure. The strength, hardness, toughness, and cracking resistance of a welded joint are set as much by what happens in the solid metal next to the weld as by the weld bead itself. Understanding that metallurgy is what separates a welding procedure that merely looks good from one that will actually perform in service.
What Happens Metallurgically During Welding
As a welding arc or other heat source passes over the joint, it creates a small pool of fully molten metal — the weld pool — that mixes filler metal (if used) with melted base metal. As the heat source moves on, the weld pool solidifies extremely quickly, typically in seconds, far faster than the cooling of a casting. This rapid solidification produces a characteristic columnar grain structure in the weld metal, growing inward from the fusion boundary toward the centerline, and it leaves little time for alloying elements or impurities to diffuse evenly, so the weld metal's composition and structure can differ noticeably from the base metal even when a matching filler is used. Beyond the pool itself, a much larger volume of solid base metal is heated by conduction to temperatures ranging from just below melting down to barely above ambient, and it is this heated-but-unmelted band — the heat-affected zone — where much of the metallurgical damage, and much of the engineering risk, actually occurs.
The Weld Zones
| Zone | Peak temperature | What happens |
|---|---|---|
| Fusion zone (weld metal) | Above melting point | Fully melted and resolidified; columnar grain growth |
| Heat-affected zone (HAZ) | Below melting, above ambient | Solid-state microstructural change without melting |
| Unaffected base metal | Near ambient | No microstructural change |
The Heat-Affected Zone and Its Sub-Zones
In steel, the HAZ is not uniform — it is a gradient of peak temperatures, and each temperature band produces a distinct microstructure on cooling:
- Coarse-grain HAZ (CGHAZ): the band immediately adjacent to the fusion line, heated well into the austenite range to a peak temperature high enough that austenite grains grow large before transforming on cooling. Coarse grain size, combined with a fast cooling rate, promotes hard, brittle transformation products (bainite or martensite depending on composition and cooling rate) and is usually the lowest-toughness location in the entire joint — the zone most fracture and fatigue assessments focus on.
- Fine-grain HAZ (FGHAZ): heated just into the austenite range, but briefly enough and at a low enough peak temperature that grain growth is limited; on cooling it produces a fine-grained, refined microstructure with good strength and toughness, often as good as or better than the base metal.
- Intercritical HAZ (ICHAZ): heated into the two-phase ferrite-plus-austenite range between the lower and upper critical temperatures, producing a mixed, partially transformed microstructure that can include localized hard, high-carbon islands (martensite-austenite constituent) if cooling is fast — a known contributor to reduced local toughness.
- Subcritical HAZ (tempered zone): heated below the lower critical temperature, too cool to re-austenitize, but hot enough to temper any existing hardened microstructure. In steels that were originally quenched and tempered or cold-worked for strength, this band can actually soften relative to the base metal, creating a local weak spot in the joint even though nothing here transformed to a hard phase.
Effects on Strength, Hardness, and Toughness
The net mechanical effect of welding on a joint depends heavily on the base metal's starting condition and the cooling rate through the weld thermal cycle. In as-rolled, normalized mild and low-alloy steels, the CGHAZ commonly ends up harder than the base metal because of grain coarsening and the formation of some bainite or martensite, and that increased hardness typically comes at the cost of reduced Charpy impact toughness — the classic trade-off that weld procedure qualification testing (Charpy V-notch testing across the weld and HAZ) is designed to catch. In quenched-and-tempered or cold-worked steels, by contrast, the subcritical HAZ can soften, and under load the joint may yield and eventually fail in that softened band rather than in the weld metal itself. Either direction — hardening or softening — is a direct consequence of the same underlying cause: the weld thermal cycle imposes a peak temperature and cooling rate on the HAZ that the base metal's original processing never anticipated.
Common Weld Defects Tied to Metallurgy
Hydrogen-Induced (Cold) Cracking
Hydrogen cracking (also called cold cracking or delayed cracking) requires three factors together: a susceptible, sufficiently hard microstructure — typically martensite or bainite in the HAZ — a source of diffusible hydrogen, most often moisture in electrode coatings, flux, or surface contamination that dissociates in the arc, and residual tensile stress from the weld thermal cycle and joint restraint. Because hydrogen diffuses through the lattice relatively slowly at room temperature, cracks can appear hours or even days after welding is complete — hence "delayed" cracking — usually initiating in the hard CGHAZ rather than in the weld metal. Control measures directly target the three contributing factors: using low-hydrogen electrodes and properly baked/dried flux, preheating to slow cooling and give hydrogen time to escape, and controlling interpass temperature and joint restraint to limit residual stress.
Hot Cracking (Solidification Cracking)
Hot cracking occurs while the weld metal is still solidifying, at temperatures near the melting point, and is essentially a casting defect that happens to occur in a weld. Impurity elements with low melting points — particularly sulfur and phosphorus — segregate to the last liquid to solidify, forming thin, weak films along the solidifying grain boundaries, typically at the weld centerline. As the weld pool cools and contracts, the shrinkage strain concentrates across these weak, still-partially-liquid films, and if they cannot accommodate the strain, a crack opens along the centerline. Hot cracking is controlled metallurgically — limiting sulfur and phosphorus in the base and filler metal, and balancing manganese-to-sulfur ratio to tie up sulfur as higher-melting manganese sulfide instead of iron sulfide films — and procedurally, by avoiding deep, narrow weld bead profiles that concentrate shrinkage strain at the centerline.
Other Metallurgically Linked Defects
Lamellar tearing occurs in the base metal beneath a weld, in the through-thickness direction, when non-metallic inclusions rolled flat during plate manufacture open up under the through-thickness shrinkage strain of a heavy, restrained weld — a joint design and base-metal cleanliness issue as much as a welding one. Reheat cracking can appear during post-weld heat treatment itself in certain alloy steels, when precipitation hardening in the CGHAZ during the PWHT temperature ramp embrittles the grain boundaries faster than stress can relax. Each defect traces back to the same theme: the weld thermal cycle interacts with the base metal's composition and prior processing in ways that a purely geometric view of "joining two pieces of metal" completely misses.
Preheat and Post-Weld Heat Treatment
Preheating the base metal before welding — commonly to 100–300 °C for carbon and low-alloy steels, depending on carbon content, section thickness, and joint restraint — slows the cooling rate of the weld and HAZ. That slower cooling does two protective things at once: it reduces the amount of hard, crack-susceptible martensite that forms in the HAZ by allowing more of the austenite to transform to softer, tougher products instead, and it gives diffusible hydrogen more time to escape from the metal before it can concentrate at trapped sites and trigger delayed cracking. The carbon equivalent (CE) formula — combining carbon content with other hardenability-raising elements like manganese, chromium, and molybdenum into a single number — is the standard tool welding engineers use to estimate how hardenable, and therefore how preheat-sensitive, a given steel will be; higher CE steels generally require higher preheat and stricter hydrogen control.
Post-weld heat treatment (PWHT), most often a stress-relief anneal performed below the lower critical transformation temperature, addresses what preheat cannot. It relieves the residual tensile stress locked into the joint by the weld thermal cycle, which lowers the risk of stress-corrosion cracking and brittle fracture in service; it tempers any hard, untempered martensite or bainite remaining in the HAZ, trading some hardness for a substantial gain in toughness, mirroring exactly the same trade-off used when tempering quenched steel; and for hydrogen-sensitive joints, extended time at PWHT temperature continues to drive out residual diffusible hydrogen that preheat and interpass control did not fully remove. Codes such as ASME Section VIII and various pipeline and structural welding standards mandate PWHT above defined thickness or CE thresholds precisely because these benefits are not optional for thick-section or high-hardenability material — they are what keeps the joint from cracking in service long after the welder has moved on.
Practical Guidance for Welding Engineers
Sound welding procedure development starts from the metallurgy, not the other way around: estimate hardenability from the carbon equivalent, set preheat and interpass temperature to control HAZ hardness and hydrogen escape, select low-hydrogen consumables and control moisture pickup, and specify PWHT wherever section thickness, restraint, or code requirements demand it. Welding procedure qualification testing — hardness surveys across the weld and HAZ, Charpy impact testing at the coarse-grain HAZ location, and macro-etch cross-sections to reveal the zones directly — exists precisely to catch the metallurgical problems described here before the procedure is ever used on a real structure. Treating the HAZ as an inevitable, controllable metallurgical consequence of welding, rather than an afterthought to bead appearance, is what separates a weld that passes visual inspection from one that will actually survive its design life.