The Welding Metallurgy Guide: Heat-Affected Zones, Preheat & Distortion

Most welders learn technique first and metallurgy never. That works until the day a weld that looked textbook cracks along the edge of the bead a week after the job shipped, and nothing in the technique explanation accounts for it. What happened was not in the weld. It was in the strip of base metal beside it — the part you never melted, never inspected, and never thought about. This guide is about that strip, and about the three forces that govern it: heat, time and restraint.

What welding actually does to steel

Welding is a very localised, very fast heat treatment that you perform accidentally while trying to do something else. In the space of a few seconds a small volume of steel goes from room temperature to molten and back again, and the metal immediately around it experiences the whole range of temperatures in between.

Steel is not one thing at all temperatures. Its internal crystal structure changes as it heats and changes again as it cools, and the structure it ends up with depends on both how hot it got and how fast it came back down. Slow cooling from high temperature tends to produce softer, tougher structures. Fast cooling from the same temperature can produce much harder, more brittle ones. Nothing about the weld's appearance tells you which happened.

So a welded joint is really three different materials sitting next to each other: the weld metal itself, a transition region, and the unaffected parent metal further out. Each has different properties. The joint is only as good as the weakest of them, and it is rarely the one you were watching.

The heat-affected zone

The heat-affected zone — HAZ in every document you will ever read — is the band of parent metal that did not melt but got hot enough to change. It runs from the fusion line outward to wherever the peak temperature fell below the point at which anything happens.

It is not uniform. Right at the fusion line the metal reached the highest temperature and typically has the coarsest grain structure, which is generally the least tough region of the whole joint. Further out, peak temperatures were lower and the effects are different and usually milder. The zone might be a few millimetres wide on a light, fast weld or considerably wider on a high heat input pass.

Cooling rate is the master variable

If you take one idea from this guide, take this one: for most steels, the single variable that most determines what you end up with is how fast the joint cooled. Peak temperature sets what is possible; cooling rate decides which possibility you get.

Cooling rate is not something you set on a dial. It is the product of several things you do control, and several you do not:

• Section thickness and mass. Thick, heavy material is an enormous heat sink. The same weld on thin sheet and on heavy plate cools at completely different rates.
• Joint geometry. A weld at the intersection of three plates loses heat in three directions. Heat escapes faster from some joint configurations than others.
• Starting temperature. Cold steel pulls heat out faster. This is exactly what preheat addresses.
• Heat input. More energy per unit length of weld means a larger heated volume, which cools more slowly.
• Ambient conditions. Welding outdoors in winter, on a windy site, or on material that has been sitting in the cold changes the cooling rate materially.
• What you do afterward. Quenching, covering the joint, or leaving it on cold concrete all change the tail of the cooling curve.

The practical shape of this: a small weld on thick, cold, heavily restrained steel is the worst-case combination. Fast cooling, hard heat-affected zone, high stress, and any hydrogen present has nowhere to go. That combination is the classic recipe for cracking, and it is exactly what happens when someone tack-welds a heavy bracket onto cold plate on a January morning and comes back to a cracked tack.

Heat input and how to think about it

Heat input describes the energy delivered per unit length of weld. It rises with current and voltage and falls as travel speed increases. You do not need to calculate it to use the concept — what matters is understanding what moving it does.

This is why welding procedures often specify a heat input range rather than a floor or a ceiling. Too little and the joint cools too fast, hardens and may crack. Too much and you get grain growth, reduced toughness and considerably more distortion. The right answer sits in a window that depends on the material and the section, which is precisely the kind of thing a qualified procedure exists to establish.

Preheat and interpass temperature

Preheat is warming the base metal before welding. Its purpose is not to make welding easier — it is to slow the cooling rate. Reducing the temperature gap between the weld and the surrounding steel means heat drains away more slowly, which gives hard structures less opportunity to form and gives any hydrogen present time to diffuse out while the metal is still warm enough for it to move.

The factors that push a job toward requiring preheat are consistent across codes: material chemistry, section thickness, joint restraint, ambient temperature and the hydrogen potential of the process and consumables. What the actual required temperature is for a given combination comes from the material specification and the applicable code — and there is no shop shortcut for it that is safe to publish.

• Heat the surrounding area, not just the joint line. Preheat is about the mass that will absorb the weld's heat, so it needs to extend well back from the joint on both sides.
• Heat through the thickness. A hot surface over a cold interior has not achieved the effect. Thick sections take time.
• Measure, do not estimate. Temperature-indicating crayons, contact thermometers and infrared instruments all work; guessing does not.
• Maintain it. Preheat that has drained away by the time you start the third pass has stopped doing its job.
• Never preheat with a sooty flame and never onto a joint you have not cleaned — you can bake contamination into the surface.

Interpass temperature is the joint's temperature immediately before each subsequent pass on multi-pass work. Procedures typically specify both a minimum — because letting it fall reintroduces the fast-cooling problem — and a maximum, because excessive accumulated heat can degrade toughness. On some materials the maximum matters as much as the minimum, which surprises people who think of heat as purely protective.

Hydrogen and delayed cracking

This is the failure mode that most alarms people who have not met it, because the weld passes inspection and then cracks hours or days later. It requires four things simultaneously, which is also the key to preventing it — remove any one and the mechanism stops.

The practical consequences are the reasons rod ovens and consumable storage exist as a product category at all. Low-hydrogen electrodes absorb moisture from the air; once damp, they are a hydrogen source rather than a hydrogen control. Storage requirements and any permitted re-drying come from the consumable manufacturer, and they are requirements rather than best practice.

Distortion: the mechanical consequence

Everything above concerns what happens inside the metal. Distortion is the same physics showing up in a form you can measure with a straightedge. Heated metal expands, is restrained by the cold metal around it, gets locally squashed by that restraint, and then contracts as it cools — ending up shorter than it started, and pulling everything attached to it.

The three ways it shows up are worth naming because they call for different responses: transverse shrinkage pulls the joint narrower across the weld, longitudinal shrinkage shortens the part along the weld, and angular distortion folds the joint as the top of the weld contracts more than the root.

• Reduce heat input where the material permits. Less energy in means less movement out.
• Balance the heat. Alternate sides, alternate ends, and weld symmetric joints in a symmetric sequence. Where a joint can be welded from both sides, doing so opposes the angular pull.
• Back-step. Weld short segments in the direction opposite to overall progress so the heat is distributed along the joint rather than following it.
• Use intermittent welds where the design allows, rather than a continuous bead you did not need.
• Pre-set the joint so the contraction pulls the part into alignment rather than out of it — the skill here is knowing how much, which comes from your own material and joints.
• Fixture thoughtfully. Restraint reduces movement and increases residual stress. On hardenable material and heavy restraint, that trade has a cracking cost.
• Tack properly. Adequate tacks, correctly spaced, sequenced from the middle outward rather than progressively along the joint.
• Let it cool. Interpass patience is not wasted time.

Straightening after the fact is possible — mechanical straightening and controlled flame straightening are both established techniques — but both add heat or stress to a part that has already had plenty of both, and flame straightening in particular is a skill rather than a fallback. Preventing distortion through sequence is cheaper than correcting it every time.

Dilution, filler and dissimilar joints

Weld metal is not the filler you fed in. It is a mixture of filler and melted parent metal, and the proportion of parent metal in that mixture is called dilution. On a joint with deep penetration into the base material, a substantial fraction of the resulting weld metal comes from the parts you were joining.

This matters because it means the weld's composition — and therefore its properties — depends on the base metal as well as the filler. It is why filler selection is a genuine engineering decision rather than a matter of what is on the shelf, and why it becomes critical when joining two different materials: the resulting weld metal is a blend of three compositions, and some blends have poor properties even though each component is perfectly good on its own.

Post-weld heat treatment

Some welded assemblies receive controlled heating after welding — broadly, to reduce residual stress, to temper hard regions in the heat-affected zone, and in some cases to allow remaining hydrogen to escape. Whether it is required, at what temperature, for how long, and with what heating and cooling rates, all come from the material specification and the applicable code.

It is worth knowing this exists even if you will never do it, for one reason: it explains why residual stress is treated as a real engineering concern rather than an abstraction. Every weld leaves stress behind. On most shop fabrication that is acceptable and nobody does anything about it. On demanding service — pressure equipment, thick section, fatigue loading, certain alloys — it is significant enough that an entire additional process step exists to deal with it.

What this means at the bench

Most of this becomes actionable as a small number of habits rather than as calculations. If you take nothing else forward, take these.

1. Know what you are welding. Unknown steel of unknown thickness and unknown history is the root of most metallurgical surprises. Mystery metal from a scrap pile deserves test welds on offcuts before it becomes a project.
2. Treat cold, thick and restrained as a warning combination. When all three are present, the cooling rate is at its worst and the stress is at its highest.
3. Keep consumables dry. It is the cheapest hydrogen control available and the one most often neglected in home shops.
4. Clean the joint properly. Oil, paint, rust and moisture are hydrogen sources as well as porosity sources.
5. Sequence deliberately. Decide the order of your welds before you strike the first arc, not after the part has moved.
6. Resist quenching. The instinct to cool a hot part quickly works against almost everything in this guide.
7. Test on scrap and break the coupons. A bend test on an offcut of the actual material tells you more than any amount of inspecting welds you will not cut apart.
8. Escalate when it matters. Structural, load-bearing, pressure-containing or safety-critical work needs a qualified procedure and a qualified welder. Knowing where that line sits is itself a metallurgical skill.

The reason any of this is worth learning is not academic. It is that once you can see the heat-affected zone in your mind's eye, a whole category of problems stops being random. Cracks that appear a day later, parts that pull out of square, joints that fail beside the weld rather than through it — all of them have causes that live in the interaction between heat, time and restraint, and all of them respond to changes you can actually make.