Steel metallurgy and weldability

Structural steel covered what the sections are and why steel is used above ground. This topic is about the material itself — because once a structure is in service, almost every decision an inspector makes about welding it, repairing it or judging a crack in it comes back to chemistry and microstructure.

The short version: the mill certificate tells you how the steel will behave when you weld it, and most people never read it that way.

A structural steel plate on a fabrication bench in a welding shop, one edge freshly cut and still carrying the straw and blue heat tint, with welding sets, electrode stubs and a welder at work behind it. — Everything in this topic is about that heat-tinted edge. The moment steel is cut or welded, its chemistry decides what the metal beside the joint turns into.

What is actually in structural steel

Iron and carbon, plus a controlled list of everything else. The quantities look trivial — tenths and hundredths of a percent — and they decide everything.

Element	What it does	Why you care
Carbon (C)	The main strengthening element	Raises strength, wrecks toughness and weldability
Manganese (Mn)	Deoxidises, ties up sulfur, refines grain	Raises strength and toughness — the useful one
Silicon (Si)	Deoxidiser	Makes the steel “killed”, i.e. sound
Phosphorus (P)	Impurity	Embrittles at low temperature. Capped, typically ≤ 0.035 %
Sulfur (S)	Impurity	Hot cracking in welds. Capped, typically ≤ 0.035 %
V, Nb, Ti	Micro-alloying	Fine grain, so strength and toughness together
Cr, Ni, Mo, Cu	Alloying	Hardenability, toughness, weathering resistance

Carbon is the trap. It is the cheapest way to make steel stronger, and every increment costs you ductility, impact toughness and weldability. Modern structural grades go the other way: keep carbon low and buy the strength from grain refinement instead.

Carbon equivalent: the number that decides preheat

You cannot judge weldability from carbon alone, because manganese, chromium, molybdenum and the rest all contribute to hardenability too. The industry rolls them into one figure. The International Institute of Welding formula is the common one:

CE(IIW) = C + Mn/6 + (Cr + Mo + V)/5 + (Ni + Cu)/15

For modern low-carbon micro-alloyed steels the IIW formula is pessimistic, and the Ito–Bessyo parameter is used instead:

Pcm = C + Si/30 + (Mn + Cu + Cr)/20 + Ni/60 + Mo/15 + V/10 + 5B

A horizontal scale of carbon equivalent from 0.30 to 0.55, divided into three bands. Below 0.35 is marked excellent weldability with no preheat needed. From 0.35 to 0.45 is marked good to moderate, requiring 100 to 150 degrees Celsius preheat and low-hydrogen electrodes. Above 0.45 is marked difficult, requiring 150 to 250 degrees Celsius preheat, interpass control and hydrogen bake-out. — The same steel section, the same welder and the same consumable — and which band the chemistry falls into decides whether the joint needs a torch on it first.

CE (IIW)	Weldability	What the procedure needs
< 0.35	Excellent	No preheat on thin sections; low cracking risk
0.35 – 0.45	Good to moderate	Preheat 100–150 °C, low-hydrogen consumables
> 0.45	Difficult	Preheat 150–250 °C, interpass control, hydrogen bake-out

The heat-affected zone

Welding is a very localised, very severe heat treatment that nobody designed. The weld metal itself is cast and usually fine. The problem is the parent metal beside it, which was never melted but was cooked — the heat-affected zone.

Cross-section through a welded joint showing the weld deposit at the centre, then moving outward the fusion line, the coarse-grained heat-affected zone above 1100 degrees Celsius with large grains, the fine-grained zone around 900 degrees with small refined grains, the intercritical zone around 750 degrees with a partly transformed mixed structure, and finally unaffected parent metal. A temperature curve above the section falls away with distance from the weld. — Peak temperature falls with distance from the fusion line, and each band ends up with a different microstructure. The coarse-grained band nearest the weld is the one that fails.

Working outwards from the weld:

Coarse-grained HAZ (CGHAZ) — above roughly 1100 °C. Grains grow large and stay large. Lowest toughness in the joint. This narrow band, often a millimetre or two wide, is where brittle fracture and cold cracking start.
Fine-grained HAZ (FGHAZ) — around 900 °C. Effectively normalised. Refined grain, good toughness. The best-behaved part of the joint.
Intercritical HAZ (ICHAZ) — around 750 °C. Only partly transformed, giving a mixed structure with local brittle zones. In multi-pass welds these can be re-heated into something worse still.
Unaffected parent metal — as delivered.

The phases, briefly

You do not need to be a metallurgist, but four names come up constantly in inspection reports and failure investigations:

Phase	Structure	Character
Ferrite	Body-centred cubic	Soft, very ductile, low strength, absorbs impact well
Pearlite	Layered ferrite + cementite	Provides hardness and tensile strength
Bainite	Fine acicular	Strong and reasonably tough — a good outcome
Martensite	Body-centred tetragonal	Very hard, heavily strained, brittle

Martensite is the one that matters. It forms when austenite is cooled too fast to transform any other way, which is exactly what an unpreheated weld on a thick, high-CE section does to its own heat-affected zone.

Hydrogen-induced cold cracking needs three things at once:

A susceptible microstructure — martensite in the HAZ.
Dissolved hydrogen — from damp electrodes, damp flux, oil, rust or humidity.
Tensile restraint stress — from the joint geometry and the weld’s own shrinkage.

Remove any one and it does not happen. Preheat attacks the first and helps the second by giving hydrogen time to diffuse out. Low-hydrogen consumables, properly stored in a heated quiver, attack the second directly.

A Charpy pendulum impact testing machine in a materials laboratory, its hammer raised on the swinging arm, three notched test specimens standing on the bench beside the anvil. — A Charpy impact test to EN ISO 148-1. Three specimens, one temperature, and the average energy in joules is what a grade designation such as S355J2 is actually promising.

Mechanical properties, and where they come from

Assessed to ASTM A370 and ASTM E8, reported on the mill certificate:

Yield strength (Fy) — where permanent deformation starts. A36 is 250 MPa (36 ksi); A572 Gr 50 and S355 are about 345–355 MPa. Steels without a sharp yield point use the 0.2 % offset method.
Ultimate tensile strength (Fu) — the maximum before necking. The gap between Fy and Fu is the reserve past first yield, and it is what lets a redundant structure redistribute load instead of unzipping.
Ductility — percentage elongation and reduction of area. This is the property that turns a sudden collapse into a visible sag that somebody reports.
Charpy V-notch impact energy — to ASTM E23. Joules absorbed at a stated temperature.

An EN 10204 3.1 mill test certificate laid on a steel bench in a fabrication shop, showing the heat number, chemical analysis and mechanical properties tables, with a pen and a work glove resting on it. — The 3.1 certificate. The heat number on it must trace back to a mark on the steel itself — a certificate that cannot be tied to the material in front of you is a piece of paper, not evidence.

What an inspector does with all this

When you pick up a mill test report on an existing structure or an incoming delivery:

Grade and standard — does it match the drawing and the specification?
Chemistry — compute the carbon equivalent yourself. Do not assume the grade implies it; two heats of the same grade can differ enough to change the preheat.
Mn/S ratio — low manganese with high sulfur is a hot cracking risk.
Impact test temperature — against the real minimum service temperature.
Heat number traceability — does it tie back to the marking on the actual member?

On an older structure with no certificate, that information does not exist and cannot be assumed. Pre-1970s steels are often higher in carbon, sulfur and phosphorus, with poor and untested toughness. This is exactly why bolted repairs are frequently preferred over welded ones on legacy steelwork — a subject taken up in repair and strengthening.

What to take away

Carbon equivalent, not carbon, is the weldability number — and it decides preheat.
The coarse-grained HAZ beside the fusion line is the weakest part of a good weld, and no soundness test will find it.
Cold cracking needs a hard microstructure, hydrogen and restraint together. Break any one.
A low-hydrogen electrode stops being one the moment it gets damp.
Impact energy without a test temperature is not a specification.
No mill certificate means no assumptions. Treat legacy steel as unknown until tested.