Materials Science in Welding: Why Metallurgy Matters

A bridge gang welded a 20 mm A572 Grade 50 bracket on a warm afternoon, ran a scheduled ultrasonic test, and walked away. Four months later, during an inspection prompted by a hairline crack report from the public, the crack was traced not to the weld metal but to the narrow strip of steel that melted and re-solidified next to it. Nobody had checked the chemistry of the plate, the heat input of the pass, or the moisture in the electrode coating. That failure is not a welding mistake — it is a materials science mistake, and it is the kind this guide exists to prevent.

Why Materials Science Decides Whether a Weld Holds

It began with a bridge gang welding a 20 mm A572 Grade 50 bracket on a warm afternoon. They ran a scheduled ultrasonic test, packed up for the day, and moved on. Four months later, a hairline crack was reported. After investigation, the culprit wasn’t the weld bead, but the narrow strip of steel directly beside it that had melted and re-solidified from the heat of welding. No one had accounted for the plate chemistry, the heat input of each pass, or the moisture in the electrode coating. This wasn’t a lapse in muscle-memory. It was a materials science mistake, and it’s exactly the type of failure this guide is built to prevent.

The reality is a simple one: most weld failures are metallurgy failures, not craftsmanship failures. For a welding audience, materials science can be boiled down to four factors. First, chemistry determines what elements are present within the metal. Second, microstructure is how those elements are arranged at the grain level after cooling. Third, thermal history is the impact of the weld heat from its peak all the way down to ambient temperature. Finally, mechanical loading is the force that will be applied to the joint years down the road. These form the four questions any weld must answer: what is the chemistry, what is the microstructure after cooling, what residual stresses were locked in, and what loads will arrive.

The Zone You Cannot See

When a joint breaks, it’s vital to know the difference between a weld metal failure and a heat-affected zone (HAZ) failure. The weld metal is only a fraction of the joint; the HAZ and fusion boundaries usually dictate the joint’s strength and toughness. The HAZ is the portion of the base metal that never reaches its melting point, yet still experiences temperatures drastic enough to austenitize, grain grow, and transform into an entirely different structure than what left the mill. Because of this, the HAZ can become brittle or weakened without any visible change to the weld bead itself. With all of this considered, the answer is an emphatic yes: materials science matters enormously, and it matters most in the zones the welder cannot see.

The Scale of the Stakes

The need for this level of control isn’t small. According to the USGS, iron and steel comprise about 95 percent of all the tonnage of metal produced annually in the United States and the world. From pressure vessels, to rail, to the bridges we drive over every day, the majority of our infrastructure relies on the metallurgy of joining. This level of production simply cannot afford the guesswork of treating welding as purely mechanical.

Because of these stakes, governing bodies have been forced to remove that guesswork entirely. AWS states it has authored more than 350 standards for welding practices and procedures, including AWS D1.1/D1.1M:2025 and AWS D1.2/D1.2M:2026. These codes, along with others such as those from ASME, exist precisely because weld properties must be controlled procedurally, not by assumption.

The Roadmap Ahead

With that foundation laid, the rest of this guide will follow the path of metallurgy from start to finish. We’ll begin with weldability and the effects of chemistry. From there, we’ll dive into the HAZ and its microstructure, before tackling heat input, preheat, and interpass temperature as the tools used to steer that structure. Next, we’ll cover hydrogen and cold cracking, followed by the metallurgy behind filler metal selection. We’ll then look at how each of the major material families behave differently, translate all of this into strength, toughness, and distortion, work through common metallurgy-driven defects, and finally, apply every concept to a shop-floor checklist to prevent the next bridge-bracket failure. Each of these pieces ties back to the same goal: control the unseen metallurgy, and the weld will hold.

Weldability: The Core Materials-Science Concept

Before a single pass is laid down, the question that matters for a structural joint is not “Can this metal be melted together?” It is “Will this metal still have the properties the design is counting on after it has been welded?” That is weldability: the ability of a material to be welded without losing the properties the design depends on.

For carbon and low-alloy steels, weldability is best understood as a race between two forces. Cooling rate decides how much time the steel has to develop softer, tougher microstructures. Hardenability is how readily that steel forms hard, brittle structures such as martensite when it is cooled quickly. When hardenability is high and the weld is allowed to cool too fast, the steel can “harden in place” inside the heat-affected zone and set the stage for cracking. That race is what makes one steel forgiving in the shop and another touchy.

The Carbon Equivalent: Measuring Hardenability

No two heats of steel are identical, so fabricators need a way to estimate hardening risk without guessing. The most widely used tool is the carbon equivalent (CE), a formula established by the International Institute of Welding (IIW) that is still used in countless shops today:

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

Every element carries weight for a reason. Carbon is the backbone of hardenability and counts at full value. Manganese contributes strength and deoxidation but also promotes hardening, so it is divided by 6. Chromium, molybdenum, and vanadium are strong carbide formers that dramatically raise hardenability, so they are divided by 5. Nickel and copper have a milder hardening effect, so they are divided by 15.

Two very common structural steels show how quickly this moves:

Steel	Typical mill chemistry	Approximate CE
A36	C 0.17, Mn 0.80	0.30
A572 Grade 50	C 0.20, Mn 1.00, V 0.04	0.37

That 0.07 increase may look like a small jump, but it is more than enough to raise the risk of a hard, crack-prone heat-affected zone, especially with faster cooling. A fabricator who never checks the mill chemistry is simply guessing at one of the most important variables.

Restraint and Section Thickness: The Other Two Levers

Chemistry is the foundation, but weldability is never chemistry alone. Two other levers can tip the scale just as hard. Restraint is the inability of a joint to move as it contracts. The more a heavy beam or a highly restrained connection is locked down, the more tensile stress builds up as the weld cools, leaving no room for a hardened area to yield before it cracks. Section thickness plays its part too. Thicker plates act as massive heat sinks and pull heat away from the weld at a much faster rate. Pair a higher-CE steel with thick material and tight restraint, and even perfectly sound welding technique becomes far less forgiving.

How Alloying Elements Shift Weldability

Every number on a mill test report has a say in how the joint behaves. Carbon is the greatest driver of hardening, followed closely by manganese. Silicon acts mainly as a deoxidizer, though excess silicon can increase hardenability in some cases. Chromium, molybdenum, vanadium, and niobium all raise hardenability and can refine grains, but when pushed too far they increase the likelihood of brittle microstructures. Nickel improves toughness at low temperatures while adding to hardenability. Copper has a modest hardening effect and can cause surface issues if not accounted for. On the flip side, phosphorus and sulfur are the troublemakers to watch for, since both increase the risk of hot cracking and, in certain plate orientations, lamellar tearing. Nitrogen can benefit grain refinement in some steels, however in excess it works against toughness. When choosing alloy steel welding electrodes, these tradeoffs are exactly what determine which filler is appropriate for the job.

Codes, Qualification, and the Reality of the Shop

With this many variables in play, the industry does not leave weldability up to guesswork. AWS D1.1 governs the structural welding of carbon and low-alloy structural steels, laying out requirements around material groups and their weldability. Meanwhile ASME BPVC Section IX governs welding and brazing procedure qualification, using P-Number groupings to determine when a procedure can be applied to similar base metals. Both sets of codes publish their own preheat, interpass, and heat-input tables, and those tables are written specifically to combat the hardening risks that come from higher CE, greater thickness, and heavier restraint.

Weldability is not a yes or no. It is a sliding scale. Drop the CE, reduce restraint, account for thickness, and slow the cooling rate with the correct procedure, and the steel becomes wonderfully cooperative. Raise any of those levers without compensating for it, and even the most experienced welder will be fighting metallurgy for the whole job. Long before the arc is struck, materials science has already decided whether the joint will hold.

The Heat-Affected Zone: Where Welding Rewrites the Microstructure

The weld bead may be the most visible part of a joint, but it is rarely the first place to fail. Directly next to it lies the heat-affected zone (HAZ): the band of base metal that experiences the welding thermal cycle without ever melting. With temperatures soaring high enough to completely re-organize its crystal structure, this unseen strip is where the metallurgy of the plate is rewritten. For most steel joints, it is the HAZ, not the weld metal itself, that governs the true integrity of the connection.

Mapping the Sub-Zones

Not every point within the HAZ undergoes the same change. As you move away from the fusion boundary, the peak temperature of the thermal cycle steadily drops, creating distinct sub-zones:

• Fusion Zone: The very edge where base metal and molten weld metal have mixed together. This boundary is often the transition point between two completely different chemistries.
• Coarse-Grained HAZ (CGHAZ): Exposed to temperatures above the upper critical transformation temperature (Ac3), this zone undergoes full austenitization. Here, austenite grains are able to grow uninhibited, leaving them extremely coarse once cooled.
• Fine-Grained HAZ (FGHAZ): Sitting between Ac3 and the lower critical temperature (Ac1), this region also fully austenitizes, however its peak temperatures are far lower. This limits grain growth, resulting in a much finer austenite grain size before transformation.
• Subcritical HAZ (SCHAZ): Heated below Ac1, this zone never forms austenite at all. Instead, any pre-existing microstructure is simply tempered or spheroidized, often resulting in a slight drop in hardness.
• Intercritical/Tempered HAZ: Found predominantly within quenched-and-tempered steels, this narrow zone is partially heated between Ac1 and Ac3. Portions of its microstructure transform to austenite, whilst others remain untransformed, creating a mixture that is notoriously tricky to temper predictably.

Cooling Rate Decides the Winner

Once austenite has formed, the real battle begins on cooling. The size of those austenite grains, paired with the cooling rate of the weld, determines which of the competing phases will form. Slower cooling favors softer, more ductile transformations such as pearlite and granular bainite. Faster cooling promotes harder phases like bainite and, at the most rapid rates, martensite. However, the phase that welding metallurgists strive for the most is acicular ferrite.

Unlike grain boundary ferrite which forms along prior austenite grain edges, acicular ferrite nucleates intragranularly, growing in chaotic, interlocking needles throughout the grain. This basket-weave arrangement is incredibly effective at deflecting cracks, making it one of the greatest allies to toughness. To achieve this, weld metals are deliberately engineered to be rich in oxide and nitride-forming elements, producing fine inclusions such as MgAl₂O₄ or TiN. These act as nucleation sites for acicular ferrite. As outlined by Bramfitt’s lattice-mismatch criterion, a crystallographic mismatch below 6% between an inclusion and ferrite is considered an effective nucleation phase, between 6–12% is medium, and above 12% is generally ineffective.

A Real-World Consequence

This isn’t theory for theory’s sake. A peer-reviewed study published within Scientific Reports on magnesium-treated EH36 ship plate showcased the drastic effect of heat input. Using laser confocal in-situ heating, austenitization was observed at 903 °C for the alloy. As heat input was raised from 150 kJ/cm to 350 kJ/cm, the t_8/5 cooling times extended from 54 seconds to 125 seconds, and the ferrite transformation temperature window widened from 98 °C to 151 °C. With this slower, broader cooling, acicular ferrite and granular bainite content fell, whilst grain boundary ferrite rose and coarsened. The result was devastating for low-temperature service: Charpy impact toughness at −40 °C plummeted from 174 J down to 67 J.

Measuring What The Eye Cannot See

Since these changes occur over mere millimeters, they cannot be judged by a pass-fail glance at the weld face. Instead, inspectors rely on quantifiable methods to expose them. Macrosections are used to reveal the width and shape of each HAZ sub-zone after etching, whilst microhardness traverses take a series of pinpoint hardness readings across the fusion line and into the base metal. Sudden spikes into hard martensitic regions, or excessive softening, are both red flags. These measurements are often the deciding factor in procedure qualification, proving that the thermal cycle was controlled well enough to preserve the properties that the design relies on. Without them, the most important metallurgy of the joint would remain entirely hidden.

Heat Input, Preheat and Interpass Temperature: Steering the Thermal Cycle

Welding doesn’t just melt metal, it forces a controlled burst of heat into the base plate. How much of that heat goes in, how fast it leaves, and what temperature the joint holds between passes will decide the fate of the heat-affected zone. The welder can control all three of these, and more often than not, this is the difference between a joint that stays tough and one that cracks.

Calculating Heat Input

The simplest way to quantify arc energy going into a joint is net heat input. For planning purposes, the formula is: Net heat input (kJ/mm) = (Voltage × Amperage × 60) ÷ (Travel speed in mm/min × 1000) × process efficiency.

To see this in action, take a common SMAW setup: 24 V and 180 A travelling at 76 mm/min. Without efficiency, that comes out to roughly 3.4 kJ/mm. However, not every amp is transferred into the plate. For planning, typical efficiencies are SMAW (~0.8), GMAW/FCAW (~0.9), SAW (~1.0), and GTAW (~0.7). These are planning values only. The governing code and WPS must always specify which efficiency to use.

It’s important to separate arc energy from net heat input. Arc energy is the electrical output, while net heat input is the adjusted value that more accurately reflects what the base metal actually experiences.

The Trade-Off Between Hot and Cold

Raising heat input slows the cooling rate, giving the HAZ more time to transform. Lowering heat input does the opposite. Both come with consequences. High heat input risks coarse grains, a softer HAZ, increased distortion, and lower toughness. Low heat input risks slag inclusion, lack of fusion, and cold cracking from fast cooling with trapped hydrogen.

This cooling rate is often tracked by t8/5, which is the time it takes for a point in the HAZ to cool from 800 °C to 500 °C. The longer the t8/5, the coarser the HAZ microstructure becomes. For this reason, AWS D1.1 sets maximum heat input values for prequalified joints based off carbon equivalent and thickness. Codes aren’t guessing here, they’re directly steering this cooling window.

Preheat, Interpass, and Keeping the Cycle Under Control

Preheat is the temperature the base metal is brought to before the first arc strikes. Its purpose is to slow cooling, allow any diffusible hydrogen more time to escape, reduce the temperature gradient, and relieve some restraint stress. Interpass temperature control is the upper limit for the joint between passes. Keeping it in range prevents cumulative heat from building up, avoids unnecessary HAZ tempering damage, and prevents re-austenitization of areas that were just carefully transformed.

In the shop, these temperatures aren’t left to a guess. Thermocouples provide the most precise readings, infrared pyrometers are fast for larger assemblies, and temperature-indicating crayons are a simple, reliable check for many joints. When working with more sensitive alloy steels, these measurements become non-negotiable.

Post-Weld Heat Treatment and Its Limits

Sometimes, steering the thermal cycle during welding isn’t enough. Post-weld heat treatment (PWHT) is used to stress relieve the weldment and temper the hardened regions of the HAZ. For many ASME Section VIII pressure vessel grades, such as SA-516 Gr 70, PWHT is specified around 590–620 °C.

However, PWHT isn’t a cure-all. On heavily restrained joints, holding at these temperatures can increase the risk of lamellar tearing, especially in plates with poor through-thickness ductility. Long cycle times, furnace availability, and distortion risks also need to be accounted for. Any PWHT must be driven by the code requirements for the service, not by habit.

At its core, heat input, preheat, interpass, and PWHT all boil down to one goal: control the cooling rate. Get that window right, and the metallurgy stays within the tight range that the design intended for it. Get it wrong, and the strongest bead in the world can still fail in the zone nobody sees.

Hydrogen, Cold Cracking and the Metallurgy of Prevention

Of all the metallurgical threats to a weld, hydrogen-assisted cold cracking is by far the sneakiest. Unlike a lack of fusion or slag inclusion that can be spotted the moment the arc goes out, cold cracking can lie dormant for hours, days, or even weeks before it reveals itself. By the time it does, the structure is often already under load, turning a joint that appeared sound into a catastrophic failure.

Where Does Hydrogen Come From?

Hydrogen never needs to be intentionally added to find its way into a weld puddle. Instead, it comes from a laundry list of everyday shop contaminants that break down under the intense heat of the arc. These include moisture in electrode coatings and flux, wet or damp filler metal, rust, mill scale, cutting oil, paint, and even humid air settling onto a cold plate. Once liberated, hydrogen takes the form of atomic hydrogen, allowing it to diffuse freely through the molten weld metal before the joint ever begins to cool.

The Three Conditions For a Crack

Hydrogen-assisted cold cracking doesn’t occur from hydrogen alone. It requires all three of the following conditions to coincide at the same time: diffusible hydrogen, a susceptible microstructure, and sufficient tensile stress from restraint. Remove any one of these factors, and the crack simply will not form.

The mechanism itself is a battle against the clock. As atomic hydrogen enters the puddle, it becomes trapped by the rapidly solidified and hardened heat-affected zone (HAZ). When the joint cools below the martensite transformation temperature, these trapped hydrogen atoms recombine into molecular hydrogen at discontinuities, such as grain boundaries or inclusions. The pressure from this gas builds to extreme levels, prying the hardened steel apart and opening internal cracks. This is exactly why the cracks form in the HAZ, not the weld bead, and why they are classified as cold cracks.

Measuring and Classifying the Threat

Not all consumables are created equal when it comes to hydrogen. ISO 3690 quantifies this by classifying diffusible hydrogen in deposited weld metal in milliliters per 100 grams (ml/100 g). These bands fall at H1 (5 ml/100 g), H2 (10 ml/100 g), H3 (15 ml/100 g), and H4 (20 ml/100 g), with lower numbers representing far less hydrogen in the deposit.

This classification directly ties into AWS electrode designations. For example, in E7018, the ‘8’ suffix denotes a low-hydrogen electrode with an iron powder coating and roughly 4% or less moisture in the coating. These consumables are specifically engineered to drastically limit the amount of hydrogen introduced into the joint, making them a staple for cracking-prone, hardenable steels.

Prevention in Practice

Stopping cold cracking is all about attacking each of those three required conditions. The most cost effective place to start is consumable control. Low-hydrogen electrodes are typically baked around 120–150 °C (250–300 °F) and stored in a holding oven, with most codes limiting the number of rebakes. From there, the joint must be mechanically cleaned to remove any hydrogen-generating contaminants. This level of cleanliness is universal, a lesson that applies just as heavily to steels as it does to the rigorous prep work required for processes such as those outlined in Aluminum Cleaner for Welding: How to Prep Aluminum Right.

With cleanliness secured, thermal control takes over. Applying the correct preheat and maintaining interpass temperature slows the cooling rate, allowing diffusible hydrogen time to escape from the HAZ before its microstructure hardens. Simultaneously, restraint must be reduced whenever possible. This can be accomplished through weld sequencing, back-stepping, or removing strongbacks before they lock excessive tensile stress into the joint.

The Delayed Danger

Perhaps the greatest hazard of hydrogen cracking is its delayed nature. Since hydrogen diffusion is time dependent, cracks often form well below the martensite transformation temperature, long after the hood has been hung up for the day. Because of this, non-destructive testing performed immediately after welding can pass, yet the joint can still fail weeks later. For this reason, soak periods and delayed-inspection requirements are written into many critical codes, especially for high restraint, hardenable steels.

In the end, hydrogen-assisted cold cracking isn’t a test of a welder’s hand speed or arc control. It’s a test of metallurgical discipline. Control hydrogen at the source, control the hardenability of the HAZ through thermal control, and control tensile stress through restraint management. Do all three, and the most deceptive crack in welding will never get the chance to start.

Filler Metal Selection and the Metallurgy of the Weld Deposit

The weld deposit is not just extra metal filling a gap. It is a deliberately engineered alloy whose strength, toughness and corrosion resistance must hold up for the whole service life of the joint. Your consumable choice either reinforces the care you took over preheat, joint preparation and heat input, or quietly undermines all of it.

Decoding the AWS Classification

To take the guesswork out of selection, the American Welding Society (AWS) uses a compact classification system for consumables. The most recognizable example for structural steel is E7018. Break it down and you know exactly what you are holding:

• E – Electrode, used for shielded metal arc welding (SMAW).
• 70 – Minimum tensile strength of the weld deposit in thousands of pounds per square inch. E7018 has a minimum of 70 ksi (484 MPa).
• 1 – Usable in all positions: flat, horizontal, vertical and overhead.
• 8 – Coating type. The ‘8’ denotes a low-hydrogen, iron powder coating, the reason E7018 is the go-to for many code-governed structural jobs.

The same logic carries into other processes. AWS A5.18 governs gas-shielded carbon steel wires, while AWS A5.9 governs solid and metal-cored stainless filler metals. Read ER70S-6 the same way: the 70 still sets the 70 ksi tensile minimum, the S marks a solid wire, and the trailing digit identifies a specific chemistry and mechanical combination. In every case the classification is a condensed spec sheet.

E7016 is worth knowing alongside it. Same 70 ksi tensile rating, same all-position capability, but a potassium-based low-hydrogen coating that gives a noticeably different arc character and handling profile from the iron powder E7018.

Matching, Undermatching and Overmatching

One of the biggest metallurgical decisions is strength matching. For common structural grades such as A36 and A572 Grade 50, matching strength fillers like E7018 are generally preferred, because you want the heat affected zone (HAZ) to govern the joint rather than the weld metal itself.

An overmatching filler has higher tensile strength than the base metal. It looks safer on paper, but raising weld metal hardness relative to the base plate creates a mismatch that can shift the weakest point to the fusion boundary, a narrow transition zone already among the most crack-prone areas of the joint when hydrogen is present.

Undermatching is used deliberately in some applications to promote ductility, but for most structural carbon steel work, matching strength is the soundest and most code-aligned approach. For consumables used on higher strength alloys, see our guide to alloy steel welding electrodes.

Dilution: The Hidden Variable

Not every pound of filler that goes into the rod ends up with its original chemistry. Dilution is the fraction of base metal melted into the pool and mixed with the filler, and it changes the effective alloy content, hardness and toughness of the deposited metal.

The effect grows in high-dilution passes such as vertical-up and overhead welds, where gravity, travel speed and arc manipulation melt more of the surrounding base metal. The more diluted the deposit, the more it inherits the chemistry of the base plate, so procedures often call for a richer or more forgiving filler to compensate.

The Role of Flux in SMAW and FCAW

With SMAW and flux-cored arc welding (FCAW), the flux is not just a shield. It is an active metallurgical ingredient: it provides gaseous protection from the atmosphere, assists deoxidation to strip oxygen from the pool, and can add alloying elements to fine-tune the final deposit. That is exactly why two rods with identical strength classifications can behave very differently from each other. Their flux formulations are not the same.

Filler Metals for Corrosion and Stainless Steels

For corrosion resistant work, the metallurgy changes. For austenitic stainless steels such as 304 and 316, ER308L is the standard matching filler for 304/304L, while ER316L serves 316/316L. The ‘L’ matters: it marks a low carbon grade (typically 0.03% maximum) that minimizes chromium carbide precipitation at grain boundaries, sharply reducing the risk of sensitization and intergranular corrosion.

These principles extend beyond austenitics. Ferritic and martensitic stainless steels need matching fillers with tight heat input control to avoid excessive grain growth or unwanted martensite. For niche alloys such as Alloy 20, very specific filler selections are required to hold up in the harshest chemical environments.

Post-Weld Heat Treatment and Practical Selection

For most carbon steel joints welded with matching, low-hydrogen fillers, PWHT is not required for strength alone. When a code does mandate PWHT, however, not all fillers respond identically. Heavily overmatched or highly alloyed deposits can show greater tempering changes through the cycle, which makes filler selection central to keeping final properties inside specification.

Filler Metal Family	Compatible Base Metal	Typical Applications
E7018 / ER70S-6	Carbon & Low Alloy (A36, A572)	Structural steel, buildings, bridges, pressure work
ER308L	304 / 304L Stainless	Food processing, tanks, general corrosion resistant fabrication
ER316L	316 / 316L Stainless	Marine environments, chemical processing, chloride exposure
E410 / ER410	410 / 420 Martensitic Stainless	Wear parts and pump components, needing tight heat input control
ER430	430 Ferritic Stainless	Exhaust systems and decorative work where weldability is prioritized

The filler metal is not a filler at all. It is an engineered bridge between two metals, and choosing it with metallurgy in mind makes that bridge as strong and as reliable as the structure it holds together.

One Weld, Five Material Families: Why Each Metal Fails Differently

Heat changes the metal, and the metal’s chemistry decides what the heat produces. The principle is identical in every family; the numbers are not. A 200 °C interpass limit in low-alloy steel is a decision about martensite, while in duplex stainless that same number decides whether the deposit holds 30 percent ferrite or 70.

A useful anchor is the magnesium-treated EH36 ship plate study in Scientific Reports, where raising heat input from 150 to 350 kJ/cm cut Charpy toughness at −40 °C from 174 J to 67 J. That is a carbon steel number, and no other family inherits it.

Carbon and low-alloy steel: the HAZ is the joint

In carbon and low-alloy steel, failure lands in the heat-affected zone, not the bead. The deposit is over-matched for toughness; the plate is not. Fast cooling through the critical range makes martensite or untempered bainite, and that hard HAZ cracks under restrained shrinkage stress. Preheat counters this: it slows cooling, lets hydrogen diffuse out, and leaves a tougher HAZ. That is why the preheat table in AWS D1.1 is not paperwork, and why a 4130 tube and an A36 plate never share an interpass limit. Choosing alloy steel welding electrodes and setting preheat are one decision made twice.

Stainless steels: four families, four different weaknesses

Austenitic grades such as 304 and 316 cannot be hardened by heat treatment, so the martensite story does not apply. Their signature failure is sensitization: between roughly 600 and 850 °C, chromium carbides precipitate at austenite grain boundaries, stripping chromium from the boundary. The result is intergranular corrosion, or weld decay, which can surface years later. The fix: L grades (304L, 316L) with extra-low carbon, and stabilized fillers such as 347 that tie carbon up in niobium carbides. The other austenitic problem is hot cracking, from sulfur and phosphorus segregation forming low-melting films in the last liquid to solidify.

Ferritic grades (430, 444) have no austenite to begin with, so they cannot produce martensite, but they grow coarse grains fast unless heat input is controlled. Martensitic grades (410, 420) are hardenable by design, so their welds need preheat and post-weld heat treatment to avoid brittle as-welded hardness.

Duplex stainless is the most demanding. 2205 balances ferrite and austenite, and that balance is the point. Excess heat input or a long interpass dwell drives the wrong phase ratio or allows sigma phase, a brittle chromium-molybdenum-nickel intermetallic forming in the 600 to 1000 °C range. Heat input and interpass temperature must be controlled together, which is why bodies such as TWI frame duplex welding around low heat input and a cool interpass.

Aluminum: the same physics, a different set of habits

Aluminum breaks carbon steel habits in four ways: it conducts heat several times faster, so heat races away instead of building a hardened zone; it expands roughly twice as much per degree, so restraint and distortion arrive quickly; it carries a tenacious oxide film that must be gone before the arc strikes; and it dissolves hydrogen readily while the pool is liquid.

The oxide is the shop-floor issue. Clean it mechanically or with AC/DC TIG cleaning before welding; this guide to prepping aluminum for welding covers the methods. Oxide left behind traps gases and creates porosity. Add hydrogen porosity from wet plate or damp filler, plus hot cracking, because aluminum alloys freeze over a wide range and shrink as they solidify, pulling the last liquid apart.

6061-T6 is the most commonly welded aluminum alloy, run with ER4043 or ER5356 filler. Since T6 comes from artificial aging at a controlled temperature, welding partly destroys the temper. The joint recovers by natural aging at room temperature over a few days, but not instantly. Process notes are collected in this practical aluminum MIG welding guide.

The 7xxx series is more temperamental: stronger, wider freezing ranges that encourage hot cracking, and stress-corrosion sensitive. Here the gap between good practice and a stress-corrosion failure can be a few hours of waiting. Welded joints are normally welded and post-weld aged with a minimum delay, since retained strength loss and stress-corrosion sensitivity grow the longer a part sits before aging.

Nickel alloys: easy to weld, easy to ruin

Inconel 625 and Hastelloy grades weld beautifully with matching filler and clean inert shielding. Their weakness is subtler: grain-boundary liquation cracking. Low-melting eutectics segregate at grain boundaries, and if the weld cools slowly through the vulnerable range, or if heat input is too low to reflow them, those films open as cracks. Oddly, the answer is to cap heat input and interpass temperature so the weld freezes quickly. Alloy 20 welding shows the same discipline in a corrosion-resistant nickel-iron-chromium grade.

Hardfacing and cladding: when the deposit is the product

Here the weld metal is not an afterthought; it is the reason for the job. Hardfacing deposits are engineered around wear mechanism: high-chromium alloys for abrasion, cobalt-based or nickel-based alloys for impact and galling, carbides for erosion. Cladding and overlay put a wear- or corrosion-resistant layer over carbon or low-alloy steel. Dilution still matters, but it is a chemistry question here rather than a strength question, and filler choice, pass count, and dilution allowance decide whether you get the alloy you paid for.

Family	Dominant failure mode	Controlling variables
Carbon/low-alloy steel	Martensitic HAZ, cold cracking	Preheat, interpass, heat input, hydrogen
Austenitic stainless	Sensitization, hot cracking	L grades, stabilized filler, S and P control
Ferritic/martensitic stainless	Coarse grain, brittle hardness	Heat input, preheat, PWHT
Duplex stainless	Ferrite imbalance, sigma phase	Heat input and interpass limits
Aluminum	Porosity, hot cracking	Cleaning, filler, cooling rate, aging timing
Nickel alloys	Liquation cracking	Heat input and interpass caps

Compare the numbers across families and the pattern is clear. The EH36 result proved that for carbon steel, more heat input meant less toughness. The same test on duplex stainless, 6061-T6, or Inconel 625 would give a different threshold, a different damage mechanism, and a different repair plan. Generalize the method, never the numbers.

Strength, Toughness and Distortion: The Properties the Design Actually Needs

When an engineer specifies a weld, the question is not whether the bead looks clean. It is whether that joint will carry its design load, remain intact in its service temperature, and hold its shape for the life of the structure. All three of those answers are determined by the metallurgy left behind after the thermal cycle, not just the appearance of the weld metal.

The Basics: Strength, Ductility and Hardness

To understand a welded joint, there are a few mechanical properties that must be distinguished from one another. Tensile strength is the maximum load a material can withstand before fracture, while yield strength is the stress at which it begins to deform permanently. A joint must meet at least the minimum yield and tensile values required by its design, but meeting them on paper is only half the battle.

Ductility is what allows a structure to deform without sudden failure, and it is most often measured by elongation and reduction of area during a tensile test. Hardness, on the other hand, is a localized resistance to indentation. In the heat-affected zone (HAZ), hardness is especially telling. An overly hard HAZ greatly increases the risk of hydrogen-assisted cold cracking, while an excessively soft HAZ can become the weak link under load and lead to low-strength or even collapse failures.

For joints that will see wear, abrasion, or high restraint, the selection of the correct alloy steel welding electrodes can help manage hardness in these sensitive zones without sacrificing the ductility the design requires.

Toughness: The Line Between Bending and Brittle Fracture

Strength keeps a structure from stretching past its limit, but notch toughness keeps it from snapping without warning. Measured by the Charpy V-notch impact test, notch toughness is a measure of a material’s ability to absorb energy in the presence of a notch, and it is the best proxy for resistance to brittle fracture below the ductile-to-brittle transition temperature.

For service in colder climates, toughness, not strength, is almost always what governs. This is proven by the peer-reviewed magnesium-treated EH36 ship plate study. As heat input increased from 150 to 350 kJ/cm, Charpy impact toughness at −40 °C fell from 174 J to 67 J. That drastic reduction shows how an increase in heat input can coarsen the HAZ microstructure and leave a joint far more vulnerable to brittle fracture, even if its tensile strength remains within spec.

Residual Stress and Distortion

Toughness can be protected, but thermal movement is unavoidable. As a weld pool solidifies, a local volume of steel rapidly expands and contracts. When that movement is restrained by the surrounding, much cooler metal, it cannot freely return to its original shape. The result is residual stress locked into the joint, often reaching very high levels in the HAZ and fusion zone.

Distortion scales with restraint and inversely with yield strength. In other words, high-strength steel distorts more for the same thermal cycle than a mild steel would, because it has less plasticity to accommodate the contraction. To combat this, fabricators rely on proven controls such as preheat where warranted, balanced symmetric sequences, back-stepping, and stiffening. For larger or more critical structures, these controls are often verified through testing.

According to TWI, its materials testing and metallurgy services include residual stress measurement, hydrogen testing and sour testing alongside standard mechanical testing. For stress relief, options such as vibration stress relief can be considered for large structures. However, caution is required: lamellar tearing is a through-thickness stress-relief cracking risk in restrained, heavily inclusion-bearing steel, so PWHT and rapid cooling cycles require extreme care.

At the end of the day, the properties that decide a weld’s fate are rarely the ones on the weld cap. They are the yield strength, ductility, notch toughness, and locked-in stress within the HAZ. By planning for all of them before striking an arc, a fabricator isn’t just making a weld, they’re ensuring it will perform exactly as the design intended.

Common Welding Metallurgy Mistakes and How to Diagnose Them

Most weld failures aren’t down to a poor bead, but to unseen changes in the base metal’s microstructure. Knowing what to look for, and when to look for it, can stop a metallurgical reaction from turning into a costly failure.

Hydrogen Cold Cracking

Hydrogen cold cracking is one of the most deceptive defects. It can form hours, days, or even weeks after welding, making an immediate visual inspection unreliable. For it to occur, three factors must align: diffusible hydrogen, a hard and susceptible microstructure in the heat-affected zone (usually untempered martensite or bainite), and enough tensile stress from restraint.

These fine, branched cracks are typically found in the hard HAZ or along the fusion boundary. Prevention comes down to control: use low-hydrogen consumables in line with their AWS classification, bake and store them correctly, apply adequate preheat and interpass temperatures to allow hydrogen to diffuse out, and reduce joint restraint where possible. For hydrogen-susceptible joints, delayed inspection is often required.

Hot Cracking (Solidification Cracking)

Unlike cold cracking, hot cracking forms while the weld pool is still solidifying. As the last liquid films between dendrites cool, segregated elements such as sulfur and phosphorus form low-melting eutectics. Under solidification shrinkage, these films tear, creating longitudinal cracks along the weld centerline or fusion boundary.

This is most common with austenitic stainless steels and many aluminum alloys. For these, filler metal selection is your strongest defense. Using the correct matching or intentionally over-alloyed filler, thorough joint cleaning (especially for tasks like aluminum MIG welding), lower restraint, and tight heat input control all help to limit risk.

Lamellar Tearing

Lamellar tearing is a through-thickness defect. Instead of a single crack, it forms jagged, stepped cracks that run parallel to the plate surface. This is caused by high through-thickness tensile stresses pulling apart planes weakened by elongated manganese sulfide (MnS) inclusions. It’s most often seen in heavy, highly restrained T-joints and corner joints made from rolled plate that wasn’t designed for short-transverse loading.

Prevention needs to be structural. Joint redesign to redirect stress, lower restraint, adequate preheat, and the use of softer capping passes can all help. As TWI notes, for thick-section steels, material selection is just as important as the welding procedure itself.

Lack of Fusion, Cold Lap, and Porosity

Not all defects come from phase changes. Lack of fusion is when weld metal fails to bond to the base metal or previous passes. Cold lap is its subtler form, where molten metal flows over a colder surface without true metallurgical bonding. Both are caused by insufficient heat input, poor travel angle, or incorrect bead placement. Changing the filler metal alone will not fix either. Raising heat input within the qualified range or reorienting the joint is the correct approach.

Porosity has several causes. Hydrogen porosity can come from damp electrodes, rusty plate, or oil. Gas-shielded porosity is often due to nitrogen or oxygen intrusion from gas leaks, incorrect flow rates, or excessive drafts. Slag or wormhole porosity is flux-related, usually from excessive arc length, wide weaves, or poor cleaning between multipass welds. For all of these, discipline is key: consumable storage, gas purity, interpass cleaning, and solid arc technique. For gas-shielded processes, proper surface prep, such as using a dedicated aluminum cleaner for welding, is often just as critical as the shielding itself.

Overheating, Burning, and Lost Toughness

There’s a narrow window between hot enough and too hot. Excessive heat input, especially over multipass welds, causes grain growth in the coarse-grained HAZ. This coarsened structure and the loss of tough, intragranular acicular ferrite in favor of grain boundary ferrite can drastically reduce notch toughness. This was quantified in a peer-reviewed Scientific Reports study on EH36 ship plate, which found Charpy impact energy dropped at sub-zero temperatures as heat input increased.

In extreme cases, this can lead to oxide entrapment or burning. The only reliable guard against it is to stay within the heat input limits of your qualified procedure, control interpass temperature, and use controlled passes over wide, slow weaves. For high-strength steels, these limits are non-negotiable. When selecting consumables, a resource such as alloy steel welding electrodes can help choose products less sensitive to this trade-off.

Diagnostic Table: Symptom to Cause to Corrective Action

Symptom	Likely Metallurgical Cause	Prevention & Corrective Action	Inspection Route
Fine, branched cracks in HAZ (delayed)	Hydrogen + hard microstructure + restraint	Low-hydrogen consumables, bake/storage, preheat, interpass, and restraint control	MT/PT, or UT with delayed inspection
Longitudinal crack in weld centerline or fusion boundary (during/at end of solidification)	Segregation & low-melting grain boundary films (
Longitudinal crack in weld centerline or fusion boundary (during/at end of solidification)	Segregation & low-melting grain boundary films (hot cracking)	Use correct filler (match or over-filler), clean base metal, lower restraint, optimize heat input	RT, PT, or macrosection
Stepped cracks perpendicular to surface, through thickness	Through-thickness stress + elongated MnS inclusions (lamellar tearing)	Joint redesign to relieve short-transverse stress, preheat, soft capping passes, lower restraint	UT (highly sensitive) or macrosection
Unbonded edges, ropey bead with lack of tie-in	Too little heat input or poor bead placement (lack of fusion/cold lap)	Raise heat input within qualified range or reorient joint. Do not rely on filler metal change.	Macrosection, RT, or UT
Scattered round voids (gas pockets)	Hydrogen, N₂/O₂ from moisture, leaks, drafts, or contaminated gas	Consumable storage, gas purity/flow, remove rust/oil/moisture, shield from drafts	RT, or cross-section for distribution
Irregular voids/trails or flux pockets	Slag entrapment or wormhole porosity (flux/arc practice)	Proper arc length, stringer beads over wide weaves, clean between passes	RT, or macrosection
Dull, coarse HAZ with hard spots and brittle fracture	Excessive heat input causing grain growth & loss of acicular ferrite	Abide by heat input limits, control interpass, use controlled passes, follow WPS	Hardness traverse, Charpy (if specified), macrosection

At the end of the day, the key to diagnosing any of these defects is to stop treating the crack, pore, or lack of bond as a “welding error”, and start treating it as a metallurgical reaction. Every one of the above can be prevented, but only if the thermal cycle, chemistry, and stress state are controlled just as deliberately as the weld bead itself. By following your procedure, qualifying your consumables, and inspecting with timing in mind, you won’t just pass inspection, you’ll prevent a failure that may never show its face until it’s far too late.

Applying Materials Science on the Shop Floor: A Practical Checklist

Metallurgy is not theory that lives in a manual. It is a sequence of decisions made before the arc is struck, decisions that show up months later in a crack report. The good news is that most of those decisions are not hard. They just have to happen in the right order, every time, and get written down where the next welder can find them.

Here is a nine-step sequence that turns the concepts in this guide into shop practice. It follows the order of the welding process itself, from identifying the steel to learning from a failure.

Step 1: Identify the Base Metal From the Mill Certificate

Never weld unidentified steel. Heat numbers and mill certificates travel with the material and tell you the grade, the specification, and the actual chemistry. That document is the difference between a controlled weld and a guess.

Match what you see on the plate to a material designation and then to its specification: ASTM A36 for general structural carbon steel, A572 Grade 50 for higher-strength structural plate, A516 for pressure vessel plate, A106 for pipe, A240 for stainless plate. Note also the thickness, because thickness changes the cooling rate and therefore the preheat requirement even when the chemistry stays the same.

Steel that arrives without documentation should be quarantined, not welded on a maybe. If you weld aluminum instead, the same discipline applies, including the joint prep work covered in our guide to prepping aluminum for welding, since surface chemistry drives porosity and cracking more than it does on steel.

Step 2: Establish the Governing Code and Its Limits

Once you know the material, you know which specification controls the job. AWS D1.1 governs structural carbon and low-alloy steel, D1.5 covers bridge welding in coordination with AASHTO, D1.2 covers aluminum structures, and AMPP publishes the materials-protection standards and qualification resources that control pipeline work under API 1104. For pressure equipment, ASME BPVC Section IX governs procedure qualification.

Then pull the three numbers the code actually specifies for your joint: minimum preheat, maximum heat input, and interpass temperature range. Code tables supersede shop rules of thumb, and the preheat table typically varies with carbon equivalent, thickness, hydrogen level and restraint. Do not interpolate from a remembered value. Read the table.

Step 3: Select the Filler Metal Deliberately

Pick the consumable for the base metal, the service environment, the position and the restraint, and write down why. Record two things the WPS should capture: dilution, the proportion of base metal that melts into the pool and dilutes the deposit chemistry, and the effect of welding out of position. Positional welds pool differently, cool differently, and often pick up more dilution, so a filler chosen for flat work may run lean in vertical or overhead. When you need a rod reference for vertical and overhead work, our all-position welding rod guide breaks down the classifications by position capability.

For sour service, the choice is driven by hydrogen-induced cracking and sulfide stress cracking, which is exactly where AMPP standards and qualification resources come into play.

Step 4: Build the WPS in Writing

Heat input, preheat, interpass range, cleaning requirements and welding sequence all belong in the written procedure. AWS and ASME both require qualified procedures, and the qualification records are the evidence that the procedure works on this material, in this position, with this filler. If the procedure is not qualified for the joint, the joint is not under control.

Step 5: Control Consumables

Storage ovens, bake logs, rebake limits and stock rotation are not housekeeping niceties. Low-hydrogen electrodes held at roughly 120 to 150 °C, properly baked and rebaked within the manufacturer’s limit, are one of the cheapest ways to keep hydrogen out of a deposit. Log what went into the oven and what came out, and rotate stock so older material gets used first rather than sitting until it ages out of specification.

Step 6: Control the Joint Condition

Grind-clean the joint. Remove mill scale, rust, paint, moisture and oil. Every one of those contaminants is a hydrogen source or a source of trapped gas, and cutting oil or a wet cold plate is enough to trigger delayed cracking in a hard HAZ. Mechanical cleaning beats chemical cleaning for most carbon steel work.

Step 7: Measure and Record

The three standard tools for verifying preheat and interpass temperature are temperature-indicating crayons, thermocouples and infrared pyrometers. Use calibrated instruments, and log the actual values during production welding, not just the planned values. Modern welding machines can log heat input directly from arc voltage, current and travel speed, which gives you an audit trail when a question comes up later.

Step 8: Inspect With the Right Method and the Right Timing

Match the method to what you are hunting. Radiography and ultrasonic testing find internal defects. Magnetic particle and dye penetrant testing find surface-breaking defects. A hardness traverse measures what a normal inspection cannot see, namely the HAZ hardness that predicts cracking susceptibility. A macrosection is a qualification tool, and macroetching a completed joint reveals the fusion boundaries, grain structure and any defects that eye inspection misses.

Step 9: Think in Metallurgy When Something Goes Wrong

When a joint fails, resist the reflex to blame the welder. Ask what the metal did. A crack in a hard HAZ hours or weeks after welding points to hydrogen and restraint. A centerline crack in stainless or aluminum points to solidification and segregation. Stepped cracks through the thickness point to lamellar tearing. Coarse grains and a soft HAZ point to too much heat. Each diagnosis sends you to a different correction, and only one of them might involve the welder.

The One-Page Checklist

1. Identify the material from the mill certificate and find its ASTM specification.
2. Determine the governing code and pull the preheat, heat-input and interpass values from its tables.
3. Select the filler metal for the base metal, service, position and dilution.
4. Use a qualified WPS with all thermal parameters written down.
5. Store and bake consumables correctly, with logs and rebake limits honored.
6. Clean the joint of scale, rust, paint, oil and moisture before welding.
7. Measure and record preheat, interpass and heat input during the weld.
8. Inspect with the correct method and allow for delayed-inspection requirements.
9. Diagnose failures metallurgically before assigning blame.

Keep Learning: Certification Tracks That Build This Knowledge

Most of what a code demands ultimately reduces to material and process control, and the people who control it best are the ones who understand why. AWS publishes qualification and certification tracks that build exactly that understanding, including the Certified Welding Inspector, Senior CWI, Certified Welding Supervisor, Associate Welding Inspector and the Welding Inspector endorsement. AWS also offers free welding resources and standards webinar series, which makes them a practical way for shop personnel to keep current without a formal course.

Materials science is what turns a code requirement into an engineering reason. Once a welder understands why preheat slows the cooling rate, why a low-hydrogen electrode matters, and why the HAZ is the weak link, the checklist stops being paperwork and starts being the reason the bridge bracket is still standing four months later.

Frequently Asked Questions

What is materials science in welding, in simple terms?

It is the study of how a metal’s chemistry, microstructure and thermal history determine how it behaves — and therefore how it must be welded. In practice it means reading the base metal’s specification, controlling heat input and temperature, choosing the right filler metal and consumable handling, so the joint ends up with the strength and toughness the design assumed.

Why does the heat-affected zone (HAZ) matter more than the weld metal?

The HAZ is base metal that was heated but not melted, so its original controlled microstructure was destroyed and replaced with whatever formed during the weld’s cooling cycle — often a harder, less tough structure. The weld metal is designed to be strong; the HAZ is an accident, and it is usually where cracking and fracture initiate.

How much heat input is too much for a structural steel weld?

It depends on carbon equivalent, thickness and the governing code. High heat input coarsens grains, widens the HAZ and lowers toughness — in one published EH36 study, toughness at −40 °C dropped from 174 J to 67 J between 150 and 350 kJ/cm. Always take the maximum heat input from the code’s prequalified-joint table rather than a generic rule of thumb.

Do I really need preheat on mild steel?

Often not, but the decision must be made from the material’s CE, thickness, restraint and hydrogen level rather than from habit or shop folklore. Thin, low-carbon, lightly restrained joints usually need no preheat, while thick, higher-carbon or heavily restrained joints frequently do. Codes publish preheat tables for exactly these variables.

What is the cheapest way to reduce hydrogen cracking risk?

Consumable discipline. Store low-hydrogen electrodes in a holding oven, bake them per the manufacturer’s and code’s instructions, respect rebake limits, rotate stock, and clean the joint of rust, mill scale, oil, paint and moisture before welding. These steps remove the hydrogen source directly.

How do I know if my weld has a hydrogen problem versus a geometry problem?

Hydrogen cracking shows up as fine, often internal cracks in a hard HAZ or at the fusion boundary, and it can appear days or weeks after welding. Lack of fusion and cold lap, by contrast, are visible at the joint or on radiographs as an unfused interface caused by insufficient heat or poor bead placement. A macrosection settles it quickly.

Is it harder to weld stainless steel or aluminum?

They fail for completely different reasons, so neither is simply harder. Stainless risks sensitization, intergranular corrosion and hot cracking, and needs matching low-carbon filler plus heat-input discipline. Aluminum needs oxide removal, protection from hydrogen, control of solidification cracking, and prompt post-weld aging of 6061-T6.

Do I need to be a materials engineer to apply this on the job?

You need the habits, not the job title: identify the material, use the code’s tables, control heat and temperature, handle consumables properly, and record what you did. AWS certification tracks such as CWI, SCWI, CWS and CAWI build and document that skill set, and most code requirements ultimately reduce to material-and-process control.

Can post-weld heat treatment fix a bad weld?

It can relieve residual stress and temper a hard HAZ, but it cannot remove hydrogen already trapped, reverse coarse grain growth, or repair lack of fusion. On restrained joints it can even cause lamellar tearing. PWHT is a designed process step, not a repair tool.

How do I calculate heat input on the shop floor?

Use net heat input in kJ/mm = (voltage × amperage × 60) ÷ (travel speed in mm/min × 1000), then apply the process efficiency specified by the governing code or WPS. For example, 24 V and 180 A at 76 mm/min gives about 3.4 kJ/mm before efficiency. Many modern welding power sources log heat input automatically, so verify the setting rather than assuming it.