What Is Slag in Welding? Definition, Types, and Removal

A welder runs a beautiful root pass on a pressure vessel, grinds it flush, and weeks later an ultrasonic test finds a 40 mm line of trapped slag where no one ever thought to chip. Slag is the ordinary glassy crust everyone scrapes off without thinking — and the reason rejects happen when it is left behind.

What Is Slag in Welding? A Clear Definition

Slag in welding is the vitreous (glassy), non-metallic byproduct that forms on and around the weld bead when flux or the electrode coating melts, reacts with oxygen and impurities in the molten pool, then solidifies on top of the cooling metal. Put plainly, it is the crust you chip, brush or peel off when the pass is done. Weld slag is not wasted filler metal, and it is not a defect by itself. It is a deliberate chemical product of the flux, and its appearance tells you whether your parameters were actually right.

Where the Word Comes From

The term long predates arc welding. In metallurgy generally, slag is the vitreous byproduct of smelting ore: the stony fraction that melts, rises and is skimmed off so the purified metal remains. Welding applies that same separation principle inside a puddle only a few millimetres wide. Flux melts, dissolves oxides and contaminants, and floats to the surface as a protective blanket. That blanket is flux slag, and structurally it is an amorphous solid — frozen glass rather than a crystalline metal, which is exactly why it shatters instead of bending.

Density explains why it sits on top. Slag typically runs about 2.5–3.5 g/cm³ against roughly 7.8 g/cm³ for steel, so it is much lighter than the pool beneath it and rises instead of sinking. That buoyancy is the whole point: it lets slag carry iron, manganese and silicon oxides up and out of the weld metal rather than leaving them buried in the bead.

What Welding Slag Looks Like

Most welders meet weld slag first as a black or dull grey crust riding on the bead, sometimes glossy, sometimes carrying a bluish heat tint. It is brittle, so a light tap cracks it and it lifts away in curved shards or glassy chunks. On flux-cored and submerged arc welds it often peels in long strips; on stick welds it usually breaks into flakes. A grey, dull, heavy crust instead of a shiny one is an early hint that heat input or travel speed was off.

Which Processes Produce Slag, and Which Don’t

• SMAW (stick): the electrode’s slag coating melts into a complete crust that must come off between every pass.
• FCAW (flux-cored): gas-shielded and self-shielded wires both leave slag, often thicker and more viscous than stick slag.
• SAW (submerged arc): granular flux melts into a continuous glassy layer, frequently liftable in one piece.
• Electroslag welding (ESW): the special case where molten slag is the resistance-heating medium itself, so slag is the process rather than a byproduct.

You will not find true slag in GTAW or in GMAW using an inert gas, because no flux is involved. What appears instead are silica islands: small glassy beads left behind by deoxidizers in the filler wire, sometimes with thin brown deposits. They resemble welding slag but are a different animal. This distinction matters most in aluminium work, where surface deposits are a routine nuisance and proper prep starts long before the arc, as our Aluminum Cleaner for Welding: How to Prep Aluminum Right guide explains.

Slag Is Not Spatter, Dross or Mill Scale

Beginners lump four different things together, and each needs its own removal method:

Material	What it is
Weld slag	Non-metallic flux-derived crust on the bead
Weld spatter	Tiny droplets of molten metal thrown from the arc onto surrounding plate
Dross	Re-solidified oxide on oxy-fuel, plasma or laser cut edges
Mill scale	The blue-black Fe₃O₄ oxide layer hot-rolled onto new steel

Spatter is metal, so it needs grinding. Slag is non-metal, so it chips and brushes. Mill scale comes off the plate before you ever strike an arc. A full side-by-side comparison follows in the next section.

Why Electrode Classification Matters

Not all slag coating behaves the same way. Under the AWS A5.1 specification for carbon steel stick electrodes, classifications group electrodes by coating type — cellulose, rutile, iron oxide and basic (low-hydrogen) — and each produces slag with a different appearance, thickness and removal effort. Rutile electrodes generally leave a thin, easy-lifting slag; basic low-hydrogen electrodes leave a denser crust that takes more work; cellulose-heavy electrodes produce very little slag at all. Recognising which coating you are running tells you what to expect before you strike the arc.

How Welding Slag Forms: Flux, Shielding Gas, and the Arc

To understand slag in welding, you have to follow what happens in the few tenths of a second when an arc strikes the plate. The temperatures involved reach several thousand degrees at the arc core, hot enough that bare iron oxidizes almost instantly if it touches air. Welding slag is essentially the controlled, deliberate product of that oxidation — the flux is engineered so the “rust” you get is a useful liquid instead of a defect. Here is the step-by-step chemistry.

Step 1: The Arc Decomposes the Flux Coating

Stick a coated electrode into an arc and the covering does not simply melt — it decomposes. Binders and carbonates in the coating break down under the heat, releasing a cocktail of shielding gases: carbon monoxide, carbon dioxide, water vapor, and fluoride vapors. These gases push the atmosphere away from the molten pool, excluding the nitrogen and oxygen that would otherwise embrittle and oxidize the weld metal. In this sense, stick welding’s coating is its own built-in shielding system, no gas cylinder required.

The same decomposition happens in flux-cored and submerged arc welding, where the flux does the job in granular or tubular form rather than as an electrode covering. The result is the same: a bubble of protective gas right where the metal is hottest.

Step 2: Molten Flux Forms a Liquid Blanket

While the gases shield the arc, part of the flux melts into a liquid — a molten silicate, titanate, or fluoride mixture that is lighter than steel, so it floats on top of the weld pool. This liquid blanket is where the real metallurgical housekeeping happens.

As the iron, manganese, and silicon in the pool react with any residual oxygen, they form oxides. Left alone, those oxides would end up as inclusions scattered through the bead. Instead, the liquid slag absorbs them, much like a sponge. It also dissolves and locks up impurities such as sulfur and phosphorus, which otherwise promote hot cracking. The clean, deoxidized steel solidifies underneath while the contaminated slag floats up and freezes on top — a refining action well documented by welding research institutes like TWI.

Step 3: The Slag Blanket Insulates the Cooling Bead

Slag is a poor conductor of heat, and that matters. By blanketing the hot bead, it slows the cooling rate, giving the solidifying metal more time to form a finer, more even grain structure. In many processes, that translates directly into better toughness in the finished weld.

The blanket also blocks hydrogen. Hydrogen is the enemy of high-strength steel welds, driving cold cracking and hydrogen-induced stress cracking, so keeping it out of the hot metal is a genuine metallurgical benefit, not a side effect. This is one reason low-hydrogen practice — dry electrodes, clean flux — dominates structural welding codes.

What’s in the Coating: Basic vs. Rutile Electrodes

Not all slag chemistry is the same, and the electrode family tells you a lot about the slag you will chip off afterward:

• Basic (low-hydrogen) electrodes rely heavily on calcium carbonate (limestone) and calcium fluoride (fluorspar) in the coating. They produce a relatively viscous, stubborn slag that must be removed carefully, but they deliver weld metal with low hydrogen content and excellent toughness — the workhorses of structural and pressure work.
• Rutile electrodes are built around titanium dioxide. They produce a more fluid slag that solidifies into a thin, glassy crust, re-solidifies smoothly, and peels off far more easily. That is why rutile rods are favorites for root passes, cosmetic welds, and beginners — the slag chemistry itself is tuned for easy removal.

Choosing between them means choosing the slag behavior you want. If you are picking a rod for a specific job, our guide to alloy steel welding electrodes covers how coating chemistry shifts with the base metal.

Why Cold Slag Is Brittle and Glassy

Once the weld cools, that liquid silicate mixture has no metallic bonding at all — it is essentially a glass. Silicate glasses are amorphous solids with a disordered atomic structure, so they fracture easily along conchoidal (shell-shaped) cracks and snap free from the bead under a light tap from a chipping hammer. This brittleness is a feature, not a flaw: it lets you break the slag off cleanly even on out-of-position welds without digging into the metal underneath. It is also why slag can shatter into sharp, hot shards that demand eye protection — a point we return to in the safety section.

So when you see that black or bluish crust peel back from a fresh bead, you are not scraping off waste. You are watching a miniature smelting operation finish its job: the flux has shielded, refined, insulated, and now leaves the clean metal behind.

Types of Welding Slag by Process: SMAW, FCAW, SAW, and Beyond

Slag is not one material with one behavior. The crust you chip off a stick weld behaves nothing like the glassy sheet you lift from a submerged arc weld, and while MIG welding, you may see a deposit that is not technically slag at all. Understanding how slag in welding behaves process by process is the difference between fighting it and predicting it.

SMAW (Stick Welding): The Classic Crust

In shielded metal arc welding, the electrode coating does double duty: it burns to create shielding gas, and the molten portion forms a complete slag crust over the bead. That crust must be chipped away before every subsequent pass, no exceptions. What surprises beginners is how much the electrode classification changes the job:

• Rutile (TiO₂-based) electrodes like E6013 produce a thin, fluid slag that usually peels off almost on its own, sometimes curling up as you finish the bead.
• Basic (low-hydrogen) electrodes like E7018 form a thicker, denser, glassier slag that adheres more firmly and needs real chipping and brushing.
• Cellulose-backed electrodes like E6010 burn hot and fast with a forceful arc, leaving minimal, flaky slag that mostly blows away as it forms.

Same process, three very different cleanup experiences. Electrode choice matters as much for weldability as it does for slag removal.

FCAW (Flux-Cored Arc Welding): Thicker and Stickier

Flux-cored wire leaves slag in both of its variants, but not equally. Gas-shielded FCAW (FCAW-G) relies on external shielding gas and tends to leave a lighter coating. Self-shielded wire (FCAW-S) generates its own shielding gas from the flux core, and its slag is typically heavier and more viscous than stick slag, a thick blanket that clings to the bead.

When polarity, voltage, and travel speed are dialed in, that heavy slag turns shiny, dark, and glassy, and it lifts cleanly in strips. When parameters drift, it comes off grey, dull, and stubborn. This parameter sensitivity is a recurring theme across every flux-bearing process, and it is well documented by technical sources such as TWI.

SAW (Submerged Arc Welding): The One-Piece Sheet

Submerged arc welding buries the arc under a mound of granular flux, so the process is literally shielded by the material that becomes slag. The melted flux solidifies into a continuous glassy layer over the bead, and with the correct current and travel speed, that layer can be pried off in a single strip rather than chipped. SAW slag chemistry can also be tailored to carry alloying elements into the weld metal, and at the right settings the slag is essentially self-releasing. It is the cleanest slag experience in welding, provided your parameters cooperate.

ESW (Electroslag Welding): The Exception

GMAW and GTAW: No True Slag

Gas metal arc welding and gas tungsten arc welding use inert or active gas shielding instead of flux, so they produce no true slag at all. What you may see are small glassy silica islands, tiny deoxide deposits left behind when silicon and manganese deoxidizers in the filler wire react with trace oxygen. They wipe or lightly brush off and are cosmetically annoying rather than structurally significant. If you are prepping aluminum, this distinction matters even more; see Aluminum MIG Welding: The Complete Practical Guide for how surface films and deposits are handled.

Thermal Cutting: Dross, Not Slag

The crust clinging to an oxy-fuel, plasma, or laser cut edge is dross, re-solidified metal oxide from the cutting reaction. It is a cutting phenomenon, not a flux product, and it has more in common with the edge conditions discussed in Alpha Laser Welding: Systems, Uses, and Safety Guide than with anything that came out of a welding arc.

Slag Behavior at a Glance

Process	Source of slag	Appearance	Typical removal
SMAW	Electrode flux coating	Complete crust; thin and fluid (rutile) to thick and glassy (basic)	Chip and wire-brush every pass
FCAW-G / FCAW-S	Flux core in the wire	Thicker, more viscous; shiny when parameters are right	Chipping, needle tool, or strip-lift
SAW	Granular flux	Continuous glassy sheet	Lifted off in one piece
ESW	Molten slag is the heat source	Fluid, resistive bath	Consumed by the process
GMAW / GTAW	None (gas shielding)	Small silica islands, brown deposits	Wipe or light brush
Thermal cutting	None (cutting reaction)	Re-solidified edge dross	Grinding or deburring

The takeaway is simple: before you decide how to remove something from a weld, know which process put it there. Slag, spatter, dross, and surface deposits each respond to entirely different tools, which is exactly what the next section breaks down.

Slag vs. Weld Spatter vs. Dross vs. Mill Scale: Know the Difference

Ask a room of new welders to point at the “slag” on a finished joint and half of them will point at a spatter droplet, a strip of blue scale, or the dross hanging off a plasma-cut edge. Slag is only one of several crusts, coatings, and contaminants that show up on steel, and misidentifying them leads to the wrong cleanup tool, a rejected inspection, or paint that peels off in sheets. Here is what each one actually is.

The Five Contenders, Defined

Weld slag is the non-metallic crust left by flux. Whether it came from a stick electrode covering, a flux-cored wire, or granular submerged-arc flux, the chemistry is the same story: melted silicates, titanates, and fluorides that floated on the puddle, absorbed oxides, then solidified on top. It is a glass, not a metal, which is why it cracks and pops off in chunks.

Weld spatter is the opposite in almost every way. Spatter consists of tiny droplets of molten base or filler metal thrown out of the arc by magnetic and gas-expansion forces, then frozen onto the plate around the bead. Because those droplets are real metal welded to the surface, they usually need grinding rather than chipping. TWI classifies spatter as a weld surface imperfection, and it is treated as such in visual acceptance criteria.

Dross is not a welding product at all. It is the re-solidified oxide film that clings to the edge after thermal cutting — oxy-fuel, plasma, laser, or air-carbon arc. Dross is harder and more tenacious than the parent metal around it and must be removed before welding, because it contaminates the fusion zone and produces lack of fusion at the root. Cutting and gouging work, such as air arc welding and carbon arc gouging, always leaves an edge that needs grinding back to bright metal.

Mill scale is the blue-black oxide layer hot-rolled onto new steel stock. It is magnetite, Fe₃O₄, formed in the rolling mill, not in your shop. It must be removed by grinding, abrasive blasting, or chemical pickling before welding or coating; AWS structural codes require base-metal cleaning that includes the removal of “mill scale, rust, moisture, and other foreign materials” before welding begins.

Silica islands are the exception that confuses people. In GMAW there is no flux, yet small glassy brown or grey deposits still appear on the bead. They come from silicon and manganese deoxidizers added to the filler wire; the silica rises to the surface and hardens into a glass bead. They are harmless to the weld itself but must be brushed or ground off before painting. If you are prepping aluminum, the same cleaning discipline applies — see Aluminum Cleaner for Welding: How to Prep Aluminum Right.

Side-by-Side Comparison

Material	What it is	Where it comes from	Appearance	Removal method
Weld slag	Non-metallic flux byproduct (silicate/titanate/fluoride glass)	SMAW, FCAW, SAW, ESW flux melting and reacting	Black or grey crust, sometimes shiny or glassy; brittle	Chipping hammer, wire brush, needle scaler; often self-releasing
Weld spatter	Molten base/filler metal droplets (a metal)	Arc forces expelling metal from the puddle	Small balls or specks bonded to the plate around the bead	Grinding or spatter removal wheels; hard to brush off
Dross	Re-solidified metal oxide film	Oxy-fuel, plasma, laser, and arc cutting edges	Rough, crusty, sometimes stringy lip on the cut edge	Grinding, filing, or machining back to sound metal
Mill scale	Fe₃O₄ oxide layer (magnetite)	Hot rolling of new steel plate and bar	Blue-black, flaky, fairly uniform sheen	Grinding, sandblasting, or acid pickling
Silica islands	Glassy silicon-oxide surface deposits	Deoxidizers in GMAW filler wire	Small brown or grey glassy spots on the bead	Wire brushing or light grinding before painting

Why the Distinction Actually Matters

Each of these materials behaves differently in three places that cost real money: removal, inspection, and coating. Slag can hide a surface-breaking crack or an arc strike underneath it, so visual testing is worthless until the crust is fully gone. Spatter and dross create stress raisers and rough profiles that interfere with ultrasonic transducer travel and with paint and coating adhesion, since coatings bridge over loose particles instead of wetting the steel. Mill scale, left on, delaminates under paint and takes the coating with it. And while porosity and slag inclusion are judged against code limits in standards maintained by AWS and ASME, spatter is largely a cosmetic and coating problem rather than a strength problem.

Getting this right at the bench saves time later. A welder who knows he is looking at silica islands will grab a wire brush, not a grinder; one who mistakes dross for slag will weld straight over oxide and wonder why the root fails radiography. Identification comes first, then the tool follows. For a broader look at process-by-process behavior, our guide to aluminum flux core welding shows how flux-bearing wires change what ends up on the bead surface.

Why Slag Matters: The Benefits of Slag and the Problems It Causes

Welding slag has a reputation problem. Most welders only ever meet it when they are chipping it off, so it feels like packaging — something you throw away to get to the product underneath. That view is half right, and half expensive. Welding slag is doing real metallurgical work in the seconds and minutes after the arc passes, and understanding both sides of that ledger is what separates a welder who consistently passes inspection from one who keeps coming back for rework.

What Welding Slag Does for You

It shields the hot metal. The molten flux blanket sits directly over the pool and seals it away from atmospheric oxygen and nitrogen. Nitrogen is the more damaging of the two in steel: it dissolves freely into molten metal and comes back out of solution as the bead cools, embrittling the weld and cutting toughness. Shielding gases released by the decomposing coating do part of this job, but the liquid slag layer is the physical barrier — the only one that stays in contact with the bead while it solidifies, which is why flux processes remain so effective at protecting weld metal.

It refines the weld metal. As the pool freezes, oxides of iron, manganese and silicon — plus the damaging impurities sulfur and phosphorus — float upward and dissolve into the liquid flux. In effect, the slag is skimming the steel, pulling contaminants out of the region that is about to become permanent structure.

It slows the cooling rate. An insulating blanket of molten silicate over a hot bead behaves like a controlled cool-down. Slower, more even cooling promotes a finer, more uniform grain structure and better toughness in many processes, and it gives dissolved hydrogen extra time to diffuse out of the weld before it can contribute to cracking.

It holds the puddle in place. The slag blanket supports the molten pool in vertical and overhead positions, helping the welder control a fluid puddle that would otherwise sag or run. Out-of-position welding is genuinely difficult, and flux processes make it far more forgiving because the crust carries part of the load. Which electrodes behave best uphill is something we walk through in our All Position Welding Rod Guide.

It enables high deposition rates and delivers alloying. In submerged arc welding, the granular flux and the slag layer it forms contain the pool even at very high currents, which is why SAW deposits metal several times faster than most manual processes. Just as important, in SAW the flux also acts as the vehicle for alloying, so slag chemistry directly controls the final weld-metal composition — change the flux and you change the weld. Welders selecting consumables for chromoly and stainless work will find the same logic laid out in our Alloy Steel Welding Electrodes guide.

For stick welders, all of these benefits ride along on every electrode, which is a large part of why SMAW survives in field and repair work — a point made at length in Advantages of Stick Welding.

What Welding Slag Costs You

It hides the weld. Because slag sits opaque over the bead, a welder cannot visually confirm surface quality until it is removed. Undercut, porosity or a sagged root can all be hiding under that crust. This is precisely why complete removal between passes is mandatory in multi-pass work rather than a courtesy: the visual check cannot happen until the crust is gone.

Left behind, it becomes a defect. Any slag trapped between passes, in a corner of a groove, or at a restart turns into slag inclusion — the subject of the next section. The trap is silent. Nothing looks wrong until a radiograph or an ultrasonic scan finds it.

Fragments cause arc strikes. Leaving hardened slag fragments on a completed pass and then striking an arc over them creates arc strikes: localized hard spots and micro-notches that raise hardness, act as stress raisers, and are rejectable under structural codes such as AWS D1.1.

It is pure labor. On structural and pressure welds, interpass chipping, brushing and grinding is a recurring line item in every welder’s day. Multiplied across hundreds of passes in a heavy fabrication, cleanup becomes one of the largest single non-welding costs on the job — and it is the cost most directly controlled by good technique.

It complicates automation. Hard, glassy, irregular slag interferes with optical and tactile weld-tracking sensors, forcing robotic and mechanized systems to stop, clean and restart, or to accept looser tolerances.

Slag-Related Weld Defects: Slag Inclusion and How It Happens

Sitting on top of a finished bead, slag is just a crust you chip off. Trapped inside the weld, it becomes a defect. A slag inclusion is non-metallic solid material — melted flux, electrode covering, tungsten, or another contaminant — that ends up locked in the weld metal, or sandwiched between the weld metal and the base metal. The TWI and welding codes describe it the same way: the key word is trapped. Slag that never left the pool can no longer protect the bead; it just occupies space where solid steel should be.

Two Shapes, Two Levels of Danger

Welding inspectors sort slag inclusions by shape first, then by whether they break the surface:

Classification	What it looks like	Where it sits	Risk level
Distributed	Rounded, dot-like specks spread through the bead, sometimes resembling fine porosity	Usually internal	Moderate — scattered, blunt, less of a stress raiser
Linear / elongated	Continuous or broken lines, shells, and seams along the weld axis	Internal or surface-breaking	High — behaves like a crack-like stress raiser

The linear form is far more damaging. An elongated inclusion presents a sharp, oriented discontinuity that concentrates stress, so it is the type that shows up as a fatigue crack origin years later. Distributed specks, by contrast, blunt each other out — which is exactly why codes treat the two forms differently.

How Slag Inclusions Form

Nearly every inclusion traces back to a small number of repeatable mistakes:

• Failure to clean between passes. A pass laid directly over an uncleaned previous bead is the single most common cause. The new puddle cannot remelt and absorb the old crust, so it seals it in.
• Wrong travel speed or arc length. Too fast, and the puddle solidifies before the lighter slag can float up and out. Too long an arc pulls turbulence and oxide back into the pool.
• Insufficient heat input. Low current makes the slag viscous and heavy; a thick, stiff blanket simply cannot escape a sluggish puddle.
• Wrong electrode angle or manipulation. Pushing the slag ahead of the arc, or washing the puddle too wide at the toes, buries the blanket at the groove walls. Out-of-position work magnifies this, and all position welding rod technique matters as much as the rod itself.
• Awkward groove geometry. Narrow, deep, or misaligned joints create tight root gaps where viscous slag wedges itself and no brush can reach.
• Contaminated flux or consumables. Damp, dirty, or damaged electrode covering breaks down unpredictably. Selecting the wrong classification for the joint is just as risky, which is why alloy steel welding electrode selection deserves real attention.

Why Trapped Slag Matters in Service

Once inside, an inclusion does three things. It acts as a stress concentration, so fatigue-loaded joints — cranes, bridges, chassis, pressure vessels cycling day and night — lose fatigue strength and can propagate a crack from the inclusion tip. In pressure-retaining welds it can form a leak path, sometimes a slow weeping one that never shows up in a hydro test until conditions change. In load-bearing members it simply removes load-carrying cross-section, reducing static strength.

Critically, structural and pressure codes do not automatically reject every speck. AWS D1.1 and the ASME Boiler and Pressure Vessel Code Section IX set limits on the size, length, clustering, and distribution of acceptable slag inclusion. Inspectors measure indications against those limits: small, isolated, distributed inclusions in a non-critical zone may be accepted, while a single 40 mm linear inclusion in a fatigue joint will not be. Judgement is always code-referenced, never “no defect allowed, ever.”

That is the practical takeaway: the crust is easy to remove while it is still on the surface, and expensive to remove once the next pass has sealed it in forever. Clean between passes, every pass.

How to Remove Slag: Tools, Techniques, and Best Practices

Slag removal is the least glamorous part of welding, and also the most consequential. The governing rule never changes: slag must be completely removed between passes on any multi-pass weld. A single leftover fragment sitting above the next bead is the classic setup for an internal slag inclusion that only shows up later on a radiograph. Cleaning is not optional housekeeping, it is part of the welding procedure.

The Slag Removal Toolkit

No single tool handles every job. Most welders carry a small kit and switch between them as the joint geometry changes:

• Pointed chipping hammer – spring steel with a narrow peen, ideal for digging slag out of a V-groove, a root pass, or a tight corner.
• Flat or needle chipping hammer – a wide flat face for knocking slag off long, flat cover beads in a few strokes.
• Wire brush – hand or power, used after chipping to clear the film that chipping leaves behind. Brush alloy matters: stainless and aluminium work needs a dedicated stainless wire brush, because plain-carbon bristles embed carbon particles in the surface that promote corrosion on stainless welds. Preparation discipline matters just as much as the welding itself, which is why proper aluminum cleaner and surface prep comes before the arc ever strikes.
• Pneumatic needle scaler – a bundle of vibrating needles that strips slag many times faster than a hand hammer. This is why they dominate fabrication shops working on submerged arc and heavy structural beads.
• Pneumatic chipping gun – similar speed advantage, with a flat chisel tip for long seams and fillet welds.
• Angle grinder with a wire cup or grinding disc – for weld toes, tie-ins, restarts, and stubborn patches. Use it sparingly on the finished bead so you do not grind away profile or gouge the base metal.
• Mechanical slag peeler – a bladed tool that lifts a continuous slag layer off submerged arc or automated welds in one piece, far faster than breaking it up.
• Peening hammer – light, controlled peening to relieve high shrinkage stress in crack-sensitive beads, always within the limits of the welding procedure.
• Chemical flux removers – for exotic alloys, crevices, and root surfaces where mechanical action would damage the parent material.

Technique: How to Strike and Scrape

Hit at a low angle, roughly 15 to 30 degrees, and drive along the direction of travel rather than straight down into the bead. Square-on blows dent the weld surface and nick the base metal, which then has to be blended out. Work the toes and the root specifically, since that is where slag hides and where a future crack or lack-of-fusion indication would start. Chip while the assembly is still warm where practical: slag releases far more easily from a warm bead than from one that has cooled and hardened completely, especially on heavy, high-heat-input passes.

Process-Specific Removal Tips

FCAW: both gas-shielded and self-shielded wire can be tuned toward self-releasing slag. Correct polarity, slightly higher voltage, and a steady travel speed produce a thinner, glassier crust that curls off on its own; too little heat or wrong polarity leaves a thick, sticky slag that has to be fought. Selection matters as much as settings, so review alloy steel welding electrode selection basics before changing wires mid-job.

SAW: raising current and travel speed within the procedure range produces a self-peeling slag. Fabricators aim for that deliberately because hand cleaning on a long submerged arc seam is pure cost.

SMAW: rutile electrodes shed their crust readily, basic electrodes take more effort, and cellulose-heavy rods leave very little to begin with. Plan your cleaning effort around the electrode family in use.

A Simple Removal Workflow

1. Let the bead cool to a warm, handleable temperature without letting it go stone cold.
2. Chip along the bead at a low angle, starting at an edge or restart where the slag has lifted.
3. Break the crust clear of the weld toes and the root on both sides.
4. Wire brush the full length of the bead, including the heat-affected zone.
5. Inspect visually for glassy flakes, pitting, or shadow lines, and re-clean any suspect spot.
6. Confirm the surface is bright and dry before laying the next pass over it.

Skip a step and the next bead simply welds over the evidence, turning a five-second cleanup into a repair. For practical reinforcement of the same habit on smaller joints, our guide to why SMAW still wins for field work walks through interpass cleaning in context.

Safety and Health: Working With Slag, Fumes, and Hot Metal

Slag looks inert once it cools: a dull grey crust, easily chipped away and swept into a bin. That appearance is misleading. Between the heat it carries, the shards it throws, the dust it creates and the fumes released underneath it, handling slag is one of the more underestimated routine tasks in a fabrication shop. Most injuries tied to slag do not come from the arc itself; they come from the cleanup that follows.

Burns and Eye Injuries From Chipping

Freshly chipped slag retains enough heat to cause contact burns long after the arc is off. Because it sticks to skin and holds heat in a dense, glassy mass, the burn is often deeper than a quick touch suggests, and welders frequently do not feel the full injury until minutes later. The safe habit is simple: treat every chipped fragment as hot work, regardless of how cooled the bead looks.

The second hazard is mechanical. A chipping hammer struck at the wrong angle sends sharp glassy fragments flying at eye and skin level. Safety glasses with side shields are the bare minimum; a face shield over them is the normal standard whenever you are hammering a heavy crust. Chips also land in collars, boot tops and hair, where they continue to burn or scratch after the work is done.

Arc Radiation and “Arc Eye”

Chipping does not happen in isolation. On multi-welder stations and production lines, you are often working beside a live arc. Ultraviolet and infrared radiation from that arc can cause photokeratitis, the painful condition welders call arc eye, where the cornea is essentially sunburned. Symptoms appear hours later as gritty pain, tearing and light sensitivity. Welding screens, curtains and proper helmet shade ratings protect both the welder and the neighbor doing cleanup.

Fumes, Slag Dust, and Metal Fume Fever

Welding on galvanized or plated steel releases zinc oxide fumes as the coating burns off. Inhaling them can cause metal fume fever, a flu-like illness with chills, nausea, headache, fever and a lingering metallic taste. It usually passes within a day, but every episode is a warning that exposure was too high, and the people chipping and grinding the coating off afterward are exposed too.

Grinding flux slag and old flux residues creates respirable dust. Depending on the electrode and base metal, that dust can carry silica, manganese, chromium or nickel compounds, which is why local exhaust ventilation and respiratory protection matter as much during cleanup as during welding. OSHA’s welding, cutting and brazing standards, 29 CFR 1910.252 for general industry and 29 CFR 1926.353 for construction, require ventilation and, where necessary, local exhaust to keep fumes and dusts within exposure limits. ANSI/AWS Z49.1, Safety in Welding, Cutting and Allied Processes, is the consensus reference most shops build their written program around, and NIOSH publishes current guidance on welding fume exposure control. Electrode choice plays into this as well; our Alloy Steel Welding Electrodes guide covers how consumable selection affects both deposition and fume output.

Fire Risk and Hot Work Control

Hot slag carried or dumped into waste is a common ignition source. Paper, oily rags, sawdust and cardboard bins smoulder quietly, sometimes hours after the shift ends. Hot-work permits, non-combustible slag containers and a documented fire watch after welding, chipping and grinding operations remain necessary, especially in occupied buildings. The same logic applies to hot processes that generate sparks and molten dross, as covered in our Air Arc Welding and Carbon Arc Gouging guide.

Slag Disposal: From Waste to Aggregate

Handled correctly, cooled slag is not hazardous waste. It is non-metallic and, in the wider metallurgical industry, is commonly crushed for use as aggregate or road base. The practical rule for the shop is to let slag cool in a steel container before disposal, keep it separated from combustible rubbish, and never transport or sweep it while still hot.

Inspection and Quality Control: Finding Trapped Slag

Finding trapped slag is a detective job with a hard deadline: once the cover pass goes over the root, any slag left behind is sealed inside solid steel. No amount of careful looking will reveal it afterward. That is precisely why nondestructive testing (NDT) exists, and why inspection is treated as a separate discipline from welding itself. Each method in the inspector’s toolbox has a blind spot, and choosing the wrong one is how a 40 mm line of slag survives all the way to service.

Visual Testing: Necessary, but Never Sufficient

Visual testing (VT) always comes first, and it starts before the inspector looks at anything. All slag must be chipped, ground or brushed off the complete weld and the adjacent base metal, because crust hides the very surface features VT is meant to judge: profile, undercut, toe cracks, arc strikes and surface-breaking inclusions. Once the bead is bare, VT checks for overlap, mismatch, pitting and visible porosity.

The limitation is simple and unforgiving: VT sees only what breaks the surface. A weld can look mirror-clean on top and still hold a string of slag along the root line underneath. VT is a gate, not a guarantee, which is why codes rarely accept it alone for load-bearing or pressure-containing work.

Liquid Penetrant Testing (PT)

Liquid penetrant testing draws a visible or fluorescent dye into any opening that reaches the surface, then reveals it after the excess is cleaned off and developer is applied. PT is excellent for surface-breaking slag, especially on non-porous materials, and it needs no magnetism or radiation. But it is strictly a surface method: the dye has no way of reaching an inclusion sealed 3 mm under the cap. Internal slag is invisible to PT, full stop.

Magnetic Particle Testing (MT)

Magnetic particle testing magnetizes the part and sprinkles ferrous particles over it. Where a discontinuity interrupts the surface, the magnetic field leaks and the particles pile up into a visible indication. MT finds both surface and near-surface flaws, making it faster and more forgiving than PT for carbon steel welds.

Two hard limits apply. It works only on ferromagnetic materials, so austenitic stainless, aluminium, nickel alloys and titanium are out. And it sees only the near-surface region, not the deep interior. MT therefore complements RT and UT rather than replacing them, as described in TWI’s overviews of NDT methods.

Radiographic Testing: Reading the Film

Radiography sends X-rays or gamma rays through the weld onto film or a digital detector. Denser material blocks more radiation and appears lighter; voids and low-density inclusions let more through and appear darker. Interpreting the shape of that darkness is how experienced radiographers tell slag from other defects.

tr>

Indication on film	What it looks like	Usual meaning
Slag inclusion	Dark, irregular, rounded, feathered or shell-shaped density, often along a pass boundary or weld toe	Flux or covering trapped between passes or against the groove wall
Lack of fusion	Straight, fine, sharp dark line running parallel to the fusion face	Metal never bonded to the wall or previous bead
Porosity	Rounded, clean-edged dark circles or spots, often scattered	Gas bubbles frozen in the bead

Codes such as AWS D1.1 and the ASME Boiler and Pressure Vessel Code do not ban every speck. Inspectors judge size, length, clustering and location against written acceptance limits, which is why interpass cleanliness matters more than any single pass looked.

Ultrasonic Testing: Following the Linear Slag

Ultrasonic testing bounces high-frequency sound through the weld and reads the echoes that come back. It is highly sensitive to internal, elongated slag and can pinpoint depth accurately without radiation, which makes it the method of choice for thick sections and root passes. The trade-off is interpretation: results depend heavily on the technician’s skill, and a long slag trail can mimic lack of fusion, or hide right beside it. Automated UT and phased array scanners are increasingly specified for pipelines and pressure vessels because they remove much of the guesswork and leave a repeatable digital record. Procedure-driven work, such as the NDT plans in Alloy 20 welding procedures, usually names the required method and acceptance standard up front.

Where Trapped Slag Hides First

Inspectors do not scan a weld uniformly; they aim at the known trouble spots. Start and stop points, root passes, and any area welded over an uncleaned bead are the highest-risk locations for trapped slag, because those are exactly the moments when a welder is most likely to leave flux residue behind. Start and stops concentrate reheating, the root is tight and hard to clean, and a fresh pass simply buries whatever the previous pass shed. That is why an inspector looks at the tacks, the tie-ins and the root before anywhere else, and why cleaning between passes remains the cheapest insurance policy in the trade. Grinding dust and hot work around live arcs add their own hazards, so examination areas need the same ventilation and controls OSHA requires for welding operations themselves.

Common Mistakes, Troubleshooting, and Key Takeaways

Few welders who fail an inspection do anything dramatic. They make six or seven small mistakes in a row, and the crust quietly stays exactly where the arc left it. The encouraging part is that slag almost always tells you what went wrong before an inspector does. If you can read its colour, thickness and how hard it fights the chipping hammer, you can correct the problem on the next pass instead of on the repair table.

Troubleshooting: Symptom, Cause, Fix

Symptom	Likely cause	Fix
Slag won’t lift from a flux-cored bead	Insufficient current, incorrect polarity, travel speed too slow, or contaminated/damp wire	Check polarity against the wire datasheet, raise amperage into the recommended range, speed up travel, and swap to dry, clean wire from sealed packaging
Grey, crusty, stubborn slag	Heat input too low or travel speed too slow, so the slag is thick and viscous	Raise current or slow just enough to get a fluid puddle; a thin shiny or blue-grey skin means parameters are back on target
Slag inclusions reported on RT	Poor interpass cleaning, arc too long, wrong electrode angle, or a tight root gap trapping viscous slag	Clean fully between passes, shorten the arc, correct work and travel angles, and gouge out the reported area before re-welding
Pitted surface under the slag	Arc length too long or current too high, allowing the arc to blast the pool	Shorten the arc, drop amperage slightly, and steady travel speed
Hard metal beads blamed on “slag”	Actual weld spatter from wrong parameters or shielding gas issues	Fix voltage, stick-out and gas flow, then grind the spatter — a chipping hammer will not lift it

Two of these deserve emphasis. Slag that refuses to release from a flux-cored weld is nearly always a parameter problem rather than a tool problem — chasing it with a sharper hammer just damages the bead. And hardened chipping tools or a grinder used carelessly can gouge the base metal or notch the weld toe, creating exactly the stress raiser that becomes the next fatigue crack.

The Mistakes That Cause Most Rejects

• Skipping interpass cleaning. This is the single most common cause of rejectable slag inclusion in structural and pressure work, and it takes only seconds to prevent.
• Grinding only what you can see. Surface cosmetology hides root and toe slag that radiography will find immediately.
• Using the wrong electrode for the position. A flat-only rod run vertical produces heavy, viscous slag that cannot float out; an all-position welding rod matches the job.
• Welding over damp electrodes. Moisture drives hydrogen into the pool, worsening porosity and cracking risk. Properly stored low-hydrogen consumables matter as much as clean passes.
• Chipping with hardened tools at the wrong angle, nicking the parent metal and the bead edges.
• Ignoring slag on the root pass because a cover pass will hide it. It will not stay hidden from RT or service.

Slag-Free Weld Checklist

1. Strike at a low angle along the direction of travel; chip, then wire brush, then inspect.
2. Clean completely before every pass, including the root, toes and start-stop points.
3. Match the electrode or wire to the weld position and the alloy.
4. Store low-hydrogen consumables dry, per manufacturer recommendations.
5. Tune polarity, amperage and travel speed until the slag lifts in one piece.
6. Use a stainless brush on stainless and a dedicated grinder on aluminium to avoid carbon contamination.
7. Never weld over a questionable previous pass because it “looks fine”.

Conclusion: One Rule Above All Others

Slag in welding is the glassy, non-metallic byproduct that forms when flux melts, shields the pool, absorbs oxides and solidifies on the bead. It is not waste. In SMAW, FCAW and submerged arc it protects hot metal from oxygen and nitrogen, refines the weld metal, supports the puddle out of position and carries alloys into the joint. The problem starts the moment it is left behind: it hides surface defects, hardens into stress concentrators, and becomes a rejectable inclusion the code judges by size, length and clustering rather than by a zero-tolerance rule.

So the discipline never changes. Chip it, brush it, look at the bare metal, and only then strike the next arc. Clean between passes — everything else in this article is detail. If you want to go further, our guides on electrode selection and identifying common weld defects will help you put the theory into practice.

Frequently Asked Questions

What is slag in welding?

Slag is the glassy, non-metallic crust that forms on a weld bead when flux or electrode coating melts and reacts with oxygen and impurities. It floats on the molten pool, solidifies on top of the bead, and must be removed before any further passes or inspection.

Is welding slag toxic?

Cool, solidified slag itself is a non-metallic silicate material and is generally handled as a nuisance dust. The real hazards are heat, flying glassy fragments, and the welding fumes generated while the arc is running — ventilation and eye/skin protection matter during chipping and grinding.

What is the difference between slag and weld spatter?

Slag comes from flux and is non-metallic, so it can usually be chipped or brushed off. Spatter consists of tiny droplets of molten metal frozen onto the plate, which bonds metallurgically and normally requires grinding.

Does MIG welding produce slag?

Gas metal arc welding uses shielding gas rather than flux, so it does not produce true slag. What it can leave are small glassy silica islands from deoxidizers in the wire, which are easily brushed or ground off before painting.

Can a weld with slag inclusion be repaired?

Yes, in most cases the defective area is gouged or ground out to sound metal and rewelded after complete cleaning, following the repair procedure allowed by the applicable code such as AWS D1.1 or ASME Section IX. Rejected areas must be re-inspected after repair.

Does slag weaken a weld?

Slag left on the surface does not weaken anything by itself, but slag trapped inside the weld does. Internal and linear slag inclusions act as stress concentrators, reducing fatigue strength and creating leak paths in pressure-containing service.

How do I know if my weld has slag inclusion?

Surface slag is visible once the crust is chipped away, but internal inclusion cannot be seen by eye. It is detected with radiographic or ultrasonic testing, which is why critical welds are inspected with volumetric methods.

What does the appearance of the slag tell me about my weld?

Shiny, dark or blue-tinted slag generally indicates correct current, voltage and travel speed. Thick, dull grey, crumbly slag usually means heat input was too low or travel speed too slow for the parameters used.

How do you remove slag from a weld?

Use a chipping hammer or needle scaler to break the crust, then a wire brush to clear the toes and any remaining residue. For heavy submerged arc or flux-cored welds, pneumatic scalers, wire wheels or mechanical slag peelers speed the job considerably.