Slag Inclusions in Welding: Causes, Detection, Acceptance Criteria, and Prevention

What a Slag Inclusion Really Is (and Why Welders Keep Meeting Them)

You chip the last crusty layer off a multi-pass fillet weld on heavy structural steel. The toes look tight, the beads overlap cleanly, and the machine ran dialed-in all shift. Then the radiographer’s film comes back dotted with dark, jagged lines running down the joint — and the welder swears the technique was fine. Where did that trapped, non-metallic junk come from?

That junk has a name, and it has been frustrating fabricators since flux was invented: a slag inclusion. Understanding exactly what it is — and how it differs from the defects it gets blamed for — is the first step toward keeping it out of your welds.

Slag Is Your Friend (Until It Isn’t)

In any flux-based process — stick (SMAW), flux-cored arc welding (FCAW), and submerged arc welding (SAW) — the flux melts and forms a molten, glassy layer that floats on top of the weld pool. AWS A3.0 defines slag as “a nonmetallic product resulting from the mutual dissolution of flux and nonmetallic impurities in some welding and brazing processes.”

That layer is doing real work. It shields the cooling metal from the atmosphere, slows the cooling rate, and carries impurities up and out of the pool. Left alone, it solidifies into a crust you chip off and brush away. Slag is a tool — until it gets trapped.

AWS A3.0 defines a slag inclusion as “a discontinuity consisting of slag entrapped in weld metal or at the weld interface.” In plain terms, it’s the same material that was supposed to float away, now locked inside the joint or smeared along the fusion line.

The ISO 6520-1 Code Family

Standards bodies categorize inclusions by shape and source. Under ISO 6520-1, slag inclusion is code 301, broken down by how the slag is distributed:

• 3011 – Linear slag inclusion: elongated, often running along the weld axis.
• 3012 – Isolated slag inclusion: a single, scattered pocket.
• 3013 – Clustered slag inclusion: a group of pockets bunched together.

That shape matters, because acceptance criteria treat a long linear indication far more harshly than a tiny isolated one. Related codes round out the family:

• 302 – Flux inclusion
• 303 – Oxide inclusion
• 304 – Metallic inclusion (including tungsten from a dipping TIG electrode)

Not the Same as Porosity or Lack of Fusion

Welders often lump these together, but inspectors don’t. Porosity is gas trapped as rounded voids — smooth-walled and generally spherical. Lack of fusion is a planar gap where the weld metal never bonded to the sidewall or previous pass. Slag inclusions are irregular, jagged, non-metallic bits of solid material, frequently angular and strung in lines. On a radiograph, porosity reads as round dots, lack of fusion as a sharp straight line, and slag as dark, ragged streaks. Oxide films are thin, layered surface oxides rolled or trapped into the joint, which is why code 303 sits alongside slag in the same standard family.

Here’s the bottom line: a slag inclusion is solid, non-metallic material that should have floated away but got buried instead. The trickier questions are why it gets trapped so reliably in certain processes, how inspectors catch it, and what the codes actually let you get away with. This guide walks through all of it — root causes, detection methods, acceptance criteria, and a prevention playbook you can run on the floor. If you want to sharpen the fundamentals first, our guide to building real welding skills is a solid starting point.

Slag shares DNA with other fusion-line defects, so it pays to know the neighborhood. Our breakdown of the overlap welding defect and guide to telling good and bad MIG welds apart cover the symptoms inspectors hunt for. Researchers continue to study how slag chemistry and cooling rates influence entrapment, and the engineering literature is a good rabbit hole if you like the metallurgy. We’ll also dig into why flux-based processes like gasless MIG (flux-cored) welding are especially prone throughout this article.

Slag Inclusions vs. Porosity, Lack of Fusion, and Oxide Films: Telling Them Apart

On a radiograph, several discontinuities look like cousins — dark, irregular, and equally alarming. Each one starts from completely different physics, though, and each demands a completely different fix. Grind out the wrong culprit and you’ll be back at the bench next week with the same reject stamp, wondering why the repair didn’t hold. In a busy fabrication shop, that guesswork doesn’t just cost time — it costs money.

Porosity: Gas Voids, Not Trapped Flux

Porosity is entrapped gas — hydrogen, nitrogen, or a shielding gap — frozen into the pool as it solidifies. On film it reads as rounded dark spots, often clustered or elongated like a worm. The culprit is escape, not mineral debris: moisture in consumables, contamination, or poor gas coverage. That’s why the fixes are about dryness and shielding, not slag chemistry. Contrast this with slag itself, which is a solid mineral byproduct of the flux or electrode coating — it always has a source you can point to. Dry your electrodes, clean the joint, and check your gas flow before you blame the flux.

Lack of Fusion: Unfused Metal, Not Non-Metal Debris

Lack of fusion is metal that never bonded to metal — a cold lap or an untied sidewall. There’s nothing foreign inside it. Radiographically it shows as straight, sharp-edged dark lines that trace the joint geometry, which is exactly why it’s mistaken for linear slag. The stakes differ, too: ISO 5817 lists lack of fusion (code 401) as not permitted at any quality level, while some inclusions are tolerated depending on size and type. Non-destructive testing can flag that line, but only cross-sectioning or careful phased-array work confirms what actually sits inside it. Remember that a slag inclusion is itself a fusion-type discontinuity that can mask or accompany incomplete fusion between passes — so linear slag on film means you should also suspect hidden lack of fusion beside it.

Oxide Films and Silicate Islands: The Solid-Wire Look-Alikes

Here’s the trap that catches experienced hands. Gas-shielded solid-wire MIG/MAG and TIG don’t make true slag inclusions in the classic sense — there is no flux to become slag. They do leave silicate islands floating on the bead, which look glassy and slag-like but sit on top, not inside. So when a gas-shielded solid-wire weld shows linear indications on radiography, treat that as a red flag for oxide films or lack of fusion rather than slag. Oxides form on dirty aluminum or a sluggish, cold pool and get trapped at the root or between passes. That’s why surface prep and travel speed matter as much as gas settings. To separate appearance from cause, compare against good and bad MIG welds and the overlap welding defect breakdown.

Tungsten Inclusions: The TIG Wildcard

Tungsten inclusions are tungsten particles trapped in the weld metal deposited with the GTAW process. They come from dipping the tungsten electrode into the molten pool, or from excessive current that melts the tungsten tip. On film they’re dense and bright — nearly the opposite of a slag’s dark blot. That is your clue to check the electrode, not the flux. Cure them with technique and settings: keep the tip clear of the pool and dial in polarity. Our TIG welding polarity guide and AC or DC for TIG steel explain the ranges, while structured welding practice builds the consistency that prevents contact dipping.

Side-by-Side: Reading the Signature

Match the signature to the mechanism and you’ll stop repairing the wrong defect — and stop paying for reject lots. The next section digs into which processes are actually prone to slag inclusions and why that junk gets trapped in the first place.

Which Welding Processes Are Prone to Slag Inclusions

Slag is a byproduct of flux. Any process that relies on flux to shield the molten pool and stabilize the arc generates a slag layer — and any process that makes slag can leave some trapped inside the joint. That single rule tells you almost everything about which processes are prone to slag inclusions.

The signature defect of flux-shielded welding, slag inclusions show up most often in three process families: shielded metal arc welding (SMAW/MMA), flux-cored arc welding (FCAW), and submerged arc welding (SAW).

How Flux-Shielded Processes Make Slag

The mechanism is the same across all three. As the arc burns, the flux or electrode coating melts and does two jobs at once: it releases shielding gas to keep the atmosphere out of the pool, and it forms a molten slag blanket that floats on top of the weld, protecting it while it solidifies. The slag then cools, hardens, and must be chipped, brushed, or ground away before the next pass. Stick and flux-cored welders know the rhythm well — weld, chip, brush, repeat.

Problems begin when cleaning is rushed or skipped, or when joint geometry gives molten slag a place to hide. Sharp notches and the deep valleys between passes act like traps: slag flows in and solidifies before it can escape. On the next pass, fresh metal freezes over the top and locks the slag inside. That is a slag inclusion.

SMAW (Stick): The Classic Slag Producer

Every stick electrode carries a flux coating, so every stick weld produces slag. It is thickest on low-hydrogen basic electrodes such as E7018, common on structural and code work. 3/16 E7018 amperage and technique call for enough heat and travel control to let slag float out rather than freeze in. Root passes in V-grooves are the highest-risk spot, because a slightly convex bead with a toe notch can trap slag at the sidewalls.

FCAW: Fast, Forgiving, and Still Slag-Prone

Flux-cored wires — whether gas-shielded or self-shielded — generate a slag layer on every bead. The self-shielded (“gasless”) variety is popular for field work, covered in our guide to gasless MIG welding. But “forgiving” does not mean immune.

An AWS Inspection Trends case study makes the point: a heavy structural fabricator reported roughly 20% fillet weld rejection and about 10% groove weld rejection with FCAW, driven largely by incomplete fusion at the root and slag inclusions. Most rejects traced back to technique and setup, not the process. Polarity matters too — running the wrong polarity changes penetration, bead profile, and how readily slag releases, as explained in polarity for flux core welding.

SAW: High Deposition, Flux Inclusions

Submerged arc welding buries the arc under granular flux, which melts into a slag that shields the pool and is then vacuumed or swept up for reuse. Its high deposition rate suits heavy plates and long seams, and automation has pushed it further — see robotic submerged arc welding.

SAW has its own quirk. Because the flux is granular, unfused grains can survive as solids and get trapped in the weld, producing a flux inclusion — catalogued under ISO 6520-1 code 302 as a specific type of inclusion. Poor flux recovery, contaminated or oversized flux, and interrupted arc starts are common contributors.

Processes That Don’t Generate Slag

Not every process produces slag, so not every process produces slag inclusions. Solid-wire GMAW (MIG), GTAW (TIG), laser, and resistance welding use gas, vacuum, or no shielding at all — no flux, no slag, no slag trapping. They carry other risks (porosity, oxidation, lack of fusion), but not this one. That matters when comparing processes: switching from FCAW to solid-wire MIG removes a whole category of inclusion risk, though it changes position capability, wind tolerance, and cost. A quick look at whether flux core is the same as MIG frames that trade-off.

Practical Implications, Process by Process

Here is how the risk stacks up across the common arc processes:

Root Causes: Why Slag Gets Trapped in the First Place

Slag inclusions rarely come from one dramatic mistake. They build from a stack of ordinary conditions — a skipped wire brush, a bead that froze too fast, an electrode angle that let molten flux run out ahead of the pool. Understanding the mechanism behind each cause is what turns guesswork into a real fix. Here are the dominant and secondary reasons slag ends up buried inside a weld.

Incomplete Interpass Slag Removal: The Number One Cause

Ask any welding inspector and you’ll get the same answer: the most common cause of slag inclusions is failure to completely remove slag from one weld pass before making another. In multi-pass work, every flux-bearing process leaves a glassy or crusty layer over the bead. If that layer isn’t fully chipped and brushed away, the next pass simply welds it into the joint. The slag doesn’t vanish — it becomes a discontinuity sandwiched between passes. That’s why consistent chipping, wire brushing, and inspecting each bead before the next one matter more than raw travel speed, and why solid habits from welding practice beat fast hands every time.

Low Heat Input and Rapid Cooling

Slag is lighter than molten steel, so given time and fluidity it floats to the surface where it can be removed. Low heat input and fast travel speed rob the pool of both. When the weld freezes quickly, slag particles are locked in place before they can rise — the same rapid cooling that produces a hard, crack-prone microstructure also seals non-metallic debris into the bead. Matching amperage to material thickness keeps the pool fluid long enough for slag to escape.

Electrode Angle, Arc Blow, and Slag Running Ahead

Welding slag can spill ahead of the arc and become covered by the incoming weld pool. The culprits are unfit joint configuration, poor joint design, incorrect electrode manipulation, or arc blow. A drag angle that’s too steep, a weave that outruns the puddle, or magnetic arc blow pushing the pool sideways all let molten flux get in front of the advancing weld metal — which then buries it. On flux-cored and stick work, polarity and angle interact, so it pays to know how DCEN versus DCEP changes puddle behavior before blaming the wire.

Geometry That Traps Slag

Slag particles tend to get caught in hollows and sharp angles. Rough or convex bead profiles, deep undercut, and narrow, deep joint gaps give debris a place to sit where the arc can’t reach. A convex bead that looks tidy on the outside can hide a valley at its toes — exactly where the next pass leaves slag behind. Undercut edges are notorious traps because the recess sits below the surface the arc is washing.

Contamination: Rust, Scale, Primer, and Moisture

Welding over rust, mill scale, primer, oil, or moisture feeds oxygen and non-metallic material straight into the pool. That contamination creates its own inclusions and thickens the slag system, making flux harder to float out. Cleaning to bright metal before each pass removes the raw material for inclusions.

Procedural Causes vs. Human Causes

Some causes are systemic: a qualified procedure that under-specifies interpass cleaning, a joint design with no access for a chipping hammer, or an equipment-side fault like a worn contact tip creating an erratic arc. Others are purely human — rushing, skipping the brush, or running too hot to save time. The distinction matters, because a procedural fix prevents recurrence, while a lecture about technique only patches one shift.

Operator technique ties it all together. A welder who long-arcs, whips the rod without watching the puddle, or starts a new pass without checking the previous one manufactures inclusions by habit. Skilled hands keep a tight arc, watch the slag line, and let the puddle do the work.

A Troubleshooting Checklist

When inclusions show up, work down this list rather than guessing:

• Slag left between passes — chip, grind, and brush every bead, then verify before the next.
• Bead freezing too fast — raise amperage or slow travel to keep the pool fluid.
• Slag running ahead of the arc — adjust work and travel angle, and check for arc blow.
• Convex or ropey beads — widen the weave and flatten the profile.
• Deep undercut — reduce travel speed and correct electrode angle.
• Rust, scale, or primer present — clean to bare metal before welding.
• Poor joint access — redesign the joint or reposition the work.
• Inconsistent arc — inspect contact tip and liner on wire-fed processes.

How Slag Inclusions Are Detected: Visual, Radiographic, and Ultrasonic Inspection

Finding trapped slag is a layered job. You start with what your eyes can catch, move to what film and sound waves reveal, and finish with destructive tests when the stakes are high or a procedure needs proving. Each method has a boundary, and knowing those boundaries is what separates a good inspector from one who signs off a defect.

Visual Inspection: The First and Bluntest Filter

Before any fancy equipment comes out, understand a simple rule: AWS D1.1 requires that slag be removed from all completed welds. That means any visible slag sitting on the surface means the weld is not finished — full stop. You don’t get to evaluate it against acceptance criteria yet, because the joint hasn’t even reached the inspection stage.

This is why pre-cleaning is mandatory. A certified welding inspector (CWI) working under the AWS D1.1 Structural Welding Code cannot apply visual acceptance criteria to a weld with crusty flux still clinging to it. Chip it, brush it, grind where needed — then look. Under D1.1 Table 6.1, surface slag is rejectable because it represents incomplete fusion between layers or between weld metal and base metal, not a cosmetic blemish.

Visual inspection also catches the telltale signs that point to internal trouble: ragged, wandering bead edges, heavy spatter, or a weave pattern where each pass looks slightly detached from the last. On flux-cored and stick work, this is common, so matching polarity for flux core welding and travel technique to the joint really does show up at the inspection table.

Radiographic Testing (RT): Seeing Inside with Film

Radiography passes X-rays or gamma rays through the weld onto film or a digital detector. Denser steel absorbs more radiation; less dense slag absorbs less, so it registers as darker, irregular, jagged indications on the image. That shape is the giveaway — porosity tends to look like round dots, while slag lines are elongated, angular, and often follow the weld axis or sit in clusters between passes.

RT is excellent for full-penetration butt welds and for capturing a permanent record. Its weaknesses are geometry and orientation. Slag that’s tightly bonded and aligned along the beam path can be missed, and complex joints or thick sections make access and sensitivity a real challenge.

Ultrasonic Testing (UT): Sound for Thick Sections

When RT can’t reach a joint — think thick heavy sections, T-joints, or awkward geometry — ultrasonic testing takes over. UT sends high-frequency sound into the weld and listens for reflections. Slag inclusions return echoes that a skilled technician maps by depth and position, revealing linear and clustered defects that film might let slip through. UT demands a trained operator and a properly calibrated procedure, but for heavy-duty, high-deposition work it’s often the only practical internal test.

Note the limits of the surface-only methods. Magnetic particle (MT) and dye penetrant (PT) testing can’t find buried slag — they only reveal surface and slightly subsurface discontinuities. Treat their clean result as necessary, not sufficient.

Destructive Testing: When Trapped Slag Shows Up in a Big Way

Bend tests, macroetch, and full-section sampling tell the honest story. A sawn cross-section can look perfectly sound to the naked eye, with no visible gap. Bend the specimen and everything changes. In GMAW work, trapped silicate or lack of fusion that was invisible on the cut face “shows up in a big way” once the sample is bent — the metal tears open along the weak bond and exposes the slag pocket.

This is why procedure qualification and welder tests rely on bend specimens. Macroetching goes further, using acid to reveal the internal structure and highlight slag bands, fusion lines, and penetration. Both methods sacrifice the part, so they’re used on test pieces rather than production welds.

Choosing the right method isn’t guesswork. ISO 17635 provides the correlation between weld quality level and the applicable acceptance level for each NDT method, so the inspection plan matches the quality required. Coupled with a credentialed inspector’s judgment, that framework turns a pile of dark film lines into decisions you can actually defend.

Get the inspection sequence right and you catch slag before it becomes a reject lot — or a field failure.

Acceptance Criteria: What AWS D1.1, ISO 5817, and ASME Actually Allow

Here’s the uncomfortable truth about slag inclusions: there is no universal “acceptable” size. Whether a trapped line of slag passes or fails depends entirely on the code written into your contract. The same 3 mm inclusion can sail through one project and get an entire weld rejected on another.

So before you argue with an inspector, find out which standard governs the job.

AWS D1.1: Structural Steel

AWS D1.1 is the go-to code for structural steel, and it draws a hard line between what you can see and what radiography reveals. Under the visual acceptance criteria in Table 6.1, any slag visible on the finished weld surface is not acceptable — period. The surface must be clean, and fusion-type discontinuities are judged by their size, length, and distribution. That surface discipline is the same skill that separates good and bad MIG welds.

For radiographic testing, D1.1 sets separate limits for fusion-type discontinuities, including slag inclusions. These aren’t a single “no” — they’re graded by size and how densely they cluster along the joint. A small, isolated inclusion may pass; a long, jagged line running down the weld will not.

When a weld fails, AWS D1.1 Section 5.26.1.3 is blunt about the fix: unacceptable portions caused by incomplete fusion, excessive porosity, or slag inclusions must be removed and rewelded. No patching over the top.

ISO 5817:2023: Three Quality Levels

ISO 5817 takes a different approach. Instead of pass/fail, it defines three quality levels — B, C, and D — where B is the highest requirement on the finished weld, C is intermediate, and D is the most permissive.

The standard applies to material thickness ≥ 0.5 mm and covers fully penetrated butt welds and all fillet welds. That scope matters: if your joint type or thickness falls outside it, ISO 5817 isn’t the right yardstick.

Limits for imperfections vary by quality level, by material thickness, and by imperfection type. A slag inclusion that’s fine in a thin section at level D might be rejectable in a thick section at level B. There’s no shortcut — you have to read Table 1 for your specific case.

One rule catches people out: under ISO 5817, systematic imperfections are only permitted in quality level D, provided the other requirements of Table 1 are met. A one-off inclusion is one thing; a repeating pattern along the length of the weld is treated far more seriously. You can review the ISO 5817:2023 standard for the full table.

ASME and API 1104: Other Frameworks

Pressure equipment falls under ASME BPVC. Section IX covers welding qualifications, while Section VIII governs the fabricated vessel and its acceptance criteria — often leaning on radiographic examination. Pipeline work typically follows API 1104, which has its own girth-weld acceptance limits for internal discontinuities. In practice, most pressure and pipeline work is examined by radiography, so internal slag is the primary concern.

Each framework treats slag differently. None of them simply says “no slag allowed anywhere.”

The Same Inclusion, Two Verdicts

This is why shop-floor arguments happen. A fabricator used to ISO level D ships a part to a D1.1 structural job and watches it get rejected — or the reverse. The inclusion didn’t change; the rulebook did.

Note: Fabricators must always work to the code specified in the contract. When documents conflict, the contract governs. Confirm it before the first arc, not after the film comes back.

Getting the acceptance criteria right is a skill built through repetition — the same way you’d refine technique in our welding practice guide. And when a repair is needed, clean removal matters as much as the original weld, whether you’re running stick, flux-cored wire, or robotic submerged arc.

Why Slag Inclusions Matter: Strength, Fatigue, Corrosion, and Failure Risk

A slag inclusion is easy to dismiss as cosmetic when it sits shallow and small. But run the numbers on a restrained, thick-section joint under cyclic load and that trapped non-metallic pocket stops being a blemish and starts behaving like a designed-in crack initiator. The engineering consequences of weld defects are well documented: reduced strength, increased stress concentration, higher probability of corrosion failures, fatigue, leaks, and outright weldment failure.

The catch is that severity isn’t uniform. As inspection guidance across the industry stresses, the consequences of a defect depend on its type, location, size, and service conditions. A small surface defect may have little effect on strength. A large internal defect — especially one oriented across the load path — can take the whole weldment down.

Stress Risers and Reduced Effective Cross-Section

Metal doesn’t fail because of the slag itself; it fails because of the geometry the slag creates. An inclusion is a discontinuity with a sharp boundary, and sharp internal geometry concentrates stress. Think of it as a notch machined inside a weld that no one planned.

• Internal notches: Brittle slag forces stress lines to divert around it, spiking local stress above nominal levels.
• Lost section: Every inclusion displaces sound metal, shrinking the load-bearing area. On a fillet throat or groove root, that reduction is direct.
• Disbonding: Because slag is non-metallic, it isn’t metallurgically bonded to the surrounding steel — there’s a gap, not a fusion line.

Fatigue: Where Inclusions Do Their Worst Damage

Under static load, a modest inclusion may survive quietly. Under cyclic loading, it becomes a crack nursery. Fatigue cracks almost always initiate at stress concentrations, and an internal slag line is a ready-made starter notch. Each load cycle opens and closes the microscopic gap, ratcheting a crack until it reaches critical size.

This is why fatigue-critical welds — bridges, crane runways, pressure vessels, machine frames — are held to tighter acceptance limits. The same logic that makes overlap welding defect investigations focus on toe geometry applies here: sharp transitions and unbonded interfaces are where failures begin.

Brittle Fracture in Restrained and Thick Sections

Thick, restrained joints — where shrinkage can’t relieve itself — carry high residual stress. Add a sharp internal notch and cold service temperatures, and you have the recipe for brittle fracture: sudden, low-energy, and catastrophic. Highly restrained details and steels with high hardness and limited toughness are the most vulnerable, because they can’t deform their way out of a stress spike.

Corrosion Cells and Leak Paths

Because slag isn’t bonded to the metal, that interface is a gap. In a corrosive environment, an unbonded crevice becomes a crevice corrosion cell — oxygen and chemistry differ inside versus outside, and the gap attacks preferentially. In pressure or fluid service, the same discontinuity can become an actual leak path straight through the joint. Thermal cycling makes it worse, flexing the gap open and shut and pumping fluid or electrolyte along the defect.

Service conditions amplify everything. Cyclic loading feeds fatigue. Pressure drives leaks. Thermal cycling widens gaps. Corrosive media accelerates attack. That’s why identical defects can be a nuisance in one application and grounds for rejection in another.

The Cost Angle: Rework, Scrap, and Schedule

Then there’s the money. Slag inclusions are expensive because they’re often found late — after radiography or ultrasonic inspection, when the part is finished. In heavy structural fabrication, high FCAW reject rates — roughly 20% for fillets and 10% for grooves — translate directly into chipping, grinding, rewelding, re-inspection, and inspection overhead. Reject lots stall schedules and tie up qualified welders on repair work.

Compare that with the cost of preventing slag in the first place through better technique, polarity, and interpass cleaning. The gap isn’t close. Whether you’re running DCEN versus DCEP on flux core or dialing in parameters on automated work such as robotic submerged arc welding, the economics favor catching slag before it’s buried. For background on how these flaws are classified, the engineering literature on slag inclusions and consensus standards like ISO 5817 frame severity in terms of type, size, and location — precisely the variables that decide whether a defect is harmless or fatal.

Bottom line: treat every inclusion as an intentional notch until proven otherwise. Type, size, location, and service conditions decide the outcome — and the cost of being wrong rarely lands on the weld alone.

Prevention Playbook: Techniques, Consumables, and Procedures That Stop Slag at the Source

Slag inclusions rarely happen because a welder is careless. They happen because the pool was too cold to let slag float, a corner trap never got cleaned, or the rod was simply too fat for the joint. The good news: every one of those causes is controllable. Using qualified welders who follow proper welding procedures and high-quality consumables significantly reduces the risk of weld defects — and most of the rest comes down to discipline between passes.

Clean Every Pass — Completely

Incomplete interpass cleaning is the number-one practical cause of trapped slag. Chip the full length of the bead with a chipping hammer, then wire brush the joint, paying attention to toes, corners, and where the bead meets the sidewall. For deep grooves, a needle gun or grinder reaches what a brush can’t.

A quick reality check on speed: blast cleaning is faster than power wire brushing but usually does not thoroughly remove weld slag. Don’t let a fast process fool you into skipping the mechanical work that actually matters. If you can still see shiny glassy islands or a faint crust, keep going.

Angle and Technique That Let Slag Escape

Set your travel and work angles so the arc pushes the pool forward while slag stays behind it. Dragging too steeply or running a weave that folds slag under the bead both trap inclusions. Keep the bead profile slightly convex and avoid an overly concave or ropey bead that creates a slag shelf at the toe. Avoiding undercut matters too — undercut lips catch slag like a gutter.

Current and heat input keep the pool fluid enough for slag to rise and release. Run below the WPS range and the pool goes sluggish; run too hot and you risk undercut and burn-through. Welding parameters should always be checked against the WPS, because many welders set parameters based on what they were taught rather than the procedure’s acceptable ranges. That habit quietly plants defects.

Joint Design, Access, and Electrode Choice

The correct welding method and avoiding excessively thick electrodes in confined joints help prevent slag inclusions. A 5/32″ rod cannot reach the root of a tight 60° included angle without dragging slag into the corner — step down, improve access, or open the bevel. Bead sequence and profile control also matter: stagger passes, keep each bead where you can see and clean it, and don’t bury a narrow root pass under wide cover passes.

Pre-clean base metal of rust, mill scale, primer, and moisture before striking an arc — those layers become slag once the arc hits them. Control arc blow with proper work-lead placement and grounding, since a wandering arc deposits slag where you didn’t intend it.

Consumables, Procedures, and Verification

Store low-hydrogen electrodes in heated or sealed holding ovens and follow the manufacturer’s exposure limits; damp flux makes slag that’s gummy and hard to remove. Flux-cored wire needs dry storage too. Correct polarity is part of the recipe — get it wrong and the arc behavior, penetration, and slag release all change. Review polarity for flux core welding and gasless MIG welding guidance before you assume your setup is right.

Beyond gear, the human loop closes the deal. Build skill with structured welding practice, verify every welder against the WPS, and audit parameters on the floor rather than trusting memory. Check electrode amperage ranges against the procedure — our 3/16″ E7018 amperage and polarity guide shows how narrow those windows can be.

• Clean every pass — chip, brush, grind, needle-gun.
• Angle correctly so slag trails the pool.
• Check parameters against the WPS, not habit.
• Right-size consumables for joint access.
• Store low-hydrogen and flux-cored consumables dry.
• Verify welders and procedures continuously.

Short Note: Per AWS D1.1, documented procedures and qualified personnel are central to defect-free work. Treat cleaning and parameter checks as part of the procedure — not optional shop habits.

Stop slag at the source and the radiograph stays clean. Next, we’ll cover removal, repair, and rework when an inclusion still slips through.

Removal, Repair, and Rework — Plus the Bottom Line

So the film comes back with that dotted dark line — now what? The good news is that a slag inclusion is rarely a death sentence for a part. It is a repair job, but a repair that has to be done properly, documented, and re-inspected. Rushing it is how a small defect becomes a rejected lot.

Removal: Cutting Back to Sound Metal

A slag inclusion is non-metallic material trapped inside the weld, so you cannot simply weld over it and expect it to vanish. The accepted path is to cut back until you reach sound metal, then rebuild the joint.

Common removal methods include:

• Grinding with a carbide burr or grinding wheel.
• Air carbon arc gouging atssso called carbon arc gouging) for larger volumes.
• Chipping with a chipping hammer for shallow defects.
• Thermal cutting followed by grinding for deep, through-thickness problems.

A rule of thumb: remove metal until the joint is clean, then take a little more. It is far safer to leave a slightly generous repair than to chase a defect and smudge it into the surrounding metal.

Air carbon arc gouging leaves a thin skin of carbon and oxidized metal on the gouged surface. Grind that layer away before rewelding, or you simply feed new inclusions into the joint.

Removing slag can be easier or harder depending on the type of electrode coating. Some coatings peel away in satisfying curls; others cling and smear, forcing extra grinding — a factor that shapes slag behavior in flux-cored wire especially. Electrode selection therefore affects cleaning effort, so weigh it before you commit to a consumable.

Grinding and gouging throw sparks, grit, and molten metal, so suit up. Proper protective eyewear and a face shield are not optional when you are cutting back a defect.

Repair Welding and Re-Inspection

Repairs must follow the approved WPS. That means the same preheat, the same filler, the same interpass limits, and the same technique the original procedure called for. AWS D1.1 requires that unacceptable portions containing incomplete fusion, excessive porosity, or slag inclusions be removed and rewelded — there is no shortcut repair invented on the shop floor.

After rewelding, the repair is re-inspected to the same acceptance criteria as the original weld. A repair is not “good enough if it looks okay” — it has to pass the same test that failed the first time.

Documentation: NCRs, Repair Maps, and Re-Test

Every repair should be captured: an NCR (nonconformance report), a repair map marking the exact location, and the re-test result. On heavy structural work — think submerged arc passes on thick plate — that traceability matters as much as the weld itself. If the same spot fails twice, the paperwork tells you it is a process problem rather than bad luck, and that is a signal to revisit your technique, not just grind harder.

The Bottom Line

Step back and the pattern is clear. Slag inclusions are almost always a cleaning-and-technique problem: an interpass that was never chipped, an electrode angle that dragged slag into the puddle, or an amperage that ran too cold. They are preventable. Knowing your governing code — AWS D1.1, ISO 5817, or ASME — turns “is this okay?” into a confident yes or no.

The two biggest levers for cutting defect-driven reject rates are welding procedure variables and welder behavior. You will not control everything, but you control your habits — your prep, your travel angle, your interpass cleaning, your arc length.

So keep practicing with intention, treat cleaning as part of the weld rather than a chore after it, and let your code — not a guess — settle the “is this okay?” question. Disciplined habits are what separate a welder who fights defects from one who rarely sees them.

Frequently Asked Questions

What is a slag inclusion in welding?

A slag inclusion is non-metallic material — solidified electrode coating or flux residue — trapped inside the weld metal or at the weld interface. It occurs almost exclusively in flux-shielded processes such as stick (SMAW), flux-cored arc welding (FCAW), and submerged arc welding (SAW). AWS A3.0 defines it as a discontinuity consisting of slag entrapped in weld metal or at the weld interface.

Can MIG or TIG welding cause slag inclusions?

Not true slag inclusions, because solid-wire MIG/MAG and TIG do not generate the flux slag that stick, FCAW, and SAW do. However, MIG and TIG can trap silicate islands, oxide films, or tungsten particles (tungsten inclusions from dipping the electrode). If a solid-wire gas-shielded weld shows linear inclusions on radiography, suspect oxide films or lack of fusion instead.

Are slag inclusions ever acceptable?

It depends on the governing code and quality level. Under AWS D1.1, slag must be removed from completed welds, so visible slag is generally not acceptable. Under ISO 5817, small isolated inclusions may be permitted at quality levels C and D, but limits tighten at level B. Always work to the code specified in the contract.

How do you detect slag inclusions?

Visual inspection catches surface slag, but internal inclusions require radiographic testing (RT), which shows them as darker irregular indications, or ultrasonic testing (UT) for thicker sections. Destructive tests like bend tests and macroetch can also reveal trapped slag that is invisible on the surface.

What is the main cause of slag inclusions?

The dominant cause is failure to completely remove slag from one weld pass before depositing the next. Contributing factors include low heat input, wrong electrode angle, travel speed that is too fast, rough or convex bead profiles, undercut, and welding over rust, scale, or primer.

How do you prevent slag inclusions?

Clean every pass completely with a chipper and wire brush, use the correct current and heat input so slag floats out, hold the right electrode angle so slag stays behind the pool, avoid excessively thick electrodes in tight joints, control bead profile, and pre-clean the joint. Following a qualified WPS and using trained welders are the biggest safeguards.

Can a slag inclusion cause a weld to fail?

Yes. Inclusions act as internal notches and stress risers that can initiate fatigue cracks, reduce load-carrying capacity, create paths for corrosion and leaks, and contribute to fracture in demanding service. The risk depends on the inclusion’s size, type, location, and the service loading.

How are slag inclusions repaired?

The defective area is removed back to sound metal by grinding, gouging, or chipping, then rewelded according to an approved repair procedure. The repair is then re-inspected to the same acceptance criteria. Repeated repairs in the same area should be minimized because they can affect base-metal properties.

Authoritative Sources & Further Reading

• https://pubs.aws.org/Download_PDFS/B1.11M-B1.11-2015-PV.pdf
• https://www.aws.org/magazines-and-media/inspection-trends/2025/august/it-aug-25-the-answer-is
• https://pubs.aws.org/Download_PDFS/D1.1-D1.1M-2015-PV.pdf
• https://law.resource.org/pub/us/cfr/ibr/003/aws.d1.1.2000.pdf
• https://www.iso.org/standard/80209.html
• https://www.iso.org/standard/54952.html
• https://www.sciencedirect.com/topics/engineering/slag-inclusion
• https://www.tws.edu/blog/welding/what-is-slag-in-welding
• https://www.osha.gov/sites/default/files/publications/OSHA_FS-3647_WELDING.pdf
• https://www.osha.gov/hexavalent-chromium
• https://www.cdc.gov/niosh/docs/88-110/pdfs/88-110.pdf
• https://stacks.cdc.gov/view/cdc/191671/cdc_191671_DS1.pdf