Non-metallic flux frozen inside the weld metal — trapped between passes, in undercut, or at the root. It is a flux-process signature, and nearly always a cleaning or technique failure you can fix on the next pass.
Slag inclusion is a solid, non-metallic discontinuity: flux from the electrode coating or wire core that never floated to the surface, frozen inside the weld metal. It forms only in flux processes — stick, FCAW, and SAW — where molten slag rides on the pool.
Slag is a planar, crack-like discontinuity, not a round void. Under load it behaves like an internal notch, which is why codes treat it seriously — and why multipass welds in flux processes live and die by interpass cleaning.
Slag entrapment, flux inclusion, trapped slag.
Read the bead before you chip it — then after. The second look is the one that counts.
Slag inclusion needs two things: a flux process, and a place for the slag to hide.
Every flux pass leaves molten slag on top of the pool — it floats because it is lighter and stays liquid longer than the steel. On a clean, open pass the slag cools behind the weld and chips off. The trouble starts when the slag has somewhere to hide: a gap, an undercut, a rough side wall, or a weave edge that folded over. The next pass melts the steel it touches but not the slag in those pockets — so the slag is never remelted and floated out. It freezes in place, and the following pass buries it.
That is why the defect is so often elongated: it follows the line of an undercut or a weave fold for the whole pass, showing up on RT as a long, irregular dark indication. And it is why interpass cleaning is not cosmetic — every pass you skip chipping and brushing is a pass with a built-in inclusion.
Slag inclusion is a habit problem before it is a weld problem. Fix the habits and it stops showing up.
Fit the electrode or wire to the groove — small enough to angle into the joint — and open the groove if it is so tight the slag has nowhere to go.
Run the amperage high enough that the slag floats behind the puddle, hold the correct angle, and keep the weave to about 2–3 times the electrode diameter with a pause at each edge.
Chip and wire-brush every pass to bright metal. Grind out undercut, surface flaws, and any pocket the chipping hammer exposes before you cover it.
Slag hides until you chip it. A bead that looks perfect on top can be sitting on a row of inclusions — the first pass you cover without cleaning is the pass that fails RT.
When slag is already in the weld, the repair is excavation — but only after the cause is fixed.
| Step | Action | Why |
|---|---|---|
| 1 | Fix the cause — cleaning, parameters, technique — before touching the weld. | Re-welding with the same habit reproduces the inclusion. |
| 2 | Locate the extent: chip and brush, then VT; use RT or UT where internal indications are suspected. | Inclusions run along the weld — the visible pocket is not the whole story. |
| 3 | Excavate to sound metal with a grinder or gouge, following the full length of the indication. | Leaving a tail behind leaves the rejection behind. |
| 4 | Verify removal with VT and PT on the groove before re-welding. | Never re-weld over unconfirmed slag. |
| 5 | Re-weld to the qualified WPS with correct interpass cleaning. | The repair pass must meet the same standard as the original. |
Never cap a slag inclusion to bury it — it is a planar discontinuity that behaves like a crack under load. Excavate fully, confirm removal, and repair within the scope the applicable code allows.
Flux processes trap slag their own way. Read yours against the list.
Low amps push the slag ahead of the arc; too steep a rod angle ploughs it into the side wall; overlapping beads without cleaning trap it at the overlap. Deep, narrow grooves are the classic hunting ground.
Stick-out too long or the wrong drag angle makes the slag run ahead of the puddle; wire feed and voltage out of balance create a cold, slag-heavy pool. Gas-shielded and self-shielded wires differ — set polarity per the wire.
Voltage too high or travel speed wrong for the flux band draws slag into the weld; damp, contaminated, or recycled flux changes the slag chemistry. Flux quality and storage are the SAW-specific checks.
Not flux processes — true slag inclusion cannot form. If a MIG or TIG weld shows embedded non-metallic material, look at lack of fusion, oxide films, or tungsten inclusions instead of blaming slag.
When slag appears, check these in order — the first three fix most jobs.
| Parameter | Check | Typical Fix |
|---|---|---|
| Interpass cleaning | Every pass chipped and brushed to bright metal | Grind out pockets and undercut; never cover an uncleaned bead |
| Amperage / voltage | Slag stays behind the puddle; SAW voltage in the flux's window | Raise amps until the slag floats; correct voltage for wire and flux |
| Weave width | About 2–3× electrode diameter, pause at the edges | Narrow the weave; use stringer beads in tight grooves |
| Travel angle | Stick ~10–15° drag; FCAW ~10–20° drag | Keep the arc pointing into the joint, slag behind the pool |
| Pass sequence | Root and fill geometry; overlap of adjacent beads | Open the groove slightly; clean between every pass |
Use the Welding Calculator to set a starting amperage for the electrode and plate, and the Heat Input Calculator when travel speed and preheat are part of the trap — slag freezes fast on cold joints.
For slag, the consumables are the story — the machine just delivers them.
For stick and FCAW the electrode and wire type must fit the groove, with the coating and core in good condition. For SAW, flux storage, drying, and recycling deserve the same attention as the machine — a moist or contaminated flux changes the slag into a defect.
Slag is found two ways: with a chipping hammer, and with NDT on critical welds.
Acceptance depends on the code and project: AWS D1.1, ISO 5817, API 1104 and project specifications each set their own limits for this discontinuity. Check the governing document before judging a weld acceptable — there is no universal pass/fail rule.
The questions welders actually ask about slag — answered straight.
No. Surface slag shows as flakes and pockets once you chip the bead, but slag trapped between passes or at the root is internal and only appears on radiography or ultrasonic testing. A weld can look clean on top and still be rejected.
Because interpass cleaning is being skipped or done too lightly — brush to bright metal on every pass. If cleaning is thorough and slag still appears, the previous pass left undercut or a rough side wall that brushing cannot reach; grind those out before covering them.
Yes — that is the normal case. The top bead can look perfect while earlier passes carry inclusions along the side walls or root. That is exactly why multipass welds on critical work get radiography instead of a surface judgment.
Both are flux processes and both trap slag when technique or cleaning fails — stick in deep grooves and root passes with low amps or the wrong angle, FCAW with too much stick-out or an unbalanced wire feed and voltage. SAW traps it too, usually from voltage or flux problems. There is no unlucky process — only skipped cleaning.