How to Deal With a Fastener That Will Not Come Out

Why this matters

A seized fastener turns a short job into a long one, and the length is decided in the first ten minutes. Reach for the aggressive method early and you can convert a stuck bolt into a scrapped housing. Work the ladder in order and most of them come out with the component intact. The ordering below is not a preference, it is a sort against one stated measure, and the measure is worth naming before the list: how much of the assembly you can no longer restore if this step goes wrong. Sorted that way, the screw extractor ends up near the bottom rather than near the top, which is the opposite of how it is usually reached for, and the reason is in step 8.

Before any of this: the fastener may be holding pressure, weight, or a live conductor. Isolate the equipment and relieve pressure to zero, verified at the component rather than at the source, and restrain or block any spring or suspended load, which is 29 CFR 1910.147 in general industry for mechanical and stored energy. Where a conductor or termination is involved, isolate under 29 CFR 1910.333(b)(2), since 1910.147 excludes exposure to electrical hazards from work on conductors and equipment in electric utilization installations at (a)(1)(ii)(C). Eye and face protection under 29 CFR 1910.133 goes on before the first blow, because a hardened fastener that lets go throws pieces.

Step 1: Find out what is actually holding it

Nothing is lost here, which is why it is first, and skipping it is what makes every later step more expensive.

Check for a left-hand thread, especially on rotating assemblies and on anything that would otherwise unwind itself in service. Check for a second fastener, a retaining clip, a set screw, a tab washer or a locking wire that is still engaged. Check whether the fastener is bottomed in a blind hole rather than clamping anything. Check whether the assembly is under load, because a fastener carrying a jammed load is not seized, it is loaded, and relieving the load will free it.

If you can rotate the fastener a few degrees in the tightening direction, the thread is not welded to the hole and you are dealing with corrosion or debris, which is a step 3 problem. If it will not move either way at reasonable effort, expect to escalate.

Step 2: Fix the grip before adding force

Cost if it goes wrong: a rounded head, which jumps you straight to step 5's damage without step 5's benefit. That is the only reason it is not first.

Use a six-point socket rather than a twelve-point, seated fully, on the largest drive that fits, and clean the recess of a socket-head or star-drive screw with a pick before inserting the key, since packed debris is the usual reason a key does not seat. Then use shock rather than a steady pull: an impact driver or a sharp hammer blow on a correctly seated tool breaks corrosion bonds that a slow lean will not, and a slow lean is what rounds heads and snaps shanks.

Step 3: Penetrant, time, and cycles

Cost if it goes wrong: time, and only time. That is why it sits this high despite being slow.

Penetrating fluid works by capillary action into the thread, and it needs both time and movement. Apply, wait, shock the head, apply again, wait, shock again. Three cycles across roughly an hour beats one application and immediate force by a wide margin.

The hazard this step creates for the next one. Penetrating fluids are flammable and are usually delivered as an aerosol. Apply them away from any ignition source, ventilate the area, and do not apply penetrant and then bring heat to the same fastener, which is the single most common way this job starts a fire. If you are going to escalate to step 4, the penetrant has to be wiped off and given time to flash off first.

Step 4: Controlled heat on the female part

Cost if it goes wrong: the temper of nearby hardened parts, any seal, coating or plastic within reach of the heat, and a fire. Recoverable, but no longer only about the fastener.

Heat expands the female member away from the male one and fractures the corrosion bond. Heat the boss or the nut, not the bolt. The clearance is greatest while the female member is hot and the fastener has not yet warmed through, so that is the moment to try it, at reduced torque, because hot metal is weaker and strips more easily than cold. If it does not move, let it cool fully and repeat: it is the cycling that fractures the corrosion product, not the cooling direction.

Everything about this step is gated:

• Never apply heat to a component containing flammable liquid, fuel gas, refrigerant, or any pressurised fluid, and never to a fastener that forms part of a pressure boundary while that boundary is pressurised. Isolate, relieve, drain and vent first.
• Hot work needs a fire watch and a cleared area. Fire prevention for welding, cutting and heating sits at 29 CFR 1910.252(a) in general industry and 29 CFR 1926.352 in construction, and NFPA 51B, in the edition your authority having jurisdiction or your insurer has adopted, is the hot work permit standard that reaches you through that adoption or through the site's own program rather than as federal law on its own.
• Heat creates an inhalation route, and the control is respiratory. A zinc-plated or galvanized fastener releases zinc oxide fume; a cadmium-plated one releases cadmium oxide, which is far more serious; thread sealant tape and paste based on fluoropolymer decompose above roughly 500 F and release hydrogen fluoride and polymer fume. Remove sealant and coating where you can, use local exhaust, and wear a respirator selected and fit-tested under a written program meeting 29 CFR 1910.134. Gloves and a face shield do not address this route.
• Protect what is nearby. Elastomer seals, wiring insulation and plastic components within a few inches of the heat need removing or shielding before the torch lights.

Step 5: Cut a new drive feature

Cost if it goes wrong: the head. Once you have slotted, notched or welded on it, socket and wrench options are gone for good.

On a rounded or broken head that still protrudes, the options are a slot cut for a large flat blade, flats ground or filed onto a round head, a nut welded onto the stub, or a gripping socket designed to bite a damaged head. Welding a nut on is the strongest of these because it delivers grip and a heat cycle at the same time, and the heat is applied exactly where step 4 wanted it.

Welding brings its own set: the same hot work requirements as step 4, welding fume needing local exhaust and respiratory protection under a 1910.134 program, arc-flash eye protection, and isolation of the equipment so welding current does not find a path through bearings or electronics.

Step 6: Drill out the fastener on centre

Cost if it goes wrong: the fastener is gone regardless, and a drill that wanders damages the female thread.

Centre-punch the fastener, start with a small pilot, and work up, keeping the drill square to the axis, which is what a drill guide or a magnetic-base drill buys you. Left-hand drill bits are worth carrying, because a left-hand bit cutting into a corroded fastener sometimes backs it out during the pilot pass and ends the job at this step. Once the fastener is drilled to near its minor diameter, the remaining shell can often be collapsed inward with a pick and unwound, leaving the original thread intact for a chaser.

Step 7: Drill oversize and repair the thread

Cost if it goes wrong: the original thread size is permanently gone. Recoverable as a working joint, but the component is no longer as-built.

Drill to the insert manufacturer's specified size, tap, and install a coil or solid insert, or tap up to the next size where the design allows and the wall thickness supports it. The result is often stronger than the original in a soft housing, and the sibling article on thread engagement explains why a steel insert in aluminum changes the engagement requirement. Record the repair on the equipment history, because the next tech meeting a steel thread in an aluminum casting needs to know it was engineered rather than original.

Step 8: The hardened screw extractor

Cost if it goes wrong: everything above stops being available.

This is the finding, and it is why the extractor sits here rather than at step 3 where most people reach for it. A screw extractor is hardened well beyond the fastener and beyond ordinary drill bits. If it snaps off inside the pilot hole, that fragment cannot be drilled with field tooling. Steps 6 and 7 are both off the table, and the job moves to step 9 or to a machine shop with electrical discharge machining.

So if you use one, use it under the conditions that keep it from breaking: a correctly sized pilot hole drilled square, penetrant and a heat cycle already applied, steady low-speed torque with no shock and no side load, and a stop the moment it stops turning. And use it knowing that the reason it is low on this list is not that it works badly, but that its failure mode removes your other options.

Step 9: Replace the component

Total, by definition, which is what puts it last. It is a legitimate call, and calling it early on a low-value part that is fighting you is often the cheapest decision on the whole list. Where the fastener is in a housing with a long lead time, that calculation runs the other way and every step above it earns its time.

The escalation, worked

Six 3/8-16 steel cap screws in a cast iron pump housing, outdoors, corroded. Four came out normally. Two did not, and one arrived with a partly rounded head.

• Steps 1 and 2, about 15 minutes. No hidden retainers, standard right-hand thread, both fasteners moved a few degrees in the tightening direction, which said corrosion rather than a welded thread. Six-point sockets seated, impact driver used. No movement.
• Step 3, about 1 hour. Penetrant, three shock cycles roughly 20 minutes apart. One of the two broke free on the third cycle. One of two resolved, for an hour of clock time and no risk taken.
• Step 4, about 40 minutes. Housing isolated and drained, an elastomer seal two inches from the boss removed, area cleared, fire watch posted, penetrant wiped and flashed off. The boss was heated and allowed to cool through two cycles. Still seized.
• Step 5, about 35 minutes. Head already rounded, so a nut was welded onto the stub, with the equipment isolated, local exhaust at the arc, and respiratory protection. It came out on the first pull after the weld cooled.

Total about 2.5 hours against a 20-minute planned task, a factor of about 7.5 on the labour for that operation. That is the real cost of the disciplined route, and it is worth stating plainly rather than pretending the ladder is free.

Set it against the alternative. Going to an extractor at minute 10, on a fastener that had seen no penetrant and no heat cycle, is the condition in which extractors break. A snapped extractor in that boss takes the job from a same-day repair to a housing replacement with whatever lead time that part carries, plus the return visit. The ladder is not slower than the shortcut. It is slower than the shortcut working, which is a different comparison.

Checking you got this right

• Before escalating a step, say what the previous step told you. "It moved in the tightening direction" and "it did not move at all" lead to different next moves.
• Confirm penetrant is wiped and flashed off before any flame or arc reaches the area, every time, not just the first time.
• On anything you heated, check what was within reach of the heat afterwards: seals, wiring, plastics, and the temper of nearby hardened parts.
• Where the female thread survived, chase it and test-fit a new fastener by hand before assembly. A hole that took the drill without complaint can still have a rolled thread.
• Write the escalation into the job notes with the step that worked. The next person on that machine gets to start at step 4 instead of step 1.

References

• 29 CFR 1910.147, control of hazardous energy, and 29 CFR 1910.333(b)(2) for the electrical case that 1910.147 excludes at (a)(1)(ii)(C)
• 29 CFR 1910.252(a) general industry and 29 CFR 1926.352 construction, fire prevention for welding, cutting and heating; NFPA 51B for hot work permitting, in the edition adopted by the authority having jurisdiction, the insurer, or the site program
• 29 CFR 1910.134, respiratory protection program requirements, for metal fume and decomposition products released by heating plated fasteners and thread sealants
• 29 CFR 1910.133, eye and face protection
• See related: Thread Engagement and Why Two Threads Are Not Enough; How a Thread Fails and What the Failure Tells You; Cross-Threading Prevention