
By Geoff Lord · 9 August 2026
Anti-seize is the most misused product on the shelf. It goes on threads that never needed it, and it is missing from the ones that did.
The reason is simple. Most people treat the paste as insurance, when it is really a friction modifier.
Change the friction in a threaded joint and you change the clamp load a given torque produces.
That is the part the tin never mentions.
This guide covers what anti-seize does, what each type is rated for, and where it should never go. The numbers come from a NASA fastener manual and from a spark plug maker, not from packaging. Nothing here is estimated.
Two metal surfaces under load do not stay separate. High spots weld together, tear, and drag fresh metal into the gap. That is galling, and on threads it ends with a fastener that will not move.
Anti-seize puts solid particles between those surfaces. The carrier oil spreads them, then burns or washes away, and the solids stay behind. What is left is a film that cannot cold weld.
NASA's fastener design manual describes one common compound as leaving non-galling oxides of nickel, copper and zinc between the threads. The joint comes apart later because those solids never bonded.
Grease lubricates while it is still there. Anti-seize is built to keep working after the liquid part has gone.
These are the default thread lubricants and the weakest link on heat. NASA puts their useful limit at 250 degrees Fahrenheit, above which the carrier melts or boils away, and none of them will hold up in a vacuum. Anything hotter than a warm gearbox needs a solid film instead.
Dry graphite is not really dry, because it needs moisture before it lubricates at all. That gives it a working limit of roughly 212 to 250 degrees Fahrenheit, set by the carrier, and past that point the graphite is simply an abrasive sitting inside your joint.
Moly is the workhorse dry lubricant, and it does work in a vacuum. Its ceiling is about 750 degrees Fahrenheit. Above that it converts to molybdenum trisulfide, which grinds rather than lubricates.
This is what most people mean by anti-seize. NASA records one proprietary compound as satisfactory to 2200 degrees Fahrenheit as a one-time lubricant, which is a little over 1200 degrees Celsius.
Because the oil boils off, you have to brush on a fresh coat at every reassembly.
Silver-bearing pastes hold up to 1500 degrees Fahrenheit. They carry 20 to 30 percent silver and cost accordingly.
NASA warns against using them on aluminium or magnesium parts.
| Lubricant type | Useful limit | Vacuum safe | Note |
|---|---|---|---|
| Oil or grease | 250 F | No | Most common, lowest ceiling |
| Graphite | 212 to 250 F | No | Abrasive past the carrier limit |
| Molybdenum disulfide | 750 F | Yes | Turns abrasive above the limit |
| Metal-filled anti-seize | 2200 F | Not stated | Reapply at every install |
| Silver-loaded paste | 1500 F | Not stated | Never on aluminium or magnesium |
A torque wrench measures resistance, not tension. Most of what it reads is friction, and anti-seize exists to cut friction.
The standard shortcut is a torque coefficient, written as K. NASA warns that the value everyone assumes by habit should not be used blindly.
Halve the friction and you nearly halve K. The same wrench setting then pulls far more tension into the bolt.
| Friction coefficient | Torque coefficient K | Effect at a fixed torque |
|---|---|---|
| 0.05 | 0.074 | Highest clamp load, yield risk |
| 0.10 | 0.133 | Well above the dry assumption |
| 0.15 | 0.189 | Realistic for steel on steel |
| 0.20 | 0.250 | Close to an unlubricated joint |
NASA's own bolt torque table assumes dry, as-manufactured threads carrying no lubricant at all.
There is no single discount that fits every compound and every finish. NASA does give one worked case, telling engineers to cut dry torque values by 50 percent on waxed silver-plated nuts. Treat that as the scale of the effect rather than as a universal rule.
Galvanic corrosion needs two dissimilar metals and an electrolyte, which in practice means moisture. The farther apart the two metals sit in the galvanic ranking, the harder that cell works.
Anti-seize adds a third metal to the joint, so the filler is part of that equation.
The active metal erodes, so it should be the larger mass in the pair. NASA calls carbon steel fasteners in a stainless or copper assembly poor practice, while stainless fasteners in a carbon steel assembly are acceptable.
Stainless galls because its own oxide film keeps rewelding under pressure. NASA notes that furnace-oxidised 300 series stainless galls less, and that passivation deters galvanic activity, yet neither treatment is available to you in the middle of a shift.
Anti-seize is the practical substitute. Stock it beside your other anti-seize compounds rather than reaching for ordinary grease.
Modern plugs arrive plated and dry, and the plating is the release agent. NGK states plainly that it ships plugs from the factory dry, with no lubrication and no anti-seize.
The paste can alter torque values by up to 20 percent, which risks thread breakage and metal shell stretch.
Shell stretch changes the heat rating of the plug and can trigger pre-ignition. That is engine damage caused by a helpful habit.
Some makers qualify an assembly around one named compound and nothing else.
A compound smeared over rust and old thread lock does nothing useful. Wire brush the threads, then wipe them.
Trapped grit turns the paste into a grinding compound.
The oil carrier boils off in service, so the joint you open next year holds solids only. NASA is explicit that each installation needs a fresh application, and skipping that step is how a good habit quietly stops working.
That discipline is what keeps anti-seize useful across the life of the plant. Keep a tin beside the penetrating lubricants so the two get used in the right order.
Yes, and by more than most workshops allow for. NASA's figures show the torque coefficient falling by almost half when thread friction halves, while NGK measures shifts of up to 20 percent on spark plug threads.
It depends entirely on the metals in the joint, rather than on which tin looks tougher on the shelf. Keep copper-rich pastes away from aluminium and magnesium, and match the filler metal to the parts you are bolting together.
They solve different problems, because grease carries load while it is present and anti-seize keeps working long after the liquid has gone.
Enough to shade the thread flanks and no more than that. Anything squeezing out of the joint is wasted product that will only attract grit and dust.
The solid fillers do not, but the oil carrier separates out of the paste over time. Stir it back to an even consistency before use, and replace the tin if the oil refuses to remix.
Anti-seize earns its place on high-temperature joints, stainless threads and anything exposed to weather. It is the difference between a ten-minute strip and a broken stud.
The cost is that it changes the joint. A lubricated thread reaches a higher clamp load at the same torque, so the setting has to come down with it.
Treat anti-seize as a specification decision rather than a habit. Check the torque figure, match the filler to the metals in front of you, then buy the right tin alongside the grease you already stock. Doing those three things in order is the whole discipline.