TorqueReference

Bolt torque, grades and vehicle specs, sourced

Try a size (“M10”), a grade (“grade 8”) or a vehicle (“Tacoma”).

Menu navigation

Anti-seize and torque: how much to reduce, by compound

With copper or metal-free anti-seize, use about 20% less torque than a plain dry spec (K 0.16 instead of 0.20). Moly paste needs about 45% less. The exact figure depends on the compound, and some makers, NGK for spark plugs and most carmakers for wheel studs, say not to use anti-seize at all.

By the Torque Reference editors 18 sources cited Last reviewed Method
Contents8 sections

Key takeaways

  • Henkel publishes measured K values: 0.16 for C5-A copper, heavy-duty and N-7000 metal-free on steel; 0.18 for marine and silver grades; 0.11 for moly paste ⁠1 ⁠2 ⁠3 ⁠4 ⁠5.
  • From a plain dry spec (K 0.20) that means 20% less torque for K 0.16, 10% less for 0.18 and 45% less for 0.11.
  • Nickel anti-seize LOCTITE LB 771 has no published K ⁠6. Don't borrow another compound's value.
  • NGK says not to use anti-seize on its spark plugs: it can change torque by up to 20% and risk thread breakage ⁠7.
  • Wheel stud torque is a dry spec. Ram: 'Do not oil wheel studs' ⁠8.

Quick facts

Copper (C5-A, LB 8007)K 0.16 ⁠1
Metal-free (N-7000, LB 8013), steelK 0.16 ⁠2
Marine (LB 8023)K 0.18 ⁠3
Moly paste (LB 8012)K 0.11 ⁠5
Nickel (LB 771)Not published ⁠6

Anti-seize is a lubricant

Anti-seize compounds are made to stop threads galling and seizing, but on the way in they also cut friction. Less friction means more of the torque turns into bolt stretch, so a torque written for dry threads clamps harder with anti-seize on them. Henkel’s data sheets state the effect as a torque coefficient, K, measured on bolts ⁠1. That is the same K as in T = K × D × F ⁠9, so the reduction is simply the ratio of the K values.

K values by compound

These are the values in Henkel’s LOCTITE technical data sheets. Each was measured on specific hardware, listed here because it matters.

LOCTITE anti-seize torque coefficients from Henkel TDS
Product Type K Test hardware Unlubricated baseline
LB 8007 (C5-A) Copper 0.16 1/2 in. Grade 8 bolt, Grade 5 nut 0.27 ⁠1
LB 8065 (C5-A stick) Copper 0.16 Grade 5 phosphate-coated bolts and nuts 0.19 ⁠10
LB 8009 Heavy Duty Metal-free (graphite, calcium fluoride) 0.16 1/2 in. Grade 8 / Grade 5 0.27 ⁠11
LB 8013 (N-7000) Metal-free, high purity 0.16 1/2 in. Grade 8 steel / Grade 5 nut 0.27 ⁠2
LB 8013 (N-7000) Metal-free, high purity 0.18 1/2 in. 304 stainless bolt and nut – ⁠2
LB 8023 Marine (calcium sulfonate) 0.18 3/8-16 steel bolts and nuts – ⁠3
LB 8060 Silver grade stick 0.18 3/8-16 steel bolts and nuts – ⁠4
LB 8012 Moly paste 0.11 1/2 in. Grade 8 / Grade 5 0.27 ⁠5
LB 771 Nickel not published – – ⁠6
LB 8150 Silver grade not published – – ⁠12

Scroll the table sideways to see every column.

How much to reduce

Keep the clamp load the same, change K: T_anti-seize = T_dry × K_anti-seize ÷ K_dry.

The answer depends on what “dry” meant in the original spec. Against this site’s plain dry K of 0.20 ⁠9:

Torque reduction from a plain dry spec (K 0.20)
Anti-seize K Multiply dry torque by Reduction
0.18 (marine, silver) 0.90 10%
0.16 (copper, metal-free on steel) 0.80 20%
0.11 (moly) 0.55 45%

Against Henkel’s own solvent-cleaned baseline of 0.27, copper anti-seize at 0.16 is a 41% reduction. Against the phosphate-coated baseline of 0.19 in the LB 8065 sheet, the same 0.16 is a 16% reduction ⁠10. That spread is why a flat “reduce by 25%” rule can’t be right for every bolt.

Worked through on a 1/2-13 Grade 8 bolt at 75% of proof (12,780 lbf): 107 ft-lb (144 N·m) plain dry becomes 85 ft-lb (116 N·m) with copper anti-seize and 59 ft-lb (79 N·m) with moly paste. The lubricated torque calculator runs any combination.

What happens if you don’t reduce

Torque a dry spec onto anti-seized threads and the clamp load rises by 0.20 ÷ 0.16, or 25%. A bolt meant to sit at 75% of proof load ends up near 94%, with the ±25% torque-wrench scatter ⁠13 now pushing the high side past proof. With moly paste the rise is 82%, about 136% of proof: past proof load before any scatter.

Where not to use anti-seize

Spark plugs (NGK). NGK plugs have trivalent plating that resists corrosion and “acts as a release agent during spark plug removal”, and they are installed at the factory dry. NGK: “Anti-seize can act as a lubricant, altering torque values up to 20 percent, increasing the risk of spark plug thread breakage and/or metal shell stretch,” and “Do not use anti-seize or lubricant on NGK spark plugs” ⁠7.

Spark plugs (Denso). Denso’s US installation page says its torque values are for new plugs with unlubricated threads, and if threads are lubricated the torque “should be reduced by approximately 1/3” ⁠14. Denso Europe says the torque must be 2/3 of normal with oil or lubricant, and that Denso does not recommend lubricants ⁠15.

When the vehicle maker says otherwise. Ford’s TSB for 4.6L, 5.4L and 6.8L 3V engines calls for new plugs to be installed with a film of Motorcraft high-temperature nickel anti-seize on the ground electrode shield, and not on the electrode strap ⁠16. The engine maker’s instruction overrides the plug maker’s general advice. See spark plug torque.

Wheel studs. Ram’s owner’s manual: “Do not oil wheel studs” ⁠8. Toyota: “Never use oil or grease on the wheel bolts or wheel nuts” ⁠17. Wheel nut torques are set for clean, dry threads, and anti-seize on them raises the clamp load the same way. See lug nut tightening.

Practical rules

  • Use the K for the compound you have. If the data sheet gives none, as for LB 771 nickel ⁠6, you don’t have a basis for a reduction. Test it on the joint or follow the equipment maker.
  • Coat both mating threads and the bearing face the same way every time. Lubricant location changes the torque-preload relationship ⁠18.
  • Don’t over-apply. Fastenal’s example of an excessively lubricated zinc-plated bolt has K falling to 0.10 ⁠9.
  • Stainless is not the same as steel. N-7000 on 304 stainless measured 0.18, not the 0.16 it gave on steel ⁠2.
  • A spec written “with anti-seize” already includes it. Don’t reduce it again.

Sources

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
  11. 11
  12. 12
  13. 13
  14. 14
  15. 15
  16. 16
  17. 17
  18. 18

Spotted a value that doesn’t match its source? Report an error