Lubricated vs Dry Bolt Torque: The K-Factor Chart (and How Much to Cut the Spec)
By WhatSizeBoltEdited 11 min read10 source links
Use the specified assembly condition and exact product instructions; no universal lubricant or threadlocker torque correction applies.
Scope and sources
Applies to: General torque/friction explanation. Calculated examples are not vehicle installation specifications.
Henkel: LOCTITE Threadlocker Manual, torque–tension discussion — Tightening friction depends on product and joint configuration. Cured break-loose torque is a different property.
Real K-factors, cited to their sources: how much to cut a dry torque spec for anti-seize or oil, a computed dry/anti-seize/oiled chart by size, and a worked example showing why the same torque over-stretches a lubricated bolt.
You just ran a bead of anti-seize onto a bolt’s threads and picked up the torque wrench. Every guide on this site — and most others — tells you to “reduce the torque.” Almost none of them tell you by how much, or where that number comes from. This page is the numbers: real K-factors, cited to the people who publish them, run through the same equation your torque wrench is built around.
The short version first, then the full derivation. If you came from oil vs anti-seize or the copper vs nickel anti-seize guide, this is the page both of them are pointing at.
Torque is a proxy for clamp load, not the goal itself
Nobody actually wants torque. What holds a joint together is clamp force — the tension the bolt is stretched into, squeezing the parts together like a spring. Torque is just the easiest thing to measure at the wrench, so every spec sheet expresses the target clamp force as a torque number instead, using the standard relationship:
K is deliberately not a clean physical property; it bundles up geometry and two separate friction coefficients — one in the threads, one under the bolt head or nut face — into a single number so a torque chart can stay a one-column lookup. That convenience is also the trap: change either friction surface and the single number stops being true.
Where the torque actually goes
Turn a wrench and the energy splits three ways: friction under the bolt head (or nut face), friction in the threads, and the sliver left over that actually stretches the bolt. Nord-Lock’s own breakdown puts the dry split at roughly 50% underhead friction, 40% thread friction, and only 10% useful clamp force — a number so lopsided it is worth repeating: nine-tenths of a dry torque reading is wasted overcoming friction before any of it stretches the bolt.
Run the same M10 bolt through the K-factor formula with a specific anti-seize paste’s published friction coefficients (Molykote’s G-n Plus, μₜ=0.12 in the threads, μₕ=0.06 under the head) and the useful slice grows from 10% to 19% — nearly double. That is the whole mechanism in one number: lubricant does not make the wrench easier to turn for free. It hands more of every turn to the bolt instead of to friction, which is exactly why the same torque reading now clamps harder.
The K-factor by condition, and who publishes it
K-factor tables disagree with each other — not wildly, but enough that quoting a single number without its source is how “reduce by 10–25%” became the industry’s favorite non-answer. Here is a table that keeps the source attached to every value:
| Condition | K-factor | Source |
|---|---|---|
| Waxed | 0.10 | Portland Bolt (T=KDP torque-chart methodology) |
| Anti-seize / moly paste | 0.13–0.20 (0.13–0.15 for premium nickel/moly pastes) | Molykote (Dow Corning) threaded-connections data sheet, 11 pastes measured on 5/8” UNF |
| Cadmium plated | 0.16 | Shigley’s Mechanical Engineering Design |
| Lubricated (engine oil) | 0.18 | Shigley’s Mechanical Engineering Design |
| ARP Ultra-Torque assembly lube | ≈0.17 | Back-computed from ARP’s own torque/clamp-load table |
| Zinc plated / as-received | 0.20 | Shigley’s; Portland Bolt (“plain, non-plated”) |
| Hot-dip galvanized | 0.25 | Portland Bolt |
| Black oxide, dry | 0.30 | Shigley’s Mechanical Engineering Design |
PTFE-coated fasteners sit at the low end even of that spread: Metal Coatings Corp publishes a kinetic friction coefficient of just 0.06–0.08 for its FluoroKote #1 fluoropolymer coating — among the lowest of any commercial thread treatment, which is why PTFE-coated bolts need the largest torque cut of anything on this page.
The chart: dry vs. anti-seize vs. oiled, by size
Computed, not typed in: tensile stress area from the ISO 898-1 / UN-thread formula, clamp target at 75% of proof load (the standard convention Portland Bolt and ARP both publish their own charts around), torque from T=K·F·d at each K-factor above. The dry column is checked against this site’s own published /standard-torque-specs/ and /resources/metric-bolt-torque-chart/ ranges and lands inside them at every size (one size, 9/16” Grade 8, comes in a few percent above the top of the published band — noted rather than quietly rounded away).
| Size (8.8) | Dry, Nm (K=0.20) | Anti-seize, Nm (K=0.14) | Oiled, Nm (K=0.17) | Cut vs. dry (anti-seize) | Cut vs. dry (oiled) |
|---|---|---|---|---|---|
| M8 | 25 | 18 | 22 | −30% | −15% |
| M10 | 50 | 35 | 43 | −30% | −15% |
| M12 | 88 | 62 | 75 | −30% | −15% |
| M14 | 141 | 98 | 120 | −30% | −15% |
| M16 | 218 | 153 | 185 | −30% | −15% |
| Size / grade | Dry, ft-lb (K=0.20) | Anti-seize, ft-lb (K=0.14) | Oiled, ft-lb (K=0.17) | Cut vs. dry (anti-seize) | Cut vs. dry (oiled) |
|---|---|---|---|---|---|
| 3/8" Grade 5 | 31 | 22 | 26 | −30% | −15% |
| 3/8" Grade 8 | 44 | 31 | 37 | −30% | −15% |
| 7/16" Grade 5 | 49 | 35 | 42 | −30% | −15% |
| 7/16" Grade 8 | 70 | 49 | 59 | −30% | −15% |
| 1/2" Grade 5 | 75 | 53 | 64 | −30% | −15% |
| 1/2" Grade 8 | 106 | 74 | 90 | −30% | −15% |
| 9/16" Grade 5 | 109 | 76 | 92 | −30% | −15% |
| 9/16" Grade 8 | 154 | 107 | 130 | −30% | −15% |
The percentage cut is constant across every size because it falls straight out of the ratio of K-factors, not the bolt’s dimensions — anti-seize cuts about 30% off the dry number regardless of whether the bolt is an M8 or a 9/16”, and oil cuts about 15%. That constant-percentage shortcut only holds as long as you are scaling a torque spec for the same grade and the same lubricant — it is not a substitute for an OEM-specific wet torque value where one is published.
Why the same torque over-stretches a lubricated bolt
A worked example, not a rule of thumb.
Take an M10 class 8.8 bolt with a dry torque spec of 50 N·m — a round number pulled straight from the chart above. That spec was written assuming K=0.20, targeting a clamp force of about 25.2 kN (75% of the bolt’s 33.6 kN proof load).
Coat the same threads in anti-seize (K≈0.14) and apply the identical 50 N·m — the number on the chart, because nobody handed you a “wet” spec. The clamp force you actually get is F=T/(K·d) = 35.7 kN.
What "torque values are for dry, clean threads" actually means
When a manufacturer’s spec sheet says a torque value assumes “dry and clean” threads, it is not a courtesy note — it is the K-factor the whole table was calculated with, spelled out in words instead of a number. The corollary that trips people up: thread-locking compound is the one common exception. Products like Loctite 243/271 add friction rather than removing it, so most manufacturers state the torque is unchanged when threadlocker is specified — do not apply an anti-seize-style reduction to a thread-locked joint.
The torque-to-yield exception
Torque-to-yield (TTY) bolts are markedly less sensitive to all of this. A standard elastic joint is tightened to a point on the straight part of its stress-strain curve, where clamp force is directly proportional to torque — exactly the regime where a K-factor error translates one-for-one into a clamp-force error. A TTY bolt is deliberately torqued past that point, onto the flat, yielded part of the curve (torque-then-angle, not torque alone), where large changes in applied load produce only small changes in stress. That flatness is what makes TTY specs comparatively forgiving of small friction changes — and also why a TTY bolt cannot simply be re-torqued or reused once it has been there. The mechanics of that stretch curve are covered in full in A Bolt Is a Spring.
Some links below are Amazon affiliate links — as an Amazon Associate we may earn a small commission at no extra cost to you.
EPAuto ½” Drive Click Torque Wrench, 10–150 ft-lb
The K-factor chart above is only as good as the wrench applying it. A calibrated click wrench in the 10–150 ft-lb band covers everything on this page.
OEMTOOLS 25424 ½” Torque Angle Gauge
Torque-to-yield bolts are set by torque-then-angle, not torque alone — this is the tool that reads the angle half of that spec accurately.
Permatex 77124 Nickel Anti-Seize, 8 oz
The premium nickel/moly pastes are the ones the Molykote data sheet clusters at K=0.13–0.15 — the number this page’s anti-seize column is built on.
Frequently asked questions
How much do I reduce torque for anti-seize?
About 30% off the dry spec for a quality nickel or moly anti-seize paste (K drops from ≈0.20 dry to ≈0.13–0.15). Heavier, less-refined pastes can sit closer to K=0.18–0.20 and need less of a cut — check the product’s own data sheet where one is published.
How much do I reduce torque for engine oil?
About 15% off the dry spec (K≈0.18 down to ≈0.16–0.17 for a light oil film). This is the assumption baked into most torque-to-yield head-bolt procedures that call for oiled threads.
What is a K-factor / nut factor?
The dimensionless number K in T=K·F·d, where T is torque, F is the target clamp force and d is the nominal bolt diameter. It bundles thread friction and under-head friction into one number so a torque chart can stay a single lookup value instead of a full friction calculation.
Why does the same torque clamp harder with anti-seize?
Because torque is a fixed budget split between friction and stretch. Lubricant lowers friction, so a larger share of the identical torque reading goes into stretching the bolt — which is exactly what clamp force is.
Do I need to reduce torque for thread-locking compound?
No — most thread-lockers add friction rather than remove it, so manufacturers typically specify the torque unchanged. Treat a thread-locked joint like a dry joint, not a lubricated one.
Are TTY (torque-to-yield) bolts affected the same way?
Less so. TTY bolts are tightened past their elastic limit onto the flat part of the stress-strain curve, where clamp force stops tracking torque one-for-one — which is why TTY procedures are comparatively forgiving of small friction changes, though they still specify oiled or dry threads explicitly.
Keep going
Bottom line: a dry torque number is calculated around a specific, published K-factor — and every gram of anti-seize or drop of oil you add moves you off that number. Cut a dry spec by about 30% for anti-seize or 15% for oil if nobody has given you a wet spec, but the manufacturer’s own wet torque — where one exists — always wins. For the exact dry spec on your vehicle, the wrench & socket finder and our vehicle database answer in one search.
Sources
- Shigley’s Mechanical Engineering Design (via AmesWeb) — the nonplated/zinc/lubricated/cadmium K-factor table.
- Portland Bolt: bolt torque chart — the T=KDP methodology and the waxed / hot-dip-galvanized / plain K-factors.
- Molykote (Dow Corning): Lubrication Solutions for Threaded Connections — the K-factor formula and the 11-paste mu-thread / mu-bearing / K table this page’s anti-seize figures are built on.
- ARP: general torque recommendations — the published torque/clamp-load table this page back-computes K≈0.17 from.
- VDI 2230 Part 1 (via AmesWeb) — the friction-coefficient ranges by surface/lubricant condition.
- Metal Coatings Corp: FluoroKote #1 torque information — the PTFE kinetic friction coefficient (0.06–0.08).
- Anti-Seize Technology (A.S.T. Industries): general torque specifications — the manufacturer-stated 25% reduction for lubricated fasteners.
- Nord-Lock: torque, preload and friction — the 50/40/10 dry torque-split figure.
- ISO 898-1 — the tensile stress-area formula and the class 8.8 stress-under-proof-load value (580 N/mm², d≤16mm) this page’s metric chart is computed from.
- SAE J429 — the Grade 5 (85,000 psi) and Grade 8 (120,000 psi) proof-stress values this page’s inch chart is computed from.
- whatsizebolt.com /standard-torque-specs/ and /resources/metric-bolt-torque-chart/ — fetched 2026-09-07 for the published dry chart ranges and lubricant-correction table this page cross-checks against.
Video plays from YouTube when clicked. Always confirm torque values against your service manual.