BlogApplication notes
Why does my threadlocker fail, or is too hard to undo?
- Type
- Application notes
- Published
- 20 November 2026
- Prepared by
- Technical team
- Product group
- Adhesives
Moving to a stronger threadlocker is the usual answer to bolts that work loose, but a stronger grade leaves under-cure or low preload where it was, and the bolt may then refuse to come apart at service. What actually drives loosening or over-locking can be oil left on the threads, the wrong strength grade, load before full cure, an inactive surface or low preload, and several can overlap at once.
On equipment that vibrates hard, inspections sometimes find bolts working loose even though they were tightened with threadlocker. At the other extreme, on a service day a bolt refuses to move under the wrench until its head twists off. One is too weak and the other too strong, so they look like opposite problems, but both lead to the same question: why the adhesive did not give that joint the locking it was meant to give.
When a bolted joint under vibration slips sideways, each slip turns the nut a little in the loosening direction, and once preload is lost the whole joint goes slack (Gong et al., 2022; Junker, 1969). The result starts with noise and leaks and can end with parts falling off and equipment stopping. A bolt locked too hard does the opposite damage: at service the head snaps or the threads strip, turning a short check into part replacement and a long stop (Woo, 2019).
An anaerobic threadlocker, one that cures in the absence of air, is a one-part liquid of methacrylate, an acrylic-type monomer, with a peroxide initiator. In the bottle, dissolved oxygen blocks the reaction and it stays liquid; trapped between threads and cut off from air, it hardens into a rigid thermoset plastic (Rooney & Malofsky, 1990; Woo, 2019). The cured resin fills the voids in the thread to stop slip and seals the joint against corrosion (Woo, 2019). Its locking effect is measured by standard tests that record breakaway torque, needed for the first movement, and prevailing torque, needed to keep turning after that (ASTM International, 2023; ISO, 1993).
In such cases the technical team sets up the five hypotheses below and rules them out one by one. Whatever the hypothesis, the starting point is the state of the adhesive on the bolt in question, set against the test bolt, cure time and temperature given in the product's technical data sheet, the TDS, and the real conditions.
Hypothesis 1: oily threads
Parts that go straight from machining to assembly, bolts delivered with rust-preventive oil, and bolts reused in field maintenance can easily meet the adhesive on threads that are not clean. Think of a housing that comes off the tapping machine onto the assembly bench without washing, where an operator coats a bolt still glossy with rust preventive and tightens it at once. Weeks later the loosened bolt comes out with the adhesive in its threads unhardened and slick.
The key is that the first step of cure needs metal. For cure to start, the hydroperoxide initiator must break down into free radicals, highly reactive molecular fragments that set off polymerisation. Copper and iron ions on the metal surface are what break it down (Raftery et al., 1997). Rust preventive or coolant left on the threads sits between the adhesive and the metal and cuts that contact (Rooney & Malofsky, 1990; Woo, 2019). Oil that mixes into the adhesive dilutes the monomer and leaves a weak boundary layer between resin and metal, so cure slows and, even once hard, breakaway torque comes out low.
Some products are made to tolerate oil, but only so far. Certain grades are formulated to cure on lightly oiled surfaces (Woo, 2019), yet that assumes a thin rust-preventive film, not threads packed with coolant or grease. Aerospace test programmes on adhesive locking also write a cleaning process into the test procedure itself (Golden et al., 2012).
Smell and feel give the first signal. If the adhesive on a loosened bolt is tacky or slick and smells of oil, put this hypothesis first. Wipe a new bolt from the same lot with a clean white cloth to see whether oil comes off, and look into blind tapped holes for pooled coolant. On a reused bolt, see whether old cured adhesive left in the thread kept the new adhesive off the metal. To confirm, apply the same adhesive to a cleaned bolt and an as-received bolt, cure both for the same time and compare breakaway torque (ASTM International, 2023; ISO, 1993).
Clean the threads before applying. Wipe both male and female threads with a residue-free cleaner and let them dry fully, and in blind tapped holes clean right to the bottom, because oil there pushes the adhesive out. On reused bolts and holes, brush out old adhesive powder with a wire brush before cleaning. Where cleaning is hard, as in field maintenance, choose a grade the maker states is oil tolerant, and confirm it on the real surface with a comparative torque test.
Hypothesis 2: wrong strength grade
Suspect the grade first where one threadlocker serves the whole plant, or where the same grade goes on joints with very different bolt diameters. This hypothesis explains both loosening and over-locking. Picture a machine where an adhesive that came off small cover bolts easily with tools was also used on large flange bolts. On service day the flange bolts will not budge under the wrench until their heads twist off, while bolts with the same adhesive on the vibrating section have worked loose.
Grades are set by the strength of the cured resin. Of low, medium and high strength, medium is the serviceable type that hand tools can undo and high is the permanent type that is hard to undo with tools. Releasing a high-strength bond takes local heating to the temperature the maker specifies to soften the resin, and forcing it without heat can damage the bolt head or shank (Woo, 2019). Put a serviceable grade where vibration is severe, on the other hand, and loosening torques larger than the resin can hold recur until the bond breaks.
The same grade also behaves differently from bolt to bolt. A larger thread has more bonded area, so a medium-strength product can reach close to high-strength breakaway torque on a large bolt (Woo, 2019). Experiments likewise showed that material, thread size and thread proportions change how a threadlocker responds (Croccolo et al., 2018), and not every combination of bolt material and adhesive met the prevailing-torque requirement (Hernandez & Hess, 2015). That is why a product that releases easily on small bolts can refuse to release on large ones.
Hard resin spread evenly on a loosened bolt means cure happened, so the grade was too weak or preload too low. Bolts that snap or strip at service mean the grade was too high, or the bolt so large that strength came out excessive. Find the test bolt size and the breakaway and prevailing torques in the TDS of the product in use, and compare them with the real bolt size. The standard test is comparative and its results depend heavily on thread clearance and air bubbles, so build in a margin when using it to accept or reject (ASTM International, 2023).
Deciding and recording a grade for each joint is the baseline. Take medium strength as the default for bolts opened regularly, high strength for severe vibration where the joint is rarely opened, and low strength for small adjusting screws. Before putting high strength on large bolts, settle how and in what order they will be heated at service. When changing grade, measure torque on the real bolt to confirm the service tools can undo it, and write the grade on the drawing or work standard. Adhesive requirements can be specified through a standard classification (ASTM International, 2016).
Hypothesis 3: load before full cure
This hypothesis comes to the front for machines test-run as soon as assembly ends, unheated sites in winter, joints near cooling equipment and work done with old stock. Consider a site where bolts were re-tightened at the tail end of a winter shutdown and the machine started that same evening. A few days later those joints loosen one after another, and taking them apart shows patches of under-cured adhesive.
A threadlocker cures in two stages. First comes the fixture stage, when it is hard to turn by hand; full cure, when it reaches design strength, comes later. Standard tests usually measure torque after 24 hours of cure (ASTM International, 2023), and research on spaceflight fasteners used 48 hours (Hernandez & Hess, 2016). Run the machine in between, and vibration and thermal expansion can crack the still-soft resin, which then hardens with that damage and stays weak. Cure time is identified as a variable with high process sensitivity (Golden et al., 2012).
Three things weigh heavily on cure. Low temperature slows initiator breakdown and polymerisation, and warmth speeds them (Rooney & Malofsky, 1990; Woo, 2019). A wide thread gap leaves air in the joint and leaves resin far from the metal surface out of the reaction's reach, which works against anaerobic cure (Woo, 2019). And product stored long at high temperature loses initiator, with breakaway and prevailing torque falling together (Bergquist & Wentworth, n.d.).
The records bring it out. See whether loosening clusters in winter, near cooling equipment or on machines test-run right after assembly, and match assembly time, start-up time and ambient temperature against the joints that loosened. Check the manufacture date and storage place of the remaining product, and ask whether it ever sat in an unconditioned store or a vehicle. If in doubt, compare a short and a full cure on the same bolt, or new product against stored product, by torque (ASTM International, 2023).
Tie restart to cure time. Read the full-cure time at the real temperature from the TDS cure curve and write it into the work standard as the condition for restart. In cold conditions, pre-warm the parts or use an activator, a surface treatment that speeds cure. For wide-gap threads, choose a higher-viscosity grade with a larger gap-fill range. Store product somewhere cool as the maker recommends, and use the oldest stock first.
Hypothesis 4: inactive surface
This hypothesis arises on surfaces other than plain steel: stainless bolts, zinc-plated bolts, anodised aluminium parts and plastic parts. Typical cases are equipment that uses stainless for hygiene and equipment moved to plated bolts against corrosion. On one machine, the plain steel bolts hold while only the stainless bolts keep working loose. Taken apart, the adhesive inside the threads is under-cured only on the stainless side, although the same person applied the same product on the same day.
The difference comes from the amount of metal ions at the surface. The hydroperoxide breakdown that starts cure is catalysed by transition-metal ions such as copper and iron on the surface (Raftery et al., 1997). Plain carbon steel, brass and copper carry plenty of these ions and are active surfaces that cure well. Stainless steel is covered by a chromium oxide film, and zinc plating, chromate finishes, anodised aluminium and plastics carry almost no iron or copper ions; on these inactive surfaces cure is slow or incomplete (Rooney & Malofsky, 1990; Woo, 2019).
On such surfaces a primer stands in for that first step of cure. The primer, also called an activator, supplies metal ions or a reducing agent to the surface to start the cure (Woo, 2019). Aerospace work on adhesive locking likewise settled the right primer for each bolt and nut material, such as nickel alloys, by test (Golden et al., 2012).
The material and plating on the drawing give the quickest check. If loosened bolts cluster on stainless, zinc-plated or aluminium parts and the adhesive inside the threads is under-cured, this hypothesis is the strong one. Look at the material and plating of both the bolt and the mating part. Applying the adhesive under the same conditions as on a steel bolt and comparing cure time and torque makes the difference plain (ASTM International, 2023; ISO, 1993).
With an inactive surface to blame, apply a primer first and the adhesive after its solvent has flashed off, or choose a grade made to cure on inactive surfaces without primer. Either way, confirm by comparing torque on the real bolt and plating. Put the adhesive specification into change control so that it is reviewed whenever the plating changes.
Hypothesis 5: low preload
If the same location keeps loosening although the adhesive cured well, look at preload. It shows up at joints with heavy vibration and changing load direction, where bolts are tightened by feel without a torque wrench, and where the tightening torque is still the one set for dry bolts. One example is a pump base whose bolts are found loose again at every service. Once the bolts are out, the adhesive in the threads is fully hard but crumbled to powder, and the bearing face carries polished slip marks.
The first force holding a bolted joint is preload, the force with which the bolt pulls the parts together. With enough preload, friction stops sideways slip; with too little, the threads and bearing face slip a little at a time, and each slip turns the bolt toward loosening (Gong et al., 2022; Junker, 1969). A threadlocker is a secondary lock that resists this rotation (Golden et al., 2012). When low preload lets slip repeat, the cured resin takes repeated shear and cracks, and from then on the bolt loosens like one with no adhesive.
The threadlocker itself shifts preload. A liquid threadlocker can lubricate the threads and lower the friction coefficient; in tests, a medium-strength product gave friction coefficients around 0.11 (Croccolo et al., 2020). In the same tests, pre-tightening speed and the interval between pre-tightening and final tightening also changed friction (Croccolo et al., 2020). So keeping a tightening torque set for dry bolts once a threadlocker is in use moves preload away from what was intended.
Ask first what basis the tightening torque was set on: dry threads, or threads coated with adhesive. Check the torque wrench calibration as well. Slip marks on the bearing face, or debris from fretting, the wear caused by parts rubbing minutely against each other, strengthen this hypothesis. Resistance to vibration loosening can be compared by a transverse vibration test, usually called the Junker test (Junker, 1969).
Correct the preload before raising the grade. Reset the tightening torque for threads coated with adhesive, and keep tightening speed, sequence and intervals constant. Check the preload design first: whether bolt diameter, property class and joint faces suit the vibration load. A threadlocker cannot replace preload, so at joints that keep loosening, preload comes before a stronger adhesive (Barrett, 1990; Golden et al., 2012).
Testing the hypotheses
A single bolt separates most of the five hypotheses. Keep the loosened bolt as found, without wiping it, and first record whether the adhesive on its threads is liquid, tacky or hard powder. Liquid or tacky means cure failed, which points to hypotheses 1, 3 and 4; hard resin points to hypotheses 2 and 5.
On the under-cured side, clear the easy checks first. Look for oil smell and residue in blind holes to test hypothesis 1, and check the material and plating of the bolt and mating part on the drawing to test hypothesis 4 (Raftery et al., 1997; Woo, 2019). If neither fits, match assembly time, start-up time, temperature, and the product's manufacture date and storage history to test hypothesis 3 (Bergquist & Wentworth, n.d.; Golden et al., 2012).
If it cured well and still loosened, test hypothesis 5 through the basis of the tightening torque and slip marks on the bearing face (Croccolo et al., 2020; Junker, 1969). If preload is right and it still loosens, or the bolt refuses to come apart at service, compare grade and bolt size with the TDS to test hypothesis 2 (Croccolo et al., 2018; Woo, 2019).
The final check is a control test. Cure two bolts for the same time, one with only the suspected condition changed and one under the old conditions, and compare breakaway torque (ASTM International, 2023; ISO, 1993); confirm the hypothesis as the cause only when the value changes on the changed side alone. After the fix, open the same joint again at the next service to check its state.
The hypotheses often overlap. An oily zinc-plated bolt assembled in winter and run at once brings hypotheses 1, 3 and 4 into play together. Even then, change one thing at a time, but fix the surface first with cleaning and primer, since only then can the effect of the other hypotheses be read clearly.
Limits and cautions
Breakaway torque, cure time, gap-fill range and the heating temperature for removal differ by product, so take the figures from that product's TDS. Where a failed joint can injure people, as in pressure vessels, lifting gear, vehicle steering and brakes or aerospace parts, follow the locking method set by the design basis and the governing rules (Barrett, 1990; NASA, 2021). Environments that may not suit anaerobic resin, such as high heat, aggressive solvents, or oxygen and high-purity gas lines, also need separate review. Hand critical joints that keep loosening to a fastening-design specialist.
Methacrylate monomers are skin sensitisers; anaerobic sealants have been reported as a common source of occupational acrylate allergy, and some safety data sheets, the SDS, lacked the warning (Aalto-Korte & Suuronen, 2021). Do not handle them bare-handed: wear nitrile gloves and eye protection, and wash skin at once after contact. When heating a high-strength bond to undo it, the resin decomposes and smokes, so ventilate and guard against burns. Check the SDS for each product.
Sources
- Aalto-Korte, K., & Suuronen, K. (2021). Ten years of contact allergy from acrylic compounds in an occupational dermatology clinic. Contact Dermatitis, 84(4), 240–246. https://doi.org/10.1111/cod.13739
- ASTM International. (2016). ASTM D5363-16, Standard specification for anaerobic single-component adhesives. https://store.astm.org/d5363-16.html
- ASTM International. (2023). ASTM D5649-15(2023), Standard test method for torque strength of adhesives used on threaded fasteners. https://store.astm.org/d5649-15r23.html
- Barrett, R. T. (1990). Fastener design manual (NASA Reference Publication 1228). NASA. https://ntrs.nasa.gov/citations/19900009424
- Bergquist, P. R., & Wentworth, S. E. (n.d.). Shelf life degradation of selected anaerobic threadlocking adhesives at various storage temperatures (ADA187712). Defense Technical Information Center. https://apps.dtic.mil/sti/html/tr/ADA187712/index.html
- Croccolo, D., De Agostinis, M., Fini, S., Olmi, G., et al. (2018). An experimental study on the response of a threadlocker, involving different materials, screw dimensions and thread proportioning. International Journal of Adhesion and Adhesives, 83, 116–122. https://doi.org/10.1016/j.ijadhadh.2018.02.024
- Croccolo, D., De Agostinis, M., Fini, S., Olmi, G., Paiardini, L., & Robusto, F. (2020). Threaded fasteners with applied medium or high strength threadlockers: Effect of different tightening procedures on the tribological response. The Journal of Adhesion, 96(1–4), 64–89. https://doi.org/10.1080/00218464.2019.1679630
- Golden, J. L., Leatherwood, M. D., Montoya, M. D., Kato, K. A., & Akers, E. (2012). Process sensitivity, performance, and direct verification testing of adhesive locking features (NASA/CR-2012-217760). NASA. https://ntrs.nasa.gov/citations/20120015467
- Gong, H., Ding, X., Liu, J., & Feng, H. (2022). Review of research on loosening of threaded fasteners. Friction, 10, 335–359. https://doi.org/10.1007/s40544-021-0497-1
- Hernandez, A., & Hess, D. P. (2015). Use of anaerobic adhesive for prevailing torque locking feature in threaded fasteners. Journal of Failure Analysis and Prevention, 15(5), 633–640. https://doi.org/10.1007/s11668-015-0008-7
- Hernandez, A., & Hess, D. P. (2016). Anaerobic adhesive as a prevailing torque locking feature. Conference version, NASA Technical Reports Server 20160008131. https://ntrs.nasa.gov/citations/20160008131
- International Organization for Standardization [ISO]. (1993). ISO 10964:1993, Adhesives: Determination of torque strength of anaerobic adhesives on threaded fasteners. https://www.iso.org/standard/18936.html
- Junker, G. H. (1969). New criteria for self-loosening of fasteners under vibration (SAE Technical Paper 690055). SAE International. https://saemobilus.sae.org/content/690055
- National Aeronautics and Space Administration [NASA]. (2021). NASA-STD-5020B, Requirements for threaded fastening systems in spaceflight hardware. https://standards.nasa.gov/standard/nasa/nasa-std-5020
- Raftery, D., et al. (1997). Investigations into the cumene hydroperoxide-based cure chemistry of anaerobic adhesives. International Journal of Adhesion and Adhesives, 17(4), 349–352.
- Rooney, J. M., & Malofsky, B. M. (1990). Anaerobic adhesives. In I. Skeist (Ed.), Handbook of adhesives. Springer US. https://www.springerprofessional.de/anaerobic-adhesives/13767264
- Woo, C. (2019). Liquid thread locking solution for machinery assembly industry. In A. Rudawska (Ed.), Adhesives and adhesive joints in industry applications. IntechOpen. https://doi.org/10.5772/intechopen.85090