English is currently available.
Metallurgical Engineering DISCUSSION
We had a batch of property class 12.9 socket head screws electroplated with zinc for corrosion protection. They were torqued to the normal value without any problem, but between a few hours and two days later several heads broke off with no extra load applied. Unplated screws from the same lot have been in service for months.
Is this hydrogen embrittlement from the plating, how can it be confirmed, and what is the correct way to give high-strength fasteners a corrosion-resistant finish?
The pattern fits hydrogen embrittlement closely: high-strength steel, an electroplating process, a sustained tensile load, and failure after a delay rather than during tightening. Acid pickling and electroplating both generate atomic hydrogen at the steel surface, and some of it diffuses into the metal. The zinc layer then acts as a barrier that keeps it in.
Under sustained stress the hydrogen migrates to the most highly stressed region, usually the head-to-shank fillet or the first engaged thread, and lowers the stress at which a crack can start. The delay is the time that diffusion takes, which is why the screws pass the torque and fail later.
Susceptibility rises steeply with strength. Class 12.9 means a nominal tensile strength of 1200 MPa and a hardness of roughly 39 to 44 HRC, which is the range where fastener standards treat the risk as serious. Class 10.9, at about 32 to 39 HRC, is much less sensitive, and 8.8 is rarely affected by plating.
That gives one simple remedy: if the joint can be designed with 10.9 screws, possibly one size larger, the problem largely goes away. For this reason many companies do not allow electroplated 12.9 fasteners at all.
If plating of high-strength parts cannot be avoided, the standard control is baking soon after plating to drive the hydrogen out before it does damage. Plating standards call for a bake at around 190 to 220 °C for a number of hours, longer for higher strength, and started within a few hours of plating; the exact time and the permitted delay come from the standard your plater works to. Baking reduces the risk but does not guarantee zero, which is why the process also has to avoid long acid pickling, using mechanical or alkaline cleaning instead.
Ask the plater for the bake record of this batch. A bake that was skipped, shortened or started the next day is the usual finding.
To confirm it, look at the fracture surface. Hydrogen-induced delayed cracks in hardened steel are typically intergranular near the origin, with a rock-candy look under a scanning electron microscope, changing to ductile dimples where the remaining section finally failed by overload. A metallurgical lab can check that and the hardness in one visit.
For the rest of the batch there are sustained-load tests: sample screws are tightened to a high preload on a test plate and held for a fixed period, typically a couple of days, then examined for cracks or broken heads. Because only a fraction of parts in an affected batch fail, a clear test on a small sample is weak evidence. With failures already seen, the safe decision is to scrap the batch.
For corrosion protection of 12.9 screws, choose finishes that do not charge the steel with hydrogen. Non-electrolytic zinc flake coatings are the common choice for high-strength fasteners, and mechanical zinc plating, where the zinc is peened on in a tumbling barrel, is another. Both are usually preceded by abrasive or alkaline cleaning rather than acid.
Remember that hydrogen can also enter in service. A sacrificial coating that is corroding generates hydrogen at any exposed steel, and very hard fasteners in wet or marine conditions can fail this way years later. That is another reason to use the lowest strength class that does the job in corrosive environments.
