Re-baking to save Hydrogen embrittled fasteners

A. On an aircraft, rocket, ordnance or auto -- no. Risk someone else's tail, not mine. There has been some work done on re-embrittlement that would make this frightening. To a certain extent, the damage has been done. The reprocess just mitigated it enough to pass. Of course, this is a personal opinion. I do not think that it is the intent of a Mil Spec to allow this.

James, Thanks for your input. The fasteners in question DO NOT go on any aircraft, rockets, ordnance or automobiles. I should have made this clear. But you are correct, I don't see how rebaking would heal any microcracks or microvoids. Rebaking is not an option we accept, and the problem has not come up, but one fastener supplier did suggest rebaking as an option to me the other day.

A. Dave,

for a non critical use, stripping, annealing and then hardening including the proper draw might be a tolerable recovery. Jim

A. Dave -

The problem you described is more typical than anyone would like to believe. The solution proposed is a common one and can be effective in salvaging material. There seems to be a misunderstanding in the way hydrogen damage or embrittlement works...

If a sample has developed microcracks or microvoids then the proposed fix is of course worthless, but typically fasteners do not show this type of damage until after they have been stressed. The damage that hydrogen causes requires a tensile stress, either residual or applied.

My job over the last 8 years has been performing failure analysis studies. Last year alone we did over 550 failures, many of them dealing with fasteners and hydrogen embrittlement.

The one caution I do have for anyone that does what was suggested is as follows: After stripping, the parts should be given an embrittlement relief because the pickling or stripping operation is a known hydrogen contributor. If this intermediate bake is not performed you can effectively worsen the situation because of the added amounts of hydrogen that can be introduced.

It is not a trivial decision and the follow-up testing needs to be agreed upon before hand....

A. The effectiveness of a delayed baking process is dependent on the severity of the embrittlement. Mr. Menendez is correct that stress is required to create nonrecoverable damage from hydrogen. If enough hydrogen is absorbed, even minor residual stresses are enough to cause unrecoverable damage. In my experience, steels with moderate hardness can recover from delayed baking. High hardness steels, 40 HRC, probably will not. The best bet is to try a rebake and redo the testing. If the testing is OK, the parts will probably be acceptable for noncritical applications.

A. Most tests for hydrogen embrittlement use a sustained static tensile force. If the test samples fracture or crack, hydrogen embrittlement is indicated. If no cracks occur, the parts are free from embrittlement.

The trick is to determine the magnitude of force and the length of time for the test. For a component like a cylindrical rod, an axial tensile force would be ideal. But without access to an appropriate machine for this, I would probably devise a three-point bend test using a vice or dead-weight loads. A notch in the test sample concentrates the stress and gives a more severe test for embrittlement.

The sample should be bent to just under the material's yield strength. You can determine this by bending and releasing a sample until it maintains a permanent set - for the test use a bend that is slightly less than the smallest bend or load causing a permanent set.

A typical time period for the test would be 200 hours.

Good Luck

A. Although neither a metallurgist nor knowledgable in testing, I sort of disagree on this one! I feel that the technician in the testing laboratory may not be privy to the thoughts & plans of the design engineer/design authority, and won't know what the designer feels is critical vs. non-critical; and is therefore not in a good position to develop a test to verify that the part is satisfactory unless the specifier is involved in the resolution. Sometimes, perhaps the opposite of expectation, the factor of safety is 3 or 5 on non-critical parts, and 1.1 on critical ones :-)

Parts are supposed to be hydrogen relieved within a specified period, whether it be 2 hours, 4 hours, or 24 hours. To say that it's okay to relieve them days or weeks or months later, as long as they pass some test sounds like dangerous business. The B-8 committee of ASTM struggled with this time-delay factor for the better part of a year, and it seemed to have ended in some acrimony (the way I read the columns in the Plating journals). Overriding the standards committees and the designer can be dangerous business; and as James' warning about potential re-embrittlement should remind us, the vast experience we have on some things can constitute a sort of factor of safety that evaporates when we move into special circumstances like baking later instead of contemporaneously.

A. If the new clips have no need for flex, you can make make them out of plain steel rather than spring steel, and not bake them.

If they must flex, they must continue to be made of spring steel, and you must continue to bake them unless you decide that (since the clips are not a critical item) you will make the economic decision to trade increased breakage for lower cost.

But I think you're better off fixing the first problem of rinsing/drying stains, rather than talking yourselves into creating a second problem of unbaked parts. Good luck.

A. I address only the staining issue mentioned by Tom Clem. You could use a commercial demineralizer which would eliminate any visible stains. This assumes you blow-dry the parts after final demineralized rinse to a reasonably dry finish. We used such a device for copper plating and the results were better than we expected. The cost was also reasonable. Hope this helps.

A. Hi Carolyne. Hopefully you'll get some detailed response, but we appended your inquiry to a thread on the same subject, so you'll have some opinions you can read in the meantime :-)

Regards,

A. Hi, Carolyne
AMSC26074 D paragraph 3.4.1 says: Within 4 hours after coating, coated class 1 parts that have a hardness of RC40 or above shall require HE relief bake...
Paragraph 4.1.1 says: All items shall meet the requirements of sections 3 and 5. etc...
If your vendor sent you non-inspected item, this item is non-conforming to the spec.
AMS2759 paragraph 7 says: Parts not heat treated in accordance with specification... will be subject to rejection...

Very sorry...

A. Please keep in mind that baking does not eliminate the risk of hydrogen embrittlement. It does significantly reduce that risk. Visual inspection IMHO will tell you nothing. If you take a couple of dozen of these parts and subject them to the stress or tension that they would be subject to in service you can be fairly sure that they won't fail in the field. A typical test for threaded fasteners would be 75% of proof load. If even one fails, they are toast.

A. Hi Carolyne,

Hydrogen embrittlement is considered a time-dependent effect and the effect happens in hours.

I work in the Aerospace industry, so if these parts were to arrive at my desk for consideration they would rapidly proceed to the scrap bin.

You can test for hydrogen embrittlement, however, the test is destructive. So, if you only have a few parts, it's just not worth testing, but if it's a batch of hundreds or thousands of fasteners, then you may be able to justify testing a percentage, although, if it is for an Aerospace customer I wouldn't bother, as I'd be very surprised if they would accept any testing regime.

A. An option would be to explain the situation to the customer. If non-structural/critical they may be willing to accept the parts at a discount. I work in aerospace. I would not accept these parts for any structural application. I would, however, be pleased with the seriousness and integrity that my vendor applied in this situation.

A. Steel part with hardness 40-42 HRC, after electroless Nickel plating has to be baked for Hydrogen Embrittlement relief as soon as possible (majority of specifications would specify within 4 hours). After that - it is too late, the damage is irreversible and the damage is not detectable visually! (Actually one smart engineer said once that inventor of a non-destructive method to detect Hydrogen Embrittlement will become a millionaire overnight).

*Even more info - spec AMSC26074 is an old spec and it is not updated anymore. It specifies short bake for 3 hours only. Currently, newer AMS specs require 8 hours of bake for such parts after electroless Nickel plating.