It’s not just the material you use, but critically, it matters how you use it.
I have been researching the possibility that the brittle fractures we are seeing at Cabo da Roca and Sesimbra in Portugal are hydrogen embrittlement (HE) fractures mediated via the presence of sulphate reducing bacteria (SRB).
Austenitic stainless steels like 304 (A2) and 316 (A4) are known to be resistant to HE because of the low hydrogen diffusion rate through the austenitic atomic lattice.
However, it is also known that the austenitic phase is metastable, and can transition to a different atomic arrangement called martensite when the material is deformed. This is the basis of the “work hardening” that occurs with these steels, and is familiar to anyone with workshop experience.
Now, it turns out that the atomic lattice of martensite allows free diffusion of atomic hydrogen, and thus opens the door to a range of HE phenomenon. So the question arises as to how purely-austenitic a formed or machined part might be following manufacture. Could there be sufficient martensite to cause a problem with HE under SRB attack?
One extra piece of information to add, and we can move from speculation to measurement. The austenitic atomic structure is not magnetic, but the martensitic structure is. So get yourself a strong rare-earth magnet and go test some bolts.
Here’s some pretty pictures from Kaoumi and Liu
So armed with my super-strong neodymium magnet from eBay, I started raking through my huge pile of bolts extracted from local cliffs over the past few years. I picked out some glue-in machine bolts that had been on the cliff for ten years or more before they were removed. There are nice and shiny because we don’t do corrosion in S E Queensland, and, besides, the glue used was terrible, so here they are, as good as new, but decidedly random. They are all of the same make and type, THE, A2, 70.
Yet three of them are quite strongly magnetic at the heads and threaded sections, and the others not at all. I suspect this has to be a manufacturing process difference.
Clearly the difference is insignificant in the mild environment of SEQ, but would we see a difference were SRB to be active?
Next up, I thought I’d look at a random selection of ring bolts I had lying around.
The first two were Raumer Super Stars in 316. They showed a trace of magnetism at the weld, but the rest of the bolt was totally free of magnetic response. That includes the tight bends, and the deep knurling.
The second was a CT 316 ring bolt with tight bends and deep deformations along the stem. Surely, this would be work hardened, but no, no trace of magnetic response.
Finally, I grabbed a selection of variations on the old standby ringbolt, Fixe #14A. All bolts showed substantial response at the deformations in the stem and at the tight radii of the ring. There was some response at the weld, but not as strong as else where.
So what can we conclude? For some this will be news, for others not so much. Alan Jarvis of UIAA Safe Com has long argued that it no good specifying just the material type if you wish to establish a corrosion standard. It is absolutely necessary to test entire assemblies in their final manufactured form.
There are two concerns: a) are manufacturers controlling their incoming bar stock with respect to composition and state of temper, and b) are manufacturers considering the impact of their manufacturing process on the corrosion resistance of the final product.
All of the above is of no consequence if you live in a part of the world where corrosion is not a problem, but it is of immense consequence for those living in the many locations plagued by the presence of abundant sulphate.