A high-strength bolt that was torqued correctly, passed inspection, and then split in two a week later is a different forensic problem from one that broke under load. Delay narrows the field immediately: something was driving a crack while the joint sat still. Two mechanisms dominate that category, hydrogen embrittlement and stress corrosion cracking, and they produce fracture surfaces similar enough that they are routinely confused. They also point at very different parties. Separating them is a matter of fractography, hardness, process history and environment, in roughly that order.
Why delay is the first clue
Ductile overload is immediate and obvious: gross plastic deformation, necking or thread stripping, shear lips, a dimpled surface under the electron microscope. Fatigue is progressive but requires cyclic loading and leaves beach marks recording it. A bolt that fractures under an essentially static clamp load, with no cycling and almost no ductility, has been cracking slowly for reasons neither mechanism explains.
That leaves an environmentally assisted mechanism, in which sustained tensile stress well below yield drives a crack that would not otherwise propagate. Both candidates fit, and timing alone will not separate them: either can surface days or months after installation.
What hydrogen embrittlement requires
Internal hydrogen embrittlement needs hydrogen already in the steel at installation, introduced during acid pickling, electroplating or melting. It needs a susceptible microstructure, generally a quenched and tempered high-strength steel. And it needs sustained tensile stress, which the preload itself supplies. Hydrogen migrates to regions of high triaxial stress, typically the first engaged thread root or the head-to-shank fillet.
The signature is delay with no external agent at all. The bolt is dry, the environment benign, and the fracture happens anyway. Where the fastener was electroplated and post-plate baking was omitted or delayed, ASTM B850 and ISO 4042 describe the relief treatment that should have followed.
What stress corrosion cracking requires
Stress corrosion needs the same sustained stress and a susceptible alloy, but also an environment in continuing contact with the metal: chlorides, sulphides, caustics, or moisture concentrated in a crevice beneath a washer. Remove the environment and the crack stops. Hydrogen already dissolved in the steel needs nothing further from outside.
That dependence is the practical discriminator. Stress corrosion leaves corrosion product in and around the crack, often branched secondary cracking away from the main fracture, and attack consistent with the service exposure. A recently installed bolt in a clean, dry interior joint is a poor candidate.
Reading the fracture surface
Both mechanisms commonly produce intergranular fracture in quenched and tempered steel, which is why appearance alone is not enough. A hydrogen fracture typically shows relatively clean intergranular facets with tear ridges and little corrosion product near the origin. Stress corrosion surfaces are usually oxidised, branched, and may be transgranular depending on alloy and environment.
Origin location matters as much as morphology. Hydrogen cracking initiates subsurface or wherever triaxial stress peaks; stress corrosion initiates where the environment reached the metal, often a pit, a crevice, or a coating breach.
Hardness and the susceptibility threshold
Sensitivity to hydrogen rises steeply with strength. Fasteners around 1,000 MPa tensile strength and above, covering SAE J429 Grade 8, ISO 898-1 property classes 10.9 and 12.9, and comparable aerospace alloys, are the population where internal hydrogen embrittlement is a live concern.
Hardness and tensile testing under ASTM F606/F606M answers a second question: whether the fastener was heat treated to the grade stamped on its head. An over-hard bolt is more brittle than the design assumed and far more sensitive to hydrogen, and that finding moves attention toward the manufacturer.
The process record
Plating line records, bake time and temperature, the interval between plating and baking, and the lot certification are the documents that make or break a hydrogen hypothesis. ASTM F1940 covers process control verification for plated fasteners and ASTM F519 the specimen-based evaluation of plating and coating processes.
Where a bolt was zinc-electroplated and installed within a shift with no relief bake recorded, the hypothesis is testable rather than speculative. Where it was mechanically galvanised, the internal hydrogen route is far less available.
Confirmatory testing
ASTM F1624 measures a hydrogen embrittlement threshold by incremental step loading and is the usual route to establishing whether the material as supplied was susceptible. ISO 15330 provides a preloading test for detecting embrittlement in finished fasteners. On the corrosion side, the ASTM G-series specimen tests characterise alloy susceptibility in a defined environment.
None of these establish what happened to the failed part. They establish whether the proposed mechanism was available at all, which is what separates a supported opinion from an inference drawn from appearance.
The overlap case
Hydrogen can also be generated in service, by corrosion reactions at the metal surface or by cathodic protection, and then embrittle the same steel. This environmentally assisted variant sits between the two categories, and the distinguishing evidence is whether hydrogen entered before service or during it. The process record and the corrosion state of the joint address that together.
Where these opinions are challenged
Predictably: that intergranular fracture was equated with hydrogen without excluding stress corrosion or temper embrittlement; that no hardness or composition data ties the part to its marked grade; that the plating and baking record was never obtained; and that one fastener was examined without comparison to unfailed siblings from the same lot.
Work that states which mechanisms were considered, which were excluded, and on what evidence survives that scrutiny. A conclusion resting on the appearance of a single fracture surface does not.
This article is general technical orientation, not a failure analysis, an engineering opinion, or advice on any specific matter. Determining the cause of a particular incident requires hands-on examination by a credentialed expert.