The fastener that broke is rarely the most informative thing recovered from a failed joint. Grip length, thread engagement, witness marks, the residual clamp force on the bolts that survived, and the condition of washers and clamped faces carry as much of the answer as the fracture surface does. Nearly all of it is destroyed within hours by ordinary, well-intentioned response: equipment is cut apart, nuts are run off, parts are cleaned, and the joint is rebuilt to restore service. Knowing what is lost, and in what order, is the difference between a determinable failure and an argument.
The joint is the specimen
A threaded fastener is one element of a clamped assembly, and most of the mechanisms that make it fail are properties of that assembly. Whether the joint was ever adequately clamped, whether the members bore across their full faces, whether grip length matched the fastener, and whether the load path ran through friction or through the bolt in shear are questions the broken bolt alone cannot answer.
The evidence boundary therefore runs around the joint and its immediate neighbours, not around the broken part. Preserving the fractured bolt while discarding the nut, the washers and the mating flange leaves the most frequently contested questions unanswerable.
Residual torque is a perishable measurement
On fasteners that did not fail, breakaway torque is a direct if imperfect indicator of the clamp force still present at the time of the event. It can be taken once. The first person to put a wrench on the nut consumes the measurement, whether or not anyone recorded a number.
Witness marks are equally fragile. Paint stripes, match marks, burnishing on the bearing face, and the impression a nut leaves in the clamped surface all record relative rotation and how hard the joint was pulled up. Backing a nut off overwrites them.
As-found documentation comes first
Before anything moves, photography with a scale, orientation marking of each fastener to its hole, and a record of which fasteners were loose, missing or intact establish a baseline nothing later can reconstruct. ASTM E1188 sets out general practice for collection and preservation by a technical investigator, and ASTM E860 addresses examining items that may become involved in litigation.
Position matters because failure patterns are directional. Which bolt in a pattern released first is often recoverable from the distribution of loosening and damage across the pattern, and that ordering disappears the moment the hardware is pooled in a bucket.
Cleaning is usually irreversible
Fracture surfaces get wire-brushed, solvent-washed or bead-blasted so that someone can see them, and each of those removes exactly what fractography reads. Corrosion product inside a crack is evidence of both environment and crack age; abrasive cleaning strips it and leaves a surface that can no longer support a stress corrosion or hydrogen conclusion.
Fracture faces should be protected from further corrosion and handled without contact, with cleaning left to a laboratory that will document each step. Where a surface must be cleaned, the sequence is a decision for the examiner rather than the recovery crew.
Extraction damage
Broken studs get drilled, heated or screw-extracted, and seized fasteners get torched off. Each of those alters the material next to the fracture: heat changes hardness and microstructure, and drilling destroys the origin. Where a fastener must come out to release an assembly, cutting well away from the fracture and preserving the full remaining length is far less destructive than working on the break itself.
The same rule governs the clamped members. Cutting through a flange to free it is usually acceptable. Cutting through the bolt holes, the bearing faces or the thread engagement is not.
The unfailed population
One fracture establishes a mechanism. Whether that mechanism was isolated or systemic is answered only by comparison: adjacent fasteners from the same joint, fasteners from the same production lot, and unused stock where any survives. ASTM F1470 addresses fastener sampling for specification conformance, and the same logic serves the forensic question.
Unused lot stock has a short life after an incident. It gets consumed on the next repair, returned to the distributor, or quarantined and then scrapped, and losing it removes the ability to distinguish a bad batch from a bad installation.
Paper evidence runs on its own clock
Torque records, calibration certificates for the tools used, assembly procedures, inspection sign-offs, the material test report and the certificate of conformance are all subject to retention schedules. So are the purchase records establishing which distributor supplied the lot, which is where substitution and counterfeit questions begin. These are usually easier to secure than the hardware, and are more often overlooked.
Destructive testing and notice
Sectioning through a fracture origin, hardness indentation, and chemical analysis all consume material. Where other parties have an interest, testing performed without notice and without an opportunity to observe invites a spoliation argument regardless of how carefully the work was done.
The workable practice is a written protocol circulated in advance, joint inspection where feasible, and full documentation of each specimen before and after every destructive step.
Restoration and preservation are separable
Pressure to return equipment to service is legitimate and rarely negotiable. The sequence is negotiable. An hour of photography, position marking and controlled removal, with each joint component bagged individually and labelled to its location, costs very little against a failure that can no longer be resolved.
That hour covers the only point at which the joint existed in the condition that caused the failure.
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.