A weld fails in service and the investigation reaches back through the paperwork to a certificate issued four years earlier by a mill in another country. That certificate is the only surviving evidence of what the steel was. If it cannot be tied to the specific material in the failed joint — the heat, the cast, the plate, the piece — then the chain breaks, and the question of whether the right material was ever supplied cannot be answered by anybody.

We edit what metallurgists and materials engineers produce — material test certificates and their interpretation, material specifications and procurement requirements, traceability and marking procedures, welding procedure specifications and qualification records, heat treatment procedures and records, positive material identification procedures, failure analysis and metallurgical examination reports, corrosion assessments and material selection reports, non-destructive examination procedures and acceptance criteria, non-conformance and concession documentation, supplier qualification and mill audit reports, and technical justifications for material substitution. Our editors work on the documents that connect a certificate to a piece of metal.

The material test certificate and its traceability chain is where metallurgical documentation earns its cost, and the failure is nearly always the link rather than the certificate. A mill certificate is usually complete; what is missing is anything connecting it to the plate that was actually cut, the offcut that was used for the repair, or the piece now in service. We write these so the traceability requirement states the level at which identity must be maintained — heat, cast, plate or piece — because a project that requires piece traceability and specifies heat traceability has bought a different thing; so marking and transfer of marking is written as a procedure with an owner, since identity is lost at the moment a plate is cut and whoever holds the marker decides whether the record survives; so the certificate's own content is checked against the specification rather than filed, given that a certificate showing a value outside the specified range is accepted by systems that only check a certificate exists; so any material released on a concession is documented with the deviation, the technical justification and the approver, as concessions are what an investigation finds years later; so positive material identification is specified where the consequence of a mix-up justifies it, with what is tested and how many; and so remnant and offcut control is written, because that is where the wrong grade enters the job. Documentation written this way answers the question that gets asked after a failure.

Everything you send is treated in confidence, including certificates, procedures and supplier information. We are editors rather than metallurgists or materials engineers, and we offer no view on material properties, specifications or acceptance. What we can do is make the chain from certificate to component unbroken on paper.