In most failure investigations the debris is an obstacle and the component underneath it is the evidence. Corrosion inverts that. The scale, tubercles and deposits on a leaking pipe or a perforated tank floor carry the chemical record of what attacked the metal, in what order and under what conditions. Strip them and what remains is a hole of known size and unknown origin. The commonest reason a corrosion mechanism cannot be established is that the surface was cleaned, blasted or descaled before anyone examined it.
The corrosion product is the record
Metal loss tells you how much was removed. What that metal turned into tells you why. Corrosion products are compounds formed from the metal and the environment, so their composition reflects the species present at the surface — not the species present in a bulk sample taken elsewhere and much later.
Scanning electron microscopy with energy-dispersive spectroscopy is the workhorse, imaging pit morphology while identifying the elements in the scale, the deposit and the pit interior separately. X-ray diffraction goes further and identifies the compounds themselves rather than only the elements present.
What the chemistry actually distinguishes
Several findings do real diagnostic work. Sulfur-bearing corrosion products where no process sulfur explains them raise microbiological involvement. Chloride concentrated inside a pit but not on the surrounding surface supports local passive-film breakdown rather than general attack. Elements from a coupled dissimilar metal at the attack site indicate a galvanic path.
None of these survive cleaning. Chloride enrichment within an occluded pit is measured in the pit, and rinsing removes exactly the soluble species that carry the signal.
Layers carry sequence
Scale is often stratified, the innermost layer formed earliest against the metal and later products deposited over it. Cross-sectioning through an intact deposit and mapping composition across those layers can show that the chemistry changed during service — a treatment change, a contamination event, an upset.
That is one of the few ways a corrosion investigation reaches sequence rather than only state. It is also the first thing lost to a wire brush, which mixes and removes the layers in one pass.
Cleaning is a laboratory step, not a site step
Corroded specimens do eventually get cleaned. Measuring mass loss or examining the metal beneath the product requires it, and ASTM G1 sets out the accepted procedures along with the repeated-cleaning approach that distinguishes product removal from attack on the base metal.
The difference is that laboratory cleaning happens after documentation and after the removed product has been sampled and analysed, using a method selected for the alloy and recorded in the report. Site cleaning happens before anyone knows what was there, with no record of what was removed.
What each cleaning method takes with it
Abrasive blasting removes the deposit entirely and works the surface underneath, rounding pit mouths and imposing a texture that obscures the original morphology. Wire-brushing and grinding do the same more selectively and can smear metal across pit openings, making depth and shape unreadable.
Acid descaling is the most damaging because it appears the gentlest. Uninhibited acid dissolves corrosion product and base metal together, deepening the very features being measured, and it removes any chance of identifying what it dissolved. Re-coating a failed section seals whatever remains under a new layer and adds a fresh contamination source to every later analysis.
Documentation before any intervention
The sequence that preserves the case is photograph, sample, then remove. In-place photography with scale and orientation, showing the deposit and the surrounding condition before anything is disturbed, is frequently the only record of the as-found state that ever exists.
ASTM E860 addresses examining and testing items that are or may become involved in litigation, and ASTM E1188 addresses collection and preservation of information and physical items by a technical investigator. Both point the same way: fix the condition in the record before altering it, and document the alteration when it happens.
What to collect while it is still available
Beyond the component itself: a sample of the deposit or scale taken dry and undisturbed; the process fluid, water or soil, sampled promptly because chemistry changes once the system is opened; coating or lining remnants; and any insulation in contact with the surface.
The paper record is perishable in its own way. Water treatment logs, thickness survey histories, cathodic protection readings, coating submittals and maintenance work orders all sit under retention schedules, and requesting them early costs nothing.
Cutting the section out without wrecking it
Removal is itself an intervention. Cuts should be made well clear of the damage, and cold cutting is preferred because flame cutting introduces heat that alters microstructure near the cut and can transform corrosion products across a wider area than expected.
Orientation must survive removal. Marking flow direction, the twelve o'clock position and which face was internal preserves the link between damage location and service condition; without it, a section showing exactly what happened can no longer show where.
Restoration pressure and the honest compromise
The pressure to clean up is legitimate. A leaking line has to be isolated and repaired, and an operator running an outage has obligations of its own. The distinction worth drawing is between restoring service and destroying the record, which deliberate sequencing usually separates.
Where a surface has already been cleaned, the analysis is narrowed rather than ended. Adjacent uncleaned areas, the removed material if it was captured, mating parts and residues in low points may still hold product. Aggressive cleaning also erases the secondary cracking that would otherwise flag stress-corrosion cracking as a separate mechanism. A report stating plainly what was cleaned, by what method and when is worth far more than one proceeding as though the surface were as found.
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.