A turbine blade that fractures by fatigue or creep is often reported as a fatigue or creep failure, and that description can be accurate about the final event while being unhelpful about the cause. In hot-section blading, the mechanism that actually consumed the margin frequently arrived years earlier and worked from the outside in: sulfidation attack, oxidation of an exhausted coating, or solid-particle erosion thinning the section. The forensic question is not only what broke the blade but what left it thin enough, hot enough or rough enough to break.
The coating is a consumable
Diffusion aluminide and overlay coatings protect the base metal by forming a stable alumina scale, and each time that scale spalls and reforms the coating loses aluminium. Thermal barrier systems add an insulating ceramic layer over a bond coat. Both are life-limited, and both are expected to be restored at overhaul rather than to last the life of the part.
Treating a coating as permanent is a recurring source of dispute, because the interval at which it is inspected and restored is a maintenance decision with a documented owner.
Two distinct hot-corrosion regimes
High-temperature hot corrosion, conventionally described as Type I, operates in the upper part of the hot-section temperature range and produces broad, internally sulfidised attack with a porous, layered scale. Low-temperature hot corrosion, Type II, operates several hundred degrees cooler and produces localised pitting under a sulfate deposit, often on cooler blade surfaces or lower stages.
The distinction matters because the two implicate different operating conditions and different parts of the machine. Reporting them together as corrosion loses the information that identifies where the contaminant was acting.
Where the contaminants come from
Sodium, potassium, vanadium, lead and sulfur are the usual actors, and they arrive through fuel, through intake air, or through injected water or steam. ASTM D2880 sets the framework for gas turbine fuel oils, including trace-metal limits and the inhibitor practice used where vanadium is present.
Air-path contamination is often the harder case: coastal salt, industrial plumes and cooling-tower drift are all plausible sources, and filtration condition and change-out history become the record that either supports or excludes them. Water and steam chemistry logs serve the same function for injected flows.
Barrier coatings and bond coats fail differently
Ceramic thermal barrier layers usually fail by spallation, driven by growth of the thermally grown oxide at the bond-coat interface and by strain accumulated through thermal cycling. The loss is visible as bare or discoloured patches rather than as corrosion products.
Bond coats and diffusion coatings fail by depletion. Once the reservoir of aluminium is exhausted the coating can no longer reform a protective scale, and attack proceeds into the base metal. A visually intact coating can be functionally spent, which is why remaining composition is measured rather than inferred from appearance.
What the laboratory actually measures
Coating condition is assessed on cross-section: remaining thickness, the depth and morphology of any internal attack, and the state of the interdiffusion zone. Preparation practice for thermal sprayed coatings is covered by ASTM E1920, and microscopical thickness measurement on a cross-section by ASTM B487.
Scanning electron microscopy with energy-dispersive spectroscopy, applied under the guidance in ASTM E1508, identifies the corrosion products and the residual aluminium content, separating a sulfidation product from a simple oxide. Where adhesion of a sprayed coating is in question, ASTM C633 is the recognised bond-strength test, though it runs on coupons rather than on the failed blade.
Erosion is a different kind of surface loss
Solid-particle erosion removes material mechanically, typically at the leading edge and tip, and in steam plant it is commonly traced to exfoliated oxide scale carried from upstream tubing. Compressor blading suffers its own erosion from ingested dust and from washing practice.
Erosion and corrosion are separated by morphology and chemistry: erosion leaves a directional, polished or faceted surface without corrosion products, while corrosion leaves scale and subsurface attack. Both thin the section, but they point at entirely different upstream problems.
Surface loss changes the failure that follows
Once the coating is breached the base metal oxidises, the section thins, the surface roughens and local metal temperature rises. Each of those raises stress or lowers material capability, and the blade eventually fails by whichever mechanism reaches its limit first.
That final mechanism is what fractography reports. Naming it without the precursor produces an opinion that is technically correct and practically misdirected, because the intervention that would have prevented the failure sits with the coating and the contaminant, not with the fracture.
The repair and overhaul record
Strip-and-recoat history is central. Which blades were refurbished, at which shop, to which coating specification, how many times, and whether wall thickness was restored or merely re-covered are all documented decisions, and they are frequently where an overhaul-shop or warranty dispute turns.
Comparison against unrepaired blades from the same set, and against the manufacturer's specification for coating type and thickness, is what converts that paperwork into a technical finding.
Where these opinions are challenged
Commonly: that the sections examined came from locations unrepresentative of the row; that remaining-coating measurements were too few to characterise a distribution; that the contaminant source was inferred from the corrosion product rather than demonstrated from fuel, air or water records; and that the comparison blades had a different service history than assumed.
Sampling plan, sample count and the basis for each comparison belong in the report. Stated plainly, they answer the challenge before it is made.
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.