Between roughly 60 and 175 °C, insulated carbon steel sits in the worst window for corrosion under insulation.
A painted pipe inside a hot insulation system is not protected by the paint alone. It is protected by paint, by the absence of water, and by luck, in roughly that order. Most of the corrosion found under thermal insulation happens on steel that was coated to specification, at a temperature that fell into the wrong band.
Why the temperature band decides everything
Below about 60 °C, water can condense and pool, but it evaporates slowly and the situation behaves like ordinary atmospheric corrosion. Above roughly 175 °C, moisture that reaches the steel is driven off quickly and the surface stays relatively dry. The dangerous interval lies between those two figures, where water arrives, stays, and evaporates repeatedly as the plant cycles up and down. That cycling is what concentrates salts, re-wets the surface and drives the corrosion rate far above what a static wet condition would produce.
Carbon steel in cyclic service across that band is the classic CUI case. Stainless steel has its own version of the problem, external chloride stress corrosion cracking, and the coating selection logic is different again.
Where conventional systems stop
A standard epoxy or polyurethane system that performs well at ambient temperature usually carries a continuous service limit well below the hot-service requirement. Above that limit the film softens, loses adhesion and may blister as water vapour passes through it. A coating that is thermally unsuitable fails not because it is a poor coating but because it was asked to work outside its range.
Read the data sheet for two numbers, not one: continuous service temperature and peak or excursion temperature, with the permitted duration of the excursion. Suppliers quote them differently, and a peak rating is not a continuous rating.
The families that are used
- Silicone-based and silicone hybrid coatings, tolerant of high continuous temperature and thermal cycling
- Aluminium-filled silicone systems, where the aluminium flake adds barrier performance and heat resistance
- Inert multipolymeric matrix coatings, formulated for cyclic high-temperature service and for CUI duty
- Thermal spray aluminium, a metallic option with a long record in hot cyclic service but a different application route and its own sealing requirement
Below the hot band, glass-flake or high-build epoxy systems remain the practical choice, and the selection question becomes one of surface preparation quality rather than polymer chemistry.
What actually kills these systems
Poor surface preparation under the insulation is the first cause, and it is invisible for years. Trapped salt at the steel interface will pull water through a marginally permeable film and start corrosion beneath it. The second cause is mechanical: insulation cladding that has been walked on, or a coating damaged during installation of the lagging, leaves a bare spot in the wettest place on the plant. The third is water ingress at penetrations and terminations, where the detailing was never designed.
This is why CUI specifications increasingly pair the coating with the insulation system and its cladding rather than treating the coating in isolation. A perfect coating behind a failed vapour barrier is a delay, not a solution.
Specifying it in practice
- Define the operating temperature range including excursions, not a single design figure
- State whether service is cyclic or continuous, because the two select different products
- Require surface preparation and salt contamination limits explicitly, and verify them before the first coat
- Specify the dry film thickness for the actual duty, not for ambient atmospheric exposure
- Include cladding, sealing and penetration detailing in the same scope of work
- Plan inspection points under insulation at defined intervals; assume the coating will be examined only when the lagging is removed
The most expensive mistake in this area is not choosing the wrong polymer. It is choosing a reasonable polymer for the wrong temperature band, and then discovering it fifteen years later when the insulation is stripped for an unrelated reason.