Blistering in a Glass-Flake Lining Six Months After Commissioning
Six months after commissioning, a 500 m³ carbon steel tank lined with 400 µm nominal glass-flake epoxy in 60 °C brine shows blistering: 3–8 mm domes, 20–40 per m², concentrated in a 200–400 mm band above the bottom seam and around a nozzle. Original QC: DFT 380–450 µm by magnetic gauge over a 75 µm blast profile; holiday test at 5 kV found nothing; pull-off adhesion 6 MPa. Re-inspection finds blister fluid at pH 11 with no chloride detected, no rust beneath peeled blisters, but a rust-stained weep at the nozzle. With the coating manually removed, magnetic gauge readings near the blisters are around 300 µm.
1. Classify the blistering. Which evidence drives the classification and which evidence is misleading?
Read the worked answer
Classification: corrosion-driven (cathodic) blistering at a defect, not osmotic blistering from salt contamination.
Why cathodic. A blister fluid at pH 11 with no chloride is the signature of the cathodic half-reaction - oxygen reduction at the steel surface under the film generating hydroxide - rather than of a soluble salt drawing water in by osmosis. Osmotic blisters driven by chloride typically show near-neutral, chloride-rich fluid. The absence of chloride supports the alkaline finding rather than contradicting it. The location is also diagnostic: around a nozzle and above a weld, where geometry creates coating defects and where the weep shows liquid has been in contact with steel.
What the original QC proves and does not prove. Holiday testing at 5 kV on a 400 µm film found no defects, and adhesion at 6 MPa exceeded a typical 5 MPa minimum - neither rules out a mechanical defect introduced after the test, or one hidden by a conductive path the test did not reach. A passing spark test is a point-in-time result, not a guarantee.
The misleading evidence is the initial DFT of 380–450 µm, which appears fully compliant but was measured over a profile that inflates magnetic readings.
2. What does the 300 µm re-inspection reading actually tell you, and what would settle whether the lining is thin at the blisters?
Read the worked answer
Magnetic gauges read over the profile peaks, not through them. Over a 75 µm blast profile, a magnetic gauge typically reads high by roughly half to all of the profile amplitude. A 300 µm gauge reading on that surface corresponds to something on the order of 225–265 µm of true film thickness - well under the 400 µm nominal, and thinner still if the profile is at the coarse end.
Why thin matters here. At 60 °C in brine, permeation scales strongly with reduced thickness. A locally thin film above a seam lets electrolyte reach the steel, corrosion starts, hydroxide builds under the film, and the resulting osmotic and gas pressure lifts the dome. The blister is the symptom; the thin film at a weld or nozzle is the cause.
What settles it:
- Re-measure at a prepared, flattened area with the profile contribution removed, or by taper-section/microscopy on a cut coupon from a blistered region.
- Map DFT specifically along the seam and around the nozzle rather than on the general wall.
- Examine the weep area destructively to establish whether steel corrosion preceded or followed lining damage.
A finding of 250 µm against a 400 µm nominal on an immersion lining is a defect in itself, independent of the blisters.
3. Define the repair scope and the preventive actions for the next tank.
Read the worked answer
Repair scope. A local patch is unlikely to hold: blistering concentrated at a weld and a nozzle means the interface at those details is compromised beyond the visible domes. Strip the band and the nozzle surround back to sound, adherent lining with a feathered edge, abrasive-blast to the specified profile, re-establish chloride-free surface, rebuild the full system thickness, and spark-test the repair and its perimeter. Repair the nozzle detail - the weep suggests either a mechanical defect or a design issue at the penetration - before relining over it.
Preventive actions for the next tank:
- Map DFT at welds, seams and nozzles during application, and correct for profile with replica tape, rather than accepting a global average that hides thin details.
- Specify and verify a stripe coat at welds and penetrations, where spray geometry typically delivers less film.
- Tighten the acceptance criterion to a minimum DFT at details, not just a nominal over the general surface.
- Treat the spark test as one data point among several, and pair it with DFT mapping and adhesion testing on witness areas at the same details.
- Record the blister fluid chemistry and location pattern in the failure report; it distinguishes corrosion-driven from osmotic mechanisms next time and points to a different root cause and a different fix.