Study models how coating thickness changes the role of roughness in aluminum corrosion

Shibaura Institute of Technology researchers reported on October 9, 2026, that their analysis of steam-coated aluminum found a change in how coating thickness and surface roughness interact in a model of pitting resistance. The research, published in August, could help guide coating experiments, but its proposed mechanism and usefulness beyond the tested specimens remain unproven.

Pitting is localized corrosion that can damage an aluminum component even when its overall corrosion rate appears low. Steam coating forms a protective boehmite film using water vapor. As that film grows, however, its thickness, surface shape and crystal structure can change together, making it difficult to isolate which features matter most for corrosion resistance.

What the aluminum coating model measured

The team studied 90 steam-coated A6061-T6 aluminum specimens. It measured pitting potential in a 5% sodium chloride solution at room temperature. Pitting potential is the voltage at which localized corrosion starts in this electrochemical test; a higher value indicates greater resistance under those conditions. It does not establish how long a coated part would last in service.

The researchers used four measured descriptors to predict that potential: film thickness, surface roughness, crystallite size and an index of dislocation density in the underlying aluminum. Their random-forest model achieved an R² value of 0.67, according to the paper's abstract. The authors identified roughness and thickness as the leading descriptors. The report also says the model predicted the test results more accurately than one based only on coating temperature and treatment time.

Why the 2,200-nanometer result has limits

The researchers suggest that roughness in thinner films may reflect the development of protective coverage, while roughness in thicker films may be associated with irregularities that allow corrosive species through. That explanation is a proposed mechanism, not a directly validated rule for making coatings. As Takahiro Ishizaki, who led the research with master's student Kei Masuhara, put it in the October report: ‘Higher roughness should not automatically be considered beneficial or harmful.’

The paper also describes interactions involving crystallite size and the other descriptors. Its abstract says the model identified relationships that vary across regions of the measured data, rather than one uniform effect. That makes the result more specific than a recommendation to increase or decrease roughness across all steam-coated aluminum surfaces.

What earlier boehmite research adds

An earlier, separate study listed by Pacific Northwest National Laboratory examined how preparing AA6061 aluminum surfaces affected thermally grown boehmite films. It found that polishing, acid etching and alkaline etching changed film properties. In its tests, acid- and alkaline-etched specimens showed shifts in corrosion potential and stable, pit-free films in potentiodynamic scans. The researchers also reported reduced oxide resistance with extended boehmite growth.

Those earlier findings provide context for why coating conditions matter, but they do not test or confirm the new model's 2,200-nanometer interaction. The two studies asked different questions. Together, they point to the difficulty of judging a coating by one visible or measured feature without considering how it was prepared and grown.

What remains to be tested

The October report says a Monte Carlo analysis attributed about 0.080 volts of prediction variability to uncertainty in descriptor measurements, against an overall model error of 0.292 volts on a root-mean-square basis. Measurement uncertainty therefore accounts for only part of the prediction error. The results leave room for influences the four descriptors do not capture.

Further experiments would be needed to test whether the proposed coverage-to-defect explanation holds and whether the modeled interaction appears with other alloys, coating processes or operating conditions. The reported results come from laboratory corrosion measurements, not a manufacturing deployment or field test. For coating researchers, the immediate finding is a reason to examine thickness and roughness together when designing the next tests, rather than treating either measurement as a stand-alone performance rule.