Machine Learning-Guided Design and High-Temperature Corrosion Behavior of Pt-Modified Aluminide Coatings
ID:92 Submission ID:94 View Protection:ATTENDEE Updated Time:2026-10-10 09:59:12 Hits:3 Oral Presentation

Start Time:2026-10-18 15:10 (Asia/Shanghai)

Duration:15min

Session:[SE] Surface Engineering Early-Career Scholars Forum » [Sv] reports

No files

Abstract
        Pt-modified aluminide coatings are widely used to protect gas turbine hot-section components against high-temperature oxidation and hot corrosion. However, their long-term durability remains a concern under aggressive marine conditions. This work combines machine learning-guided compositional design with experimental investigations of the high-temperature corrosion behavior of Pt-modified aluminide coatings. A Gaussian process regression model was trained on a hot-corrosion dataset compiled from published studies and experimental measurements to establish quantitative correlations among coating composition, test conditions, and corrosion-induced mass change. Based on a multilayer screening strategy, the target Al and Pt concentration ranges were identified, and a β-(Ni, Pt)Al coating with a composition of 48.71Al–3.12Co–2.22Cr–37.11Ni–8.84Pt (at.%) was successfully prepared. The coating exhibited good resistance to NaCl/Na2SO4-induced hot corrosion at 900℃. Microstructural characterization revealed the formation of a heterostructure oxide scale, in which the dense inner Al2O3 layer effectively inhibited the inward penetration of corrosive species and played a key role in the excellent hot corrosion resistance of the coating.
        Further investigations under high-temperature salt-fog conditions revealed substantially improved corrosion resistance of the β-(Ni, Pt)Al coating compared with the uncoated substrate, with additional improvements achieved through pre-oxidation. At 900℃, coating degradation was dominated by hot corrosion, while the evolution of the surface oxide scale was governed by the competition between oxidation and salt-induced corrosion, resulting in a layered structure consisting of an Al2O3-rich outer layer and an underlying nitride-containing inner layer. At 1000℃, scale growth and elemental diffusion became increasingly important in governing coating degradation, suggesting a shift in the dominant corrosion processes with increasing temperature. These results highlight the potential of data-driven compositional design for improving the hot-corrosion resistance of Pt-modified aluminide coatings and provide insights into their degradation mechanisms in marine environments.
Keywords
High-temperature corrosion,β-(Ni, Pt)Al coating,Machine learning,Corrosion behavior,Simulation marine environment
Speaker
Peng Deng
Engineer Institute of new materials, Guangdong academy of sciences

Submission Author
Peng Deng Institute of new materials, Guangdong academy of sciences
Bin Yin Institute of new materials, Guangdong academy of sciences
Jianbao Gao Central South University
Lijun Zhang Central South University
Chunming Deng Institute of new materials, Guangdong academy of sciences
Comment submit
Verification code Change another
All comments
Registration Submission