TY - JOUR
T1 - Reverse design of non-equimolar rare-earth monosilicates
T2 - Data-driven and molecular dynamics insights into CMAS corrosion mechanism
AU - Qian, Bin
AU - Wang, Yu
AU - Xu, Keyuan
AU - Zu, Jiahao
AU - Liu, Jiaqi
AU - Liang, Wei
AU - Yu, Fangli
AU - Li, Yan
AU - Bai, Yu
N1 - Publisher Copyright:
© 2026 Elsevier Ltd
PY - 2026/8
Y1 - 2026/8
N2 - Rare-earth monosilicates (REMs) exhibit excellent resistance to calcium-magnesium-alumino-silicate (CMAS) corrosion, but the vast compositional space, particularly for non-equimolar ratios, poses significant challenges for rapid development and mechanistic understanding. In this work, a reverse design strategy was employed to tailor non-equimolar REMs, coupled with a multiscale investigation framework. Density functional theory trained neuroevolution potential molecular dynamics simulations, revealed the atomic-scale CMAS corrosion mechanism: initial interfacial contact, interdiffusion with bridging oxygen formation and substitution channel migration, and final apatite phase precipitation. Lattice distortion in the REMs mainly from RE cation sites and spatial reorientation of [SiO4] tetrahedra. Data-driven models for residual melt viscosity and phase evolution demonstrated strong cross-scale consistency, identifying Sc-rich compositions as exhibiting superior corrosion resistance during the initial stages of CMAS attack at 1400 °C. This combined data-driven, simulation, and experimental workflow offers a robust pathway for building multiscale frameworks to alleviate CMAS degradation in environmental barrier coatings.
AB - Rare-earth monosilicates (REMs) exhibit excellent resistance to calcium-magnesium-alumino-silicate (CMAS) corrosion, but the vast compositional space, particularly for non-equimolar ratios, poses significant challenges for rapid development and mechanistic understanding. In this work, a reverse design strategy was employed to tailor non-equimolar REMs, coupled with a multiscale investigation framework. Density functional theory trained neuroevolution potential molecular dynamics simulations, revealed the atomic-scale CMAS corrosion mechanism: initial interfacial contact, interdiffusion with bridging oxygen formation and substitution channel migration, and final apatite phase precipitation. Lattice distortion in the REMs mainly from RE cation sites and spatial reorientation of [SiO4] tetrahedra. Data-driven models for residual melt viscosity and phase evolution demonstrated strong cross-scale consistency, identifying Sc-rich compositions as exhibiting superior corrosion resistance during the initial stages of CMAS attack at 1400 °C. This combined data-driven, simulation, and experimental workflow offers a robust pathway for building multiscale frameworks to alleviate CMAS degradation in environmental barrier coatings.
KW - CMAS corrosion
KW - Data-driven
KW - Environmental barrier coatings
KW - Molecular dynamics
KW - Rare-earth monosilicates
UR - https://www.scopus.com/pages/publications/105039147883
U2 - 10.1016/j.corsci.2026.113908
DO - 10.1016/j.corsci.2026.113908
M3 - 文章
AN - SCOPUS:105039147883
SN - 0010-938X
VL - 268
JO - Corrosion Science
JF - Corrosion Science
M1 - 113908
ER -