High‐Entropy Perovskite Oxide With Ultralow Thermal Conductivity via Computational‐Experimental Synergy

ABSTRACT High‐entropy perovskite ceramics have emerged as a promising new generation of thermal protection coating material due to their exceptional thermal protection performance. However, the vast compositional space of high‐entropy systems challenges the identification of optimal compositions through conventional trial‐and‐error experimental methods. This work integrates machine learning and strategic doping. A random forest model with a high predictive accuracy ( R 2 = 0.86) was first employed to identify an optimal A‐site composition. Subsequently, B‐site doping was introduced to further reduce thermal conductivity and enhance the high‐entropy ceramic material's stability. The synthesized high‐entropy ceramic, (Ca 0.2 Sr 0.2 Ba 0.2 La 0.3 K 0.1 )(Ti 0.5 Hf 0.5 )O 3 (HETH), exhibits significantly intensified lattice distortion and mass fluctuation owing to Hf incorporation, which synergistically strengthens phonon scattering. Consequently, the HETH demonstrates monotonically decreasing thermal conductivity with temperature, achieving an ultralow value of 0.759 W·m − 1 ·K − 1 at 1200°C. This study establishes an efficient and accelerated strategy for developing advanced high‐entropy thermal protection coating materials.