Investigating the Hydroxyl Reorientation in Hydroxyapatite Using Machine Learning Potentials

The chain or network of hydroxylgroups (OH-)is crucial in determining the structure and function of materials,especially in hydroxyapatite (HAP), a mineral essential for humanbones. HAP exhibits a linear arrangement of OH- alongthe c-axis, which determines its phase transition,dielectric, and piezoelectric properties. However, the mechanism underlyingOH(-) reorientation with temperature remains elusiveusing traditional experimental and theoretical methods. To addressthis, we developed a machine learning atomistic potential for HAPusing an active learning algorithm, which achieved density functionaltheory-level accuracy in describing OH- of HAP.The machine learning molecular dynamics simulations revealed thatthe reorientation of OH- in HAP with temperatureoccurs through '' flip-flop '' motion, rather than protontransfer. This process starts at about 473 K and accelerates withincreasing temperature, consistent with the experimentally observedtransformation from the monoclinic to hexagonal phase. At 973 K andabove, the rapid "flip-flop" reorientation process leadsto an undetermined orientation of OH- along the c-axis. These findings highlight the potential of machinelearning-accelerated molecular dynamics simulations in unravelingthe microscopic mechanisms underlying the hydrogen bond network incomplex multicomponent materials at the atomic level.