On January 20, 2023, American Physical Society (APS) notified Assistant Professor of Geosciences, Jie Deng, that his journal article “Melting of MgSiO3 determined by machine-learned potentials" has been selected for the Physical Review B (PRB) Editors’ Suggestion, as well as their Outreach-to-the-Press program. Deng is featured as the first author, along with Haiyang Niu, Junwei Hu, Mingyi Chen, and Lars Stixrude.
Since 1970, the PRB peer-reviewed journal has evolved into the world’s largest publication dedicated to physics, inclusive of condensed matter and materials physics. It has become an essential resource for scientists focused on established and emerging research in condensed matter, materials physics, and their related topics. This particular journal has been known to publish up to 100 significant papers in one week, if not more. The importance of highlighting the most innovative science cannot be understated.
Deng’s paper was published and highlighted on 13 February 2023. It is listed as being in the 6th issue of the February 2023, Vol. 107 edition of the PRB journal. The Editor’s Suggestions appear within the "PRB Highlights." It is the first menu item on the PRB website at journals.aps.org/prb/highlights. What is meant by the Outreach-to-the-Press program is that the APS Physics press office sends out its most press-worthy papers to authors and journalists in the scientific community. This is done on a weekly basis.
The melting temperature of MgSiO3 is crucial in controlling the interior structures and dynamics of Earth and super-Earths. Deng and his coworkers propose an iterative learning scheme that combines enhanced sampling, feature selection, and deep learning, and develop a unified machine learning potential of ab initio quality valid over a wide pressure-temperature range to determine the melting temperature of MgSiO3. Modeling based on these results shows that heat flux from the core to the mantle is favorable for generating magnetic fields.
(Above) Two-phase simulations of MgSiO3 bridgmanite (the most abundant mineral of Earth) and liquid at the Earth’s core-mantle boundary pressure (i.e., 140 GPa) and 6000 K. This temperature is lower than the melting point of bridgmanite, and thus the system crystalized. The system contains 3000 atoms, and the simulation lasts for 1 ns (nanosecond) with a 1 fs (femtosecond) timestep. This simulation is made possible with carefully designed machine learning potential of ab initio quality.
Professor Deng is one of the newest faculty members at The Department of Geosciences arriving in Fall of 2021. He received his Ph.D. in geophysics from Yale University in 2019 and performed post-doctoral scholar work at UCLA with Professor Lars Stixrude. Deng is currently teaching “Introductory Geophysics” and “Fundamentals of the Geosciences” for the Department of Geosciences.
