On the liquid–liquid phase transition of dense hydrogen

Journal Article · 15 December 2021 · Nature (London)

DOI: https://doi.org/10.1038/s41586-021-04078-x · OSTI ID:1837432

Karasiev, Valentin V.  ^[1]; Hinz, Joshua ^[2]; Hu, S. X.  ^[2]; Trickey, S. B.  ^[3]

---

1. Univ. of Rochester, NY (United States); Laboratory for Laser Energetics, University of Rochester
2. Univ. of Rochester, NY (United States)
3. Univ. of Florida, Gainesville, FL (United States)

Until recently the consensus theory/computation interpretation of the challenging liquid liquid phase transition (LLPT) of high-pressure hydrogen was first order. Cheng et al. developed a machine-learnt potential (MLP) that, in larger molecular dynamics (MD) simulations, gives a continuous transition instead. Here, we show that the MLP does not reproduce our still larger MD density-functional theory (MD-DFT) calculations as it should. Since the MLP is not a faithful surrogate for the MD-DFT, the Ref. 6 prediction of a supercritical atomic liquid is unfounded.

• Research Organization: Univ. of Rochester, NY (United States). Lab. for Laser Energetics; University of Florida, Gainesville, FL (United States)
• Sponsoring Organization: USDOE National Nuclear Security Administration (NNSA); USDOE Office of Science (SC), Basic Energy Sciences (BES)
• Grant/Contract Number: NA0003856; SC0002139
• OSTI ID: 1837432
• Journal Information: Nature (London), Journal Name: Nature (London) Journal Issue: 7889 Vol. 600; ISSN 0028-0836
• Publisher: Nature Publishing GroupCopyright Statement
• Country of Publication: United States
• Language: English