Lawrence Livermore National Laboratory



Back to top

Back to top

Superionic Water Properties Deciphered

At a specific combination of extreme temperature and pressure, water can enter a “superionic” phase in which oxygen atoms retain a lattice structure, but hydrogen ions move fluidly with liquidlike behavior. Superionic water remained a hypothetical state for decades, thought to exist within massive planets such as Uranus and Neptune that contain substantially more water than all of Earth’s oceans. Only recently have scientists confirmed the superionic phase’s existence and accurately assessed its properties.

New research published September 23, 2021, in Nature Physics, applies machine learning to understand the behavior of atoms in water’s superionic state. Experimentation with superionic water is complex, so quantum-based simulations of molecular dynamics are often used to help design an experiment and evaluate its results. However, these simulations become prohibitively expensive with large system sizes and timescales greater than a matter of picoseconds. Machine learning programmatically determines the nature of atomic interactions via quantum mechanical calculations so the team can infer atomic behavior over longer timescales with greater precision.

Identifying the phase boundaries of water in such an environment allows scientists to differentiate multiple superionic phases present within ice giants. Co-author Sebastien Hamel says, “Our quantitative understanding of superionic water sheds light into the interior structure, evolution, and magnetic fields of planets such as Uranus and Neptune and also of the increasing number of icy exoplanets.”

Contact: Sebastien Hamel (925) 423-8048 (hamel2@llnl.gov).


Livermore Optics for World’s Newest Telescope

For the past decade, Livermore researchers have collaborated with international partners to design and fabricate major optical components for the world’s newest telescope. Now complete, the 8.4-meter Simonyi Survey Telescope will take digital images of the southern sky using the Legacy Survey of Space and Time Camera (LSSTCam) at the Vera C. Rubin Observatory in Northern Chile. Livermore researchers made essential contributions to the optical design of LSSTCam’s lenses and the Simonyi Survey Telescope’s mirrors such as determining how the camera and telescope surveys the sky and how these components work together to compensate for temperature and gravity.

The telescope’s camera weighs more than 3 tons. Its six 76-centimeter-diameter filters are among the largest produced, and one of its three optical lenses is the world’s largest, high-performance optical lens at over 1.5 meters in diameter. Inside the National Ignition Facility’s optical assembly building, industrial partners fabricated the lenses and filters, which were then placed into Laboratory-developed mounts. Each filter transmits light from a segment of the electromagnetic spectrum, progressing throughout the entire visible range and moving from near-ultraviolet to near-infrared.

“The successful fabrication of these optical filters and lens assemblies is a testament to the Laboratory’s world-leading expertise in large optics, built on decades of experience constructing the world’s largest and most powerful laser systems,” says Livermore physicist Scot Olivier. LSSTCam data will help researchers better understand the makeup of the universe, detect and study about 20 billion galaxies over a 10-year span, track potentially hazardous asteroids, and observe exploding stars.

Contact: Scot Olivier (925) 423-6483 (olivier1@llnl.gov).


Algorithmic Blackbox Takes on Black Holes

Black hole mergers are the only cosmic phenomena explosive enough to produce gravitational waves detectable by current instrumentation. Such a merger—in which binary black holes (BBH) fuse and expel mass as energy—was the source of the revelatory direct observation of gravitational waves at the Laser Interferometer Gravitational-Wave Observatory in 2016. Findings published November 9, 2021, in Physical Review Research, present new computational methods to rapidly decode the dynamics of the black hole systems responsible for gravitational disruptions.

Albert Einstein’s field equations to describe black hole mergers require resource-intensive computation to solve for each unique combination of physical parameters, leading researchers to reduce them to more manageable, yet less precise forms for describing BBH motion throughout each merger stage. Livermore mathematician and computational scientist Brendan Keith led a collaborative effort with researchers from the University of Massachusetts, Dartmouth College, and the University of Mississippi to rethink the problem. The team devised a machine-learning model that intakes raw gravitational wave data to systematically learn and produce equations describing BBH motion.

Starting with a nonrelativistic physics model and a system of differential equations adjusted and refined by neural networks, the method algorithmically fills in the relativistic aspects of motion not accounted for in the basic model. “Our model takes astronomically fewer resources than conventional computational methods and can return highly accurate descriptive equations in mere minutes,” says Keith. This predominantly data-driven approach may provide for accurate simulations of astrophysical dynamics even in scenarios with limited or low-resolution data.

Contact: Brendan Keith (keith10@llnl.gov).