Surprising Results Illuminate Hot Spots
A Lawrence Livermore team has advanced understanding of microscopic hot spot formation in insensitive high explosives based on 1,3,5-triamino-2,4,6-trinitrobenzene (TATB). Their findings, published February 25, 2025, in Journal of Physical Chemistry C, simplify hot spot modeling and offer guidance for applying molecular dynamics (MD) simulations at micrometer-length scales. Hot spots—regions of intense heat—can form at microstructural defects and strongly affect an explosive’s detonation response. Expanding knowledge of hot spot formation mechanisms across length scales is essential for developing predictive models of explosive safety and performance.
Led by computational chemist Matt Kroonblawd, the team ran the largest MD simulations of an explosive material to date on Livermore’s Sierra supercomputer, ultimately reaching a multimicrometer domain containing a 300-nanometer pore defect that was fully resolved with 600 million atoms. Complementary continuum-based ALE3D simulations analyzed even larger pores and bridged the gap between atomistic and microstructural scale modeling.
Hot spot temperature distributions and structural features were found to scale directly with system size for pores larger than 20 nanometers. This scale invariance was traced to TATB’s mechanical strength (its stress–strain response) being unaffected by strain rate under ultrafast shock conditions. Kroonblawd says, “Such work positions molecular dynamics simulations as a foundation for developing more general multiscale models of insensitive high explosives and can potentially guide the development of new explosive materials with improved safety and performance properties.”
Contact: Matt Kroonblawd (925) 422-2221 (kroonblawd1 [at] llnl.gov (kroonblawd1[at]llnl[dot]gov)).
Audible Enclaves for Directed Sound
Researchers at Lawrence Livermore National Laboratory and The Pennsylvania State University have pioneered a method of delivering sound precisely and exclusively to a location without requiring an on-ear device. Specifically, localized audio spots created through the nonlinear interaction of self-bending ultrasonic beams create a type of audible enclave that renders sound inaudible outside the target area. A paper, published in the March 17, 2025, issue of Proceedings of the National Academy of Sciences, describes the researchers’ application of 3D-printed acoustic surfaces and acoustic principles to achieve directed sound.
The path to self-bending ultrasonic beams started with Livermore engineer Xiaoxing Xia, who fabricated intricate acoustic metasurfaces with a high-resolution stereolithography 3D printer, achieving complex air-channel networks to shape the phase and amplitude of sound waves. Individual meta-atoms constructed in the design and printing process were arranged to modify the wavefront (propagating wave) for tight control of ultrasonic wave trajectories. Furthering the work, researchers applied nonlinear acoustic interactions—in which two self-bending waves of different spectra merge to vibrate at a unique frequency—to enable audible sound while ensuring ultrasonic wave components were inaudible. The team demonstrated the ultrabroadband capabilities of their approach, spanning six octave bands (from 125 hertz to 4 kilohertz) and most of the audible frequency range. Xia says, “Future applications may include encrypted information communication and broader applications of 3D-printed metasurfaces beyond acoustics, such as medical imaging, detection, and wireless communications.”
Contact: Xiaoxing Xia (925) 423-6489 (xia7 [at] llnl.gov (xia7[at]llnl[dot]gov)).
AI Optimizes Antibodies
A team including several Laboratory researchers applied an AI-driven platform to preemptively optimize an antibody to neutralize a range of SARS-CoV-2 variants. The team’s research builds on earlier work to tailor antibodies for infectious disease management using AI tools. The platform and approach are described in a paper appearing in the March 28, 2025, issue of Science Advances; the paper was authored by more than 40 scientists and the Lawrence Livermore National Laboratory Generative Unconstrained Intelligent Drug Engineering (GUIDE) Consortium.
SARS-CoV-2, the virus that causes COVID-19, evolves rapidly, making antibody treatments designed for earlier variants obsolete. Researchers working with pharmaceutical company AstraZeneca studied viral mutations that could reduce the effectiveness of AZD3152, a medicine approved in some nations for COVID-19 pre-exposure protection. Using the GUIDE computational platform, the team analyzed more than 10 billion antibody modifications to determine which candidates might enhance binding with SARS-CoV-2 variants. Candidates viewed to have the highest potential for success were then tested in a laboratory to confirm their efficacy. Following two design cycles, the optimized antibody 3152-1142 emerged as the most promising, demonstrating a 100-times improvement in potency against a SARS-CoV-2 variant for which AZD3152 had not been effective. Livermore computational physicist Fangqiang Zhu, the paper’s first author, says, “By looking ahead to address how the virus might evolve, we are not just responding to current threats, we are proactively developing therapeutics to combat potential future viral evolution.” The team aims to build the capability for rapid antibody redesign, supported by an expedited review cycle, similar to the system for developing and approving influenza vaccines.
Contact: Fangqiang Zhu (925) 724-9897 (zhu13 [at] llnl.gov (zhu13[at]llnl[dot]gov)).
