LDRD Day Cultivating Next-Generation Science
Los Alamos National Laboratory scientists and engineers participated in the Laboratory's second annual Laboratory-Directed Research and Development (LDRD) Day, at which the public viewed advanced research currently underway. This year's LDRD Day, held near Santa Fe on September 28, consisted of poster presentations that covered subjects spanning global security, space science, and biofuels development. Materials science—a focal point in several research fields—and its reliance on computers was highlighted in many of the presentations.
Advancing Our Understanding of Materials and Achieving Exascale Supercomputing
Charles McMillan, featured speaker at LDRD Day and head of LANL's weapons program, noted that sustaining the nation's nuclear deterrent over the long term requires investing in advanced science and technology. He highlighted how nanoscale materials synthesis and advanced simulation are transforming materials science, which has broad applications (including energy independence). Initially funded by LDRD, LANL researchers discovered new mechanisms for substantially increasing materials strength and radiation resistance. According to McMillan, two elements of interest to LDRD researchers are better understanding of materials in extreme environments and achieving exascale computing—supercomputers capable of handling a million trillion calculations per second.
In addition to sustaining the nation's nuclear deterrent, exascale computing and better understanding of materials at the mesoscale have applications that can benefit the world, according to McMillan. For example, it would be possible for researchers to design technologies to better protect electric-transmission infrastructures from attack, create novel metals for industrial applications, and develop technologies that enhance energy production, transmission, and storage.
MaRIE and Cielo: Two Powerful Allies
Advanced computing capabilities or an extensive knowledge of materials often lies at the core of breakthrough science. This is a great strength for LANL and its LDRD-funded researchers. The first in a proposed new generation of scientific facilities for the materials community, MaRIE (Matter-Radiation Interactions in Extremes) will enable scientists to develop next-generation materials that will perform predictably and on demand for currently unattainable lifetimes in extreme environments.
Cielo is a new supercomputing platform at LANL with petascale capabilities—that's more that one quadrillion floating-point operations per second. The role of Cielo will be to run the largest and most demanding workloads involving modeling and simulation. Cielo will support large single jobs that can take advantage of the entire platform.
Combined, facilities such as MaRIE and supercomputers such as Cielo (in conjunction with advances in computational science and theory) will serve as tools that scientists will use to make breakthroughs in predictability that in turn will enable materials science to move from a focus on empirical observation to one of predictive design and control.
LDRD Day Presentations: Work on Scientific Breakthroughs in Predictability Already Underway
At this year's LDRD Day, scientists from Los Alamos presented 45 posters covering the projects sponsored by LDRD. Following are highlights of LDRD Day posters demonstrating efforts in predictability specific to materials science and advanced computing.
Understanding Explosives Initiation for Threat Evaluation and Mitigation
Understanding how an explosive is set off at a molecular and granular scale is critical for scientists designing safer explosives. Such knowledge may even help find ways to prevent explosives from initiating.
To determine elements such as initiation thresholds, mechanisms, and initiation pressure dependencies of relevant explosive formulations, scientists are using in situ measurements of shock and reactive-wave profiles under controlled shock-compression conditions. Liquid explosives (homogenous) display different initiation behaviors than do multicomponent (heterogeneous) explosive mixtures. Whereas homogenous explosives initiate as a result of a thermal explosion derived from bulk shock heating, heterogeneous explosives initiate with the creation of "hot spots" (localized regions of high temperature and pressure).
As a result of this LDRD project, Los Alamos researchers now have a thorough understanding of the initiation behaviors of several important liquid explosives. They have compared their relative sensitivities and have discovered that pressure dependencies of their run-distances-to-detonation appear to be similar across the series. Scientists have also learned that incorporating hot-spot "seeds" influences initiation (critical size and separations, for example).
Scientists have evaluated the relative effectiveness of solid vs. hollow particles in creating hot spots and have discovered that a balance exists between hot-spot-driven and thermal-driven burn mechanisms. By studying explosives under controlled shock conditions, scientists are establishing a foundational knowledge of the important features that dictate initiation behaviors. Predictive models derived from this knowledge will help researchers create the next-generation of explosive-modeling capabilities for the United States.
Los Alamos scientists working on this LDRD project are D. Dattelbaum (principal investigator), S. Sheffield, R. Engelke, D. Stahl, and L. Gibson.
LANL scientist Al Migliori talks with weapons director Charles McMillan and former Congresswoman Heather Wilson.
First-Principles Molecular Dynamics for Extended Length and Time Scales
Designing new materials is at best a hit-and-miss approach using traditional laboratory processes. Because such efforts waste time and effort, scientists have turned to computers, where they can analyze and experiment with materials without any experimental input. In the world of computing, it is possible to make a model of a million atoms as they move and vibrate at unimaginable speeds.
Simulations of molecular dynamics enable researchers to conduct detailed studies of fundamental material properties, such as phase transitions, chemical reactions, and molecular structures. Currently, the "gold" standard for atomistic simulations consists of molecular dynamic simulations in which the molecular motion is derived directly from the first principles of quantum mechanics.
First-principles simulations require no input from experiments, thus enabling analysis and prediction of material properties under conditions that are either too expensive or cannot be naturally achieved under laboratory conditions. Such simulations also yield more complete and detailed descriptions that can lead to a better understanding and sometimes the successful prediction of materials with new tailored properties.
Los Alamos scientists have taken a major step forward in the ability to study and predict complex material properties directly from theory. In conjunction with rapid increases in computing power, basic research conducted so far may find applications in almost every field of materials science, chemistry, and molecular biology. Applications include developing high-performance steel, new and safer explosives, nanodevices, novel polymers, and enzymes for producing new biofuels.
Los Alamos scientists working on this LDRD project are Edward Sanville, A. Niklasson (principal investigator), M. Cawkwell, D. Dattelbaum, and S. Sheffield.
Novel Materials for Detection of Radioactive Materials
A scintillator is in essence a substance that glows when struck by high-energy particles or photons. By better understanding how scintillators work, scientists can develop better detection devices and thus minimize the proliferation of radioactive materials.
Using a combined approach that consists of materials synthesis and theoretical modeling, Los Alamos scientists are designing new, more efficient optical materials and modeling their luminescent behavior. Synthesizing modular molecular and supramolecular materials containing phosphors enables researchers to control the size and dimensionality of such materials, as well as providing a better understanding of the scintillation process. By using synthesis, characterization, and theoretical modeling to better understand how molecular phosphors work, scientists can create a platform from which to design more efficient detector materials.
Los Alamos scientists working on this LDRD project are Rico Del Sesto (principal investigator), D. Ortiz-Acosta, and R. Feller.
Petascale Synthetic Visual Cognition: Large-Scale, Real-Time Models of Human Visual Cortex on the Roadrunner
Humans take sight as a given, but the human eye and brain—and their interactions—are extremely complex. Thus, it appears easy for a human to spot a vehicle hidden under camouflage, yet not even the fastest supercomputer in the world can analyze a similar image and reliably find the vehicle.
The goal of this LDRD project is to understand how the human visual cortex works so that researchers can teach computers to "see." To achieve this goal, Los Alamos scientists are using one of the most powerful supercomputers, Roadrunner, to build the world's first full-scale, real-time model of the part of the human brain known as the visual cortex. Also of interest are small mammals capable of excellent visual acuity and objective recognition with orders-of-magnitude smaller brains. New computing technology based on graphic cards could enable the widespread use of brain models for many computer vision tasks.
Los Alamos scientists working on this LDRD project are Luís Bettencourt (principal investigator), S. Brumby, G. Kenyon, J. George, C. Rasmussen, A. Galbraith, M. Anghel, and M. Ham.
Sergei Tretiak discusses his materials research. Tretiak won a 2010 LANL Fellows Prize for Outstanding Research in Science or Engineering, in part for his development of organic light-emitting diodes for flexible displays, organic lasers, and light-harvesting energy devices.
This year's LDRD Day posters addressed four focus areas: energy security, national security, global security, and scientific discovery.
The United States is currently facing security, energy, and environmental challenges that, with respect to their scope and complexity, are perhaps unparalleled in the nation's history. So, how does the nation go about developing long-term solutions to such challenges?
"Scientific breakthroughs seldom arise spontaneously from individual talent but from a critical mass of talented individuals who are supported by an institution committed to basic research," noted Los Alamos LDRD Program Director William Priedhorsky.
The LDRD program is also a powerful means of attracting and retaining top researchers from around the world. Such researchers foster collaborations with other prominent scientific and technological institutions and leverage some of the world's most technologically advanced assets.
By investing in high-risk and potentially high-payoff research, the LDRD program works to create innovative technical solutions for some of the most difficult challenges faced by the United States and the world. In many instances, projects started with LDRD funding grow into much larger projects that in some cases attract interest from other government agencies, academia, and private industry.
"Much of the Laboratory's scientific capabilities, from energy security to large-scale infrastructure modeling, from actinide science to nuclear nonproliferation and detection, can be traced to LDRD investment," said Priedhorsky.