A Major Advance in Understanding Metals
A principal element of Lawrence Livermore’s national security mission is stockpile stewardship, the program to ensure that the nation’s nuclear stockpile is safe and reliable. Many of the key components of nuclear warheads are crystalline metals. The fundamental defect governing the mechanical behavior of a crystalline metal is a dislocation line—that is, a displaced row of atoms. The interaction among thousands of dislocation lines, called dislocation dynamics, is the key to a metal’s plasticity and strain hardening. Plastic deformations confer strength that is proportional to the number of deformation lines and their entanglements.
For the past decade, Livermore researchers have been developing increasingly realistic three-dimensional simulations of dislocation dynamics because this phenomenon is the missing link in multiscale modeling. Dislocation dynamics lies at the microscale, between atomistic descriptions of materials that incorporate atomic-scale physics and mesoscale and continuum scales that describe the engineering performance of a material component. By understanding how dislocations behave at all scales, we can predict how materials respond under many different conditions and why and how they fail.
As the article Materials Scientists Discover the Power of ParaDiS describes, a team of Livermore researchers has developed a new code, called Parallel Dislocation Simulator (ParaDiS), that models the behavior of large ensembles of dislocations in unprecedented detail and under a wide range of conditions. The development of ParaDiS is significant because the code will help us to better predict the performance of materials in the stockpile as a function of age and under extreme conditions. In short, ParaDiS allows us, for the first time, to use a strong scientific basis to understand and predict dislocation dynamics performance of crystalline metals.
In developing ParaDiS, we built on two foundations. The first is an improved understanding of the physics of crystalline metals obtained using atomic-scale simulations and laboratory experiments. Our findings were the subject of several papers published in scientific journals over the past few years. This work provided the data about atomistic phenomena used in ParaDiS. The second foundation is the availability of massively parallel supercomputers, such as Livermore’s Thunder and BlueGene/L, which use thousands of microprocessors working in tandem. Developing a code to work on these machines was an enormous challenge requiring the combined efforts of Livermore materials scientists, computational scientists, and physicists. Over a period of 4 years, the team, led by materials scientist Vasily Bulatov, worked to produce the world’s highest performance dislocation dynamics code—a code based on the physics of dislocation lines that also takes advantage of the latest advances in supercomputers.
The capability of the code was demonstrated in its unexpected discovery of a new type of dislocation microstructure, called a multijunction. Multijunctions are thought to play an essential role in the strength evolution of crystalline metals as they deform. As the article describes, the existence of multijunctions was confirmed in images taken with a transmission electron microscope. We are currently exploring the full practical significance of multijunctions.
ParaDiS will allow us to examine other issues of materials phenomena, especially plasticity, that have been unexplained for decades. In particular, we now have a major new tool for predicting the performance of the aging components in our current nuclear stockpile and for possibly helping to guide the development of advanced materials for the National Nuclear Security Administration’s Reliable Replacement Warhead Program. This program is evaluating the feasibility of providing replacements for existing stockpile weapons that are more reliable, less expensive to maintain, and more easily manufactured using readily available and environmentally benign materials.
In addition, we anticipate that ParaDiS will be useful in strengthening America’s energy security. For example, it could predict the performance of materials for fission power plants and for magnetic fusion energy and inertial confinement fusion plants. We are confident that other applications will be found for ParaDiS as it becomes adopted throughout the materials science community.