This award to Californaia Institute of Technology is from the Chemistry Division and the Division of Materials Research in The Mathematical and Physical Sciences and the Office of Cyberinfrastructure. The funds support entitled ?Workshop on Computational Scattering Science 2010? and will be held on July 6-9 at Argonne National Laboratory. The Workshop is to identify, and to reassess opportunities to connect the fields of scattering science and computational science. The goal is to develop a roadmap showing how modern computing can advance scattering science, to offer new opportunities for scientific discovery, and to identify opportunities. The workshop will also develop ideas to organize future software development, manage maintenance, upgrades, and user support at the Nation?s neutron and x-ray facilities. The outcome of the workshop will be a report that will critically assess strengths, weaknesses, and cost-effectiveness and present possible paths forward. The workshop will also identify the role of computational scattering science in education, and in promoting public awareness of scattering research in the U.S. Lowering the barrier for entry into scattering research will also be discussed. Finally, the workshop will help in the formation of a nucleus of a community for computational scattering science, and such a community is needed for optimizing the future effort in this field.
Over the past two decades, large national user facilities such as synchrotron x-ray sources, spallation and steady-state neutron sources, and electron microscopy facilities in the United States have had a major impact on basic research into many of our critical technologies. Discoveries are changing the way we understand the world around us from the perspectives of physics, materials, chemistry, biology, geosciences, engineering, life sciences and medicine, and the energy sciences. In general, the sophistication of experimental work at these national user facilities has grown along with the improved performance of these instruments. Simultaneously, the developments in computational materials science have been at least as remarkable, with game-changing advances in both hardware and in computational methods. Many state-of-the-art computations of structure and dynamics now address the same phenomena, at the same scales of space and time, as state-of-the-art experimental measurements at national user facilities. Far less well developed are the interfaces between computational science and scattering experiments. While measurement capabilities are uncovering information on atomic length scales and on macroscopic functional scales, and discovering a myriad of excitations in condensed matter, the theoretical interpretations of these measurements lag far behind what is possible. This is a loss of opportunity for scientific discovery. The workshop attendees, with the support and endorsement from the NSF and DOE Basic Energy Sciences, assessed the relationship between computing and scattering science. Their focus was on how modern computation can leverage and grow the scientific output from scattering experiments by linking theory, modeling and simulation more closely with experimental scattering research. Those involved in the workshop were experimentalists in materials research, theorists and computational materials scientists, and computer scientists. The interactions between these different groups were often new, and stimulating. It seems likely that follow-on workshops and discussion groups will be organized to identify better the possible needs and scientific collaborations. The outcome of the workshop was a report that offers assessments and specific recommendations in many topical areas. Some ideas, such as making standard computational methods available to interpret experimental data, were widely accepted. These include methods of density functional theory calculations of electronic structure, and molecular dynamics calculations of the motions of atoms. Simulations of experiments, methods for calculating internal strains in materials, and optimization of complex physical models are other types of computations that could improve the sophistication of interpretations of data acquired today. Other computational methods would be most helpful for interpreting modern experimental measurements if they could be perfected in the near term, such as methods for handling correlated electrons, for handling chemical processes far from equilibrium, and computational methdos for fast and ultrafast phenomena in materials. The workshop and its report discussed issues of software development, software maintenance, community expectations, support for multicore computing, education and career paths, and financial support. Some suggestions for the structure of a national organization were also presented. The finished workshop report can be downloaded here: www.its.caltech.edu/~matsci/Publish/CompScatWkshp_2010.html The workshop had intellectual merit because it identified and organized the opportunities to connect the fields of scattering science and computational science. It has been demonstrated that high performance computing can facilitate and elevate the science of neutron, x-ray and electron scattering studies on materials. The workshop helped reassess the relationship between computing and experimental scattering science. The workshop had broader impact by identifying roles for computational scattering science in education, promoting public awareness of scattering research in the U.S., and lowering the barrier for entry into scattering research. More broadly, advances in computing, driven by experimental need, can help show the path forward for cyberinfrastructure development in other scientific communities and technical groups. Computational scattering science may offer career paths for young scientists, and the workshop helped assess these opportunities. Finally, the workshop attendees seem to be forming the nucleus of a community for computational scattering science, and such a community is needed for optimizing the future effort in this field.
