Science enabled by DANSE
From DANSE
DANSE is enabling new science in several areas :
| Table of contents |
Computational Neutron Scattering Science
The DANSE system for doing neutron scattering research by computer will deliver new science, better science, and ease of use to neutron scattering researchers.
The DANSE project will help the SNS by providing some traditional approaches for obtaining neutron spectra from raw data. The bigger opportunities, however, are in the confluence of multi-teraflop computing with the SNS experimental facilities. The scientific goals of neutron scattering research are the determinations of atomic structure and dynamics of solids, liquids, and condensed matter, for which greater detail and new phenomena can be found when computing augments the measurements. The details differ for different fields of research, but typically involve the fitting of experimental data to underlying physical models, simulations, or constraints imposed by other measurements. Many of the important models and simulations have come from advances in materials theory over the past two decades.
Deliverables from the DANSE project should prove useful for other fields of experimental materials research, such as synchrotron radiation research. DANSE will broaden participation in neutron scattering research by developing intuitive, science-based interfaces to the software. It will also help educate new users through realistic virtual experiments.
- The DANSE system for doing neutron scattering research by computer will deliver new science, better science, and ease of use to neutron scattering researchers.
- Many of the new capabilities originate from combining experimental analyses with modern materials theory.
- Additional capabilities originate from new software that will exploit the higher quality data from SNS instruments, or their higher rates of data acquisition.
Scientific Software Development
The exploitation of these new scientific opportunities requires the collaborative production of software on a scale unfamiliar to x-ray and neutron researchers. The technical approach of the DANSE project is to develop software components with independence, integrating them into a common runtime framework. This framework supports the lifecycles of software components and applications, and gives them methods for distributed computing. The project has adapted methods for commercial software management and has applied leading-edge software engineering to establish a model for software productivity and reliability. Commercial software management practices are not typical for scientific software development, but this approach is both necessary and overdue. Its quantitative aspects fit well into the culture of analytical scientists, as do its early reviews of design, inspections of code, and extensive testing. Quality assurance practices are recognized by the DANSE team as improving efficiency over the course of the project. Many of these practices are automated, and are not an excessive burden to the young scientists who develop the software.
- Professional standards and processes for software development are in place for the DANSE project.
- A common runtime framework facilitates the integration of software components developed by different groups.
- Common algorithms and methods for distributed computing are shared by all scientific subgroups, ensuring consistency and reliability.
Diffraction
Knowledge of atomic-scale structure underlies our understanding of materials properties. Much of the interest today is in materials of unprecedented complexity, often involving physical dimensions on the nanometer length scale or larger. Conventional crystallographic methods fail at the nanoscale, and new approaches are required. High-performance computing hardware and DANSE software will allow us to combine data from multiple experimental techniques, and integrate theory with data analysis, enabling robust solutions to nanostructure problems.
Rapid data collection from SNS diffractometers will allow new types of time-dependent and parametric studies on materials. The flexible and scalable deployment of DANSE software will enable complex analyses in near real-time at the data rates expected from parametric diffraction experiments at the SNS.
- For diffraction research to extend to the nanometer length scale and beyond, new software is needed to augment neutron data with information from other sources such as different experimental methods or ab-initio calculations of materials structure.
- DANSE provides a flexible architecture for these comparisons between computation and experiment, often making them possible for the first time.
Diffraction Nugget (2008) (http://wiki.cacr.caltech.edu/danse/images/4/48/Nugget_Diff_2008.pdf)
Diffraction Nugget (2007) (http://wiki.cacr.caltech.edu/danse/images/5/5a/PDF-Refinement_GPRA_Nugget_2007.pdf)
Diffraction Nugget (2006) (http://wiki.cacr.caltech.edu/danse/images/1/11/SB_nuggetDANSEnsf0601.pdf)
Engineering Diffraction
Neutron diffraction is a powerful probe of the internal solid mechanics of heterogeneous materials, giving information that underlies mechanical properties and predictions of lifetimes of engineering structures. This work requires an integrated approach combining diffraction experiments with mechanics modeling, but robust integration has proved elusive to date. For the first time, DANSE will allow a direct simulation of an engineering diffraction experiment by combining realistic mechanics models of microstructure and Monte Carlo simulations of engineering diffractometers. The models and software to be developed in this project will re-define diffraction stress analysis and open new venues of research, not only with neutrons, but also with synchrotron radiation.
- Engineering diffraction requires comparisons to predictions from large-scale solid mechanics calculations.
- DANSE provides a flexible architecture for these comparisons between computation and experiment, often making them possible for the first time.
Engineering Diffraction Nugget (2007) (http://wiki.cacr.caltech.edu/danse/images/7/76/Diffraction-Peak-Splitting_GPRA_Nugget_2007.pdf)
Engineering Diffraction Nugget (2006) (http://wiki.cacr.caltech.edu/danse/images/5/53/EU_NSF_nugget_EngND_01_2006.pdf)
Small-Angle Neutron Scattering
SANS has seen widespread use for measuring mesoscopic structures in research fields from biology to physics, but to date the analysis of these data has been limited to simple uniform shapes and simple types of interactions. In too many studies of complex materials, especially in extreme environments, the richness of the measured data is not fully recovered by the analysis. For example, although SANS has been touted as an ideal tool for examining conformational changes in proteins (whose shapes are anything but regular) as they perform their function, few successful studies exist. DANSE will provide the tools needed for current research on complex nanoscale structural questions.
- Software ease-of-use is especially important for many users of these techniques.
- SANS measurements probe mesoscopic-scale structures of materials and surfaces, but have largely been confined to one-dimensional averages.
- DANSE will allow fuller utilization of the features in the data of today, extending the analysis to more complex models of interest in biology and soft condensed matter.
SANS Nugget (2008) (http://wiki.cacr.caltech.edu/danse/images/2/28/Nugget_SANS_2008.pdf)
SANS Nugget (2006) (http://wiki.cacr.caltech.edu/danse/images/7/77/PB_Nugget-2006.pdf)
Reflectometry
Reflectometry is a technique for the study of surfaces and interfaces within a sample. The limiting factor on the productivity of reflectometry instruments is the software. DANSE will provide a framework for robust global optimization on flexible models, allowing systems to be refined during the course of an experiment rather than requiring months of effort, as is sometimes the case today. The richer modeling environment will enable users to work directly with the parameters of interest, such as conformations of proteins or micromagnetic structure, rather than the averaged planar information of today. Measurements of diffusely scattered neutrons will provide insight into the in-plane structures of magnetic and non-magnetic thin film systems, such as the secondary structures of small proteins vectorially oriented in single-membrane systems in biology, or magnetic domains in patterned structures with applications to high-density storage media.
- Software ease-of-use is especially important for many users of these techniques.
- Reflectometry measurements probe mesoscopic-scale structures of materials and surfaces, but have largely been confined to one-dimensional averages.
- DANSE will allow fuller utilization of the features in the data of today, extending the analysis to more complex models of interest in biology and soft condensed matter.
Reflectometry Nugget (2007) (http://wiki.cacr.caltech.edu/danse/images/2/23/Alzheimers-peptide_GPRA_Nugget_2007.doc.pdf)
Reflectometry Nugget (2007) (http://wiki.cacr.caltech.edu/danse/images/5/5e/Magnetic-Structure_GPRA_Nuggets_2007.pdf)
Reflectometry Nugget (2006) (http://wiki.cacr.caltech.edu/danse/images/4/42/PK_nugget_refl3-2006.pdf)
Inelastic Neutron Scattering
The dynamical motions of electron spins and atom vibrations in materials are largely understood for simple periodic systems, but analytical theories become only qualitative for solids with even moderate disorder or complexity. Computer simulations are a most promising approach for understanding basic dynamics phenomena in real materials, and are needed today for interpreting experimental data from inelastic neutron scattering. For the first time, DANSE will enable direct simulations of experimental data by merging molecular dynamics methods with Monte Carlo simulations of neutron spectrometers. In addition, DANSE will provide a system for comparing results from modern electronic structure calculations with data from inelastic neutron scattering experiments, allowing much more rigorous interpretations of experimental data.
- New types of data analysis developed under the DANSE project will exploit the crisp momentum transfer information from SNS inelastic instruments that is not usable today.
- Materials theory is especially useful for leveraging materials dynamics information from SNS spectrometers.
Inelastic Nugget (2008) (http://wiki.cacr.caltech.edu/danse/images/a/a9/Nugget_Inelastic_2008.pdf)
Inelastic Nugget (2008) (http://wiki.cacr.caltech.edu/danse/images/5/54/ARCS_DANSE_Highlight.pdf)
Inelastic Nugget (2007) (http://wiki.cacr.caltech.edu/danse/images/e/e5/Data-Reduction-GPRA_Nugget_2007.pdf)
Inelastic Nugget (2007) (http://wiki.cacr.caltech.edu/danse/images/1/14/Phonon-Scattering_GPRA_Nugget_2007.pdf)
Inelastic Nugget (2006) (http://wiki.cacr.caltech.edu/danse/images/7/7f/DANSE_demo-2006.pdf)
