In partnership with colleagues in the Life Sciences Division, Chemical Sciences Division, Materials Science Division, and the Advanced Light Source, members of CCMC are looking at a number of different problems in materials science.
High-throughput Discovery of Improved Scintillator Materials
This project, involving first-principles calculations of modest computing requirements is designed as a systematic, high-throughput method to aid in the discovery of new bright scintillator materials by prioritization and down-selection on the large number of potential new materials. Some new bright scintillators such as BaCsBr:Eu have been produced in microcrystal form based on the theory work. This is joint work with with the Life Sciences Division. For more information contact Andrew Canning.
Developing f-electron Soft X-ray Spectroscopy Simulation, Theory, and Experiments for Clean Energy Materials
This project is developing and validating new user-friendly methods and codes for interpreting soft X-ray f-electron (lanthanide 4f and actinide 5f) spectroscopic experiments with light atoms since current methods do not handle f-electrons well. It is also connecting f-elements to the Materials Project at LBNL for discovery, property prediction, synthesis, spectroscopy, and validation using the Materials Project online software for structure prediction etc. and supporting clean energy efforts and materials (relevant to CC2.0). This is joint work with the Chemical Sciences Division (David Shuh) and the Materials Science Division (David PrenderGast). For more information contact Andrew Canning.
Informatics Tools for Analysis of Void Space of Porous Materials
Porous materials such as zeolites and metal organic frameworks have been of growing importance as materials for energy-related applications such as CO2 capture, hydrogen storage, and as catalysts for oil refinement. Very large databases of these structures are being developed, and so there is a requirement for tools to analyze and screen structure libraries – which can contain millions of entries – to discover materials with certain properties important to these applications. Critical to the success of this endeavor are tools and approaches which enable automatic, unsupervised analysis of these materials, as well as development of suitable descriptors which encode chemical information, and which can be used for screening. We developed a suite of such algorithms and released them as Zeo++ - a software suite to analyze structures and calculate relevant descriptors in a high-throughput manner. Zeo++ relies on computational geometry technique, Voronoi decomposition. This is joint work with the Materials Science Division. For more information contact Maciej Haranczyk.
Similarity-Driven Discovery of Porous Materials for Gas Separations
Promising zeolite materials for adsorption of CO2 and other molecules can be efficiently discovered through geometry-based identification of guest-molecule binding sites. The best known materials for CO2 separation exhibit local binding cavities of the approximate shape to envelope a guest molecule; these sites can contribute over 90% of a material’s total adsorption strength. Novel algorithms are developed to detect these binding sites in materials contained in a database, enabling efficient discoveries, avoiding computationally expensive molecular simulations. This is joint work with the Materials Science Division. For more information contact Maciej Haranczyk.
Exploring Frontiers of High Surface Area Materials
The vast chemical space of metal-organic framework (MOF) materials was selectively explored using a gradient-based optimization algorithm. MOF crystal structures were computationally assembled and refined through iterative modification of organic components. Constraining the search space to synthetically realistic organic ligands, materials with higher surface area than the current record-holding materials were shown to be theoretically achievable for a variety of crystal structure topologies; in particular, branching of organic molecules was shown to be an aspect of structure design with the capacity to expand the present estimated limits of synthesizable high surface area materials. For more information contact Maciej Haranczyk.
In partnership with colleagues in the Chemical Sciences Division, Materials Science Division, Kitware, and Intel members of CCMC are looking at a number of different problems in chemical science.
Quantum Chemistry on Quantum Computers
Quantum simulations on quantum computers have the potential to revolutionize the field of computational chemistry and physics. The goal of this project is to design and demonstrate quantum simulations of quantum chemistry problems using superconducting circuit simulation systems developed at Berkeley. Starting with small systems consisting of 2 or 3 qubits, the near future target is the use 10s of qubits to simulate chemical problems of a scale that has not been achievable until now. This project will be a close partnership with the groups of Siddiqi and Whaley from the Materials Science Division/UC Berkeley who have been focusing on building, controlling and analyzing the capabilities of the superconducting qubit circuit technology. This partnership has the potential to establish LBNL at the forefront of quantum computing from hardware to simulation. For more information contact Jarrod McClean.
Towards Predictive Chemistry of Heavy Elements
Development of new nuclear fuels for carbon free power production and the cleanup of nuclear waste from WWII nuclear weapons production relies on controlling heavy elements (such as uranium and plutonium) in chemical processes. To date the chemistry of radioactive heavy elements is complex and not fully understood. Experiments to learn more about the chemistry of heavy elements are difficult and hazardous. Predictive computational methods and approaches are being developed to complement the experimental investigations and grow the heavy element chemistry knowledgebase. This is joint work with the Seaborg Center at the Chemical Science Division. For more information contact Bert de Jong.
Knowledge Discovery Tools for Chemical Transformations
Integration of chemical information of molecules and their interaction with the environment from simulation and experimental data has the promise to discover new knowledge and insights that will enable researchers to develop new and better catalysis, or better processes to make chemicals needed in everyday life. In collaboration with open source software company Kitware new tools and ways to represent the chemical information are being developed that enable researchers to build, analyze and query databases full of chemical data. For more information contact Bert de Jong.
Accelerating Chemistry Simulations on Intel's New Knights Landing Architecture
Computing hardware architectures are undergoing a transformation from smaller and faster processors to many slower processors. To use these new architectures efficiently, requires the development of new and novel computational chemistry algorithms that take full advantage of all the features in the hardware. Together with Intel the NWChem computational chemistry software is being modernized to better utilize the new Knights Landing Architecture. For more information about the LBNL Intel Parallel Computing Center, contact Bert de Jong.
As people and institutions become more aware of climate change, they are asking whether extreme weather and climate events have already become more or less frequent. Using computing resources at the National Energy Research Scientific Computing Center (NERSC) and Oak Ridge National Laboratory, we are creating a large number of simulations of the CAM5.1 climate model, resolving features down to 25km globally, to quantify how the chances of regional and local extreme events are changing because of our past and current emissions. This is joint work with the LBNL Earth Sciences Division. For more information contact Michael Wehner and Daithi Stone.