Accelerating materials research in the information age
New materials are crucial to cleaner energy technologies, from batteries to photovoltaics to lighter weight vehicles. But today the development cycle is too slow: around 18 years from conception to commercialization.
That may be about to change, thanks to a new science gateway called the Materials Project. The gateway was designed by a team of researchers from the United States Department of Energy’s Lawrence Berkeley National Laboratory (Berkeley Lab) and the Massachusetts Institute of Technology to speed up that process.
“Our vision is for this tool to become a dynamic ‘Google’ of material properties, which continually grows and changes as more users come on board to analyze the results, verify against experiments and increase their knowledge,” said Kristin Persson, a Berkeley Lab chemist and one of the founding scientists behind the Materials Project. “So many scientists can benefit from this type of screening. Considering the demand for innovative clean energy technology, we needed most of these materials yesterday.”
Taking a genome-like approach
The Materials Project employs an approach to materials science inspired by genomics. But rather than sequencing genomes, researchers are using high performance computing to characterize the properties of inorganic compounds, such as their stability, voltage, capacity, and oxidation state. The results are then organized into a database with a user-friendly web interface that gives all researchers free and easy access and searching.
“First-principles calculations have reached the point of accuracy where many materials properties, relevant for photovoltaics, batteries, and thermoelectrics, can be reliably predicted,” said Gerbrand Ceder, an MIT materials scientist and founder of the Materials Project.
A better battery—one that is cheaper and has more power and energy while being safe—could finally make possible the dream of an electric vehicle reaching performance and cost parity with a gasoline-powered car. But beyond batteries, novel materials could transform a host of other industries, from food packaging to buildings. For example, the Materials Project is working with several entities interested in making stronger, corrosion-resistant, lightweight aluminium alloys, which could make possible lighter vehicles and airplanes.
“Materials innovation today is largely done by intuition, which is based on the experience of single investigators,” said Persson. “The lack of comprehensive knowledge of materials, organized for easy analysis and rational design, is one of the foremost reasons for the long process time in materials discovery.”
The US president, Barack Obama, has recognized the importance of advanced materials with his announcement in June of the Materials Genome Initiative “to double the speed with which we discover, develop, and manufacture new materials.” Many of the concepts of that initiative were inspired by the Materials Project, Persson said.
With the help of high performance computing systems at the US National Energy Research Scientific Computing Center (NERSC), the Berkeley Lab Lawrencium cluster, and systems at the University of Kentucky, the Materials Project database currently contains the structural and energetic properties of more than 15,000 inorganic compounds, and up to hundreds more are added every day. Researchers are continuously adding new properties to enable true rational design of new materials for a wide variety of applications.
A gateway for science
To build the Materials Project web tool, the team approached computer systems engineers at NERSC who have extensive experience building web-based interfaces and technologies. These science gateways make it easier for researchers to access computational resources and share data with the rest of their community.
“The Materials Project represents the next generation of the original Materials Genome Project, developed by Ceder's team at MIT,” said Shreyas Cholia, a NERSC computer engineer who helped develop the Materials Project tool. “The core science team worked with developers from NERSC and Berkeley Lab’s Computational Research Division to expand this tool into a more permanent, flexible and scalable data service built on top of rich modern web interfaces and state-of-the-art NoSQL database technology.”
The project's front-end uses JQuery and AJAX to interact with a back-end built using the Django framework. The gateway uses a REST-like model to query resources and return data as JSON objects. As a whole, the Materials Project will be hosted on NERSC's science gateway infrastructure.
This is only the beginning for the Materials Project, funding permitting. Persson described future plans, which involve expanding both the database content and the portal capabilities, as ambitious.
“Currently we are only featuring known inorganic materials, but we are planning to go beyond the known world by data mining and substitution algorithms,” Persson said. “We are also planning significant expansion in terms of materials properties (electronic properties, elastic constants, defect energies etc) that the user can search on, targeting different technological applications.”
Upgrades to the portal will create a more dynamic user environment in which the user can create 'sand boxes' to upload their own data, submit calculations, and combine their results with the Project's data to explore potential new materials solutions, Persson said.
All of these features, plus a growing user base, are bound to create a need for more computing power. Currently they are using only a few terabytes of data storage, but they expect their data needs to grow, perhaps beyond the computing grants they currently have for NERSC and the University of Kentucky's Supercomputing Center.
“Our calculations are very diverse in terms of memory and computing time as well as not highly scalable,” Persson said. “That goes a bit against the current computing trends that favor high concurrency jobs.”
A version of this article first appeared on the Berkeley Lab website; this version with additional content by Miriam Boon, iSGTW.
Comments
Post new comment