A key tenet of the Materials Genome Initiative for Global Competitiveness (pdf) is using computation to reduce the time for materials development by 75%, from 20 years to 5 years. A recent press release from Fujitsu Laboratories in Japan gives an early clue about the feasibility of this approach.
Fujitsu is interested in developing materials for novel nanodevices to replace silicon large scale integration devices. The drive to shrink electronic devices is starting to run up against the physical limits of the material to be miniaturized.
Turning to computational methods, Fujitsu used a first-principles method to calculate the electrical properties of a 1,000-atom device based on carbon nanotubes and graphene electrodes. In the press release the company says the significance of this breakthrough is that “The new technology opens the door to the design of exceptionally high-speed, energy-efficient nanodevices that break totally new ground with their development.”
First-principles computation is based on the quantum mechanics of a material’s electrons and atoms, thus experimental data or empirical parameters are not needed. It is useful for simulating the properties of materials like carbon where small differences in atomic arrangement results in large property differences. Consider, for example, how different the electrical properties of charcoal, graphite and diamond are.
Electrical properties were calculated using software developed by the Japan Advanced Institute for Science and Technology and the computational power of a supercomputer at the Information Technology Center at Nagoya University. First-principles calculations are iterative and tend to need a lot of computing time and memory. Each iteration updates input values and the computation continues until the output values converge. It took about three days to calculate the electrical properties of a 1,000-atom nanodevice using about one-third of the supercomputer’s capacity.
In the press release, Fujitsu explained that they worked with JAIST to tweak the software somewhat, and that they also used a “hybrid parallel processing technique.” As a result, Fujitsu was able to include the modeling of several times more atoms than it had previously be able to do.
The nanodevice modeled is a carbon nanotube with graphene electrodes. Lithium atoms occupied the inner edges of the graphene electrodes and several hydrogen atoms bridge the atomic layer between the electrodes and the nanotube. This is a very simple system, atomically, compared to most commercial engineered materials, which often have complex compositions or atomic structures. However, the company said its success in this instance “significantly paves the way to designing novel nanodevices.”
Because the Materials Genome Initiative is aimed at elevating the United States’ national competitiveness, there is some irony of discussing the efforts of a Japanese enterprise. However, this example illustrates the type of technology—also available in the US—that the MGI intends on leveraging.
And, while Fujitsu’s work shows great promise for designing a type of nanodevice, it also demonstrates that this route to materials design requires sizeable computational investment (hardware and software). Even at the speed attained by Fujitsu, the span of potential materials compositions, crystalline structures, properties and applications make it clear that a lot of computational capacity and agility will be needed in the US.
The Fujitsu work was published in the Aug. 11 online edition of Applied Physics Express.