You Break it, You Understand It

PSC, XSEDE Help Quantum Chemists Understand Break-up of Atmospheric Chemicals

March 6, 2018

Why It’s Important:

Hydrazine and samarium are two chemicals with little in common, other than that humans sometimes introduce them into the atmosphere. Hydrazine is a rocket fuel. Samarium is an element that, when it reacts with oxygen atoms in the upper atmosphere, can damp down plasma fluctuations caused by solar activity. The U.S. Air Force is interested in both of them. Hydrazine, because they’d like to remotely identify its combustion products when the Bad Guys launch a rocket. Samarium, because if a national crisis ever coincided with sunspot activity, it would be nice to release a small amount into the ionosphere to maintain vital communications. To better understand the chemistry of these very different substances and how they will act once in the atmosphere, the Air Force Office of Scientific Research engaged quantum chemist Peter Armentrout and his colleagues at the University of Utah.

“For [the hydrazine] work, we were engaged by ENSCO Corp, which contracts with the Air Force. ENSCO was interested in modeling hydrazine, its compounds and IR signatures, so any information they could obtain that might help benchmark their modeling was useful. That way they could model other species potentially important in the launches so that all species formed might be identified. If all you have is the recorded signature, then you don’t necessarily know what species you are looking at. And if the bad guys use a different fuel mix, then the species formed could be very different.”

“The Air Force has done three trial releases of samarium into the ionosphere. The idea is that samarium atoms react with oxygen atoms to form the SmO molecule with enough energy to lose an electron, forming a samarium oxide cation, and both charged particles mediate plasma fluctuations. There aren’t many molecules [that can do this].”
—P. B. Armentrout, University of Utah

Armentrout and his team used a simple method to better understand both hydrazine and samarium compounds: they broke them. By studying the energy required to split the chemical bonds holding them together in the lab, and the way they interacted with components in the atmosphere, the Utah scientists could compare the results to what chemists thought they knew about both. But they needed intensive computing resources to simulate the exact results predicted by the theory. They turned to the XSEDE-allocated supercomputers Bridges at PSC and Comet at the San Diego Supercomputer Center.

How PSC and XSEDE Helped:

Bridges and Comet brought to the table a computational capacity that even a large university “cluster” supercomputer couldn’t match. Both studies employed the regular-memory and large-memory (LM) nodes of Bridges. The complex, multi-faceted “matrix” calculations needed for the work would not have been possible on other supercomputers. The 3-terabyte LM nodes offered a unique ability to provide “spare room” for the computations, avoiding the data traffic jam that would have occurred with more conventional 386 gigabyte (equal to 0.386 terabytes) conventional nodes.

The bond between the two nitrogen atoms in hydrazine rocket fuel (left) broke to produce fragments that weren’t at the lowest possible energy level. While the result was a little surprising, a comparison with the theory via Bridges helped explain it.

In the case of hydrazine, the bond between the two nitrogen atoms in the rocket fuel broke to produce fragments that weren’t at the lowest possible energy level. That was a little surprising, because generally molecules find the lowest level that they can. But the simulations on Bridges and Comet explained the result. In not going to the lowest level possible, the fragments had maintained the orbitals of the electrons within the fragments. This conserved the character of the “binding” electrons that had kept the nitrogens together.

“We turned to [Bridges] because … the electron-correlation calculations … [were] fundamentally just too big to do on a traditional supercomputer … We wanted to do a really high-level calculation, as good … as one could conceivably do. PSC and XSEDE allowed us to do that pretty routinely.”—P. B. Armentrout, University of Utah

In the case of samarium, the picture wasn’t as clear. Working with computational chemistry specialist Kirk Peterson of Washington State University, the team studied a compound similar to samarium oxide, samarium sulfide (SmS). They started with positively charged SmS+ cations, because generation of the oxygen-containing equivalent, SmO+, in the ionosphere is an important part of samarium’s ability to reduce plasma fluctuations. Also, the cations can be easily accelerated and detected in the lab (unlike the neutral SmS molecules). Finally, chemists understand a lot about the relationship between SmS and SmS+, so the results with the cation will tell them much about neutral SmS as well.

By hitting SmS+ with various components of the atmosphere and comparing the lab results with simulations on Bridges and Comet, the scientists found a discrepancy between measurement and theory. While initially a setback, the result will allow scientists to adjust their computational approaches to better predict the properties of heavy elements like Sm. This will improve their ability to use such elements in maintaining critical communications.

The scientists published their results in two papers. They reported the hydrazine results in the Journal of Chemical Physics in September 2017, and the samarium results in the same journal the following December.