SBU’s Artem Oganov and crystal structure prediction

SBU’s Artem Oganov and crystal structure prediction

by -
0 1074

Computer model predicted that sodium would become transparent under pressure, and it does

Finding new materials has been a field where industrious and determined workers mixed elements, hoping to come up with the right combination to form a structure that might meet their needs.

While their choices of ingredients weren’t random, their results often proved disappointing, as the process produced considerably more failures than breakthroughs.

About eight years ago, however, Artem Oganov tried to change that. He didn’t want to build a better workbench or come up with a way to test more materials in a lab. He wanted to come up with a more efficient approach. Armed with a computer and working in an office at Stony Brook University, Oganov hoped to improve the process.

Predicting the crystal structure of an element or molecule in its lowest possible energy state presented an enormous challenge.

“Mathematically, when you formulate this problem, it looks intractable,” explained Oganov.

“The number of possible structures can be boiled down to ten to the power of 20 or 10 to the power of 50. Technically, you can’t sample all those structures,” he continued.

In 2003, Swiss scientists developed a computational method called metadynamics. It provided the first hope that the problem of crystal structure prediction might not be totally hopeless. Still, the process had significant limitations, Oganov said.

Oganov and several graduate students over the years, including Colin Glass, Andriy Lyakhov, Qiang Zhu, Guangrui Qian and Salah Eddine Boulfelfel, attempted to create a computer program that would narrow down those possibilities.

When their first several efforts were unsuccessful, “we were ready to give up.”

By combining several innovative approaches, including uniting global optimization (looking for the most likely solution in the big picture) with local optimization (narrowing the choices down among more subtle differences), they came up with a program that worked.

“The whole trick is to invent an algorithm which can work efficiently and reliably for a nearly infinite size,” he offered. “The problem was so big that we were dreaming without really hoping to get it.”

Oganov’s work has become “the gold standard,” suggested Stony Brook Geosciences Chairman Richard Reeder. “Discovering structures before was kind of random trial and error. There’s no systematic way to do it.”

Using the computer model, Oganov and his team predicted that sodium would become transparent under pressure. They found a collaborator who would conduct the test and, as they predicted, the metal became transparent.

The Oganov lab also became involved in an important discovery about carbon. Under high temperatures and pressure, carbon becomes diamonds. Under the same pressure, but at room temperature, carbon becomes superhard, without turning into diamonds.

Scientists had made guesses about the structure of this superhard carbon, but had trouble narrowing down the list in part because of the low resolution of experimental data. Using his computer model, Oganov predicted its structure. With some experimental support, Oganov’s prediction of a so-called M Carbon proved accurate.

Because other forms of carbon have had applications in technology, Oganov suggested this form might become instrumental in future manufacturing breakthroughs.

Oganov’s discoveries “won’t be seen to be applicable immediately,” Reeder explained, but could impact a wide range of fields, from planetary sciences to drug design.

In his presentations about his work, Oganov includes numerous historical references.

Indeed, if he hadn’t become a scientist, the Russian-born Oganov would have become a historian.

“History,” he explained, “gives very valuable lessons of wisdom: what were the good decisions and the bad decisions people made. How does progress work? History gives you good perspective on that.”

It’s important, he suggested, for people to have historical role models. Some of his include Linus Pauling, whom he described as being the “greatest chemist of the 20th century,” as well as Lev Landau. A physicist, Landau could “throw more ideas on one page than in a whole book written by other people.”

Laudau worked on his couch, scribbling notes that became the basis for papers and books, including one that physicists are still using, Oganov said.

A resident of East Setauket, Oganov would like to become a role model to future generations.

Oganov’s chairman believes he already stands out in his field, not only for his accomplishments but also for his intelligence.

“He’s well on his way. His code is the best out there,” Reeder said. “He has the motivation and the drive to do it.”