Physicists develop method to detect single atomic defects in semiconductors

One of the challenges of cramming smarter, more powerful electronics into ever-shrinking devices is developing tools and techniques to analyze the materials that make up these devices with ever-greater precision.

Physicists at Michigan State University have taken a long-awaited step in that direction with an approach that combines high-resolution microscopy with ultrashort lasers.

The technique, described in the magazine Nature Photonicsallows researchers to discover misfit atoms in semiconductors with unprecedented precision. Semiconductor physics labels these atoms as “defects,” which sounds negative, but they are usually added to materials on purpose and are critical to the performance of semiconductors in today’s and tomorrow’s devices.

“This is especially relevant for nanostructured devices,” said Tyler Cocker, the Jerry Cowen Endowed Chair in Experimental Physics and leader of the new study.

That includes things like computer chips, which routinely use semiconductors with nanoscale features. And researchers are working to push nanoscale architecture to the limit by developing materials that are one atom thick.

“These nanoscopic materials are the future of semiconductors,” said Cocker, who also directs the Ultrafast Terahertz Nanoscopy Laboratory in the Department of Physics and Astronomy at MSU. “When you have nanoscale electronics, it’s really important to make sure that electrons can move the way you want them to.”

Defects play a big role in that electron movement, which is why scientists like Cocker are eager to know exactly where they are and how they behave. Cocker’s colleagues were excited to learn that his team’s new technique would allow them to obtain that information easily.

“One of my colleagues said, ‘I hope you went out and celebrated,’” Cocker said.

Vedran Jelic, who led the project as a postdoctoral researcher in Cocker’s group and now works at the National Research Council Canada, is the first author of the new paper. The research team also included doctoral students Stefanie Adams, Eve Ammerman and Mohamed Hassan, as well as undergraduate researcher Kaedon Cleland-Host.

Cocker added that the technique is simple to implement with the right equipment and that his team is already applying the technique to atomically thin materials such as graphene nanoribbons.

“We have a number of open projects where we’re using the technique with more materials and more exotic materials,” Cocker said. “We’re basically applying it to everything we do and using it as a standard technique.”

A light (almost) touch

Tools already exist, notably scanning tunneling microscopes (STMs), that allow scientists to detect defects in individual atoms.

Unlike the microscopes that many people recognize from high school science class, STMs don’t use lenses and light bulbs to magnify objects. Instead, STMs scan the surface of a sample with an atomically sharp point, much like the stylus on a record player.

However, the STM tip does not touch the surface of the sample. It only comes close enough for electrons to hop or tunnel between the tip and the sample.

STMs record how many electrons jump and where they jump from, among other information. In this way, they can provide information about samples at the atomic scale (which is why Cocker’s lab calls this nanoscopy instead of microscopy).

But STM data alone are not always sufficient to unambiguously detect defects in a sample, especially in gallium arsenide, an important semiconductor material used in radar systems, high-efficiency solar cells and modern telecommunications equipment.

For their latest paper, Cocker and his team focused on gallium arsenide samples to which defective silicon atoms had been deliberately added to influence the way electrons move through the semiconductor.

“To the electrons, the silicon atom looks like a deep hole in the road,” Cocker said.

Although theoreticians have been studying this type of defect for decades, experimentalists have not been able to detect these individual atoms directly. In their new technique, Cocker and his team still use an STM, but the researchers also shine laser pulses directly onto the tip of the STM.

These pulses consist of light waves with terahertz frequencies, meaning they move up and down a trillion times per second. Recently, theorists showed that this is the same frequency at which silicon atom defects should move back and forth in a gallium arsenide sample.

By coupling STM and terahertz light, the MSU team created a probe with unparalleled sensitivity to the defects.

When the STM tip encountered a silicon defect on the surface of the gallium arsenide, an intense signal suddenly appeared in the team’s data. When the researchers moved the tip one atom away from the defect, the signal disappeared.

“Here was a defect that people had been looking for for over 40 years, and we could hear it ringing like a bell,” Cocker said.

“At first it was hard to believe because it’s so obvious,” he continued. “We had to measure it in every way possible to make sure this was real.”

However, once they were convinced that the signal was real, it was easily explained thanks to the years of theoretical study devoted to the subject.

“When you discover something like that, it’s really helpful that there’s decades of theoretical research that describes it thoroughly,” said Jelic, who co-authored the new paper with Cocker.

While Cocker’s lab is at the forefront of this field, there are groups around the world that are currently combining STMs and terahertz light. There are also several other materials that could benefit from this technique for applications beyond defect detection.

Now that his team has shared its approach with the community, Cocker is curious to see what other discoveries await him.