Â鶹ÒùÔº

Materials scientists can learn a lot about a sample material by shooting lasers at it. With nonlinear optical microscopy—a specialized imaging technique that looks for a change in the color of intense laser light—researchers can collect data on how the light interacts with the sample, and through time-consuming and sometimes expensive analyses, characterize the material's structure and other properties.

Now, researchers at Pennsylvania State University have developed a computational framework that can interpret the nonlinear optical microscopy images to characterize the material in microscopic detail.

The team has its approach in the journal Optica.

"Nonlinear optical microscopy is an important tool that can reveal structural information about different materials," said lead author Albert Suceava, doctoral student in materials science and engineering at Penn State.

"The method, which looks for exotic interactions between matter and light, can be used to see things in material samples normally invisible to us otherwise. The samples you can study with this technique can come from anywhere. The method can be used in many fields ranging from biology to even quantum computing."

The way our eyes see the world is through linear optical interactions like reflection, refraction and absorption, Suceava explained.

"Whereas in nonlinear optical microscopy, we use focused laser beams to get light that is more intense than what you can get with everyday light sources like sunlight," Suceava said. "And this intense light can produce new kinds of optical signals that are detected to form an image. We can understand something about the structure of the material by looking at how these new signals change across a sample, or how they change with something like the polarization of the laser source.

"From there, we used our understanding of classical optical microscopes to develop a computational tool to interpret these images, which enables the determination of material properties at the microscopic scale."

The work came about, the researchers said, when they observed unexpected phenomena in microscopy images and questioned whether it was due to the sample or the microscope.

"This whole project started when we were doing nonlinear microscopy on a sample that we thought we understood very well but we were seeing things in our images that we couldn't explain, almost like an optical illusion," Suceava said.

"So, we took a very long time to ensure the observations were not just an optical illusion but accurate data. We had to make sure that we were able to break down exactly what the microscope is doing to the light and to our probe when it's focused very tightly. Our approach focuses on modeling the effects that tight laser focusing has on the polarization of light that is interacting with the sample."

Light travels in the form of electromagnetic waves with unique frequencies, and the interaction of atoms and molecules with light—also known as electromagnetic radiation—provides information about their structure.

"Light is really central to seeing our world; in fact, our sense of physical reality is dominated by what we see," said Venkatraman Gopalan, professor of materials science and engineering at Penn State and co-author on the paper. "Imaging with light is very fundamental and we are constantly looking at new ways of imaging things. It's all light interacting with atoms and scattering."

The electromagnetic spectrum has many types of light waves, ranging from radio waves to gamma rays. Each type of light has a different wavelength and frequency, and scientists can use the information on how objects and materials emit, absorb, transmit or reflect light to investigate their properties.

"Atoms vibrate differently and make music; they dance to different beats, and light is like music," Gopalan said. "From electrons to nuclei to clusters of atoms to their spins, they all sort of dance at different frequencies. It's almost like an opera. And when, for example, you want to know how atoms are vibrating, you may send in one color of light, and the atoms may vibrate and absorb some of that light.

"The light that's reflected back is slightly shorter and different in color. It has a slightly longer wavelength and smaller frequency, because that reflects the little bit of energy it gave off. Looking at the atomic scale structure and vibrations in molecules gives a very good signature of the material."

There are many techniques of using light to study the properties of materials ranging from X-rays to thermal imaging. For this research, the team employed a technique known as second harmonic generation microscopy.

"Second harmonic generation is where a material changes the color of light by doubling its frequency," Gopalan said. "It can detect signals that indicate a lopsided dance of electrons, which can reveal the polarity of materials. This doubling of frequency can turn an infrared into blue, which comes from this lopsided dance of electrons inside the atoms in these solids."

The scientists say they can make an image from the signals, but truly characterizing a material requires more than just creating an image.

"We need to know what is going on, what the atoms are doing, what's going on with local properties, but what the image is telling us has been a challenge because there's a lot more information than show and tell," Gopalan said.

The goal was to develop a framework that accurately models the interaction of tightly focused light with samples in nonlinear optical microscopy, providing reliable quantitative information, according to the researchers.

The team tested their framework on a variety of reference materials, comparing the results to known properties. Suceava noted that by doing this, they were also able to extract quantitative information from samples. Understanding the specific features, along with the quantitative information, is critical for developing new materials and understanding their properties, Suceava said.

"Our framework tries to move beyond 'look-and-see' to actually say why an image looks the way it does," Suceava said. "We want to know what additional information could be buried in the way images change with different light sources or different optics. We envision this framework as helping standardize the approach for data analysis in the non-linear optics community to improve the consistency and reproducibility of characterizing materials.

"We think that we found a way to look at this problem that's simpler than how other people have done it and still gives very good agreement with known samples. By mapping the material properties instead of just snapping a photo, we can help build a library of material properties that can be used in various applications."