Satellite communications, military radar systems and those 5-G networks you will soon depend on have something in common. Scientists base these systems on microelectronics – the design, manufacture, and use of integrated circuits.
One bottleneck in the integrated circuits is self-heating. These devices get hot because of the Joule heating occurring in it. Like your phone, all electronics get hot, and the device’s lifetime consequently decreases.
It is therefore necessary to measure the temperature these devices and circuits generate to understand the component mean-time-to-failure.
The problem is the devices exist on a sub-micron level. Engineers have shrunk these devices below the limits of physical thermometers and thermocouples that are used to measure temperatures at larger scales.
They have to find a way to measure that temperature on a small scale.
Sukwon Choi, Ph.D., and Assistant Professor of Mechanical Engineering at Penn State University
Sukwon Choi, Ph.D., and Assistant Professor of Mechanical Engineering at Penn State University has found a way.
“Failure is a local phenomenon,” Choi said. “Like our car, if the tire breaks, the whole car breaks. If you get like a weak spot in the circuit and that spot fails, then the whole system fails. Similarly, these transistors, if you have a hotspot and failure occurs at that local hotspot, then the transistor doesn't work anymore.”
Those thermal hotspots are on the submicron scale in length.
“These microelectronic devices are prone to overheating,” Choi said. “You have to be able to measure that temperature in order to design cooling systems or to assess or predict the component lifetime.”
Local peak temperature and device lifetime are directly correlated. Thus, scientists need to know the temperature to estimate the mean-time-to-failure or component lifetime.
Physical thermometers are too large to measure the local temperature in a circuit. These next generation devices based on microcircuits for power conversion and wireless communications operate under high voltage and high current conditions. You can’t touch a device for safety reasons. Physical contact will also interfere with the device’s operation.
Choi uses light to measure temperature.
“We want a high spatial resolution, noncontact and noninvasive thermography or thermometry technique,” he said. “And we can do that with light using micro-Raman spectroscopy. You're basically inventing an optical thermometer. We’re leveraging Raman's capabilities for optical thermography.”
How does that work?
Raman spectroscopy looks into the frequency or energy of phonons, or quantized lattice vibrations. A phonon is a discrete unit or quantum of vibrational energy, just as a photon is a quantum of electromagnetic or light energy. Phonons and electrons are the two main types of elementary particles or excitations in solids.
“They're looking into phonon frequency - the frequency of the atoms that oscillate within a crystal or amorphous solid. Then that oscillation frequency is directly related to energy. Raman spectroscopy measures that frequency or energy,” he said.
Raman spectroscopy looks into a phonon frequency or energy. Researchers ordinarily use that capability to study the structure of materials. Yet lattice vibrations or phonons are related to thermal transport in solids and thus temperature. Heat conduction in a solid happens in the form of vibrational waves, according to Choi.
“For conventional electronics, transistors and diodes going into all of these consumer electronics are based on silicon,” he said. “But for applications that demand higher power, higher current, higher voltage operation, silicon has its own limits regarding its material properties. One important thing is the electronic band gap.”
For these high frequency high power electronics, microelectronics researchers have been seeking materials with wider band gaps for higher frequency, high voltage and higher power operation. These researchers have been pursuing gallium nitride and silicon carbide for the past 10 years.
“Researchers have made huge successes in terms of applications like electric vehicles or wireless communications, satellite communications and military applications. We mainly focus on gallium nitride.”
Now scientists are looking into what they refer to as ultra-wide band gap semiconductors as the next generation for these high power electronic devices. That includes materials like aluminum gallium nitride and gallium oxide.
Larger band gap materials conduct more energy that permits the operation of devices under higher voltage conditions. Wider bandgap translates into a larger critical electric field of the material, which means for a given thickness, it can tolerate higher voltage without burning out.
That opens the door for many favorable aspects, such as development of smaller and lighter systems with higher efficiency, Choi said.
For things like electric cars, researchers want to make smaller power converters that consume less battery power.
“However, a major bottleneck to the commercial success of wide bandgap and ultra-wide bandgap semiconductors is overheating,” he said.
LabRAM HR Evolution Confocal Raman Microscope examines a slicone wafer
Choi uses a HORIBA LabRAM HR Evolution Confocal Raman Microscope with an ultra-low frequency (ULF) filter and customized pulsed laser source. The filter allows him to capture Stokes and anti-Stokes Raman scattering, giving him more accurate temperature readings. Choi uses the ULF filter measurements with other methods to get the most precise readings possible.
“You can go down to low wave numbers and you can see more peaks that can be leveraged to measure temperature and study sub-continuum-scale thermal transport,” he said.
The gentle, soft-spoken professor gestures with his hands, drawing imaginary waves on a graph against a black curtain. Black barriers surround the instrument to eliminate ambient light for more accurate measurements.
LabRAM HR Evolution protruding outside of blackout box in Sukwon Choi’s Penn State University lab.
“The LabRAM has a longer focal length compared to other Raman systems, which translates into higher spectral resolution, which in turn gives us a higher temperature resolution. It also has very good mapping capabilities,” Choi said.
His research uses Raman spectroscopy as a unique solution to study thermal problems in micro/nanosystems.
“We're leveraging the Raman capability for temperature measurement instead of structural characterization that all others do, so we can understand the self-heating or overheating behavior of next generation microelectronic devices,” he said. “It’s usually based on wide band gap semiconductors like gallium nitride or beyond that are going to be used in high power, high frequency applications, like power conversion for electric vehicles, renewable energy sources, and RF power amplifiers for wireless communication applications.”
Do you have any questions or requests? Use this form to contact our specialists.
