by Steve Miller
As electronic devices become increasing common in our lives, they simultaneously shrink in size. Microelectronic devices whose dimensions are measured in micrometers (millionths of a meter) are common in military applications and likely to expand into other uses, such as communication and medical technologies. As is common with miniaturization, building devices at these dimensions introduces challenges that do not exist on larger scales.
“New materials are being developed for their high power capability so we can build devices that are smaller and smaller,” says Sukwon Choi, assistant professor of mechanical engineering. “As the power density increases, it gets hotter and hotter. As with all electronic devices, high temperature decreases reliability.”
Measuring the operating temperature in small scale devices
The problem with using conventional temperature measurement techniques is that the devices have been shrunk below the limits of the thermometers and thermocouples that are used to measure heat on a larger scale. Thermocouple use is also restricted by the high voltage environment of the devices when they are in operation. Optical measurement techniques allow measurements in a noncontact, nondestructive way on the scale of current devices that typically measure about 5 µm in length.
Choi’s group is developing techniques and instruments for thermal measurements at the micrometer scale in the Thermal Characterization Lab in Hammond Building on the University Park campus.
By incorporating four different thermal analysis techniques, they are able to look at multiple aspects of the thermal characteristics of microelectronic structures as they operate. According to Choi, his lab is the only facility in the United States that implements all four techniques in one place. Most university and government labs tend to focus on one of the four analytical methods.
Chemical engineers have long used Raman spectroscopy, based on the change in energy of photons reflected from a surface, to determine the structure of crystalline materials. Because the change in energy is related to the lattice vibrations of atoms in the structure, changes in the vibration rate can be measured using Raman techniques. The relationship of vibration to temperature of a material allows the technique to be applied to temperature measurements of crystalline structures, such as semiconductors. The ability to focus the laser source of the incident photons gives this technique a high spatial resolution. Choi is able to resolve temperature readings at a length of about 1 micrometer or less. This allows him to profile the temperature characteristics across a transistor channel with dimensions of several micrometers.
While Raman optical methods are well suited to measurements in crystalline semiconductor materials, they are not applicable to metal. Choi applies a second technique to characterize the temperature of metals. Thermoreflective thermoimaging measures the reflectivity of a material as a function of its temperature. Because metallic surfaces are highly reflective, this analysis is particularly useful in devices that have a metal surface above the semiconductor. In addition, most semiconductor devices have metallic components, such as electrical interconnections. Combining these two optical techniques can increase the temperature profile information available for these devices, leading to a 2D image of the entire structure.
Infrared thermography is widely used by industry to measure temperature on larger scales. Among other applications, IR thermography measures heat loss from buildings, finds hot spots in industrial processes, and is even used to search for disaster survivors. Choi uses this analytical tool on a much smaller scale, but the principle is the same.
IR themography offers several advantages in microelectronic thermal characterization. While its resolution at about 3 microns is less precise than that of the Raman and thermoreflective techniques, it can be used to characterize the overall device temperature. Because infrared radiation is a characteristic of temperature, measurement is not dependent on the nature of the surface material. In some cases, it may measure the temperature of layers beneath the surface.
For some applications, the decreased resolution is not critical for the analysis. On the scale of a microelectronic device, the measurement tends to give the average temperature of the entire device, not a profile across its surface. However, since the lifetime or reliability of such a device is frequently related not to the profile, but instead to the peak temperature during operation, IR thermography is an excellent method for failure analysis in microelectronics.
The final tool in Choi’s lab is photoluminescence in which photons are absorbed by a surface, followed by release of a photon of a different wavelength. The instrument measures the difference between the energy absorbed and the energy emitted. It provides an additional tool for temperature measurement. The technique can also be used to measure the surface strength of a material.
Choi works closely with several industrial partners to design and improve the instruments used in his lab. While the instruments may eventually be available commercially, the instruments themselves are very expensive and specialized. His current focus is on providing a service to users. By providing multiple techniques, he is able to reduce error in temperature measurements at the micrometer scale. He points out that the users of many current methods operate using rules of thumb that were developed on larger scale applications and may not transfer directly to small scale analyses.
Bridging a Gap
Choi is interested in combining expertise in multiple disciplines to develop new engineering solutions. This approach comes from his work in optoelectonics at Sandia National Laboratory in New Mexico. “Because we pushed the limits so much and were developing new materials, we had some exotic configurations,” he says. “We had to solve new problems in device design to make things work, and to do so we brought together experts from a wide range of specialties to work together.”
Choi’s goal is to train his students to think in ways that find solutions to the unique problems of microelectronics. As industries continue to develop new materials and push them to the limits of their capabilities, more and more complex problems will develop, he says. Choi plans to develop a hub to provide experts who are ready to tackle these problems related to new materials and higher power density.
“Right now we don’t have those types of specialists,” he says. “This will provide an opportunity to students to learn these techniques while also providing solutions to industry. There is a real need to solve microelectronics cooling issues through measurement.”
While the role of a thermal engineer in physics has traditionally focused on heat transfer in solids and on computational fluid flow, this new focus couples thermal science with semiconductor science. Real devices have electric field profiles and heat generation profiles that depend on what voltage is applied, on the gain, and on the variables in materials. Choi wants his students to be able to calculate those heat generation profiles and electric fields and use that information to get a better understanding of what is going on inside the device, determine precisely how hot it is getting, and find the location of the heating.
“In industry labs, they don’t have the people who are able to couple the two physics together,” he says. “Sometimes they have two or three people who work together to figure out how to do that, but the language is different for people from different fields. I am trying to educate my students to understand the terminologies from electrical engineering and from material science and to come up with some basic simulations for coupling all these physics. I am bridging the gap.”
Sukwon Choi received his Ph.D. in mechanical engineering from the Georgia Institute of Technology and was a post-doc at Sandia National Laboratories from 2013 to 2015. He came to Penn State in 2015.
Contact Dr. Choi at Sukwon.Choi@psu.edu