Microscale IR thermography
Infrared cameras are used to visualize thermal conditions at IR wavelengths. They have several applications that include night vision and nondestructive testing of composite materials used in aircraft and building construction. The sensors in IR cameras are assembled within focal-plane arrays (FPAs) and can be categorized as either photon (indium antimonide—InSb, or mercury cadmium telluride) or thermal (pyroelectric or ferroelectric crystals, amorphous silicon, or vanadium oxide—VOx) detectors.1 Photon detectors have better speed and sensitivity in thermal and long-IR (LWIR) wavelengths than thermal detectors. Traditionally, therefore, microscale thermography has been achieved by operating InSb detectors in combination with an optical lens that has an optimized design for 3–5μm wavelengths. However, photon detectors must be cooled using liquid nitrogen or a Stirling cycle cooler to achieve high-quality images with high sensitivity and low noise levels. Photon detectors are thus costly, and the associated cooling systems take up valuable space.
Thermal images, obtained with microscopic IR cameras at sub-0°C temperatures, have been used to visualize the latent heat of freezing for biological cells (see Figure 1).2–5 This method has had a large impact on cryopreservation techniques in biotechnology, from both a scientific and engineering standpoint. The thermal microscope approach has also been used to observe the crystallization of organic materials (see Figure 2).6–8 When used with a pulsed or modulated light source, the method can be used to detect invisible points of failure and defects in semiconductors and microstructures of composite materials (see Figure 3).9,10
We have designed a new mobile quantitative microscale thermography instrument that incorporates a thermal detector with improved performance.11–13 We apply our new hardware and software to a commercial IR microbolometer camera that itself has poor time and temperature resolution, but is small and cheap. Our new approach includes an improved time response and a microlens that is optimized for the wavelength of the microbolometers. Our equipment is widely applicable in the field of microscale thermography to achieve high-quality thermal imaging.
With our technology, a noise-equivalent temperature difference of 50mK and a spatial resolution of 11μm can be obtained for thermal imaging even when uncooled IR detectors are used. Cheaper IR imaging systems can thus be realized. Our design includes an achromatic lens for microscale imaging, video signal superimposed for real-time emissivity corrections, and pseudo-acceleration of the time frame. It is designed to fit inside a 17×28×26cm carrying box. The precise emissivity correction is required for quantitative analysis of the thermal image that is obtained. This is especially true for certain thermal phenomena (e.g., phase transitions) because the emissivity change during these transitions is not a simple function of temperature.
The video signal synthesizer in our instrument allows the direct digital signal of monitored temperature or positioning data to be recorded. We use our own original encode/decode protocols to read out the encoded digital signal data that is embedded within each image. This process involves applying the mixed IR signals and the superimposed data to the individual pixel emissivity corrections and pseudo-accelerations of the periodic thermal phenomena. Each pixel has a different temperature dependence because the emissivity of industrial materials (as well as biological tissues) is usually heterogeneous. The pseudo-acceleration algorithm improves the timescale resolution for measurements of periodic thermal events and thus integrates multiple image data and reduces instrument noise.
An emissivity-corrected thermal image of freezing onion skin cells that was captured using a microbolometer VOx FPA is shown in Figure 4(a). These images indicate that our microbolometer system can be used successfully to perform quantification of microscale thermal images (e.g., visual differential thermal analysis). Figure 4(b) shows the pseudo-accelerated frame rate that was applied to modulated microscale spot heating of a polyimide composite film (uniaxially oriented) at a frequency of 0.97Hz. The pseudo-acceleration procedure improves the frame rate from 30 frames per second (fps)—at which the raw data was captured—to 690fps, which allows the in-plane thermal diffusivity to be precisely determined.
We have developed a quantitative microscale thermal imaging system that can be used as a portable instrument. Our apparatus has the potential to improve the spatial and time resolution of uncooled IR cameras. We are now planning to market our instrument design for the general-purpose thermal characterization of materials, with energy transport and storage applications. We are targeting spectroscopic and 3D thermal imaging for the next generation of our technology.
Junko Morikawa has been a professor in the Department of Organic and Polymeric Materials since 2013.