Sub-diffraction thermoreflectance thermal imaging using image reconstruction

Amirkoushyar Ziabari,Yi Xuan,Je-Hyeong Bahk,Maryam Parsa,Peide Ye,Ali Shakouri
DOI: https://doi.org/10.1109/itherm.2017.7992461
2017-05-01
Abstract:Thermoreflectance thermal imaging technique uses light in the visible wavelength range and has a diffraction limit of ∼250nm. Despite that TR is still capable of acquiring temperature signal from devices smaller in size down to ∼3x below diffraction limit. Below diffraction limit, the detected thermoreflectance signal underestimates the true measured temperature by 360%. Image blurring was used in the forward problem to explain the apparent temperature of the device quite accurately. In most applications, there is no unambiguous model of the device temperature for forward problem and one needs to reconstruct the true temperature profiles of the sub-diffraction devices from their measured TR images. This is an ill-posed inverse problem which may not have a unique solution. Here, a maximum-a-posteriori (MAP) image reconstruction technique is used along with an Iterative Coordinate Descent (ICD) Optimization approach to solve this inverse problem and restore the true temperature profile of the devices. Preliminary results show that temperature of sub-diffraction heater lines down to ∼150nm can be accurately estimated.
What problem does this paper attempt to address?