Advantage and Core Benefit
- Spectral zoom function (hardware): Spectral resolution improved by a factor of 5 or more simply by attaching two slits to an existing spectrometer.
- Significant reduction in the number of measurement data points (software): Machine learning using sparse principal component analysis reduces the number of interference spectrum data points to be measured to less than 10, resulting in ultra-fast data acquisition and analysis time.
- Demonstrated that thin film thickness can be estimated with high accuracy using only 3 data points of interference spectra.
- Enables real-time imaging of film thickness on semiconductor wafers, films, FPDs, etc.
Background and Technology
There is a need for instantaneous mapping of film thickness on semiconductor wafers, films, FPDs, etc. Currently, however, it takes 0.1-10 seconds per measurement location (data acquisition and analysis), so it takes nearly 30 seconds to measure 25 locations on a single sample.
The inventor was able to improve the spectral resolution by a factor of 5 or more by simply applying the Moire effect (*) and attaching slits at the entrance and exit based on an existing compact and inexpensive multi-channel spectrometer. Furthermore, machine learning using sparse principal component analysis was performed from multiple interference spectrum data acquired in advance from the target sample to reduce the number of interference spectrum data points to be measured to less than 10, resulting in ultra-fast data acquisition and analysis time..
This enables real-time imaging of semiconductor wafers, films, FPDs, etc., which has been difficult in the past, and enables inspection of all areas and all numbers of film thicknesses.
* T. Konishi et al., “Super spectral resolution beyond pixel Nyquist limits on multi-channel spectrometer,” Optics Express, Vol. 24, No. 23, pp. 26583-26598, 2016 (November 9, 2016)
Publications & Patents
PCT/JP2014/001316、2014-185152JP、PCT/JP2017/024479、PCT/JP2020/013495
Principal Investigator & Academic Institution
- Tsuyoshi KONISHI (Osaka University, Graduate School of Engineering, Associate Professor)
Expectations
- We are seeking companies to in-license this technology or work with our researchers (with technical guidance from the inventor) to further develop this technology.
Project No.HK-00258c