Near-Field Semiconductor Imaging and Classification Enabled by Signal Cancellation Using Truncated mm-Wave Silicon Waveguide

We demonstrate that a straightforward truncated silicon dielectric waveguide serves as an effective millimeter-wave (mm-wave) near-field semiconductor imaging sensor. The strong localization of modal fields at the truncated silicon tip interface results in fine resolution below the diffraction limit. It is found that, at a specific probing distance, high conductance of a metallized sample exhibits lower reflection than bare silicon substrate due to the signal cancellation with the reflection generated from the significant mismatch at the truncated interface. We experimentally validate that this cancellation effect is dependent upon conductivity and it allows us to differentiate multiple levels of conductivity. With this technique, we achieve the resolution of ~1/3 λ0 and a high image contrast of ~20 dB between conductors and the bare silicon substrate. In addition, a cancellation effect is employed that allows for accurate calibration of working distance, providing stable responses across the scanning area. We demonstrate how this effect can be exploited for high-accuracy material classification with broadband spectral correlation analysis enabled by the silicon sensor. The proposed method can be easily incorporated with integrated all-silicon circuits, showing promise for cost-effective, compact imaging modules, which is highly beneficial for real-world deployment in the industry.