CC BY
3Tunable metasurface for ultrafast Fourier filtering
V.V. Yuskov'*, P.D. Kiryanov1, V.A. Sitnyanski1, A.S. Shorokhov1, A.A. Fedyanin1
1-Faculty of Physics, Lomonosov Moscow State University, Moscow 119991, Russian Federation
Due to the limitations of the electronic devices for information processing, approaches allowing analog processing in optics are being investigated [1-3]. The most traditional approaches are the Green's function method, the Fourier approach and the use of diffraction neural networks. Each of these approaches has its own advantages and disadvantages. The Green's function method allows to simplify the optical scheme in comparison with the Fourier method. However, it can be very difficult to realize the desired but arbitrary complex transmission function by this method. A rapidly developing new optical processing method using diffraction networks looks promising, but it seems to be more suitable for classification tasks. In addition, the method is currently limited due to the challenges of creating an optical analog of the nonlinear activation function.
The above approaches use amplitude-phase filters to transform the light beams. Metasurfaces are promising candidates for such devices due to their compactness and lack of power consumption. Their application simplifies the optical scheme compared to the use of, for example, spatial light modulators, but the main limitation in their use is static response.
In our study, a gallium arsenide metasurface has been modeled for analog optical image processing using a Fourier approach with all-optical tuning. The choice of material can be explained by the large refractive index variation at a wavelength of 850 nm due to bandgap effects under photoinduced carrier generation. The complex transmittance of the metasurface is chosen so that the metasurface switches between different image processing modes when femtosecond optical pumping is used. Following this approach, the second derivative of the analyzed image without pumping and the first derivative under optical pumping have been numerically demonstrated.
[1] W. Fu, D. Zhao, Z. Li, S. Liu, C. Tian, K. Huang, Ultracompact meta-imagers for arbitrary all-optical convolution. Light Sci. Appl., 11, 62 (2022).
[2] M. Cotrufo, S. Singh, A. Arora, A. Majewski, A. Alu, Polarization imaging and edge detection with image-processing metasurfaces, Optica, 10, 1331-1338 (2023).
[3] T. Badloe, S. Lee, J. Rho, Computation at the speed of light: metamaterials for all-optical calculations and neural networks, Adv. Photon, 4, 064002 (2022).
