Metasurface-Enabled Analog Optical Computing

Analog optical computing harnesses the wave nature of light to perform mathematical operations directly on optical signals. Recent advances in metasurface design have enabled compact, flat optical elements that manipulate amplitude, phase and polarisation at subwavelength scales. By engineering the spatial transfer function of metasurfaces, fundamental operations such as differentiation, integration and convolution can be achieved at the speed of light with minimal energy consumption. This paradigm offers an alternative to bulky lens–filter assemblies and electronic processing, promising ultrafast, parallel and low-power data processing for imaging, signal analysis and machine learning.

Research from Nature Portfolio

Broadband manipulation of the angular spectrum has been achieved using dielectric metasurfaces operating across the visible range, demonstrating direct differentiation of angular components and enhanced feature contrast in complex images. A single-layer Huygens’ metasurface has been shown to implement Fourier-domain transfer functions for operations such as differentiation and cross-correlation, enabling on-chip image processing with subwavelength footprint. In parallel, topological concepts have been applied to optical computing, where nontrivial singularities in the transfer function yield isotropic two-dimensional differentiation over broad bandwidths, and topologically protected analog processors solve differential equations with robustness against disorder and fabrication tolerances.

Research from all publishers

Dielectric metasurfaces have been used to realise all-optical object identification and three-dimensional reconstruction by encoding mathematical operations directly into subwavelength nanostructures, facilitating real-time imaging without digital post-processing. On the frontier of photonic neural networks, metasurface-based diffractive architectures have been multiplexed in the visible spectrum to perform multi-channel classification tasks, achieving chip-scale integration with high neuron density and energy-efficient inference. Additionally, broadband two-dimensional spatial differentiation and high-contrast edge imaging have been demonstrated across the visible spectrum, enabling phase and intensity object analysis with a simple metasurface insert in conventional microscopes.

Metasurface-Enabled Analog Optical Computing publication trend

The graph below shows the total number of articles in metasurface-enabled analog optical computing across all publications each year (not limited to Nature Index journals).

Technical terms

Metasurface: A planar array of subwavelength scatterers that imposes spatially varying phase, amplitude or polarisation changes on incident light.

Analog optical computing: The direct execution of mathematical operations on optical wavefronts without conversion to electronic signals.

Angular spectrum: The distribution of plane-wave components in an optical field, characterised by spatial frequency and propagation angle.

Spatial differentiation: A mathematical operation that extracts edge and feature information by computing gradients across an optical field.

Diffractive neural network: A neural-network-like architecture realised with cascaded diffractive layers performing linear optical transformations for machine learning tasks.

References

1. Meta-optics for spatial optical analog computing. Nanophotonics (2020).
2. Broadband angular spectrum differentiation using dielectric metasurfaces. Nature Communications (2024).
3. Single-layer spatial analog meta-processor for imaging processing. Nature Communications (2022).
4. Topological optical differentiator. Nature Communications (2021).
5. Topological analog signal processing. Nature Communications (2019).
6. All-optical object identification and three-dimensional reconstruction based on optical computing metasurface. Opto-Electronic Advances (2023).
7. Metasurface-enabled on-chip multiplexed diffractive neural networks in the visible. Light: Science & Applications (2022).
8. Two-dimensional optical spatial differentiation and high-contrast imaging. National Science Review (2020).