Tuesday 11 March 2025
The quest for better photonic integrated circuits (PICs) has led researchers to develop new ways of characterizing and reducing losses in waveguides, crucial components of these devices. A recent study published in a scientific journal presents an innovative approach that leverages advanced image processing techniques to estimate the sidewall roughness of waveguides at a wafer scale.
In traditional PICs, optical loss is a major hurdle to overcome. Waveguide edges can exhibit significant roughness, which scatters light and increases losses. Accurately measuring this roughness has been a challenge due to the small scales involved. Researchers have relied on atomic force microscopy (AFM), but this method is time-consuming and impractical for large-scale PICs.
The authors of the study employed scanning electron microscope (SEM) images to estimate sidewall roughness. They used edge detection algorithms to identify the waveguide edges, which were then analyzed using autocorrelation functions (ACFs). The ACFs provided valuable insights into the roughness parameters, such as amplitude and correlation length.
The researchers fabricated two silicon nitride waveguides with different fabrication processes. Each waveguide was cut into 12 individual devices, and optical losses were measured at a wavelength of 633 nanometers. The SEM images were taken from various locations along the waveguide edges, yielding a total of 90 images for each wafer.
The authors found that the sidewall roughness of the two wafers differed significantly. One wafer had a mean roughness of approximately 5.86 nanometers, while the other wafer exhibited a mean roughness of around 8.58 nanometers. The results showed a direct correlation between the roughness and optical loss.
The study’s findings have important implications for PIC design and fabrication. By accurately estimating sidewall roughness at an early stage, designers can optimize their designs to minimize losses and improve overall performance. This approach could also enable faster development of new PIC architectures and more efficient use of materials.
In addition to its practical applications, the research highlights the potential of advanced image processing techniques in characterizing complex optical systems. As PICs continue to play a crucial role in emerging technologies such as artificial intelligence and quantum computing, the need for innovative characterization methods will only grow.
The authors’ approach offers a promising solution to the challenges faced by PIC designers and researchers. By leveraging SEM images and ACF analysis, they have demonstrated the feasibility of wafer-scale sidewall roughness estimation.
Cite this article: “Characterizing Sidewall Roughness in Photonic Integrated Circuits using Advanced Image Processing Techniques”, The Science Archive, 2025.
Photonic Integrated Circuits, Waveguides, Optical Loss, Sidewall Roughness, Scanning Electron Microscope, Edge Detection Algorithms, Autocorrelation Functions, Silicon Nitride, Pic Design, Material Optimization
