Programmable photonic processor for discrete fractional Fourier transform with π/8-order resolution

Area of Science:

• Photonics
• Optical Signal Processing
• Integrated Optics

Background:

• The fractional Fourier transform (FrFT) is vital for signal processing, but conventional methods lack reconfigurability and integrability.
• Existing implementations using lens systems or fiber arrays are bulky and difficult to adapt.

Purpose of the Study:

• To present a novel programmable discrete fractional Fourier transform (DFrFT) processor.
• To overcome limitations of conventional FrFT implementations by enabling dynamic reconfigurability and integration.

Main Methods:

• Designed a processor using a fixed array of basic transformation units (BTUs) with a dynamically reconfigurable architecture.
• Leveraged DFrFT order additivity to synthesize arbitrary DFrFT matrices.
• Employed inverse design to create four BTUs (π/8, π/4, π/2, π) with high simulated fidelity (>0.995).

Main Results:

• Achieved high fidelity (>0.989) for assembled DFrFT matrices across 16 transformation orders in numerical analysis.
• Experimental results on a silicon photonic platform showed individual BTU fidelity (>0.85) and assembled DFrFT fidelity (>0.8).
• Confirmed modularity, reconfigurability, and robustness through order mapping and fabrication tolerance analysis.

Conclusions:

• The developed processor offers a scalable and programmable platform for integrated optical signal processing.
• This technology is particularly suitable for time-frequency transformations and non-stationary signal analysis in optical computing.
• The system demonstrates significant potential for advancing future optical computing systems.