Efficient all-optical switching remains a significant challenge due to the bosonic nature of photons, requiring nonlinear optical interactions. The Kerr gate, a classical example, suffers from limited fan-out capability, necessitating strong pulses to switch weaker ones. Various methods, primarily relying on resonant enhancement of light-matter interactions, have been explored to address this limitation. However, resonant enhancement leads to trade-offs, including reduced switching speed, increased intracavity dispersion, and challenges in channel separation. Slow light approaches also face similar limitations. This research proposes and experimentally validates an alternative method that leverages switching between different portions of soliton fission-induced supercontinua, exploiting the concept of an optical event horizon. This approach promises high switching efficiency, contrast, fan-out capability, insensitivity to critical biasing, and partial cascadability, addressing shortcomings of previous methods and opening new possibilities for all-optical logic gates.

Existing all-optical switching systems are often hampered by poor fan-out capabilities. Resonant optical cavities are frequently proposed to improve switching efficiency by enhancing local field strength; however, this enhancement necessitates collinear launching of interacting fields and introduces limitations in channel separation, speed, and dispersion. Slow light techniques, while aiming to increase interaction time, share similar drawbacks. Integrated photonic silicon nitride waveguides, offering higher nonlinearities than photonic crystal fibers, present an alternative platform for supercontinuum (SC) generation, potentially enabling high-speed optical communications. However, a device enabling time-division multiplexing at data rates exceeding 100 Gbit/s remains a significant gap. Previous research has suggested using an effective χ(3) nonlinearity enhancement without resonant cavities, exploiting the matched wavelengths of light pulses in opposite dispersion regimes. This involves the interaction of solitons and dispersive waves (DWs), leading to deceleration or acceleration of the DW. Analogies have been drawn to fluid dynamics and event horizons, but early experimental realizations fell short of expectations. This work addresses these shortcomings by employing incommensurate frequencies to ensure sustained interaction between pulses.

The experimental setup used an optical parametric chirped pulse amplifier (OPCPA) to generate broadband pulses, which were then shaped to produce a dispersive wave (DW) and a higher-order soliton (S) with adjustable frequencies and relative delay. These pulses were launched into a photonic crystal fiber (PCF) and their output analyzed using a spectrometer. The PCF was chosen to have two zero-dispersion wavelengths (ZDWs) to create a barrier against further soliton red-shifting and minimize sensitivity to energy fluctuations. Numerical simulations were performed using the generalized nonlinear Schrödinger equation (GNLSE) and the forward unidirectional propagation equation to model pulse propagation in the fiber, including effects like dispersion, self-phase modulation, self-steepening, and Raman scattering. Simulations incorporated realistic experimental parameters and accounted for the inherent sensitivity of soliton fission to quantum noise by using pulse ensembles with a spread in pulse energy. Cross-correlation frequency-resolved optical gating (XFROG) was used to analyze the temporal and spectral evolution of the interaction.

The experiments successfully demonstrated all-optical switching by manipulating the soliton fission process using a DW. The initial time delay between the soliton and DW controlled the output spectrum, allowing for deterministic modification of the supercontinuum. Three distinct all-optical AND gate functionalities were observed, with switching times in the femtosecond regime. The system exhibits a fan-out capability, meaning a weaker DW can switch a stronger soliton. The use of a fiber with two ZDWs mitigated the need for critical biasing, and one of the output channels showed cascadability. Numerical simulations corroborated the experimental results, showing excellent agreement with the observed switching behavior. The simulations further revealed the role of four-wave mixing processes in the switching mechanism and demonstrated the ability to suppress soliton collisions using a DW. The simulations also confirmed that the switching behaviour is robust, even in the presence of noise.

The experimental results validate the proposed all-optical switching concept, offering a significant improvement over previous methods by avoiding the drawbacks associated with resonant enhancement. The demonstrated fan-out capability, insensitivity to critical biasing, and partial cascadability fulfill key criteria for practical all-optical switching. The use of higher-order solitons significantly enhances switching efficiency compared to previous studies involving fundamental solitons. The integration of this technique with silicon-based platforms is projected to further reduce pulse energies, potentially enabling energy-efficient high-speed optical communications. Although further advancements are needed to fully realize the potential of optical logic circuits, the findings represent a significant step towards realizing high-performance, energy-efficient all-optical switching.

This research successfully demonstrated a novel all-optical switching mechanism using soliton fission-induced supercontinua. The method achieves high efficiency, fan-out, and cascadability while eliminating the need for critical biasing. Future research should focus on optimizing the system for specific applications, such as THz time-division multiplexing, and exploring the integration of this technique with other optical components to build more complex all-optical logic circuits. The potential for integrating this technology onto silicon-based platforms to reduce power consumption is a promising avenue for future research.

The current experimental demonstration uses a photonic crystal fiber with specific dispersion properties. The performance might vary with different fibers. The partial cascadability demonstrated needs further development to ensure full cascadability across all output channels. Further research is necessary to fully explore the scaling of this switching mechanism to larger arrays and more complex logic functions.