Demonstration of on-chip all-optical switching of magnetization in integrated photonics

This article presents the first demonstration of single-pulse all-optical switching (AOS) for magnetization control within an integrated photonic platform, specifically using silicon nitride (Si3N4) waveguides. The goal is to explore the integration of ultrafast spintronic and photonic systems. The authors use femtosecond laser pulses to achieve deterministic magnetization toggling in a Co/Gd-based ferrimagnetic material patterned into a Hall cross structure, integrated directly on the photonic waveguide. The switching is controlled optically with precise electrical readout via the anomalous Hall effect (AHE).

Key findings and insights include:

1. On-chip AOS: The demonstration relies on fs-laser pulses at a 1 Hz repetition rate, successfully toggling the magnetization within sub-micron Hall cross structures. For smaller Hall crosses (500 nm arms), switching contrasts up to 90% are achieved. As device size increases, the switching becomes more stochastic due to non-uniform light absorption, as confirmed by simulations.

2. Device Scaling: A major challenge is the spatially non-uniform absorption profile in larger devices, leading to partial or stochastic switching. This behavior is linked to domain wall dynamics and thermal processes. As the device scales down, the deterministic switching improves, highlighting the importance of miniaturization for reliable operation.

3. Simulation and Analysis: Simulations reveal that the absorption of laser pulses is uneven across the Hall cross, exacerbated by edge effects and material inhomogeneities. This influences the switching contrast and contributes to domain wall pinning, which limits efficient switching in larger devices.

4. Magnetic Domain Relaxation: The study also investigates the role of domain wall (DW) relaxation in stochastic switching behavior. At near-threshold energies, DWs are pinned at sharp device corners, contributing to intermediate switching states.

5. Future Applications and Scalability: The research paves the way for integrating AOS in spintronic-photonic platforms for ultrafast, energy-efficient data processing, especially in applications requiring high-speed memory or neuromorphic computing.

This work marks a significant step in advancing integrated, on-chip AOS and lays the foundation for future hybrid spintronic-photonic systems.