In coupled semiconductor lasers, synchronized outputs do not always behave the same way across the full signal bandwidth. This dynamics-dependent synchronization can change as the lasers move from chaotic oscillations to low-frequency fluctuation (LFF) behavior. A numerical study of mutually coupled lasers shows how asymmetric optical feedback can produce in-phase synchronization in the original signals while low-pass-filtered signals become anti-phase synchronized during LFF dynamics.
For industrial automation and photonics engineers, the key lesson is that a single overall correlation value may not describe the relationship between coupled nonlinear signals. The frequency range being measured, the operating dynamics, and moments of optical frequency locking or unlocking can all affect the observed synchronization state.
How asymmetric feedback shapes coupled-laser dynamics
The modeled system contains two semiconductor lasers coupled through a partially transparent common external mirror. A semiconductor optical amplifier (SOA) is placed in the feedback path of one laser to create an imbalance between the two feedback strengths. The SOA also affects the coupling between the lasers, but not by the same factor: light returning to the laser with the SOA passes through it twice, whereas coupling light passes through once.
As a result, changing the SOA amplification coefficient changes both the feedback to one laser and the coupling strength, while feedback to the other laser remains fixed. At lower amplification, coupling-induced behavior is more prominent. At higher amplification, feedback-driven dynamics in the laser connected through the SOA become more influential.
The simulations use Lang-Kobayashi rate equations, which represent the laser electric fields and carrier densities while accounting for delayed optical feedback and coupling. This framework allows the researchers to examine how changing the feedback configuration affects the lasers’ intensity signals and synchronization.
Why synchronization depends on the frequency band
The study distinguishes between the full output signals and their slower components by applying a low-pass filter with a 1 GHz cutoff. This separates the fast chaotic oscillations from slower intensity variations associated with LFF dropouts. Cross-correlation is then used to assess the relationship between the two laser outputs, with positive correlation indicating in-phase behavior and negative correlation indicating anti-phase behavior.
At an SOA amplification coefficient of 1.3, the lasers show chaotic oscillations and positive correlation for both the original and filtered signals. Their optical frequencies remain close to one another, consistent with injection locking in the reported analysis.
At a coefficient of 2.6, the behavior is different. The original signals retain in-phase synchronization, but with weaker correlation. After low-pass filtering, the signals show anti-phase synchronization. The laser outputs also exhibit LFF dynamics: fast chaotic fluctuations coexist with slower, sudden intensity dropouts.
This contrast is the central feature of dynamics-dependent synchronization. The synchronization state is not determined only by whether two complete waveforms appear alike. It can vary with the frequency content being considered and with the underlying laser dynamics.
What happens during low-frequency fluctuation dropouts
Short-term correlation analysis helps explain the change. During periods when both lasers produce pulsations, the original signals can remain positively correlated. When an LFF dropout occurs, one or both laser outputs fall sharply, and the short-term correlation can change. The timing of these dropouts is not identical between the two lasers.
The study also tracks local optical frequency detuning—the instantaneous difference between the lasers’ optical frequencies. During pulsation periods, the frequency difference can remain near zero, indicating that the lasers are frequency locked. When dropouts occur, the detuning increases, indicating that the optical frequencies become unlocked. The reported interpretation is that the dropout reduces the coupling power available to maintain injection locking.
Because the two lasers have asymmetric feedback, their trajectories through the dynamics are not identical. One laser may enter a dropout before the other. This mismatch affects the slower signal components particularly strongly, producing intervals of negative short-term correlation even when the full signals show positive correlation over a broader measurement window.
Implications for analyzing coupled nonlinear systems
These findings have practical implications for anyone evaluating synchronization in nonlinear photonic systems. A measurement method that averages over long periods or combines all frequency components may conceal changes that occur locally in time or in a specific band.
- Separate frequency ranges: Compare broadband signals with filtered signals when the system contains both fast oscillations and slower events.
- Track synchronization over time: Short-term correlation can reveal local changes that an overall correlation value may obscure.
- Consider frequency locking: Monitoring local optical frequency detuning can help relate synchronization changes to periods of locking and unlocking.
- Account for asymmetry: Unequal feedback can lead to different dynamics in coupled elements, including different dropout timing.
The study is numerical and focuses on semiconductor lasers; its results should be understood in that context. However, the analysis illustrates a broader methodological point for coupled dynamical systems: synchronization can depend on the operating regime, the signal bandwidth, and the timing of transitions within the dynamics.
Conclusion
Asymmetric optical feedback can change both the dynamics of mutually coupled semiconductor lasers and how synchronization appears across different frequency components. In the reported simulations, chaotic operation supports in-phase synchronization in both original and filtered signals, while LFF behavior is associated with in-phase synchronization in the original signals and anti-phase synchronization in the filtered signals. Short-term correlation and local optical frequency detuning connect this difference to dropout timing and episodes of frequency unlocking. For engineers analyzing coupled nonlinear systems, the practical takeaway is to examine synchronization across relevant frequency bands and time scales rather than relying on a single aggregate measure.