TL;DR
This paper explores optical injection locking as a superior alternative to phase-locked loops (PLLs) for reducing phase noise in semiconductor lasers. The authors demonstrate that injection locking can achieve state-of-the-art phase noise levels (-125 dBc/Hz at 1 MHz offset) with a simpler, more practical implementation, paving the way for more robust coherent optical systems.
What Problem Does This Paper Address?
Coherent optical systems offer exciting possibilities for advanced functionality and performance, but they are notoriously susceptible to optical phase noise. This noise can severely degrade the signal, especially in applications requiring precise phase relationships between lasers, such as advanced analog signal processing or optical channelizers. The traditional solution has been to use phase-locked loops (PLLs) to force one laser to track the phase of another. However, semiconductor lasers, while desirable for their cost and integration potential, have relatively large linewidths. This necessitates very wide PLL bandwidths, which in turn imposes stringent requirements on loop time delays, making practical implementation difficult and limiting performance. This paper addresses the challenge of achieving low phase noise in semiconductor laser systems by investigating a more promising approach: optical injection locking.
Methods
The authors compare two main methods for achieving coherent optical detection and carrier tracking: phase-locked loops (PLLs) and optical injection locking (IL). They theoretically analyze the phase noise performance of both, deriving transfer functions that describe how effectively they suppress laser phase noise. For PLLs, they highlight the critical role of loop time delay and its impact on stability and bandwidth. For optical injection locking, they derive a first-order transfer function and estimate its achievable bandwidth.
The experimental setup involves an external cavity master laser and a distributed feedback (DFB) slave laser. The master laser's light is injected into the slave laser's cavity. A heterodyne beat note is formed between the injection-locked slave laser and a frequency-shifted portion of the master laser. This beat note's phase noise is then measured using a dedicated phase noise test set. A key experimental consideration is the precise matching of path lengths in the interferometer to minimize excess noise due to delay decorrelation.
Datasets
Not discussed in the paper.
Key Insights and Ideas
- Injection Locking as a Superior Alternative: Optical injection locking offers a significantly simpler and more practical approach to carrier tracking compared to PLLs, largely circumventing the problematic time delay constraints.
- High Bandwidth Potential: Injection locking can achieve very high effective bandwidths (on the order of 10^10 rad/s), which is crucial for effectively suppressing the phase noise of semiconductor lasers.
- Path Length Matching is Critical: For accurate measurements and effective injection locking, the optical path lengths from the master laser to the slave laser and to the detector must be carefully matched to avoid excess noise from delay decorrelation.
- Acoustic Noise is a Significant Factor: At lower offset frequencies (< 20 kHz), environmental acoustic noise can become a dominant source of phase noise, even in a well-designed system.
- Injection Ratio Matters: The phase noise performance of injection-locked lasers is dependent on the injection ratio (the ratio of injected power to free-running slave power), with higher injection ratios generally leading to lower phase noise.
Key Results
The authors experimentally demonstrated phase noise levels as low as -125 dBc/Hz at a 1 MHz offset frequency using optical injection locking. This result is comparable to state-of-the-art performance achieved with PLLs but with a simpler implementation. They observed that phase noise generally increases as the offset frequency decreases. The phase noise also decreased as the injection ratio increased, consistent with theoretical predictions.
Limitations and Future Work
The authors acknowledge that the measured phase noise at low offset frequencies (< 20 kHz) appears to be dominated by environmental acoustic noise rather than limitations of the injection locking mechanism itself. They suggest that further improvements could be made by acoustically isolating the fibers. They also note that injection locking alone may not be a complete carrier tracking solution for all applications and may need to be supplemented with a drift control mechanism, such as an optical PLL or a Pound-Drever-Hall servo loop, to achieve the full potential performance. Future work could involve independent measurements of the slave laser linewidth to validate theoretical models and exploring hybrid schemes that combine injection locking with drift control.
Summary updated