TL;DR
This paper introduces a novel fiber-based coherent detection system that combines an optical preamplifier with both a spectrum bandpass filter and a time-domain filter. This approach effectively mitigates Amplified Spontaneous Emission (ASE) beat noise, a significant challenge in preamplified coherent systems, leading to improved detection sensitivity.
What Problem Does This Paper Address?
In high-speed optical communication and sensing systems like LiDAR, achieving high receiver sensitivity is crucial for extending link distances and improving performance. While coherent detection theoretically offers superior sensitivity over direct detection by extracting phase and frequency information, its practical implementation is often hampered by noise. Specifically, using optical preamplifiers, which are beneficial for boosting weak signals, introduces Amplified Spontaneous Emission (ASE) noise. This ASE noise degrades the optical Signal-to-Noise Ratio (SNR) and can vary with data or pulse rates, making preamplified coherent detection less common than direct detection. This paper aims to overcome this limitation by developing a preamplified coherent detection system that effectively suppresses ASE noise.
Methods
The proposed system integrates three key components:
- Optical Preamplifier (EDFA): This component amplifies the incoming optical signal. However, it also introduces ASE noise.
- Spectrum Bandpass Filter: A narrow bandpass filter (100 GHz in this study) is used to filter out unwanted spectral components of the amplified signal, including much of the ASE.
- Time-Domain Filter (TDF): This is the novel element, implemented using a Mach-Zehnder Electro-Optic Modulator (MZM) acting as a synchronous time gate. The TDF selectively passes signal pulses within a defined time window, further reducing the impact of residual in-band ASE that the bandpass filter couldn't eliminate. The TDF's gate width (duty cycle) is a key parameter.
The authors also provide a detailed noise analysis, deriving equations for various noise components (shot noise, signal-ASE beating, ASE-ASE beating, local oscillation-ASE beating) and the overall receiver SNR.
Datasets
Not discussed in the paper. The experiments were conducted using a custom-built experimental setup.
Key Insights and Ideas
- Synergistic Noise Reduction: The combination of a spectrum bandpass filter and a time-domain filter is more effective at reducing ASE noise than either filter alone. The bandpass filter removes out-of-band ASE, while the time-domain filter targets residual in-band ASE.
- Time-Domain Filtering for ASE: The time-domain filter, acting as a synchronous gate, is particularly effective at reducing the in-band ASE beat noise, which is a persistent problem in preamplified coherent systems.
- Trade-off in Time-Domain Filtering: The effectiveness of the time-domain filter is dependent on its duty cycle. Narrower duty cycles (e.g., 5%) lead to greater ASE reduction and SNR improvement, but this must be balanced with the need to capture the entire signal pulse.
- Preamplification Benefits: Even with the added complexity, optically preamplified coherent detection, when properly filtered, can achieve high sensitivity, comparable to pure coherent detection, without degradation.
Key Results
- The system successfully demonstrated a 2.78 dB SNR improvement with a 5% time gating duty cycle and a 1 dB improvement with a 50% duty cycle, compared to an open (100%) duty cycle, at 10 pW input power.
- For a shot noise-dominated system, the time-domain filter provided minimal improvement. However, when ASE-induced noise was comparable to shot noise, the TDF significantly enhanced detection performance.
- The sensitivity of the coherent detection system achieved was as high as 0.7 pW (approximately 10.9 photons within a 10 ns pulse).
- The direct detection sensitivity was found to be over 20 dB worse than the coherent detection.
Limitations and Future Work
- Shot Noise Domination: The paper notes that when the system is dominated by shot noise, the time-domain filter's improvement is less significant. This suggests that further optimization might be needed for such scenarios.
- Duty Cycle Optimization: While narrower duty cycles showed better results, the paper implies a need to balance this with signal capture. Future work could explore optimal duty cycle settings for different pulse widths and repetition rates.
- Bandwidth Considerations: The paper mentions that a wider digital bandpass filter leads to stronger shot noise, indicating that the interplay between optical and digital filtering bandwidths is an area for further investigation.
- Experimental Validation: The authors successfully demonstrated the concept experimentally, but further characterization and optimization across a wider range of parameters (e.g., different preamplifier gains, filter bandwidths, and pulse characteristics) would be beneficial.
Summary updated