Periodic Reporting for period 2 - COLOR-UP (All-optical sub-THz signal filtering with multi-COLOR lasers)
Reporting period: 2022-05-01 to 2023-10-31
Information has always been a key element in society. The ability to exchange large amounts of data at high-speed has been a life-changing progress even though the public only saw its indirect consequences and didn’t always realize how much it relied on telecom and information. Despite its already amazing improvements seen in the last 20 years, the demand keeps going up and no one can really anticipate how a ten or hundred-fold bandwidth increase would further change our lives: from autonomous cars, remote surgeries, real-time diagnostic solutions or solutions helping fight climate change e.g. through improvement of logistic processes. All these are possible and are exciting opportunities.
From a physics point of view, to make wireless communication faster, a general trend has been to go towards higher frequencies. Following this idea, the technology is now looking at the mm-wave (> 100 GHz) and up to the THz range. To work at such a high-frequency, integrated microwave photonics solutions have shown crucial advantages supported by the development of ultra-high-bandwidth modulators, receivers, photomixers. Nevertheless, key components are still missing in this eco-system: in particular, signal processing solutions like filtering or conversion need further development. A central difficulty here being the lack of scalability of current approaches, i.e. to make them work at frequencies 1 or 2 orders of magnitude higher.
In this project, we are developing multi-wavelength lasers and are exploring how they could be used for all-optical signal processing. The crucial advantage of our approach is its intrinsic scalability: processing capability relies on the coupling between wavelengths taking place in the laser which takes place for any frequency separation from tens of GHz to several THz.
The overall objectives of COLORUP are therefore three-fold:
- Develop a robust and flexible design of multi-wavelength lasers including efficient control mechanisms.
- Demonstrate the signal processing capability of multi-wavelength lasers, with a first focus on signal filtering.
- Gain further insight into the wavelength coupling mechanism to understand how it can be tailored to improve signal processing functionalities.
From a physics point of view, to make wireless communication faster, a general trend has been to go towards higher frequencies. Following this idea, the technology is now looking at the mm-wave (> 100 GHz) and up to the THz range. To work at such a high-frequency, integrated microwave photonics solutions have shown crucial advantages supported by the development of ultra-high-bandwidth modulators, receivers, photomixers. Nevertheless, key components are still missing in this eco-system: in particular, signal processing solutions like filtering or conversion need further development. A central difficulty here being the lack of scalability of current approaches, i.e. to make them work at frequencies 1 or 2 orders of magnitude higher.
In this project, we are developing multi-wavelength lasers and are exploring how they could be used for all-optical signal processing. The crucial advantage of our approach is its intrinsic scalability: processing capability relies on the coupling between wavelengths taking place in the laser which takes place for any frequency separation from tens of GHz to several THz.
The overall objectives of COLORUP are therefore three-fold:
- Develop a robust and flexible design of multi-wavelength lasers including efficient control mechanisms.
- Demonstrate the signal processing capability of multi-wavelength lasers, with a first focus on signal filtering.
- Gain further insight into the wavelength coupling mechanism to understand how it can be tailored to improve signal processing functionalities.
Since the beginning of this project in November 2020, an important part of the work has been dedicated to setting the scene before the real scientific effort could be made. This includes the recruitment of two PhD students and one post-doctoral researcher. In the mist of the covid crisis which was still having a significant impact in 2020-2021, the process was unfortunately slower than usual and the whole team was only up and running by September 2021. This initial delay was however not a loss as we have exploited it to purchase the important pieces of equipment necessary for our project, including a new optical table, two high-performance tunable lasers, a real-time large bandwidth electrical spectrum analyser and a large bandwidth arbitrary waveform generator. With the complete team and available equipment, it was then possible to kick off the scientific work both experimentally and theoretically.
For the latter, we started simulating a multi-mode rate equation model that we then expanded to include up to 4 different wavelengths, (modulated) optical injection and phase-controlled optical feedback to match the characteristics of our experimental system. Though simulations were also used to confirm a good agreement with experimental results, we primarily focused on the effect of optical feedback and the constraint on laser and feedback parameters to achieve good control of the laser multi-wavelength emission. Thus, we used random sampling techniques to explore the system behaviour in a comprehensive way despite the large number of parameters. We are now processing these data to extract a clear set of constraints to achieve good control of the laser. In a next step, we will verify that these constraints scale well when additional modes are considered.
Experimentally, taking advantage of the multi-wavelength lasers already available in the lab, we performed a detailed characterization of their emission properties and behaviour. This was obviously a very important step to a have a clear baseline for further investigations. We then started investigating their response to optical injection, i.e. their behaviour when coherent light from another laser is sent to the device. Afterwards, we were then ready to add a modulation to the optical injection which led to two research breakthroughs detailed in the next section, namely the demonstration of spectral multiplication and wavelength conversion capability.
Finally, we also prepared several new laser designs and sent them for manufacturing to our partner foundry platform. At this stage, we are unfortunately still waiting to receive these new devices.
For the latter, we started simulating a multi-mode rate equation model that we then expanded to include up to 4 different wavelengths, (modulated) optical injection and phase-controlled optical feedback to match the characteristics of our experimental system. Though simulations were also used to confirm a good agreement with experimental results, we primarily focused on the effect of optical feedback and the constraint on laser and feedback parameters to achieve good control of the laser multi-wavelength emission. Thus, we used random sampling techniques to explore the system behaviour in a comprehensive way despite the large number of parameters. We are now processing these data to extract a clear set of constraints to achieve good control of the laser. In a next step, we will verify that these constraints scale well when additional modes are considered.
Experimentally, taking advantage of the multi-wavelength lasers already available in the lab, we performed a detailed characterization of their emission properties and behaviour. This was obviously a very important step to a have a clear baseline for further investigations. We then started investigating their response to optical injection, i.e. their behaviour when coherent light from another laser is sent to the device. Afterwards, we were then ready to add a modulation to the optical injection which led to two research breakthroughs detailed in the next section, namely the demonstration of spectral multiplication and wavelength conversion capability.
Finally, we also prepared several new laser designs and sent them for manufacturing to our partner foundry platform. At this stage, we are unfortunately still waiting to receive these new devices.
The work performed so far, led to a few important breakthroughs. Beside an improved understanding of the multi-wavelength lasers and their response to optical injection, we demonstrated two important new features:
1. We showed that sending a narrow optical frequency comb at a wavelength close to a depressed mode of the multi-wavelength could be used to trigger “spectral multiplication” of the comb on the other modes of the laser. In other words, the initial comb is precisely copied at different wavelengths. Importantly, key comb characteristics such as the comb linewidth and phase properties are perfectly preserved. Moreover, we showed that this process doesn’t interfere at all with the feedback control: by simply tuning the phase in the feedback cavity used as control mechanism, we achieve rapid switching between wavelengths (less than 10 ns) thus leading to agile spectral multiplication of the input comb from tens of GHz to several THz.
2. Using a similar approach, we demonstrated wavelength conversion of data signals with excellent Bit Error Ratio. Besides its potential interest for optical communication, this result also show that the “spectral multiplication” mechanism is not limited to harmonic signals, and can be efficiently exploited for data/random signals. Further investigations are ongoing to determine the limitations of this process in terms of bit rate and data modulation format.
Beyond these two important results and further development to understand limitations and underlying mechanisms, we expect to make progress on two other aspects in the next two years and a half before the end of the project.
On one hand, it is quite clear experimentally and numerically that the coupling mechanism between modes is essential when considering the multi-wavelength dynamics of the laser. With the new chips that we are now expecting from the foundry, we expect to further clarify this role and, ideally, demonstrate that this coupling can be tuned to some extent thus influencing the multi-wavelength dynamics of the laser.
On the other hand, we will also focus on the core aspect of the project: the signal filtering capability of multi-wavelength lasers and their scalability. Again relying on new devices, we expect to show that all-optical filtering can be achieve using these lasers and aim to achieve cutting-edge performances thanks to design improvements of the laser cavity.
1. We showed that sending a narrow optical frequency comb at a wavelength close to a depressed mode of the multi-wavelength could be used to trigger “spectral multiplication” of the comb on the other modes of the laser. In other words, the initial comb is precisely copied at different wavelengths. Importantly, key comb characteristics such as the comb linewidth and phase properties are perfectly preserved. Moreover, we showed that this process doesn’t interfere at all with the feedback control: by simply tuning the phase in the feedback cavity used as control mechanism, we achieve rapid switching between wavelengths (less than 10 ns) thus leading to agile spectral multiplication of the input comb from tens of GHz to several THz.
2. Using a similar approach, we demonstrated wavelength conversion of data signals with excellent Bit Error Ratio. Besides its potential interest for optical communication, this result also show that the “spectral multiplication” mechanism is not limited to harmonic signals, and can be efficiently exploited for data/random signals. Further investigations are ongoing to determine the limitations of this process in terms of bit rate and data modulation format.
Beyond these two important results and further development to understand limitations and underlying mechanisms, we expect to make progress on two other aspects in the next two years and a half before the end of the project.
On one hand, it is quite clear experimentally and numerically that the coupling mechanism between modes is essential when considering the multi-wavelength dynamics of the laser. With the new chips that we are now expecting from the foundry, we expect to further clarify this role and, ideally, demonstrate that this coupling can be tuned to some extent thus influencing the multi-wavelength dynamics of the laser.
On the other hand, we will also focus on the core aspect of the project: the signal filtering capability of multi-wavelength lasers and their scalability. Again relying on new devices, we expect to show that all-optical filtering can be achieve using these lasers and aim to achieve cutting-edge performances thanks to design improvements of the laser cavity.